FINANCIAL STABILITY REVIEW DECEMBER 2005€¦ · INDEBTED EURO AREA HOUSEHOLDS USING MICRO-LEVEL...

F INANC IAL STAB I L I TY REV IEWDECEMBER 2005

EURO

PEAN

CEN

TRAL

BAN

KFI

NAN

CIAL

STA

BIL

ITY

REVI

EWD

ECEM

BER

200

5

In 2005 all ECB publications will feature

a motif taken from the

€50 banknote.

FINANCIAL STABILITY REVIEWDECEMBER 2005

© European Central Bank, 2005

AddressKaiserstrasse 2960311 Frankfurt am MainGermany

Postal addressPostfach 16 03 1960066 Frankfurt am MainGermany

Telephone+49 69 1344 0

Websitehttp://www.ecb.int

Fax+49 69 1344 6000

Telex411 144 ecb d

All rights reserved. Reproduction foreducational and non-commercial purposesis permitted provided that the source isacknowledged.

Unless otherwise stated, this documentuses data available as at 4 November 2005.

ISSN 1830-2017 (print)ISSN 1830-2025 (online)

3ECB

Financial Stability ReviewDecember 2005

PREFACE 7

I OVERVIEW OF RISKS TO FINANCIALSTABILITY 9

II THE MACRO-FINANCIAL ENVIRONMENT 17

1 THE EXTERNAL ENVIRONMENT 171.1 Risks and financial imbalances

in the external environment 171.2 Key developments in international

financial markets 331.3 Conditions of non-euro area

financial institutions 41

2 THE EURO AREA ENVIRONMENT 462.1 Economic outlook and risks 462.2 Balance sheet conditions of non-

financial corporations 472.3 Balance sheet conditions of the

household sector 54

III THE EURO AREA FINANCIAL SYSTEM 63

3 EURO AREA FINANCIAL MARKETS 633.1 Key developments in money

markets 633.2 Key developments in capital

markets 64

4 THE EURO AREA BANKING SECTOR 744.1 Financial conditions in the

banking sector 744.2 Risks facing the banking sector 844.3 Shock absorption capacity of the

banking sector on the basis ofmarket indicators 101

4.4 Overall assessment 106

5 OTHER EURO AREA FINANCIALINSTITUTIONS 1085.1 Financial conditions in the

insurance sector 1085.2 Risks facing the insurance sector 1145.3 Overall assessment 119

C ON T EN T S6 STRENGTHENING FINANCIAL SYSTEM

INFRASTRUCTURES 1216.1 Payment systems 1216.2 Securities clearing and

settlement systems 127

IV SPECIAL FEATURES 131

A MEASUREMENT CHALLENGES INASSESSING FINANCIAL STABILITY 131

B FINANCIAL MARKET CONTAGION 142

C ASSESSING THE FINANCIALVULNERABILITY OF MORTGAGE-INDEBTED EURO AREA HOUSEHOLDSUSING MICRO-LEVEL DATA 150

D WHAT DETERMINES EURO AREA BANKPROFITABILITY? 159

E MAIN EFFECTS FROM THE NEWACCOUNTING FRAMEWORK ON BANKS 168

F CENTRAL COUNTERPARTY CLEARINGHOUSES AND FINANCIAL STABILITY 177

STATISTICAL ANNEX S1

BOXES1 The recent surge in US share buybacks:

causes and possible financial stabilityimplications 19

2 Falling savings and rising debt in thehousehold sector: a financialstability risk? 24

3 Credit developments in the newnon-euro area EU Member States 26

4 The reform of the renminbiexchange rate regime 31

5 Sectoral profit and leveragedevelopments of euro area listednon-financial corporations: evidencebased on micro data 47

6 Corporate earnings and sectoralexposure at risk in the euro area 53

4ECB cFinancial Stability ReviewDecember 2005

7 The influence of mortgage productinnovations on risks to household debtsustainability 59

8 Structural trends in the euro moneymarket 65

9 Developments in the European creditderivatives markets 69

10 Financial conditions of large euroarea banks 75

11 Assessing financial stabilityimplications of recent findings fromthe ECB Bank Lending Survey 79

12 Survey on major EU banks’ perceptionof risks in the year ahead 85

13 The syndicated loan market in the euroarea matures into a distinct asset class 92

14 Large EU banks’ exposures tohedge funds 97

15 A decomposition of euro area bankstock volatility 103

16 Ageing population and longevity risk 11117 Low interest rates and balance sheet

vulnerabilities of life insurancecompanies and pension funds 114

18 The oversight of SWIFT: objectives,scope and structure 123

19 The oversight of retail paymentsystems 124

20 The CPSS-IOSCO Recommendationsfor Central Counterparties 128

CHARTS1.1 Net lending/borrowing of the US

economy 171.2 US non-farm, non-financial

corporate sector financing gap 181.3 Credit market liabilities of US

non-farm non-financial corporations 191.4 US corporate sector rating

downgrades, upgrades and balance 211.5 US housing affordability 231.6 US household net worth 231.7 Share of foreign currency loans in

total loans in the new EU MemberStates 26

1.8 Emerging market bond spreadsrelative to US short and long-terminterest rates 29

1.9 Global current account positions 301.10 China monthly foreign exchange

reserve increase, trade balanceand FDI 30

1.11 Speculative positions and theUSD/EUR exchange rate 33

1.12 Implied volatility on USD/EUR 331.13 Risk-neutral probability density

function of the USD/EURexchange rate 34

1.14 US six-month TED spread 341.15 US commercial paper, total amount

outstanding 351.16 Indicators of positioning and

leverage in the US bond markets 361.17 Net non-commercial positions on

futures and options, and the yieldspread between ten and five-yeargovernment bonds 36

1.18 Net percentage of equity fundmanagers responding that globalbond markets are overvalued 37

1.19 US industrial bond spread and thedownward rating revision ratio 37

1.20 US “earnings yield premium” andthe real oil price 38

1.21 Speculative positions and oil prices 391.22 Brent crude oil futures prices 391.23 Options-implied risk-neutral

densities 391.24 Non-commercial positions and total

open interest in gold futurescontracts and the gold price 40

1.25 Emerging market bond spreads 401.26 Emerging market local currency

bond yields 411.27 Trading revenues as a proportion

of total net revenues 431.28 Changes in Value at Risk (VaR)

levels 442.1 Survey-based estimates of the

downside risk of weak real GDPgrowth in the euro area 46

2.2 Profit ratios of euro area listednon-financial corporations 47

2.3 Actual and expected corporateearnings in the euro area 49

5cECB


CONTENTS

2.4 Breakdown of the real annual rateof growth of external financing tonon-financial corporations in theeuro area 50

2.5 Cost of external financing of non-financial corporations 50

2.6 Demand for loans and credit lines toenterprises, and contributing factors 50

2.7 Debt securities issued by euro areanon-financial corporations – fixedversus floating rate 51

2.8 New business loans to euro areanon-financial corporations withshort-term interest rates and termspreads 51

2.9 Credit standards applied to theapproval of loans to non-financialcorporations – small and medium-sized versus large enterprises 51

2.10 European non-financial corporatesector downgrades, upgrades andbalance 52

2.11 Euro area households’ financialsituation and unemploymentexpectations 56

2.12 Residential property price growthrate in the euro area 56

2.13 Building permits and residentialinvestment in the euro area 57

2.14 House price-to-rent ratio for theeuro area 57

3.1 Euro area 2015 real bond yield andoil price 64

3.2 Euro area and US yield curves 643.3 Net purchases of long-term bonds

and equities by euro area insurancecorporations and pension funds 68

3.4 Euro area BBB-rated corporatebond spread and the real short-terminterest rate 69

3.5 Dow Jones EURO STOXX totalreturn index 72

3.6 Dow Jones EURO STOXX earningsrevisions ratio 72

3.7 Relative performance betweeneuro area and US stock prices andEUR/USD exchange rate 73

4.1 Profitability and cost-to-incomeratios of euro area banks 74

4.2 Frequency distribution of the returnon equity for euro area banks 74

4.3 Euro area banks’ total provisions 834.4 Frequency distribution of overall

solvency ratios for euro area banks 834.5 Euro area banks’ liquid asset ratios 844.6 Annual growth of euro area MFI

loans to non-financial corporationsand changes in credit standards 90

4.7 Expected default frequency fordifferent euro area industrialsectors 91

4.8 Interest rate VaR for selectedlarge euro area banks 94

4.9 Distribution of hedge fund returns 964.10 Share of hedge funds breaching

triggers of net asset value (NAV)decline 96

4.11 Ratio of the Dow Jones EUROSTOXX bank index to the overallmarket index for the euro area 101

4.12 Risk-neutral probability densityfunction on the Dow Jones EUROSTOXX bank index 101

4.13 Large euro area banks’ earnings pershare (EPS) and 12-month-aheadforecasts 102

4.14 European financial and non-financialinstitutions’ credit default swaps 103

4.15 Western European banking sectorrating downgrades, upgradesand balance 106

5.1 Return on equity in the euro arealife, non-life and reinsurancesectors 108

5.2 Expense, loss and combined ratiosof the euro area non-life insuranceindustry 109

5.3 Frequency distribution of return onequity of euro area life insurancecompanies 109

5.4 Expense, loss and combined ratiosof the euro area non-life reinsuranceindustry 110

5.5 Capital positions in the euro arealife, non-life and reinsurancesectors 111

5.6 Dow Jones EURO STOXX insuranceindex and its implied volatility 118


5.7 Risk-neutral probability densityfunction on the Dow Jones EUROSTOXX insurance index 119

5.8 Cumulative change in the euro areainsurance stock price indicesrelative to the Dow JonesEURO STOXX 119

6.1 Large-value payments processedvia TARGET 122

6.2 Large-value payments processedvia TARGET by country 122

6.3 Volumes and values of foreignexchange trades settled via CLS inUSD billion equivalent 123

TABLE6.1 Euro area CCPs for financial

instruments 127

7ECB


P R E FA C EFinancial system stability requires that theprincipal components of the system – includingfinancial institutions, markets andinfrastructures – are jointly capable ofabsorbing adverse disturbances. It alsorequires that the financial system facilitates asmooth and efficient reallocation of financialresources from savers to investors, thatfinancial risk is assessed and priced reasonablyaccurately, and that risks are efficientlymanaged. By laying foundations for futurevulnerabilities, inefficiencies in the allocationof capital or shortcomings in the pricing of riskcan compromise future financial systemstability. This review assesses the stability ofthe euro area financial system both with regardto the role it plays in facilitating economicprocesses and considering its ability to preventadverse shocks from having inordinatelydisruptive impacts.

The purpose of publishing this review is topromote awareness in the financial industryand among the public at large of issues that arerelevant for safeguarding the stability of theeuro area financial system. By providing anoverview of sources of risk and vulnerability tofinancial stability, the review also seeks to playa role in preventing financial crises.

The analysis contained in this review wasprepared with the close involvement of, andcontribution by, the Banking SupervisionCommittee (BSC). The BSC is a forum forcooperation among the national central banksand supervisory authorities of the EU and theECB.

9ECB


The strength and resilience of the euro areafinancial system has further improved over thepast six months, contributing to a positiveoutlook for financial stability. Nevertheless, atthe same time financial imbalances have grownlarger and seem likely to continue expanding,primarily at the global, but also at the euro area,level. With shock-absorption capacitiesimproving, but risks and vulnerabilities rising,the financial stability outlook continues to restupon a delicate balance. While the likelyoutcomes could, at this stage, best be describedas bi-modal, a positive outcome remains themost likely prospect in the period ahead.

Continued strength in the pace of globaleconomic growth in 2005, despite further oilprice rises, low interest rates in the euro area aswell as indications of further improvements incorporate sector credit quality, provided afavourable environment for financialinstitutions and markets. Global credit marketssuccessfully weathered a test of their resiliencein the first half of the year, prompted by thedowngrading of two large US automobilemanufacturers. The downgrade only resulted ina short-lived disturbance and the pricing ofcredit risks in the euro area financial marketsgenerally remained very favourable. Inaddition, the conditions for raising funds inequity markets remained favourable, andfinancial market volatility stayed very lowacross most asset classes. In this environment,there was further and broad-basedimprovement in the profitability of banks, andthe balance sheets of insurance companies werestrengthened. In addition, key financialinfrastructures – including payments systemssuch as TARGET, and securities clearing andsettlement systems – remained robust andcontinued to operate smoothly.

Within the euro area financial system, the mainsource of vulnerability in the period aheadappears to be associated with concerns that thesearch for yield, which began in 2003, mayhave led investors to underestimate risk,pushing asset prices beyond their intrinsicvalue, especially in fixed income and credit

I O V E RV I EW O F R I S K S TO F I N ANC I A LS TA B I L I T Y

markets. To the extent that very low long-termrisk-free interest rates have driven investors toseek higher expected returns in exchange forcommensurately higher risks in corporate bondmarkets, credit derivatives and collateraliseddebt obligations (CDOs), pricing in thesemarkets could prove vulnerable to anyunexpected upturn in long-term interest rates.Moreover, there is uncertainty about theways in which markets for credit risk transfer(CRT) products would function – includingthe settlement of complex contractualarrangements – in a stress situation. Anycrowding of trades – where many investorsenter into similar strategies – in these marketscould further aggravate their vulnerability.

Concerning the sources of risk andvulnerability outside the euro area financialsystem, large and growing global financialimbalances continue to pose medium-termrisks to the stability of foreign exchange andother financial markets, especially bondmarkets. By raising costs, the further surge inoil prices over the past six months could dentfuture corporate sector profit growth, if itproves to be as lasting as futures pricescurrently suggest. Concerns also remain aboutthe credit and wealth risk implications of risinghousehold sector debt and house prices in someeuro area countries.

It should be stressed that calling attention tosources of risk and vulnerability to financialstability such as these does not mean seeking toidentify the most probable outcome. Rather, itmerely entails highlighting potential andplausible sources of downside risk, even ifthese are relatively remote. The remainder ofthis chapter examines the main sources of riskand vulnerability to euro area financial systemstability. The chapter concludes with an overallassessment of the outlook.

RISKS FROM GLOBAL FINANCIAL IMBALANCES

Because of the pressure they place on globalcapital markets, the magnitude of capital flows


required to finance the large and growing UScurrent account deficit continues to pose risksfor global financial stability. In 2005, the UScurrent account deficit is widely expected toreach more than 6% of US GDP, a new post-Bretton Woods record. The prolongedaccumulation of expanding deficits means thatthe US has become the world’s largestinternational debtor. While there have beenconcerns about the medium-term sustainabilityof the US external position, this ultimatelydepends upon the ability and willingness ofexternal investors in surplus economies tocontinue financing the US deficit. So far, therehas been little indication of any financingchallenges or diminished willingness on thepart of foreign investors to increase theirholdings of US assets.

Overall, sight should not be lost of the fact thatthere are two sides to the significant wideningof global imbalances since 2000. Significantcapital inflows into the US have been mirroredby outflows from surplus economies in Asia,especially China and Japan, and, morerecently, from oil-exporting countries, whichhave benefited from the surge in oil prices. Theforeign exchange reserves accumulated bysome Asian economies have continued to berecycled into the US bond markets. Thisappears to have underpinned a further wideningof US imbalances, thereby delaying anyadjustment, and contributing to holding USlong-term interest rates down.

Looking ahead, several factors suggest thatglobal imbalances could yet widen further.These include the recent strengthening of theUS dollar, the robust pace of US economicactivity, and the recent further rise in oil prices.Within the US, the main sources of domesticsavings-investment imbalances for much of thetime since 2000 have been growing fiscalimbalances and heavy household borrowing.Given the costs associated with HurricaneKatrina, it seems unlikely that the fiscal deficitwill significantly contract in the period ahead.At the same time, the US household sectorsavings rate reached very low levels in recent

months, and it cannot be excluded thatcorporate sector financing surpluses couldsoon turn negative, as firms may be encouragedto increase leverage.

As far as risks from these large and growingglobal imbalances are concerned, theirsustainability appears to be closely connectedwith the ability and willingness of foreigninvestors – both official and private – toaugment their holdings of US dollar assets stillfurther. From a financial stability viewpoint, ifconcerns were to surface that overseas demandfor US dollar assets could slow markedly, thelikelihood of a disorderly rebalancing wouldincrease, involving capital account adjustment.This could bring with it the possibility ofsignificant downward pressure on the USdollar, coupled with significant upwardpressure on long-term interest rates whichcould spill over into other financial assetclasses. A positive development in this respectwas that the decision of the Chinese authoritiesto introduce greater exchange rate flexibility inJuly 2005 did not appear to lead to expectationsthat Asian central banks would diversify thecurrency composition of their reserves out ofUS dollars in the short term. Indeed, the reformdid not have any lasting effect on the majorexchange rates. Nevertheless, if the recentfurther widening of global imbalances is notcorrected over the medium term, importantrisks will remain.

RISKS IN GLOBAL CAPITAL MARKETS

In the course of 2003, long-term bond yields inthe US and the euro area dropped to very lowlevels, and exhibited little discernible trendthereafter. Initially, low short-term interestrates had encouraged “carry trades” – wherefunds are borrowed for the short term andinvested in long-term maturity instruments –along market yield curves, apparentlycontributing to the decline in yields. However,despite widespread fears of a repeat of theturmoil that occurred in global bond markets in1994, the measured and broadly anticipated

11cECB


1 OVERVIEWOF RISKS TO

FINANCIALSTABILITY

increases in the US Federal Funds rate fromJune 2004 onwards only had a relativelylimited impact on pricing in the fixed incomemarkets. Even though the tightening of USmonetary policy seemed to prompt an orderlyunwinding of these leveraged positions, USlong-term nominal interest rates still remainedwell below consensus expectations for nominalGDP growth over the same horizon. Theresilience of pricing in the fixed incomemarkets to changing fundamentals was alsonotable in view of the strength of globaleconomic activity, growing twin (fiscal andcurrent account) – deficits, and institutionalinvestor surveys that persistently revealedconcerns about the possibility of an abruptupturn in long-term bond yields.

Several potential explanations have beenadvanced for the very low level of global long-term interest rates. Foremost among these hasbeen the credibility of monetary policy indelivering low and stable rates of inflation. Inaddition, as far as the US is concerned, anotherfactor has been strong official sector demandfor US assets, particularly in Asia; and highlevels of global saving, especially in emergingmarket economies (EMEs) and in the corporatesectors of mature economies. Turning to theeuro area, an additional factor has been thedemand for fixed income assets by institutionalinvestors eager to close balance sheetmismatches, a portfolio reallocation stimulatedby recent regulatory and accounting changes.In addition, the strength of oil prices is thoughtto have played a role in pushing down globalbond yields, both by lowering global growthexpectations and through a recycling ofexpanding revenues of oil-exporting countriesinto fixed income assets. While it remainsdifficult to gauge the relative importance ofthese influences on the pricing of long-termbonds, a pick-up in external funding by thecorporate sector, for instance, or a move byAsian central banks to diversify the currencycomposition of their foreign external reserveportfolios, could trigger an upturn in long-termyields.

Regardless of the causes, the consequences oflow long-term interest rates have been clear.By flattening market yield curves, low long-term interest rates encouraged financial marketinvestors to seek higher returns withcommensurately higher risk in credit andemerging markets. While improvingfundamentals undoubtedly played an importantrole in compressing spreads in markets forcorporate and emerging market bonds, therehave been some concerns that very low interestrates and abundant liquidity may have ledinvestors to perceive risks as being very lowand/or to accept less compensation for holdingrisky assets. Lower returns on risk-free assetsis thought to have favoured substantial growthin markets for complex and leveragedinstruments such as credit derivatives andCDOs, and in the global hedge fund industry.While these developments can be seen aspositive for market efficiency and liquidity,there are uncertainties about the ways in whichmarkets for credit risk transfer products wouldperform in a stress situation. Any crowding oftrades in these markets, especially by hedgefunds, could aggravate their vulnerability.

Low long-term interest rates and the search foryield also seem to have had consequences fornon-financial sectors and financial institutionsin the euro area. It allowed firms to restructuredebt, thereby improving credit quality, and itappears to have supported the strength ofmortgage lending, both of which have favouredthe financial condition of euro area banks.However, to the extent that the search for yieldhas more broadly pushed asset prices abovetheir intrinsic value, this may have sown theseeds of future vulnerability for financialinstitutions, especially as it implies thatinvestors may have underestimated risks.Hence, financial institutions, including euroarea banks, that hold fixed income and creditsecurities may yet face greater than normalmarket risk.

Looking ahead, not withstanding thesuccessful weathering of a well anticipatedcredit event in the first half of 2005 by credit


markets, there is no room for complacency as itcannot be excluded that a pick-up in corporatebond issuance activity, together with anunexpected upturn in longer-term rates and/or abroad-based reappraisal of credit risks, couldpose strains and widen spreads in corporatebond and credit derivatives markets. Eventhough long-term rates in the euro area haveremained close to nominal growth expectationsand therefore do not appear to represent anindependent source of vulnerability, riskscould still arise through correlation betweenUS and euro area long-term bond yields, whichtends to be very high at times of market stress.The vulnerability of pricing in credit markets toreappraisal was recently demonstrated in earlyMay 2005 when the widely expecteddowngrading of two large US automobilemanufacturers led to some dislocation instructured credit markets and losses for somehedge funds. While the turbulence passedquickly and credit markets proved to beresilient, the incident highlighted thedependence of pricing in credit derivative andCDO markets on relatively untestedassumptions about default correlations.Although no clear signs have yet emerged ascorporate earnings growth has remainedrelatively strong, these markets may yet betested by further credit events similar to thoseseen in May when the credit cycle, which isrelatively mature, begins to show signs ofturning.

EXPOSURES TO EURO AREA NON-FINANCIALSECTORS

An evaluation of the credit risks posed by firmsand households depends upon both the natureof the exposures of banks and financial marketparticipants – including investors in corporatebonds and participants in CRT markets – andbalance sheet conditions in the two sectors.Over the past six months, private sectorbalance sheet conditions in the euro area havecontinued to diverge, with firms continuing tostrengthen their balance sheets, but householdsexpanding their balance sheets further.

A further strengthening in the profitability offirms, together with ongoing debt restructuringefforts, has contributed to improving thecondition of euro area corporate sector balancesheets still further over the past six months.Moreover, in an environment in which interestrates remained very low, the debt financingburdens of firms generally remained contained.Even though corporate debt-to-GDP ratiosrose somewhat in early 2005, there was anacross-the-board improvement in traditionalindicators of corporate sector creditworthiness– including better credit ratings, tight spreadson corporate bonds, lower expected defaultfrequencies, and an overall easing of banks’credit standards on the approval of loans toenterprises in the first half of 2005. Creditassessments improved for both large and smallenterprises, although to a far lesser extent forthe latter.

Notwithstanding the relatively benign outlook,there are some risks and vulnerabilities facingfirms in the period ahead. Uncertaintiessurrounding the broad economic outlook haveincreased somewhat over the past six months,primarily owing to the further rise in oil prices.This has cut into analysts’ expectations offuture corporate sector profit growth. Althoughat this stage there is uncertainty about theprospect of a general turn in the credit cycle,the profit cycle has shown some signs ofmaturing. In particular, while rates of profitgrowth have been strong, in some moredomestically oriented sectors, there was amarked slowdown in the course of 2005. Anunexpected disturbance to future profit growthor increase in leverage would most likelytranslate into higher corporate sector creditrisk. In addition, while balance sheet and debtrestructuring has generally contributed toimproving the creditworthiness of thecorporate sector, there are some concerns thatfirms have shortened the effective maturities oftheir debts, thereby making balance sheetsmore interest rate-sensitive.

As rates of lending growth to households havestrengthened further, euro area household

13cECB



FINANCIALSTABILITY

sector indebtedness has continued to rise,prompting questions about sustainability andthe degree of credit risk that banks face.Ultimately, the sustainability of householdsector indebtedness depends on the ability ofhouseholds to service outstanding obligationsout of income and, if necessary, assets in caseof adverse disturbances to income. In thisrespect it is notable that, while climbing,household debt-to-GDP ratios have remainedlow by international standards; debt servicingburdens have remained stable; and aggregatehousehold sector solvency – gauged bydebt-to-financial asset ratios – has remainedcomfortable. There are also some indicationsthat the most heavily indebted householdsin the euro area tend to be those in thehighest income categories. Furthermore, thecounterpart of rising indebtedness in some euroarea countries has been an expansion of theasset side of household balance sheets, thanksprimarily to further house price appreciation.

Looking ahead, the risks facing the euro areahousehold sector include risks to householdincomes, as well as interest rate risks and, insome countries, house price risks. Concerningrisks for household incomes, there are someuncertainties surrounding the euro areaeconomic outlook which, if they were tocrystallise, could have implications foremployment prospects and householddisposable incomes. As for interest rate risk,the greater part of this is borne by banks ratherthan households, as the bulk of mortgagecontracts in the euro area are agreed at fixedrates or quasi-fixed rates.1 Hence, it is unlikelythat, on aggregate, an unexpected disturbanceto interest rates would diminish the strength ofaggregate household balance sheets to the pointof significantly raising the credit risks faced bybanks across the euro area. Nevertheless, theinterest rate risks facing households are notevenly spread across the euro area, givendifferences across countries in terms of debtlevels, contractual interest rate variability, andtypical loan maturities. As for exposures to therisks of property price reversals, banks appear,by and large, to have carefully managed the

risks to collateral behind mortgages by settingloan-to-value ratios at conservative levels,even though signs of intensifying competitionin mortgage markets may have led to aloosening of credit standards. This means thathouseholds would probably bear the brunt ofany property price reversal. The implicationsfor financial stability would ultimately dependupon the strength of any wealth effect onhousehold consumption. Since some of thecountries that have experienced substantialincreases in house prices in recent years areones where the preponderance of variable ratedebt is high, this may amplify the effect ofany interest rate changes, especially forhouseholds with high levels of outstandingdebt, low housing equity, low financial assetbuffers and/or uncertain employment andincome prospects.

PERFORMANCE OF THE EURO AREA BANKINGSECTOR

Consolidating upon the recovery in euro areabanking sector profitability that began in 2003,there was further broad-based improvement in2004, including among national bankingsectors that had significantly underperformedeuro area averages in previous years.Moreover, the financial results of large euroarea banks suggest that profitabilitystrengthened further in the first half of 2005,although these results should be interpretedcautiously in view of the introduction ofInternational Financial Reporting Standards(IFRS) at the start of the year. All in all, theseimprovements in performance were notable inview of the continued sluggishness in the paceof euro area economic activity. The factorsdriving the improvement in the profitability ofbanks included significant growth in lending tohouseholds, mostly for housing purposes, andan incipient recovery in lending to thecorporate sector, including to small andmedium-sized enterprises (SMEs). Markedreductions in loan loss provisioning also

1 Some mortgage contracts have periods of f ixation that areshorter than the term of the contract.


contributed to this improvement. All in all,given stronger profitability and with solvencyratios remaining broadly unchanged atcomfortable levels, the shock-absorbingcapacity of the euro area banking sector hasimproved.

Looking ahead, the profitability of euro areabanks is expected to improve further,benefiting from generally benign creditand liquidity conditions. However,notwithstanding the recent improvement inprofitability and the generally favourableoutlook, there are some risks andvulnerabilities that the euro area bankingsector could face in the period ahead. Some ofthese vulnerabilities lie within the bankingsystem. Banks have faced challenges overrecent years in maintaining, or increasing,interest income because of margin erosion,driven by persistently low interest rates and thestrength of competition from their peersregarding the granting of loans. Hence, there isa risk that banks may have started to loosentheir credit standards, thereby possiblyincreasing their future exposure to credit risk,or that they have sought out alternative,possibly riskier, sources of income. A furtherrisk is that historically low provisioning forloan losses could prove inadequate in the faceof an unexpected deterioration in the economicoutlook, even though provisioning patterns inmost countries have reflected readjustmentsfrom greater than normal levels, credit risk hasimproved, and coverage ratios have increased.

The most prominent sources of risk andvulnerabilities outside the banking sectorappear to be further oil price increases as wellas the risks posed by large and growing globalimbalances. Although the direct exposures ofeuro area banks to these sources of risk arelikely to be limited, if such risks were tocrystallise, they could manifest themselvesindirectly in increasing credit risk as thefinancial condition of non-financial sectorsdeteriorates. Concerning the interest rate risksfaced by banks, the risk of an abrupt upturn inlong-term bond yields, while not priced into

market yield curves, has continued to be pricedinto options markets as a low probability event.If this risk were to crystallise, banks could beexposed to greater market risks than normal,including the possibility of capital losses onfixed income securities – especially if it were totrigger a reassessment of the appropriateness ofpricing in credit markets. However, banks alsoface risks if long-term interest rates were toremain low for a protracted period, especially ifyield curves were to flatten, since this wouldincrease the challenges facing the sector ingenerating interest income from core business.Moreover, as low interest rates may havesustained tight credit spreads, to the extent thatbanks’ pricing of credit risks is market-sensitive, longer-term vulnerabilities could bebuilding up. These vulnerabilities would beexposed in the event of an upturn in long-terminterest rates or an unexpected credit event.

Forward-looking indicators based on assetprices generally suggest that the outlook for theeuro area banking sector remains bright. Theperformance of banks’ stock prices over thepast six months suggests that marketparticipants expect further improvements inprofitability. This outlook is confirmed byprivate sector analysts’ forecasts of futurebanking sector profitability, as well as bypatterns in banking sector credit riskindicators. Nevertheless, some options market-based indicators point towards increasinguncertainty about the outlook for bankingsector profitability. However, the messageemerging from market-based indicators shouldalways be interpreted with caution owing to thepossibility of pricing misalignments.

PERFORMANCE OF THE EURO AREA INSURANCESECTOR

Profitability in the euro area insurance industryimproved further in 2004, driven mainly bystrong underwriting results. Investmentincome, however, continued to be subdued inan environment where interest rates remainedvery low. Capital bases improved in both the

15cECB



FINANCIALSTABILITY

non-life and reinsurance industries in 2004,whereas indicators of solvency in the lifeinsurance industry remained largelyunchanged. Looking ahead, by late 2005 theoutlook for the euro area insurance sectorseemed broadly favourable, although market-based indicators did not reveal a clear pictureconcerning market participants’ assessment ofthe risks facing the insurance sector, therebycalling for ongoing monitoring andsurveillance.

The main risk facing the insurance industryrelates to uncertainty about the likely futurepath of global and euro area long-term interestrates. The persistently low level of long-terminterest rates has continued to impose strainsboth through the valuation of future liabilitiesand, for some life insurers which have raisedtheir holdings of fixed income assets, byimpinging on investment returns. For thesefirms, a rise in long-term interest rates wouldmost likely lead to significant balance sheetstrengthening. Concerning individual sub-sectors, the prospects for the euro area lifeinsurance industry have been improving.Decisions taken by some Member States tooffload pension funding from fiscal budgets arelikely to promote the pension business of thelife insurance industry. The outlook for thenon-life insurance industry in the euro area alsoappears broadly positive: the risk of asignificant decline in premium prices appearscontained in the period ahead, as investmentincome should remain modest given the currentlow investment return environment. Moreover,premium written should expand further as thepace of economic activity picks up. As for thereinsurance sector, underwriting results for2005 are generally expected to be strong, withpremium prices declining only slowly. In theperiod ahead, owing to important claims fromHurricane Katrina, reinsurance premium pricesmay halt declining and possibly even increaseslightly, depending on the final amount ofcapital depletion in the sector worldwide. Inthe euro area, the losses incurred byreinsurance from the hurricane are expected toonly dent capital positions.

OVERALL ASSESSMENT

Continuing strength in the pace of globaleconomic activity in 2005, together with afurther strengthening of the balance sheets oflarge euro area firms and financial institutions,has contributed positively to the financialstability outlook of the euro area, despite furtheroil price rises. However, several potential risksand vulnerabilities have grown in importance inthe past six months. Within the financial system,despite recent improvements, the durability ofbanking sector profitability could be tested in theperiod ahead, especially if long-term interestrates remain low for a protracted period.Declining loan loss provisioning flows couldalso adversely affect the ability of banks to copewith unforeseen disturbances. Althoughfinancial markets successfully weathered a wellanticipated credit event in the first half of theyear, this should not lead to complacency: thepossibility still exists that a reappraisal couldtake place with regard to far-reaching marketrisks stemming from the aggressive search foryield that began in the course of 2003. This hasleft some financial markets and institutionsvulnerable to changes in global liquidityconditions and unexpected credit events. Adisorderly correction in the level of long-termyields could potentially disrupt theintermediation of funds through global capitalmarkets, which would have implications for theeuro area. Moreover, some euro area financialinstitutions, including banks, would likelyendure losses – at least in the short term – fromany upturn in long-term interest rates. On theother hand, the life insurance industry wouldmost likely benefit as this would help in relievingremaining balance sheet vulnerabilities.

Looking ahead, the risk of an abrupt unwindingof global imbalances remains, especiallybecause these imbalances may yet widenfurther. It also cannot be excluded that furtheroil price increases could test the resilience offirms’ balance sheets, especially those ofSMEs. Household balance sheets may also bevulnerable in countries where house pricesseem to have risen beyond their intrinsic value.

17ECB


I I T H E MA CRO - F I N ANC I A L E NV I RONMENT1 THE EXTERNAL ENVIRONMENT

Although the pace of global economic activityhas remained strong, large and growingfinancial imbalances continue to pose risks forglobal financial stability. There are severalfactors that suggest that global imbalancescould widen further in the period ahead. Theseinclude the recent strengthening of the USdollar, the stronger pace of US growthcompared with its trade partners, and thefurther rise in oil prices. From a financialstability viewpoint, if concerns were to surfacethat overseas demand for US assets could slowmarkedly, this could entail the possibility ofsevere downward pressure on the US dollar,coupled with significant upward pressure onlong-term interest rates that could spill over toother financial asset classes. While foreigninvestors have shown no signs of reluctance toincrease their holdings of US assets, importantrisks will remain if the recent further wideningof global imbalances is not corrected over themedium term.

1.1 RISKS AND FINANCIAL IMBALANCES IN THEEXTERNAL ENVIRONMENT

In the first half of 2005, the pace of globaleconomic activity remained vigorous andproved resilient to further oil price increases.Growth was particularly robust in the US,notwithstanding increasing policy interestrates, as well as in some countries in non-JapanAsia. While the outlook for global growthremains favourable, some risks do remain,including those posed by large and wideningglobal financial imbalances and the furtherincreases in oil prices over the past six months.

US CURRENT ACCOUNT AND FINANCINGThe magnitude of capital flows required tofinance large and growing US current accountdeficits has continued to pose significant risksfor global financial stability. With theprolonged accumulation of these deficits,questions remain about medium-termsustainability, thereby posing a potentialsource of vulnerability for global currency

markets that could also spill over to otherfinancial asset classes. The possiblecrystallisation of these risks mainly depends onthe willingness of foreign investors, bothofficial and private, to increase their holdingsof US securities. As US external imbalancesmay yet expand further – given the recent USdollar appreciation, faster growth in the USthan in many of its major export markets, andrising oil prices – pressure on the internationalcapital markets is likely to remain.

In the first half of 2005, the US current accountdeficit continued to expand. After reachingnew post-Bretton-Woods records of 6.5% ofGDP in the first quarter of 2005, it decreased to6.3% of GDP in the second quarter. For muchof the time since 2000, the main domesticsources of the US current account deficit havebeen growing fiscal imbalances and heavyhousehold sector borrowing. However, ratheratypically, the corporate sector – includingboth financial and non-financial firms – hasenjoyed a financial surplus since the lastquarter of 2001 (see Chart 1.1).

Notwithstanding the further widening of theUS current account deficit, there has been no

Chart 1.1 Net lending/borrowing of theUS economy

(Q1 1973 - Q2 2005, % of GDP)

Source: US Bureau of Economic Analysis.Note: Net lending/borrowing equals gross saving (net savingplus consumption of f ixed capital) minus gross investment.The contributions of the three domestic sectors do not add upto the total owing to capital account transactions andstatistical discrepancies.

-8

-6

-4

-2

0

2

4

6

-8

-6

-4

-2

0

2

4

6

1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003

householdcorporategovernmentnet lending/borrowing


Chart 1.2 US non-farm, non-f inancialcorporate sector f inancing gap

(Q1 1980 - Q2 2005, % of GDP)

Source: US Federal Reserve Board.Note: The financing gap equals capital expenditures lessinternal funds and inventory valuation adjustments.

-1

0

1

2

3

4

-1

0

1

2

3

4

1980 1984 1988 1992 1996 2000 2004

evidence of financing difficulties. In thesecond quarter, US-owned assets abroad,which amounted to USD 250.8 billion, weremore than sufficient to finance the deficit ofUSD 195.7 billion. While foreign officialinflows continued to diminish in importance in2005 in overall portfolio investment, lookingahead, at least in the near term, there are fewsigns of diminishing willingness on the part offoreign investors to increase their holdings ofUS securities.

US CORPORATE SECTOR BALANCESThe strength of US corporate sector balancesheets can have a bearing on euro area financialstability, as many euro area financialinstitutions have direct exposures to US firmsthrough lending. Furthermore, the financingneeds of US corporations may affect the costsfaced by large euro area firms in global capitalmarkets, both through competing demands forfunds as well as in the global pricing ofcorporate sector credit and equity market risks.

Against a background of robust economicgrowth, US corporate sector profitabilitystrengthened further in the first half of 2005,and especially in the non-farm, non-financialcorporate sector, which represents the bulk ofUS corporate activity. Compared with thefourth quarter of 2001, when US non-farm,non-financial corporate profits (after corporateincome taxes and after inventory valuation andcapital consumption adjustments) reached atrough, by the second quarter of 2005 profitshad increased by more than 40%. Furthermore,in the first six months of 2005, corporate sectorcash flows also strengthened as dividendpayout ratios fell back from the very high levelsobserved in the final quarter of 2004 – a risewhich was mostly explained by a one-offdividend payout made by a large informationtechnology firm. Indeed, the improvement ininternally generated funds, in conjunction witha deceleration in fixed investment spending,led the non-farm, non-financial corporatesector to a financial surplus in the secondquarter of 2005 (see Chart 1.2). In the last fewyears, the non-farm, non-financial corporate

sector appears to have used an increasinglylarge percentage of internal funds to repurchaseshares from the marketplace, this phenomenonbecoming marked over the last twelve months,and in particular in the first half of 2005, whenthe issuance of new shares was largelysurpassed by share retirements (see Chart S2and Box 1). Some companies financed, at leastin part, share buybacks with external liquidity,and the growing activity in cash-financedmergers and acquisitions (M&As) led toadditional demand for funds. However, onaggregate, non-farm, non-financial corporatesector credit market debt as a proportion ofGDP has remained virtually unchanged sinceearly 2004 (see Chart S1).

Regarding the composition and maturityprofile of the debt of the US non-farm, non-financial corporate sector, there wereindications that firms increased their short-term liabilities after the second quarter of 2004,while the rise in long-term credit marketliabilities remained relatively stable. This wasdue to a shift in the composition of new creditfrom corporate bonds to bank loans, probablypartly related to an easing of credit standardson commercial and industrial loans, as reportedin the July 2005 Federal Reserve Senior LoanOfficer Opinion Survey on Bank LendingPractices. This may also have affectedcommercial paper issuance which, after

19cECB


II THE MACRO-FINANCIAL

ENVIRONMENTpicking up in the first quarter of 2005 against abackground of significant improvement in thecredit ratings of large US firms, decelerated in thesecond quarter. Notwithstanding indications ofimproved access to credit markets, the shifttowards short-term liabilities exposes US firms togreater interest rate risk. Nevertheless, short-term indebtedness remained low compared withthe long-term debt (see Chart 1.3).

Regarding the asset side of US corporate sectorbalance sheets, the market value of theoutstanding assets of the non-farm non-financial corporate sector edged up in mid-2005 by 6.7% compared with the secondquarter of 2004. This was mostly due to anincrease in the market value of tangible assets.

Chart 1.3 Credit market l iabi l it ies of USnon-farm non-f inancial corporations

(Q1 1990 - Q2 2005, USD billions)

Source: US Federal Reserve Board.

800

1,000

1,200

1,400

1,600

1,800

2,000

1990 1992 1994 1996 1998 2000 2002 20041,000

1,500

2,000

2,500

3,000

3,500

4,000

short-term liabilities (left-hand scale)long-term liabilities (right-hand scale)

Box 1

THE RECENT SURGE IN US SHARE BUYBACKS: CAUSES AND POSSIBLE FINANCIAL STABILITYIMPLICATIONS

From the second half of 2004 onwards, US corporations retired an extraordinary volume ofequity from the market. Throughout 2004 and the first half of 2005, equity retirements in theUS non-financial corporate sector exceeded gross equity issuance by USD 1,149 billion. In thefirst half of 2005, share repurchases by public companies listed in the S&P500 index reached ahistorical record (see Chart B1.1). This Box discusses some of the causes behind the surge inequity repurchases by US firms, and highlights some of the possible financial stabilityimplications of this.

Companies decide to buy back their shares for a number of reasons. Managers may believe thattheir best investment option is in the company itself; they may think that the company’s sharesare undervalued, as a reduction in the number of shares outstanding raises expected earningsper share (EPS), thereby possibly boosting share prices; and finally, they may fear hostiletakeovers. As buybacks imply a distribution of profits to shareholders, US companies haverepurchased their shares in addition to, or as an alternative to, distributing dividends.Repurchased shares may be either retired, or they may be held within the company. In this case,they may be reissued for mergers and acquisitions, or to meet employee stock options (ESOs)and benefit plan obligations.

Standard and Poor’s has largely attributed the recent pick-up in share repurchases to anincrease in exercised ESOs and to US companies’ desire to reduce their share countoutstanding. Among S&P500 companies, buybacks occurred in conjunction with the greatestnumber of dividend increases since 1998, although on the other hand, in the US corporatesector as a whole, profits have risen at a faster pace than dividends, with the share of netdividends in after-tax profits gradually declining in recent quarters (see Chart B1.2).


Chart B1.1 Share buybacks by S&P 500corporat ions

(USD billions)

Source: Standard and Poor’s.

25

35

45

55

65

75

85

25

35

45

55

65

75

85

1998 1999 2000 2001 2002 2003 2004 2005

Chart B1.2 Net dividends paid out by UScorporat ions

(% of profits after tax)

Source: US Bureau of Economic Analysis.Note: Prof its include inventory valuation and capitalconsumption adjustments.

30

35

40

45

50

55

60

65

70

75

30

35

40

45

50

55

60

65

70

75

1970 1973 1976 1979 1991 1994 1997 2000 200319851982 1988

Starting in the mid-1990s, the use of non-qualified ESOs to compensate labour has becomeincreasingly common in the US. According to the US National Center for EmployeeOwnership, in 2003 16.3% of companies granted stock options to at least 50% of theiremployees. By early 2005, there were an estimated 10 million non-qualified stock optionholders, with ESO plans valued at several hundred billion US dollars.

A non-qualified ESO gives the employee the right, but not the obligation, to purchase acompany share at a set strike price – which typically coincides with the market price of theshare on the day the option is granted – over a specific time after an initial vesting period.Companies award their employees stock options as part of their labour compensation. Optionsare often seen as a mechanism for increasing employee motivation and retention, as well as away of better aligning the incentives of employees with those of the shareholders. Furthermore,options provide fiscal benefits and, until recently, accounting-related advantages.

The granting of non-qualified ESOs has led to corporate income tax savings, as the differencebetween share current market and strike prices when employees exercise the options isdeducted from corporate income tax. Such deductions have been extremely large for firmswhich have made intensive use of ESOs, and whose stock prices have risen. Sullivan (2002)1

estimates that US corporate tax savings from the deduction of stock options totalled aroundUSD 56 billion in 2000, when options tax deductions exceeded net income for eight of the40 highest market-capitalised US companies. Graham et al. (2004)2 show that in 2000, stockoption deductions reduced the median marginal tax rate of firms within the Nasdaq from 31% to5%, and argue that the tax benefits associated with ESOs may help explain the recentdownward trend in debt issuance by US corporations, since options tax savings may outpacethe value of interest rate deductions.

Until recently, US companies were allowed to decide whether to subtract the value ofoutstanding ESOs from their income statement reported to the Securities and ExchangeCommission – thereby reducing reported profits – or simply to note the expense in a footnote.

1 M. Sullivan (2002), “Stock Options Take $50 Billion Bite out of Corporate Taxes”, Tax Notes, 18 March, pp. 1396-1401.2 J. R. Graham, M. H. Lang and D. A. Shackelford (2004), “Employee Stock Options, Corporate Taxes, and Debt Policy”, Journal of

Finance, Vol. LIX, No 4.

21cECB



ENVIRONMENTAs a consequence, the reported profits (and EPS) of companies not expensing their optionswould turn out higher than the actual ones, possibly increasing the risk of asset pricemisalignments. Starting from June 2005, the revised Financial Accounting Standards Board(FASB) Statement No 123 requires all US public companies to recognise the value of ESOs(estimated by a single fair-value-based method – option-pricing models –) on the date of issue,and to expense it through the vesting period. In such a way, all US public companies will beforced to treat stock options uniformly, enhancing transparency and comparability of profits.Standard and Poor’s estimate that the expensing of stock options would have reduced reportedEPS of the companies within the S&P500 by 21.5% in 2001 and by 19% in 2002. Thereafter,following the episodes of corporate malfeasance in 2002, some firms began to expense options.Hence the underlying reported EPS of S&P 500 would have fallen by a smaller amount of 8.6%in 2003 and 7.4% in 2004.

All in all, it appears that by increasing the award of stock options to pay for labour services, UScompanies appear to have gained some control over their reported corporate performance.Specifically, to the extent that companies have significantly improved their announcedprofitability, ESOs may have altered the efficient functioning of financial markets, andquestioned the reliability of EPS as an indicator of corporate performance. Looking ahead, thenew accounting standards may temporarily lead to smaller returns for investors should theylead to a fall in reported profits and EPS. Furthermore, should the leveraged buybacks thatsome companies are engaged in trigger a fall in the credit ratings of their debt, bond investorsmay incur unexpected losses.

Notwithstanding the general strength of the USeconomy, coupled with increased profitability,favourable financing conditions, balance sheetrestructuring and improved solvency (as

indicated by a further decline in the liability-to-financial assets ratio (see Chart S1)), thefrequency of credit rating downgrades of UScorporations has remained higher than that ofupgrades (see Chart 1.4).

All in all, the strengthening of corporate sectorbalance sheets in the aftermath of the burstingof the equity price bubble has meant that thecredit risks posed by the US corporate sectorhave improved considerably since early 2003.However, some risks do lie ahead. The upturnin short-term rates since mid-2004 can beexpected to cut into US corporate sectorprofitability, especially given indications thatUS corporate sector balance sheets areincreasingly sensitive to short-term interestrates. Recent oil price rises also pose a risk forfirms’ cost bases, especially if these increasesprove to be lasting. Sectors particularly at riskinclude the already troubled energy-intensiveand energy-sensitive industries such as theairline and the automobile industry.

Chart 1.4 US corporate sector rat ingdowngrades, upgrades and balance

(Jan. 1987 - Sep. 2005, 12-month moving average, number)

Source: Moody’s.

-50

-40

-30

-20

-10

0

10

20

30

-50

-40

-30

-20

-10

0

10

20

30

1987 1989 1991 1993 1995 1997 1999 2001 2003 2005

downgradesupgradesbalance


US HOUSEHOLD BALANCESRising US household sector indebtednesscould, if it proves unsustainable, pose risks andcreate vulnerabilities for euro area financialstability. This is because some euro area banksare directly exposed to this sector, or to UScredit institutions exposed to US households.Exposures also arise through holdings ofmortgage-backed securities issued by UScredit institutions.

Continued strength in the demand of the UShousehold sector for credit raised debt-to-disposable income ratios to new heights in thesecond quarter of 2005. Mortgage creditremained the predominant source of risinghousehold sector indebtedness, and alsoaffected consumer credit, the growth of whichhas recently slowed owing to significant homeequity withdrawal (see Chart S3).

Despite the rise in US household sectorindebtedness, there have been few signs offinancing difficulties. According to the July2005 Federal Reserve Senior Loan OfficerOpinion Survey on Bank Lending Practices,lending standards and terms on residentialmortgages and other consumer loans remainedessentially unchanged from the April survey.Meanwhile, the household debt service ratio(DSR), although at a high level relative to theearly 1990s, has remained broadly stable sinceearly 2002. The wider financial obligationsratio (FOR)1 has also remained relatively stableover the same period (see Chart S4). At thesame time, delinquency rates on credit carddebt and auto loans have continued to decline.

The bulk of outstanding US mortgage debt iscontracted at relatively low fixed interest ratesfollowing unparalleled mortgage refinancingin 2003, and is thus sheltered from interest rateincreases. While the share of adjustable ratemortgages (ARMs) in new mortgages hadbegun to rise steadily after mid-2003, through2005 it declined steadily from March to earlySeptember, to rebound thereafter reachingaround 45% of the total loan dollar volume byend October (see Chart S5). An important

factor driving these changes was the evolutionof the spread between interest rates on fixedand adjustable mortgage contracts; it narroweduntil September and has widened since then.The significant proportion of ARMs in totalnew mortgages raised some concerns, inparticular about the rising popularity ofinterest-only (I-O) mortgage contracts. Thesemortgages require very low monthly paymentsduring the first years of the loan, because theprincipal is being repaid over the last few yearsof the contract. Whereas in 2001 only about10% of new mortgages were contracted in thisform, a significant fraction of all newmortgages extended in 2005 were thought to beinterest-only. This may imply rising creditrisks for mortgage lenders, as interest-onlymortgages are often extended to lower-incomehouseholds and because they facilitate greaterleverage from borrowers (who are thereforeexposed to more interest rate and house pricerisk). Furthermore, because they are relativelynew products, the pricing of the credit riskembedded in them is challenged by the lack ofsufficient data histories to conduct stress tests.2

US home prices rose on average by 13.4%between mid-2004 and mid-2005, the largestyear-on-year increase since mid-1979. As thisrise outstripped the growth in personaldisposable income over the same period, theincome-to-house price ratio – a measure ofaffordability – continued to decline (seeChart 1.5).

The main contributor to rising household sectorasset valuations in recent quarters has beenincreasing home values. Moreover, despitefurther rises in household sector indebtedness,coupled with very low household sector

1 The DSR is an estimate of the ratio of debt payments todisposable personal income. Debt payments consist of theestimated required payments on outstanding mortgage andconsumer debt. The FOR, a broader measure, adds automobilelease payments, rental payments on tenant-occupied property,homeowners’ insurance and property tax payments to the DSR.

2 See Real Estate @ Wharton (2005), “Could Risky MortgageLending Practices Prick the Housing Bubble?”, University ofPennsylvania, September.

23cECB



ENVIRONMENTChart 1.5 US housing af fordabi l ity

(Q1 1975 - Q2 2005, index: Q1 1980 = 100)

Sources: OFHEO and US Bureau of Economic Analysis.

90

95

100

105

110

115

120

125

130

90

95

100

105

110

115

120

125

130

1975 1980 1985 1990 1995 2000 2005

Chart 1.6 US household net worth

(Q1 1975 - Q2 2005, % of disposable income)


400

450

500

550

600

650

400

450

500

550

600

650

1975 1980 1985 1990 1995 2000 2005

savings (see Box 2), surging house pricescontinued to produce sizeable gains inhousehold sector net worth in the first half of2005 (see Chart 1.6).

Overall, US households appear to face risks onboth sides of their balance sheets. The strengthof house price inflation – which has continuedto outstrip growth in rents and disposableincomes – poses questions about UShouseholds’ vulnerability to rising interestrates. On the liability side, the sustainability ofincreasing household sector indebtednesscould also be tested by rising short-terminterest rates. However, there are severalmitigating factors, including furtherimprovements in household sector net worth,stable debt servicing ratios, and very low ratesof unemployment.

US FISCAL IMBALANCESAgainst a background of wide current accountimbalances and very low real interest rates,large US budget deficits may pose additionalrisks for global financial stability. By raisingthe financing needs of the public sector, thestrength of federal bond issuance poses notonly the risk of crowding out US private sectordebt issuance, but also of pressuring global realinterest rates.

In 2004, the US general government deficitwas 4.0% of GDP, down from 4.6% in 2003.3

Hence, the debt-to-GDP ratio continued risingto 62.5% in Q2 2005, up from 57.7% at the endof 2001.

In 2005 there was a marked improvement in USfederal finances which resulted from a pick upin revenues thanks to robust economicconditions, although federal spendingcontinued to rise at a brisk pace, mainly onaccount of a sharp acceleration in net outlaysfor interest on the public debt.

Looking ahead, the US administration’s budgetproposals for fiscal year 2006 aim at reducingthe deficit to 1.5% of GDP by 2009. If theserather optimistic expectations are met, thisshould contribute to easing US imbalancesmore generally. However, in the short termthere are risks for the fiscal outlook resultingfrom important administration priorities,including the recent relief bills associated withHurricane Katrina and future spending onmilitary operations in Iraq and Afghanistan.More generally, on account of the growingstructural nature of the US fiscal deficit, thefiscal outlook is likely to remain an ongoingsource of concern.

3 By contrast, the US federal def icit deteriorated further in 2004.It reached 3.5% of GDP, which was the largest in eleven years.


Box 2

FALLING SAVINGS AND RISING DEBT IN THE HOUSEHOLD SECTOR: A FINANCIAL STABILITY RISK?

The personal savings rate in the United States has declined steadily over recent decades and fellto a negative value in the third quarter of 2005 (see Chart B2.1). At the same time, householdsector indebtedness has risen to historically unprecedented heights (see Chart B2.2).Somewhat similar patterns have also been observed in other mature economies, such asAustralia and Canada. These developments have occurred against a background of risinghousehold sector net worth, an important part of which has been due to valuation gains onwealth holdings. This Box discusses some of the financial stability risks that could arise froman increased dependence on asset valuation along with any associated increase in leverage ofhousehold balance sheets.

1 In this sense, household wealth building can be broken down into an active component (the national accounts notion of the savingratio) and a passive one (capital gains on existing assets). See F. Juster, J. Lupton, J. Smith and F. Stafford (2004), “The Decline inHousehold Saving and the Wealth Effect”, Federal Reserve Board of Governors Working Paper, April.

Chart B2.2 US household l iabi l it ies andassets

(% of disposable income)

Source: US Federal Reserve Board.Note: Household sector asset and liability holdings refers tohouseholds and non-prof it organisations and, for liabilities,also incorporates personal trusts.

400440480520560600640680720760800

1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 200430405060708090100110120130

assets (left-hand scale)liabilities (right-hand scale)

Chart B2.1 US household saving and networth

(% of disposable income)

Sources: US Federal Reserve Board and Bureau of EconomicAnalysis.Note: Household sector net worth refers to households andnon-prof it organisations.

-2

0

2

4

6

8

10

12

14

1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004

340

380

420

460

500

540

580

620

660

personal saving (left-hand scale)household net worth (inverted, right-hand scale)

Rising asset valuations may have contributed to the observed drop in savings throughhouseholds’ decisions related to lifetime wealth building. While the standard measure of thehousehold sector saving rate as reported in the national accounts has fallen considerably overthe past decade, it does not take into account changes in the market valuation of existinghousehold assets. Market valuation has been influenced by higher than historical averagereturns on asset holdings over this period and in this way may have supported households’expectations of continued strong wealth valuation on their financial and non-financial wealthand hence weighed negatively on savings.1

Rising valuations on existing household asset holdings may also have favoured the observedstrong rise in indebtedness through financial innovation and the associated availability ofcredit. This factor may be especially important in residential housing markets, whereby a surge

25cECB



ENVIRONMENTin leverage has occurred owing to an expansion of homeownership – also to households whichmay have been credit-constrained in the past – along with enhanced access to equity built up inhousing. This higher degree of leverage implies an increased sensitivity of household balancesheets to changes in asset prices.

An increased asset dependence and any associated leverage imply financial stability risksthrough both a direct channel (increasing credit risk for the financial sector) and an indirectchannel (spillover effects from the broader macroeconomy). Concerning the direct channel,increased credit risk may arise directly given the risk inherent in leveraged acquisition ofassets. In this respect, asset values can vary considerably whilst the nominal value of debt isfixed – and although household debt expansion has been roughly matched by a similarexpansion in household assets, household debt has historically displayed significantly lessvolatility than household assets (see Chart B2.2). Any strains associated with a rapiddeterioration in the value of debt collateral could be manifested in increasing default rates inthe household sector, including strategic defaults. This could entail potential spillover effectsto financial sector balance sheets through their exposures to household sector defaults.Concerning the indirect channel, such a spillover effect could be amplified by a deterioratingmacroeconomic environment that places strains on households’ ability to repay accumulateddebt, possibly due higher unemployment rates or arising from a curtailing of householdspending in favour of increased personal saving. In those economies where wealth valuationeffects have increasingly been used as a substitute for personal saving in lifetime wealthbuilding, a marked fall in asset values has the potential to trigger a compensatory increase inpersonal saving – implying a slowdown in household consumption – if households also revisedownwards expectations regarding future returns on asset holdings. This in turn could have asignificant effect on the economy, given the importance of household consumption in nationalincome, thereby also possibly adding to any strain on financial sector balance sheets.

RISKS IN NON-EURO AREA EU MEMBER STATESThe general economic outlook for non-euroarea EU Member States has remainedfavourable in the six months following the June2005 Financial Stability Review (FSR).

Since the beginning of the 2005, the pace ofeconomic growth in the United Kingdomweakened, largely due to subdued domesticdemand, and the Bank of England lowered itspolicy rate. A slowdown in privateconsumption growth was mainly driven by theincreased cost of debt servicing, following thefour interest rate hikes in 2004, and a notabledeceleration in house price inflation. Againstthis background, mortgage lending growthappears to have stabilised, while unsecuredlending growth has slowed down. Whilecorporate sector capital gearing – the ratio of

firms’ indebtedness to the market value of theirassets – remained high, it has also edged downand profitability remains strong. However, thenear-term outlook for economic growth hasrecently deteriorated, with the balance of riskstilted towards the downside, and there havebeen some signs of labour market easing.

In Sweden, quarterly real GDP growthaccelerated in the second quarter of 2005 after aweak first quarter. In June the SwedishRiksbank cut its policy rate – for the eighthtime since November 2002 – to 1.5%. InDenmark, quarterly output growth alsorebounded substantially in the second quarterof 2005, after having remained subdued in thefirst quarter. In both countries, economicactivity was mainly supported by a recovery inprivate consumption and investment.


The bulk of the banks operating in the other tennon-euro area EU Member States4 are owned byeuro area banks. These banks are thereforeexposed to credit risks arising from theirlending to the household and corporate sectorsin the new Member States.

In these countries, strong output growth – withthe exception of Poland – and prospects ofstrong equity returns have attracted capitalflows from abroad. For instance, since thebeginning of 2005 stock market returns in allthese countries except Poland and Sloveniahave significantly outperformed the Dow JonesEURO STOXX Index.5 These inflows, coupledwith growing foreign currency indebtedness,have fuelled upward pressures on thecurrencies in some of the countries withflexible exchange rates.6

The proportion of foreign currency-denominated loans is important in most of thenew non-euro area EU Member States (seeChart 1.7). While banks manage their directexposures by matching foreign currency assetsand liabilities, they still have indirectexposures, the importance of which depends onthe ability of households and corporations tomanage their foreign currency exposures.Furthermore, in those countries that witnessedbrisk credit growth, the share of non-

performing loans in total loans ceased todecline, although it still remained rather low(see Box 3).

4 The new Member States comprise the ten countries that joinedthe EU on 1 May 2004: the Czech Republic, Estonia, Cyprus,Latvia, Lithuania, Hungary, Malta, Poland, Slovenia andSlovakia.

5 In Estonia, Lithuania and Slovakia, stock exchange indicesincreased by around 50%, 70% and 40% respectively since thebeginning of 2005. In the Czech Republic, Hungary and Latvia,stock prices registered signif icant gains of around 35% in thesame period.

6 The Czech koruna and the Polish zloty appreciated since thebeginning of 2005 by around 2.5%. The Hungarian forint and theSlovak koruna, however, depreciated by 2.0% and 1%respectively in the same period.

Box 3

CREDIT DEVELOPMENTS IN THE NEW NON-EURO AREA EU MEMBER STATES

Macroeconomic stabilisation and banking sector restructuring are two of the factors that havefostered financial sector development in several of the new EU Member States. Regulatoryreforms, leading to increased competition and supply of new products have, together withimprovements in domestic legal systems, also supported dynamic credit activity in thesecountries. In addition, favourable financing conditions, supported by low-inflation policiesand higher incomes and income expectations, have encouraged strong credit demand, up fromrelatively low levels. This Box discusses some of the financial stability implications that couldarise from the strength of credit growth in these countries.

Over recent years, credit growth to the private sector has been very robust in most new EUMember States (see Table B3.1). This growth seems to have been driven primarily by

Chart 1.7 Share of foreign currency loans intotal loans in the new EU Member States

(2002 - 2004, % of total loans)

Source: National central banks.

0

10

20

30

40

50

60

70

80

90

0

10

20

30

40

50

60

70

80

90

EE LV LT HU SI CY PL SK CZ MT

200220032004

27cECB



ENVIRONMENT

household lending, largely in the form of mortgage lending. Annual growth in housing loans inSeptember 2005 exceeded 30% in six new EU Member States. Mortgage lending wasparticularly buoyant in the Baltic countries, where growth rates above 70% were recorded. Incomparison, the growth of loans to non-financial corporations has been more moderate in mostcountries, albeit faster than in the euro area. In many countries, the share of foreign currency-denominated loans in lending has been high and continues to grow. These loans are mostlyeuro-denominated, although other currencies have also been gaining in importance recently.They have been typically granted to the non-financial corporate sector, although in the Balticcountries and Poland, to households as well. Borrowing in foreign currency is mostlyassociated with lower borrowing costs. In addition, the bulk of borrowing in foreign currency isgenerally undertaken by larger multinational firms, which generate the greater part of theirrevenues in foreign currency, and can therefore be seen as a natural hedge. In general,borrowing in foreign currency has been more typical in those countries with fixed exchangerate regimes or exchange rate targets (particularly the Baltic countries), where borrowing hadalready picked up as early as the mid-1990s.

From a financial stability viewpoint, rapid credit growth deserves careful monitoring. This isbecause many banking crises have been preceded by episodes of rapid or excessive creditgrowth, although the opposite has not always held.1 Several theoretical explanations exist forwhy credit booms tend to be associated with a higher probability of banking distress.According to one main strand of the literature, this relationship may be attributed to the pro-cyclicality of bank lending behaviour. Risks may be underestimated during expansionaryphases of the business cycle, thereby resulting in loosening credit standards and a loweraverage quality of borrowers. This may lead to higher credit losses when the next economicdownturn occurs. Another often cited theory related to the over-expansion of credit is the“financial accelerator” mechanism. Over-optimism about future returns could boost assetvaluations and thus firms’ net worth, which then feeds back into higher investment and creditdemand and a further increase in asset prices. Consequently, this self-reinforcing mechanism

1 On the first point, see, for instance, D. Ottens and E. Lambregts (2005), “Credit Booms in Emerging Market Economies: A Recipefor Banking Crises?”, DNB Working Paper No 46, De Nederlandsche Bank, June. On the latter point, see, for instance, A. Tornell andF. Westermann (2002), “Boom-bust Cycles: Facts and Explanation”, IMF Staff Papers, 49 (Special Issue).

Table B3.1 Credit to the private sector

(% per annum; period average)

2002 2003 2004 Q4 2004 Q1 2005 Q2 2005 July 05 Aug. 05

Czech Republic . . 2.0 11.3 13.0 13.8 15.1 17.0 18.0Estonia 23.3 26.0 27.5 30.8 31.2 32.2 34.3 35.4Cyprus 7.7 6.3 5.0 5.8 5.2 4.0 4.2 4.4Latvia 43.6 39.3 43.1 46.8 48.7 50.3 54.9 55.5Lithuania 27.4 43.1 51.1 42.4 37.1 37.6 42.7 44.4Hungary 21.5 24.2 24.7 18.5 18.2 16.8 15.1 14.5Malta 1.7 4.9 12.6 9.2 7.8 6.1 4.2 4.2Poland 4.6 5.4 5.2 5.0 3.5 6.0 6.5 6.1Slovenia 13.4 13.3 18.8 19.3 22.9 24.3 22.9 23.7Slovakia -1.6 8.3 9.7 8.4 10.6 14.9 19.0 19.5

EU-101) . . 11.2 13.8 12.7 12.4 13.8 14.6 14.7

euro area 4.9 4.9 5.9 7.0 7.8 9.0 8.6 8.9

Sources: ECB and national central banks.1) The EU-10 aggregate comprises the ten new EU Member States, 2003 GDP weights. EU-10 aggregate data are not included for2002 due to a structural break in the data series of the Czech Republic.


2 See, for instance, C. Cottarelli, G. Dell’Ariccia and I. Vladkova-Hollar (2003), “Early Birds, Late Risers, and Sleeping Beauties:Bank Credit Growth to the Private Sector in Central and Eastern Europe and in the Balkans”, IMF Working Paper No 03/213.

3 P. Hilbers, C. Pazarbasioglu, G. Johnsen and I. Ötker (2005), “Assessing and Managing Rapid Credit Growth and the Role ofSupervisory and Prudential Policies”, IMF Working Paper No 05/151.

4 See ECB (2004), “Aggregated EU Household Indebtedness: Financial Stability Implications”, Financial Stability Review,December, pp. 147-153.

may lead to excesses in the growth of credit and asset prices. A change in expectations couldthen precipitate a reverse process with falling asset prices and a credit crunch, which maysignificantly increase repayment difficulties for borrowers and may ultimately lead to higherloan losses for banks.

When assessing the nature of credit growth in the specific case of the new EU Member States,it is important to recall that the initial depth of financial intermediation was low in thesecountries compared to their level of economic development. Some studies concluded thatcredit-to-GDP ratios in the central and eastern European and Baltic countries weresignificantly lower than what could be justified by their fundamentals.2 Thus, rapid creditgrowth can, to a considerable extent, be attributed to a catching-up effect. There is as yet onlylimited empirical evidence on whether credit growth in the region has been excessive or not.According to recent empirical findings, there may be, at best, only a small sub-group of new EUMember States in which rapid credit expansion might have reached the proportions of a lendingboom.3 However, even in those countries, the pace of credit growth is not out of line with thatexperienced in former “converging” countries, such as Ireland or Portugal.

It is also important to bear in mind that the pace of credit growth alone may not be a sufficientguide to assess its riskiness. Considering other important aspects of credit growth in the newMember States, there are concerns that rapid credit growth expansion may have beenaccompanied by the build-up of some vulnerabilities, of which four main concerns can beidentified. First, this may put a strain on banks’ ability to monitor and assess risks, especiallybecause risk assessment by banks in these countries is burdened by measurement difficulties inforecasting future credit losses owing to the lack of sufficiently long credit histories. Thisproblem may be even more pronounced in the case of previously under-serviced customers ofbanks such as households and SMEs. Second, foreign currency lending to the domestic privatesector has been strong in some countries. This may have also contributed to increasing thevulnerability of households or unhedged non-financial corporations to unexpectedly largeadverse exchange rate movements. Third, due to relatively moderate growth in domesticdeposits, banks in some countries increasingly rely on foreign interbank borrowing to financecredit growth. This may have left these countries more vulnerable to potential changes in thecurrent favourable external financing conditions. Fourth, since in several new Member Statesmortgage loan contracts typically have floating interest rates, this implies that rising interestrates would weigh mostly on households’ debt servicing ability in those countries.4

There are, however, some mitigating factors. Credit risk is contained by the fact that the ratio ofdebt servicing burdens of firms and households relative to income remains considerably lowerin these countries than in the euro area. Moreover, a favourable growth outlook and improvingincome prospects owing to continuing real convergence as well as a low interest rateenvironment are likely to support the debt servicing ability of private sector borrowers. Somecomfort can also be drawn from the fact that a strengthening of banks’ profitability in most newMember States has helped to maintain a solid capital base, thereby helping to increase banks’shock absorbing capacity. With regard to asset quality, the ratio of non-performing loans

29cECB



ENVIRONMENT

SOURCES OF RISKS AND VULNERABILITY INEMERGING MARKETS ECONOMIESAn evaluation of the risks posed by emergingmarket economies (EMEs) for euro area banksand investors in these countries’ capital marketsdepends upon both the nature of their exposuresand financial imbalances within and acrossEMEs. Risks to euro area financial stabilitystemming from EMEs appear to be contained inthe short term on two grounds. First, the overalleconomic outlook for EMEs as a group remainspositive. Although economic activity acrossEMEs showed signs of moderation in 2005 –except in China – the pace of growth remainedrobust. Coupled with a still favourable externalenvironment, this led to a further improvementin standard EME vulnerability indicators (seeTable S1). Second, the emerging market sharein the total consolidated exposures of euro areabanks – the most direct channel through whichthe euro area banking system may be affected –remained stable and limited at under 25% in Q12005 (or 17.0% excluding offshore centres, seeTable S3).

However, although limited in nature, risksfrom EMEs appear to be more skewed towardsthe downside since the June 2005 FSR,especially in the short to medium term. Threemain risks to the emerging market outlook maybe identified. First, in the short term, EMEsremain vulnerable to sudden shifts in globalliquidity. EME bond spreads have benefitedfrom the hunt for yield in global financialmarkets and, if global long-term governmentbond yields were to rise, they would probablyface a tightening of financing conditions (seeChart 1.8). So far EME bond spreads haveproven resilient to a slight rise in US Treasury

yields. However, they could be tested in theevent of a sharper and more pronounced rise.

Second, inter-regional and intra-regionaldifferences in economic performance acrossEMEs are likely to be accentuated by risinginternational energy prices amid a less dynamicevolution of hitherto compensating factors fora higher energy-related bill (such as non-fuelcommodity exports for some EMEs). And third,external demand could be more sluggish thancurrently anticipated.

Region or country-specific risks also remain amatter of concern. Among the emerging regionsneighbouring the euro area, these risks include acredit boom in south-eastern Europeaneconomies, which in some cases are combinedwith high current account deficits. In LatinAmerica, vulnerabilities stem mostly from thepolitical realm, including a major political

decreased or remained at a low level in most countries. It should be stressed, however, thatbanks’ asset quality indicators are typically backward-looking. Since risks may build upduring boom phases, a deterioration in credit quality might only become visible with asignificant time-lag if general economic conditions were to worsen. As most new MemberStates have been enjoying relatively high rates of economic growth in recent years, theresilience of loan portfolios to negative output shocks remains untested up to now.

Chart 1.8 Emerging market bond spreadsrelat ive to US short and long-term interestrates(Jan. 1991 - Nov. 2005)

Sources: JP Morgan Chase & Co. and Bloomberg.

0

400

800

1,200

1,600

2,000

1991 1993 1995 1997 1999 2001 2003 20050

2

4

6

8

10

EMBI+ brady (basis points, left-hand scale)US 10-year Treasury yield (%, right-hand scale)US Fed Funds rate (%, right-hand scale)


Chart 1.9 Global current account posit ions

(1996 - 2005, USD billions)

Source: IMF (WEO database).Note: 2005 f igures are IMF projections.

-800

-600

-400

-200

0

200

400

600

800

-800

-600

-400

-200

0

200

400

600

800

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

oil exportersAsia excl. Japan and ChinaChina, mainlandJapaneuro areaUS

Chart 1.10 China monthly foreign exchangereserve increase, trade balance and FDI

(Jan. 2004 - Sep. 2005, USD billions)

Source: CEIC.

-10-505

10152025303540

-10-50510152025303540

Jan. July2004

Jan. July2005

FDItrade balanceforeign exchange reserves

scandal in Brazil (the largest international issuerof sovereign bonds among the EMEs) whichmight evolve into a governance crisis, continuedinstability in some Andean countries, and aheavy electoral cycle including presidentialelections in major economies in 2006. Inemerging Asia, risks include the possibility thatexcessive curbs in investment could lead to asharp slowdown in growth in China, althoughthere are as yet few indications of this.

EMERGING MARKET ECONOMIES AND GLOBALIMBALANCESIn the medium term, risks to euro area financialstability stemming from emerging markets areprimarily associated with the potential for adisorderly correction of current accountpositions worldwide. With regard to the roleplayed by EMEs in the widening of globalimbalances, three elements are relevant. First,the main counterparts of the large and growingUS current account deficits have been largesurpluses in Asia – especially China and Japan– and increasingly in oil exporting economies,including many EMEs as well (see Chart 1.9).The current account surplus in China alone isexpected to widen to over USD 115 billion,which could amount for as much as 15% of theUS deficit for 2005.

Second, although some exchange rateadjustment has taken place in the six monthssince the June 2005 FSR, it has been mixed.The exchange rate policy framework in Chinaand Malaysia was significantly reformed at theend of July 2005 (see Box 4). However, inpractice the appreciation of both the renminbiand the ringgit was contained to little more thanthe initial 2% that followed these changes.Upward pressures on other Asian currencies –which were significant in some cases duringH1 2005 – largely subsided in the wake of theChinese policy reforms, although this alsoreflected concerns about the impact of higheroil prices on the pace of economic activity inthe region.

Third, reserve accumulation in China continuedunabated in the year to September 2005 at USD159 billion, up from USD 111 billion over thesame period in 2004 (see Chart 1.10). Reserveaccumulation remained higher than the sum ofthe trade balance and foreign direct investment(FDI), indicating significant speculative capitalinflows into the country betting on a renminbiappreciation. Although this gap has narrowed inrecent months, the extent to which the Chineseauthorities can contain speculative capitalinflows will continue to be an important

31cECB



ENVIRONMENTdeterminant behind the country’s foreignexchange reserve build-up.

Overall, sight should not be lost of the fact thatthere are two sides to the widening of globalimbalances insofar as EMEs are concerned.EMEs have been one of the sources – but not theonly source – of the build-up of globalimbalances in recent years. In particular,relatively stable exchange rates in many Asianeconomies (including emerging ones) vis-à-visthe US dollar have been associated with largecurrent account surpluses that have mirroredthe growing US deficit in recent years. Inaddition, the strategy of reserve accumulation,

which has underpinned monetary and exchangerate policies in Asia, has helped in preventing amore pronounced upward adjustment in matureeconomy long-term government bond yields.However, the difficulties inherent in sustainingthe pace of reserve accumulation of recentyears and the delicate equilibrium betweencurrent account surplus and current accountdeficit regions worldwide implies that EMEsare also vulnerable to a disruptive unwinding ofglobal imbalances. In this case, risks to EMEsmight be associated with disruption in globalcapital markets, tighter financing conditions,and, possibly, more sluggish demand frommature economies.

Box 4

THE REFORM OF THE RENMINBI EXCHANGE RATE REGIME

Against a background of large and growing global financial imbalances and concerns about theassociated risks for global financial stability, international pressure mounted on the Chineseauthorities to adopt a more flexible exchange rate regime in order to help curb growing globalcurrent account imbalances and to alleviate upward pressure on more flexible internationalcurrencies. Many analysts also expected that any revaluation of the Chinese currency wouldtrigger greater exchange rate flexibility in other Asian countries. On 21 July 2005, the renminbiwas revalued by 2% against the US dollar, from 8.2765 to 8.11, and the Chinese authoritiesannounced that they had moved to a “managed floating exchange rate regime based on marketsupply and demand with reference to a basket of currencies”. However, since the reforms, therenminbi has continued to be tightly managed against the US dollar, so that little impact on thescale of global imbalances can be expected in the short term.

The Peoples Bank of China (PBC) has emphasised that managing the exchange rate “withreference to” a basket of currencies does not mean that the renminbi will be pegged to a basket ofcurrencies. The objective of the new regime is to keep the renminbi exchange rate basically stableat an adaptive equilibrium level. While the weights of the currencies in the reference basket werenot disclosed, the currencies were selected mainly on the basis of the relative shares of China’strading partners in goods and services. However, other variables considered included the sourcesof FDI into China and the currency composition of Chinese debt. The currencies with the highestweight in the basket are the US dollar, the euro, the Japanese yen and the Korean won. Othercurrencies in the basket include the Singapore dollar, the UK pound sterling, the Malaysianringgit, the Russian rouble, the Australian dollar, the Thai baht, and the Canadian dollar. Under thenew regime, the PBC announces at the end of each working day the closing price of the foreigncurrencies traded against the renminbi in the interbank market. This closing price then serves asthe central parity for trading against the renminbi on the following day. As in the past, the dailytrading price of the US dollar against the renminbi is allowed to float within a ±0.3% band aroundthe central parity announced each day; however, the daily trading band of the renminbi againstnon-US dollar currencies has twice been widened to the current ±3.0%.


Since 21 July 2005, the PBC has continued to manage the exchange rate tightly, andmovements of the renminbi against the US dollar have remained well within the statutory bands(see Chart B4.1). Thus, although the new regime technically creates room for further market-driven appreciation, the renminbi had, as of early November 2005, appreciated by a maximumof 0.3% against the US dollar in addition to the initial one-off revaluation. The management ofthe transition to a more flexible regime had raised concerns for financial stability in China, inparticular with regard to the possibility that the revaluation would be perceived as beinginsufficient by the markets, possibly precipitating expectations of further revaluation andtriggering an increase in speculative capital inflows. So far the Chinese authorities appear tohave been relatively successful in managing market expectations of a further renminbiappreciation, and these expectations have actually receded since the initial revaluation (seeChart B4.2). By early November, the non-deliverable forward market was pricing in a 0.97%appreciation of the renminbi against the US dollar over the following three months, and a 3.6%appreciation over the following 12 months.

Foreign exchange reserves have, however, continued to grow at a monthly rate of around USD20 billion since July, indicating that intervention by the PBC has remained substantial in theaftermath of the reforms. In spite of this, and unlike patterns seen in 2004, most of the increasein reserves was accounted for by a surge in the trade surplus and not by non-FDI and non-trade-related inflows (a proxy for speculative inflows).

As of early November 2005 the Chinese reforms have had little impact on exchange rates in therest of the region, with most other currencies remaining either relatively stable or continuing todepreciate against the US dollar against a background of declining trade surpluses andconcerns about the impact of high oil prices on their economies.

Source: Bloomberg.

Chart B4.1 Intraday renminbi exchange rateagainst the dol lar

(index: 21 July 2005 = 100, upward movement = appreciationof the renminbi)

2005

99.9

100.0

100.1

100.2

100.3

100.4

99.9

100.0

100.1

100.2

100.3

100.4

21 July 5 Aug. 20 Aug. 4 Sep. 19 Sep. 4 Oct. 19 Oct. 3 Nov.

Chart B4.2 RMB/USD non-del iverableforwards

(inverted scale)

Source: Bloomberg.

7.7

7.8

7.9

8.0

8.1

8.2

8.3

8.4

7.7

7.8

7.9

8.0

8.1

8.2

8.3

8.4

3 months12 monthsrenminbi spot

Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov.2005

33cECB



ENVIRONMENT1.2 KEY DEVELOPMENTS IN INTERNATIONALFINANCIAL MARKETS

FOREIGN EXCHANGE DEVELOPMENTSNotwithstanding a further widening of the UScurrent account deficit, there was a broad-based appreciation of the US dollar in the firsthalf of 2005 (see Chart S11), supported byhigher money market interest rates andcontinued signs of robust economic growth.After a period of broad weakness between Julyand early September, the US dollar continuedto strengthen in October.

A persistent issue for foreign exchange ratemarkets has been the extent to which Asianauthorities have remained active in preventingtheir currencies from appreciating vis-à-vis theUS dollar. As reported in the June 2005 FSR, bythe end of 2004 the reserve holdings of majorcentral banks7 had increased by more than USD500 billion compared with a year earlier.Although only a proxy of foreign exchangeintervention activity, it is indicative that in thefirst six months of 2005, the same countriesdecreased their pace of accumulation of foreignholdings to about USD 270 billion inannualised terms. This appears to be mainlyrelated to the Japanese authorities’ policy ofnot intervening in foreign exchange markets,

whereas the rate of accumulation of foreignholdings in China remained robust.

Turning to forward-looking indicators, netspeculative positions – the difference betweennon-commercial long positions and non-commercial short positions – were mostly innegative territory between May and August2005, in line with the overall broad weakness ofthe dollar vis-à-vis the euro. They did howeverpick up in the aftermath of Hurricane Katrina,although this effect turned out to be short-lived(see Chart 1.11).

Expected short-term exchange rate volatilityvis-à-vis the euro implied in options pricesremained low for much of the time, and thedecision of the Chinese authorities to revaluethe renminbi had little discernible impact onexpected exchange rate volatility (see Box 4).Short-term volatility edged up instead, thoughonly slightly, in the aftermath of HurricaneKatrina in late August (see Chart 1.12).

Between early May 2005 and early November,the risk-neutral density (RND) functionimplied in options prices – which provides an

7 Japan, China, Taiwan, South Korea, Hong Kong, India,Singapore, Thailand and Malaysia.

Chart 1.11 Speculative posit ions and theUSD/EUR exchange rate

(May 2005 - Oct. 2005)

Source: Bloomberg.

-30,000

-20,000

-10,000

0

10,000

20,000

30,000

1.18

1.20

1.22

1.24

1.26

1.28

1.30

May June July Aug. Sep. Oct.2005

USD/EUR non-commercial net position (number of contracts, left-hand scale)USD/EUR exchange rate (right-hand scale)

Chart 1.12 Impl ied volat i l ity on USD/EUR

(Jan. 2005 - Nov. 2005, one-month implied volatility)

Source: Reuters.

7

8

9

10

11

7

8

9

10

11

USD/EUR USD/EUR average since 1999

Jan. Feb. Mar. Apr. June July Aug. Sep. Oct. Nov.May2005


Chart 1.13 Risk-neutral probabi l ity densityfunction of the USD/EUR exchange rate

Sources: Citibank and ECB calculations.

0

2

4

6

8

10

12

14

16

0

2

4

6

8

10

12

14

16

1.09 1.13 1.17 1.20 1.24 1.28 1.31 1.35 1.39

6 May 20052 Nov. 2005

USD/EUR exchange rate

marginal probability

estimate of the full range of marketexpectations of future exchange rate outcomestogether with associated likelihoods – for theUSD/EUR exchange rate over the next monthshifted to the left as the US dollar strengthened(see Chart 1.13). At the same time, thedistribution did not change significantly. Thissuggested that market participants continued toattach little likelihood to the possibility ofunusually large, or disorderly, exchange ratemovements in the short term.

US MONEY MARKETSUS money market conditions are importantfrom a euro area financial stability perspectivebecause financial institutions – includingcounterparties of euro area banks – usuallysecure their daily liquidity needs in thesemarkets. Therefore, any disturbances in thefunctioning of the US money markets couldgive rise to liquidity problems with thepotential for spillover effects to the euro areafinancial system.

Against a background of continued strength inthe US economy, the Federal Reserveproceeded with a measured removal of earliermonetary policy accommodation. The FederalOpen Market Committee (FOMC) increasedinterest rates by 25 basis points at each of itsmeetings from 30 June 2004 onwards. All in

Chart 1.14 US s ix-month TED spread

(Jan. 1999 - Nov. 2005, basis points)

Source: Bloomberg.

0

15

30

45

60

75

90

105

0

15

30

45

60

75

90

105

1999 2000 2001 2002 2003 2004 2005

all, by early November 2005, the Federal Fundstarget rate had been raised 12 times, and by atotal of 300 basis points, bringing it to 4.0%, alevel last seen in June 2001.

By early November 2005, the pricing ofFederal Funds futures contracts showed thatmarkets were expecting the Federal Fundstarget rate to be raised to 4.25% by the end of2005, and to 4.75% by summer 2006.Therefore, expectations had been significantlyscaled up since the publication of the June 2005FSR.

The so-called TED spread – the differencebetween uncollateralised money marketinterest rates and risk-free Treasury bill rates ofsimilar maturities – can provide indicationsabout the degree of concern among moneymarket participants with regard to counterpartycredit risks. While this spread remainedbroadly stable between late 2001 and early2005, it widened significantly after March2005 (see Chart 1.14). The main explanationfor this appears to have been concerns aboutabnormally high exposures of some financialinstitutions to Ford and GM, following theirdowngrading by credit rating agencies.However, as these concerns subsequentlyproved to be unfounded, the TED spreadquickly settled back to lower levels. Overall,

35cECB



ENVIRONMENT

this suggests that market participants considerthe financial position of the maincounterparties in the US money markets to berobust.

Concerning the functioning of money markets,the issuance of commercial paper, which beganto recover after mid-2003, continued tostrengthen in the course of 2005 (seeChart 1.15).

The return of issuance activity in this markettowards the peak observed in 2001 generallymirrored the improvement in the credit ratingsof US corporations. By improving the access ofcorporations to direct short-term financing, therecovery of this market should be seen as apositive indication of US financial systemstability.

US GOVERNMENT BOND MARKETSLong-term government bond yields in the USremained unusually low by historical standardsin the six months after the June 2005 FSR.However, the low level of ten-year bond yields,which stood at 4.7% by early November 2005,remained a conundrum that was difficult toreconcile with underlying macroeconomicfundamentals. At the same time, and bycorollary, concerns about the risk of anunexpected and abrupt upturn in these yieldshave remained.

Chart 1.15 US commercial paper, totalamount outstanding

(Jan. 1991 - Oct. 2005, USD billions)


0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

0

200

400

600

800

1,000

1,200

1,400

1,600

1,800

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Although their precise influence is difficult toquantify, several demand factors appear tohave played a role in holding US long-termyields down. First and foremost, the demandfor US Treasuries among foreign investors,especially in Japan and other Asian countries,has been significant since 2003. Furthermore,institutional demand to reduce balance sheetmismatches, which has been stimulated byrecent regulatory and accounting changes, alsoappears to have played a role in holding long-term yields down. More recently, oil-exportingcountries have sought to invest high and risingoil revenues, leading to additional demand forUS bonds. By contrast, the demand for USTreasuries from US retail investors hasremained rather modest, as indicated by mutualfund inflow patterns (see Chart S18).

Another potential reason why US long-termbond yields have remained as low as they haverelates to high levels of global saving. Not onlyhas saving been high in EMEs in recent years,but lower investment by the corporate sector,especially in the US, in the wake of the burstingof the equity market bubble around the turn ofthe century, may have played a role in pushinglong-term real yields down. High levels ofsaving vis-à-vis investment in the US non-farm,non-financial corporate sector, as evidenced bythe progressive narrowing of the financing gapafter 2000 – including a relatively exceptionaltransition from financing deficit into surplus –were accompanied by a further drop in ex postreal long-term bond yields – defined as the ten-year nominal yield minus actual consumerprice index (CPI) inflation. However, theinfluence of this factor may start to wane asdemand from corporations for external sourcesof funding edges up. This could precipitate anupturn in long-term real yields.

While the low level of short-term interest ratesin the US had been seen as another importantfactor in keeping long-term yields low beforemid-2004, the progressive tightening of USmonetary policy since then has had little impacton the level of long-term yields. As a result,there has been a substantial flattening of the


yield curve, as measured by the spread betweenthe ten-year government bond yield and thethree-month money market rate, from about300 basis points in the second half of 2003towards 40 basis points in early November2005. Nevertheless, it cannot be excluded thatthe protracted period of ample liquidityconditions in US money markets may havegenerated an overhang that has continued toinfluence the level of nominal long-termyields.

With the flattening of the US yield curve, theattractiveness of so-called carry trades – wherefunds are borrowed short-term and invested inlong-term maturity instruments – hasdiminished. In fact, in early 2005, speculativeinvestors began to position themselves for anupturn in long-term bond yields (see Chart1.16).8 However, as long-term yields continuedto decline, positioning shifted again around themiddle of the year towards betting on furtherdeclines in long-term yields.

Differences in the positioning of speculativeinvestors at different points in the maturityspectrum can reveal information onexpectations regarding future changes in theslope of the yield curve. Whereas investors hadbeen positioning themselves for a steepening of

8 For an analysis of the effects of speculative positioning on USbond yields, see ECB (2004), Financial Stability Review,December, Box 2.

the yield curve throughout 2004, thispositioning shifted towards flattening in thecourse of 2005 (see Chart 1.17). This wouldsuggest that speculative investors do not expecta sudden upturn in long-term yields.

At the same time, however, on balanceinstitutional investors have exhibited concernsabout overvaluation in global bond markets formuch of 2005 (see Chart 1.18).

Looking ahead, an unexpected and abruptupturn in US long-term bond yields cannot beexcluded. While several factors can beidentified as explanations for holding yieldsdown, there are some uncertainties about therole they may play in the future. For instance, ifshort-term interest rates follow the futuretrajectory implicit in interest rate futures pricesand if the US non-farm, non-financialcorporate sector financing surplus becomes adeficit, two important supporting factors forthe low level of yields will fade. Moreover,there is uncertainty about whether the strengthof demand for US bonds will be sustained,especially from (Asian) official accounts.

Chart 1.16 Indicators of posit ioning andleverage in the US bond markets

(Jan. 2002 - Oct. 2005, four-week moving average)

Sources: Federal Reserve Bank of New York and Bloomberg.

-220

-180

-140

-100

-60

-20

20

60

100

140

2002 2003 2004 2005120

140

160

180

200

220

240

260

280

300

net non-commercial positions in ten-year Treasury futures (thousands of contracts, left-hand scale)net secured financing of US primary dealers (USD billions, right-hand scale)

Chart 1.17 Net non-commercial positions onfutures and options, and the yield spreadbetween ten and five-year government bonds(Jan. 1996 - Oct. 2005)

Sources: CFTC reports and Bloomberg.

-300

-200

-100

0

100

200

300

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

5-year (thousands of contracts, left-hand scale)10-year (thousands of contracts, left-hand scale)yield spread (% points, right-hand scale)

37cECB



ENVIRONMENTChart 1.18 Net percentage of equity fundmanagers responding that global bondmarkets are overvalued(Mar. 2003 - Oct. 2005, net %)

Source: Merrill Lynch Global Fund Manager Surveys.

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

2003 2004 2005

Chart 1.19 US industria l bond spread andthe downward rating revis ion ratio

(Jan. 1987 - Sep. 2005)

Source: Moody’s.Note: The downward rating revision ratio is computed as thenumber of downward industrial rating revisions vis-à-vis thetotal number of industrial rating revisions.

20

40

60

80

100

120

140

160

1987 1989 1991 1993 1995 1997 1999 2001 2003 200530

40

50

60

70

80

90

100

Moody’s Baa-Aa industrial bond yield spread (basis points, left-hand scale)total downward industrial rating revisions ratio (number, right-hand scale)

US CORPORATE BOND MARKETSWith the downgrading of two very largeissuers, GM and Ford, to sub-investment grade,the US corporate bond market was subject to asignificant “credit event” in early May 2005,but proved to be resilient. Although there wassome dislocation in structured credit markets,the corporate bond market remained contained,and the forced selling of corporate bonds byinstitutional investors did not materialise inquite the way that had been previously feared.Corporate bond spreads at the lowest end of therating class spectrum widened only slightly,and spreads remained close to historically lows(see Chart S22). The factors contributing togenerally tight corporate bond spreadscontinued to be the robust pace of corporateearnings growth, low default rates, balancesheet restructuring, and low stock marketvolatility (see Chart S16). An ongoing hunt foryield may also have played a complementaryrole.

With some signs that the US credit cycle maybe maturing, the main risk facing corporatebond markets in the period ahead would appearto be an unanticipated turn in the cycle,possibly triggering other large credit events. Inthis respect, it has been notable that the number

of US industrial corporate credit ratingdowngrades vis-à-vis upgrades has remainedrelatively high, despite an environment ofrobust economic growth (see Chart 1.19). Therecent expansion of the spread betweenMoody’s Baa and Aa industrial corporate bondyields may be a harbinger of a turn in the UScredit cycle.

US EQUITY MARKETSUS stock prices increased slightly in the sixmonths after the June 2005 FSR (see ChartS14). This rise took place against a backgroundof historically low stock market volatility (seeChart S16), robust earnings growth andearnings outcomes that were consistentlystronger than market participants hadanticipated. In addition, there was a further risein the funding of equity investments throughborrowing (see Chart S19). Rising short-terminterest rates and the further upturn in oil pricesmay, however, have worked in the oppositedirection. In particular, oil prices and ameasure that captures the appetite for investingin risky equity – the spread between theS&P500 operating earnings yield and the (expost) real ten-year government bond yield –have tended to co-move, and the recent surge inoil prices may have raised this premium (see


Chart 1.20 US “earnings yie ld premium” andthe real oi l pr ice

(Jan. 1980 - Sep. 2005)

Source: Global Financial Data.Note: The “earnings yield premium” is the differencebetween the operating earnings yield and the real long-terminterest rate (nominal long-term interest rate minus actualCPI inflation).

-5

0

5

10

15

20

0

10

20

30

40

50

“earnings yield premium” (% points, left-hand scale)CPI deflated oil price (West Texas intermediate, USD per barrel, right-hand scale)

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004

Chart 1.20). Against this background, therewas a slight reversal in the strength of equitymutual fund inflows (see Chart S18).Nevertheless, by early November 2005, theprice-earnings ratio for the US stock market,based on ten-year trailing earnings, remainedhigh (see Chart S15). At the same time,investors’ willingness to invest in risky USassets such as equity, as opposed to risk-freesecurities, have remained in neutral territorysince May 2005 (see Chart S13).

Although the overall increase in equity priceswas limited, performance remained variedacross different segments of the market. Whilesmall and mid-cap stock price indices reachedall-time highs, large-cap stock prices remainedwell below the peaks seen in early 2001.

Looking ahead, the US stock market may facesome vulnerabilities. With high valuations andindications of increased borrowing in order topurchase stocks, the market could bevulnerable to adverse market dynamics. Withindications that the US corporate earningscycle may have peaked, coupled with risingshort-term interest rates and higher oil prices,downside risks for US equity prices may haverisen. In this respect, valuation indicators

based on options prices in October 2005suggested that market participants wereslightly more concerned about the likelihood ofabnormally large stock price changes than theywere in May. The skewness of the RND for USstock prices remained tilted towards thedownside, and the associated probabilitiesof large declines increased slightly (seeChart S17).

With high price-earnings ratios and low equitymarket volatility, financing conditions in theUS equity markets remained favourable.Nevertheless, while remaining high, the valueof realised initial public offering (IPO) dealstended to level off in the course of the year toSeptember 2005 (see Chart S21). Secondarypublic offerings (SPOs), after having declinedslightly in the first half of the year, picked upthereafter. With indications that some firmshave stepped up repurchases of their ownequity through borrowing, corporate sectordebt-equity ratios may yet rise. Just as lowerdebt-equity ratios have contributed to thelowering of equity market volatility over recentyears, lower equity issuance could trigger a risein volatility towards more normal levels.9

COMMODITY MARKETSFrom a financial stability viewpoint, risks incommodity markets, especially oil markets,tend to operate largely through indirect, ormacroeconomic, channels. High and volatileoil price levels can pose risks for generaleconomic activity and inflation and couldcontribute to financial sector stress.Furthermore, asset prices, such as stock prices,can be adversely affected by sharp oil priceincreases, especially if they persist. There areindications that speculative activity in the oilderivatives markets has been increasing inrecent years, so that the importance of directchannels (i.e. the exposures of financialinstitutions to oil price movements) may haverisen commensurately.

9 For an analysis of the factors driving recent declines in equitymarket volatility, see ECB (2004), Financial Stability Review,December, Box 3.

39cECB



ENVIRONMENTChart 1.21 Speculative posit ions and oi lpr ices

(Jan. 2005 - Nov. 2005, net futures commitments ofnon-commercials on the New York Mercantile Exchange)

Source: Bloomberg.Note: Net commitment = number of long-short contracts,where each contract represents 1,000 barrels of WTI crudeoil. “Non-commercials” denotes entities not engaged incrude oil production or ref ining.

-60

-40

-20

0

20

40

60

80

100

35

40

45

50

55

60

65

70

75

net commitments (thousands, left-hand scale)WTI oil price (USD per barrel, right-hand scale)

Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct.2005

Oil prices rose sharply after the June 2005 FSR,against a background of relatively robustdemand, especially from North America, theMiddle East and Asia. Meanwhile, OPEC hasbeen producing at near-capacity, while non-OPEC supply growth remained short ofexpectations. Moreover, global spare refiningcapacity has shrunk as a result of unexpectedlystrong growth in demand in recent years as wellas under-investment in refining capacity. Aseries of refinery disruptions in the US in Julyand August 2005, largely due to an

Chart 1.22 Brent crude oi l futures prices

(July 2005 - Dec. 2010, USD per barrel)

Source: Bloomberg.

20062005 2007 2008 2009 2010

1 June 2005 4 Nov. 2005

48

50

52

54

56

58

60

62

48

50

52

54

56

58

60

62

Chart 1.23 Options- impl ied r isk-neutraldensit ies

(Jan. 2004 - June 2006, USD per barrel, 4 November 2005,10%, 20%, 50%, 80% and 90% confidence intervals)

Sources: Bloomberg and ECB calculations.

30

40

50

60

70

80

90

30

40

50

60

70

80

90

2004 2005 2006

overstretched refinery system, coupled withheightened geopolitical concerns over thesecurity of oil supplies, put further upwardpressure on prices in those months. The effectsof hurricanes Katrina and Rita furtheraggravated constraints in an already stretchedoil supply chain. At the same time, speculativeactivity does not appear to have played a majorrole in driving oil prices higher (seeChart 1.21).

The surge in oil prices since June 2005coincided with large increases in the prices ofoil futures contracts with expiry dates at longhorizons. The futures price for short horizonsremained above spot prices – a feature seldomassociated with high spot prices, reflectingconcerns over the adequacy of future suppliesand inventories in the near future (seeChart 1.22). Looking ahead, given the limitedspare capacity all along the oil supply chain, oilprices are likely to remain sensitive to anyunanticipated changes in the supply-demandbalance.

The high degree of uncertainty in oil marketshas been reflected in the pricing of options onoil. Implied distributions for future oil prices,which are extracted from options prices,exhibited very wide confidence intervals inearly November 2005, and the balance of riskswas tilted towards the upside (see Chart 1.23).


Turning to other commodity prices, there havebeen signs of increasing speculation in theprecious metals markets, which drove goldprices to their highest levels since 1988 (seeChart S25). Evidence of speculation wassuggested by the rising share of non-commercial positions in total open interest ingold futures contracts (see Chart 1.24).Furthermore, net non-commercial longpositions reached their highest-ever level inearly October. Other factors may also haveplayed some role in the recent increase in goldprices, such as growing physical demand forgold and, to a lesser extent, concerns about therisks of rising inflation.

EMERGING MARKET FINANCING CONDITIONSFinancing conditions in EMEs have remainedbenign in the period since the June 2005 FSR.In the interim period to early November 2005,emerging market bond spreads have evennarrowed (by about 60 basis points)notwithstanding the downward leveladjustments to the benchmark EmergingMarkets Bond Index (EMBI) Global indexassociated with the completion of Argentina’ssovereign debt restructuring. At about 240basis points, emerging market bond spreadsthus remained at historic lows in early

Chart 1.24 Non-commercial posit ions andtotal open interest in gold futurescontracts and the gold price(Jan. 2000 - Oct. 2005)

Source: Bloomberg.

10

20

30

40

50

2000 2001 2002 2003 2004 2005100

200

300

400

500

non-commercial gold futures contracts (% of total contracts, 4-months moving average, left-hand scale)open interest in gold futures contracts (thousands of contracts, right-hand scale)gold price (USD per troy ounce, right-hand scale)

Chart 1.25 Emerging market bond spreads


Source: JP Morgan Chase & Co.

0100200300400500600700800900

1,000

01002003004005006007008009001,000

2002 2003 2004 2005

EMBI GlobalEMBI Global adjusted for Argentina

November 2005 (see Chart 1.25 and ChartS23).

Favourable financing conditions for EMEscontinued to be explained by an ongoing huntfor yield by international investors, whichfavoured the broad EME asset class. In the yearto early November 2005, returns on EME bondshad exceeded those across several other assetclasses. For instance, over this period theEMBI Global index posted a 6.5% return,compared to a 1.2% return on an index of high-yield US corporate bonds. However, as theupside potential of traditional EME assetclasses became more limited, there have beensigns that international investors have shiftedtheir attention elsewhere – notably to EMEbonds denominated in local currency. Thisbrought yields down in some regions (see Chart1.26). The strength or potential forappreciation of certain emerging marketcurrencies – which had already enabled somesovereigns to undertake internationalissuances in domestic currency – was also afactor in this context.

The relative attractiveness of the emergingmarket asset class has also been supported bythe generalised improvement in EME

41cECB



ENVIRONMENTChart 1.26 Emerging market local currencybond yie lds

(Jan. 2002 - Nov. 2005, %)

Source: JP Morgan Chase & Co.

0

2

4

6

8

10

12

0

2

4

6

8

10

12

2002 2003 2004 2005

GBI emerging marketsGBI emerging Latin AmericaGBI emerging EuropeGBI emerging Asia

fundamentals in recent years. This process hasbeen helped by external factors, including highcommodities prices and hence favourablecredit dynamics for EME exporters of thoseproducts, but also by fundamental policy andstructural reforms undertaken by somedomestic authorities. Improvements in theoverall underlying credit quality of benchmarkbond indexes following rating upgrades (someto investment-grade status) were rewarded byinstitutional investors. Coupled with the huntfor yield, this has led to a wider investor basefor the broad EME asset class.

Against this background, international bondissuance by EMEs has remained brisk. Thepace of issuance has only moderated slightly incomparison to the record levels reached in2004, as both sovereigns and corporatesanticipated obligations for the remainder of theyear and into 2006 (see Table S2). Themomentum behind investor demand fordomestic currency debt has also promptedsome sovereigns to make growing use of thismarket segment to meet or anticipate financingneeds. Nonetheless, heavily indebted EMEs,particularly in Latin America, have continuedto take advantage of the benign financialenvironment to engage in strategic debtmanagement in order to attain a morefavourable amortisation and maturity profile.

Brazil’s buyback of its outstandingcapitalisation bonds (C-bonds) issued underthe Brady plan – hitherto the benchmark bondamong the EME asset class – stands out amongthese operations.

The emerging market financing outlooktherefore remains favourable, although somedownside risks remain. The main cause forconcern is the possibility of a pronounced andsustained upturn in mature economy long-termgovernment bond yields, and that this would beassociated (as has been often the case) withtighter financing conditions. If such a negativescenario were to materialise, policymakersmay take comfort in the fact that EMEs’vulnerability to external shocks has beensignificantly reduced in recent years. However,any rebound in long-term government bondyields would probably be associated withgreater discrimination on the part ofinternational investors owing to changes inunderlying risk preference.

1.3 CONDITIONS OF NON-EURO AREAFINANCIAL INSTITUTIONS

CONDITIONS IN NON-EURO AREA EU BANKINGSECTORSThe financial performance of non-euro area EUMember States’ banking sectors was strong in2004. An improvement in profitability wasmainly driven by enhanced cost efficiency,strong growth in lending, mostly for housingpurposes, as well as reductions in provisioningflows relative to total assets. Asset qualityimproved given a benign credit environment inmost non-euro area EU banking sectors. Whilesolvency indicators did deteriorate to a certainextent,10 they still remained at relativelycomfortable levels.

The profitability of the EU-1311 banking sectorsimproved across the board in 2004. The averagereturn on equity (ROE) for EU-13 banks stood at

10 Owing to rapid credit growth, risk weighted assets increased ata higher pace than regulatory capital.

11 The EU-13 grouping comprises all EU Member States that arenot in the euro area.


16.4% at end-2004, up from 15.2% in 2003. Thedegree of improvement in profitability variedacross EU-13 countries, with banking sectors inhigh-growth catching-up economies generallyshowing the largest gains.

Operating income (as a percentage of totalassets) fell slightly on aggregate in 2004 for theEU-13 banking sectors, owing to declining netinterest margins. The pressure on net interestmargins in several countries stemmed from lowinterest rates and strengthened competition,especially in the market for housing loans. Thisnotwithstanding, net interest income continuedto rise, as the impact of narrowing margins wasoutweighed by brisk growth in lendingvolumes. Concerning lending to the privatesector, a common pattern across non-euro areaEU Member States was strong growth inlending to households, in particular for housingpurposes. Lending to non-financialcorporations was characterised by highergrowth in lending volumes in the EU-13countries than in the euro area, although thiswas outpaced by the growth in householdlending.

Contrary to developments in net interestincome, non-interest income increased at afaster rate than total assets in 2004 for EU-13banks. Consequently, as a share of totaloperating income, non-interest income alsorose. This notwithstanding, the relativeimportance of non-interest income ingenerating income remained significantlylower for EU-13 banks than for theircounterparts in the euro area. In 2004, the shareof net non-interest income in total incomereached an average of 32.4% in the EU-13banking sectors compared to 47.8% in the euroarea.

An important factor underlying the improvedprofitability in 2004 across the non-euro areaEU banking sectors was cost containment. In2004, banks in these countries further reducedtheir cost-to-income ratios, and generallyregistered lower ratios than the average foreuro area banks. Looking at different sub-

groups, countries with rapid growth in creditand operating income recorded the mostmarked improvements in cost-to-incomeratios.

In 2004, general economic conditions in thenon-euro area EU Member States weresupportive of banks’ activities. On account ofthe benign credit environment, both the flowand the stock of provisions fell, as a percentageof total assets, between 2003 and 2004. The lowlevel of provisioning may have been areflection of improved asset quality in mostcountries. The improvement in loan quality,however, was not uniform across all the EU-13banking sectors. In those countries wherecredit growth was strong, the share of non-performing and doubtful assets in total loansedged up slightly. The coverage ratio, i.e. theratio of provisioning stocks over total non-performing and doubtful assets, increased in2004 for EU-13 banks as a whole. At first sight,this might seem to alleviate potential concernscaused by lower provisioning flows. Thispositive assessment should be qualified,however, as the coverage ratio in several EU-13 countries remains below the average of euroarea banks.

Notwithstanding improved profitability,banks’ capital adequacy ratios tended todecrease in the EU-13 countries, with both theoverall solvency ratio and the Tier 1 ratiodeclining. In general, however, solvency ratiosremained comfortable and still comparefavourably with those of euro area banks.

Looking ahead, available data for the first halfof 2005 indicate that the positive trendsexperienced in 2004 are likely to continuefurther. In an environment of favourablecyclical conditions and low interest rates,growth in lending to the private sectorremained robust in the first half of 2005,suggesting that banks’ profitability might notbe threatened by a slowdown in credit growth.Looking at the downside risks to banks’profitability outlook, concerns related to thesustainability of improved profitability might

43cECB



ENVIRONMENTarise in some EU-13 banking sectors whererecent improvements in banks’ financialconditions have mainly resulted from decliningprovisions and enhanced cost control. In thosenon-euro area EU Member States where rapidgrowth in credit to the private sector has beenaccompanied by a marked increase inborrowers’ exposure to foreign exchange raterisk, unexpectedly large adverse exchange ratemovements might be a source of rising loanlosses for banks.

GLOBAL BANKSThe condition of global financial institutionsmatters for euro area financial stability becauseof the important role they play in financialmarkets and their function as counterparties toother financial institutions. Potential problemsin global financial institutions could lead todisruptions in euro area financial markets andcould pose counterparty risks for some euroarea financial institutions.

Most global financial institutions enjoyed goodfinancial results for 2004, albeit slightlyweaker than the year before. Performance forthe first two quarters of 2005, although stillhealthy, deteriorated somewhat for several ofthese institutions. The simple average ROEwas 13.6% for the second quarter of 2005,compared with about 17% for 2004 as awhole.12 Adequate levels of profitability weremaintained and costs remained under control.In addition, legal risks have been significantlyreduced as some institutions have settledoutstanding Enron-related litigation, removinga source of uncertainty for future earnings.Furthermore, M&A and investment bankingoperations, which have tended to provelucrative in the past by generating significantfee income, continued to contribute positively.In some cases, this was accompanied byincreased staff compensation costs, but by andlarge these costs remained relatively well-contained.

However, conditions in global capital marketshad an adverse impact on financial results fromtrading, which has been a major contributor to

Chart 1.27 Trading revenues as a proportionof total net revenues

(1998 - Q2 2005, %)

Sources: SEC f ilings and ECB calculations.Note: The institutions included are Goldman Sachs, LehmanBrothers, JP Morgan Chase & Co., Morgan Stanley, MerrillLynch, Citigroup, UBS and CSFB.

0

5

10

15

20

25

30

35

0

5

10

15

20

25

30

35

1998 1999 2000 2001 2002 2003 Q2Q1 Q32004

Q4 Q12005

Q2

average median

profitability over the last two years. For mostinstitutions, trading revenues decreased in thefirst quarter of 2005 both in absolute amountsand as a share of total revenues compared with2004, although they showed a slightimprovement in the second quarter (seeChart 1.27).

There are three main reasons for this. Firstly,the flattening of market yield curves,especially in the US, made carry trades lessprofitable for most institutions. Secondly, forsome institutions, the widening of spreadsprovoked by the Ford/GM downgrade resultedin some trading losses. Thirdly, the subsequentdecline in volatility in US bonds, combinedwith subdued conditions in equity markets,lessened the opportunities for these institutionsto risk their own capital and provide market-making services to clients.

Value at Risk (VaR) exposures increasedmarginally, with commodities recording thelargest increase as institutions had takenvarious positions in cash and derivativesmarkets by June 2005 compared with June 2004(see Chart 1.28). However, owing to a lack of

12 See Fitch Ratings (2005), “2Q05 Peer Data for SecuritiesFirms”, Special Report. Data exclude UBS and include CharlesSchwab and Lazard.


Chart 1.28 Changes in Value at Risk (VaR)levels

(changes in levels between Q2 2004 and Q2 2005)

Sources: SEC f ilings and ECB calculations.Note: The institutions included are Goldman Sachs, LehmanBrothers, JP Morgan Chase & Co., Morgan Stanley, MerrillLynch, Citigroup, UBS, and CSFB.

-60-40-20

020406080

100120140160

-60-40-20020406080100120140160

max-min rangemedianaverage

total interest equity FX

data, it is difficult to determine whether or notthese firms have been speculating on oil pricedevelopments.

Foreign exchange or currency-related riskexposures rose somewhat. Average equity-related risk exposures decreased slightly, whilethe average change in interest rate riskexposures remained essentially zero. However,since these changes are calculated as anaverage change compared with 2004, this mayunderstate the fact that some institutionsreduced their interest rate spread exposure atcertain points during the first half of 2005owing to the widening of spreads in corporatedebt markets. Notwithstanding, in overallterms of actual exposure, the largest marketrisk remains interest rate-related. The apparentincrease in risk-taking appetite by theseinstitutions also led to a slight increase in thenumber of trading days with recorded losses inthe first half of 2005 compared with the sameperiod in 2004. However, in cases wherenegative trading days occurred, the realisedloss was not greater than the amount predictedby the VaR models, indicating that risks havebeen managed relatively well.13

Looking ahead, global banks can expectcontinued profits, although at a more moderate

pace. This is primarily due to the decline intrading revenues experienced by several ofthese institutions. One institution was placedon negative ratings watch in April 2005 forreasons related to changes in management andthe possible sale of one of its business units;however, the overall assessment of ratingagencies for this group as a whole is positive.

While the risk appetite of some of theseinstitutions may have increased in the course of2005, indications are that risks have been well-managed. All in all, the financial condition ofthese institutions and their outlook give littlecause for concern from a financial systemstability viewpoint.

JAPANESE BANKSWhile the direct links between euro area andJapanese financial institutions appear to belimited, individual euro area banks may haveexposures to the Japanese banking sectorthrough direct claims or through financialmarkets.

As discussed in the June 2005 FSR, the balancesheet positions of Japanese banks havesignificantly improved in recent years. Thisimprovement has, to some extent, reflected theimplementation of a broad spectrum of policiesaimed at guaranteeing the stability of thecountry’s banking sector. These includeincreased regulatory and supervisory pressureand public capital injections. More recently,the banking sector has also benefited from thegradual recovery in the macroeconomicenvironment and from improved credit riskconditions.

The decline in non-performing loan ratios (i.e.the ratio of non-performing loans to totaloutstanding loans) and the rise in capitaladequacy ratios observed in recent years areamong the most visible signs of improvement.The non-performing loan ratio of all banksstood at 4% at the end of March 2005, compared

13 This is based on information made in public SEC f ilings bythese institutions. Not all institutions disclose the number ofnegative trading days or VaR exceptions.

45cECB



ENVIRONMENT

14 See Fitch Ratings (2005), “Japan Major Banks 2005 – HappyDays Are Here Again?”, Special Report, July; and FitchRatings (2005), “Japanese Large Regional Banks: Results for2004/2005”, Special Report, July. The Tier 1 f igures for majorJapanese banks exclude Resona.

15 For a comprehensive analysis, see A. Wolfson (2005), “WhyIs Japanese Banking Sector Prof itability So Low?”, IMFArticle IV Consultation Discussions with Japan – SelectedIssues, pp. 54-64.

16 See Bank of Japan (2005), Financial System Report: AnAssessment of Financial System Stability, Focusing on theBanking Sector, August.

with 5.3% at the end of September 2004 (seeChart S7). This reduction has contributedsomewhat to the improvement in profitability.

This improvement in profitability has alsocontributed, to a limited extent, to increasingsolvency ratios, which have improved for bothmajor and smaller regional banks. The averagecore Tier 1 capital adequacy ratio of the majorbanks improved from 6.0% at the end of March2004 to 6.2% at the end of March 2005.Meanwhile, the smaller regional banksincreased their average Tier 1 ratio from 7.5%in March 2004 to 7.9% in March 2005.14 Itshould be noted that some of the increase insolvency ratios is due to injections of publicfunds to re-capitalise the banking system.Furthermore, some Japanese banks continue touse (net) deferred tax assets (DTA) to maintaintheir regulatory capital above minimumthresholds. These are discretionary accountingitems included in balance sheets in order tobridge the gap between accounting and taxableincome. A continuation of the improvementsnoted over the past few years should lead toboth an increase in overall solvency levels andin the quality of capital.

From a financial stability perspective, twomain concerns remain. Firstly, the recentincreases in the profitability of Japanese bankscan be attributed to a reduction in credit costsand a rise in non-interest income, rather than toan improvement in interest income.15 Theprofitability of the Japanese banking sectorcontinues however to remain low byinternational standards.

Secondly, the large holdings of Japanesegovernment bonds by domestic banks may giverise to potential losses arising from market riskin the event of an increase in long-term interestrates. However, there is some evidence thatJapanese banks, especially the major banks,have in recent years reduced the duration ofthese portfolios, which should limit potentiallosses.16


2 THE EURO AREA ENVIRONMENT

Notwithstanding a background of risinguncertainty concerning the euro area economicoutlook, the balance sheet positions of largecorporations and households appear to bereasonably solid. However, potentialvulnerabilities remain that may pose risks foreuro area financial stability if they were tointensify. Despite improved profitability,corporate sector indebtedness remains high. Asudden and sharp reversal in economic growth– although very unlikely – could, along withpersistently high oil prices, erode profitability.The sustained dynamic – and sometimesdivergent – developments in residentialproperty prices in individual Member Statesalso continue to call for careful monitoring.

2.1 ECONOMIC OUTLOOK AND RISKS

The macroeconomic environment in whichfinancial institutions operate is an importantdeterminant of the creditworthiness ofhouseholds and firms, and of banking sectorprofitability in general. This means thatgeneral macroeconomic conditions can be animportant exogenous source of risk forfinancial stability.1 In the six months since theJune 2005 FSR, there have been someindications that the pace of economic growthhas lost some momentum since mid-2004. BothECB staff projections and the expectations ofinternational and private sector organisationsregarding economic activity in the euro areahave been revised downwards over the past sixmonths. Nevertheless, the slowdown in growthseems to be temporary, and longer-termexpectations point towards the pace ofeconomic activity rising towards potentialgrowth rates.

The risks that surrounded the outlook for euroarea growth at the time of the June 2005 FSRappear to have become more pronounced overthe past six months. On the external side, thesustained rise in oil prices poses risks forcorporate profit margins, thereby raising

corporate sector credit risk, and for consumerspending. Moreover, widening globalimbalances have continued to pose a risk ofsharp exchange rate movements. On thedomestic side, the main risk to a strengtheningof growth in the euro area stems from lowconsumer confidence and consequently lowdomestic demand, which can be attributable toboth higher oil prices and only graduallyimproving labour market conditions.

The possibility of increased risk to theeconomic outlook appears to be shared byprivate sector forecasters, as revealed forinstance in the Survey of ProfessionalForecasters (SPF). The percentage of theprobability distribution of one-year-aheadforecasts for euro area real GDP growth below2% in the SPF has risen since the June 2005FSR (see Chart 2.1). Moreover, there has been aslight rise in the percentage of the probabilitydistribution of growth below 1%.

1 See ECB (2005), “Indicators of Financial Distress in MatureEconomies”, Financial Stability Review, June, pp. 126-131.

Chart 2.1 Survey-based est imates of thedownside r isk of weak real GDP growth inthe euro area(Q1 1999 - Q4 2005, %)

Source: ECB.Note: Measured as the percentage of the probabilitydistribution for real GDP growth below the threshold of euroarea growth of 1% and 2% in the SPF, one year ahead.

0

20

40

60

80

100

0

20

40

60

80

100

1999 2000 2001 2002 2003 2004 2005

below 1%below 2%

47cECB



ENVIRONMENT2.2 BALANCE SHEET CONDITIONS OFNON-FINANCIAL CORPORATIONS

From a financial stability viewpoint, thecondition of non-financial corporate sectorbalance sheets is crucial for an evaluation ofthe credit risks posed by firms for banks andinvestors in corporate bond markets.Furthermore, the condition of non-financialcorporate sector balance sheets is fundamentalto the performance of stock markets and capitalmarkets.

The outlook for the euro area non-financialcorporate sector has remained benign over thepast six months. The balance sheets of firmshave been further strengthened by robust profitgrowth and further debt restructuring efforts. Inaddition, in a very low interest rateenvironment, the debt financing burden of non-financial corporations has remained contained.Reflecting these developments, credit spreadshave remained tight and banks’ creditstandards on the approval of loans toenterprises have continued to ease.

By late 2005, the main risks facing the non-financial corporate sector continued to beuncertainties surrounding the broad economicoutlook – including the potential adverseeffects of higher oil prices on corporate sectorprofitability. In addition, relatively high levelsof indebtedness and continued reliance onborrowing at floating and short-term interestrate fixation has raised the interest rate risk oncorporate sector balance sheets.

After mid-2002, the profitability of large non-financial euro area firms had strengthenedconsiderably. In the first half of 2005, theaggregate ROE of listed firms reached four-yearhighs (see Chart 2.2), indicating that profitabilityremained strong. The ongoing improvement incorporate sector profitability was partly due tostrong sales growth. However, towards the end of2004 and into early 2005, costs increased, mostlikely owing to rising oil and commodity prices.Although the recent overall financialperformance of the non-financial corporatesector has improved considerably, performancehas varied across sectors, with profitabilityimproving most significantly in the more export-oriented sectors (see also Box 5).

Chart 2.2 Prof it rat ios of euro area l istednon-f inancial corporations

(Q1 2002 - Q2 2005, %)

Sources: Thomson Financial (Worldscope) and ECBcalculations.Note: The calculation is based on an unbalanced sample ofquarterly data over time covering around 600 f irms for ROEand return on assets (ROA), and around 1,100 f irms for netincome to sales. Figures for Q2 2005 are based on a limiteddata set.

-6

-4

-2

0

2

4

6

8

-6

-4

-2

0

2

4

6

8

2002 2003 2004 2005

net income to salesreturn on assetsreturn on equity

Box 5

SECTORAL PROFIT AND LEVERAGE DEVELOPMENTS OF EURO AREA LISTED NON-FINANCIALCORPORATIONS: EVIDENCE BASED ON MICRO DATA

Following the slump in aggregate euro area corporate sector profitability in 2001 and 2002,there was a significant turnaround (see also Chart 2.2). At the same time, the accumulation ofdebt slowed down as companies sought to restructure their balance sheets. Since aggregatefigures may hide differences at the sectoral level, and because banks may have differentexposures to different corporate sectors, it is of interest to analyse measures of corporate


financial performance at a sectoral level. This Box examines recent developments in profit andleverage indicators for the non-financial corporate sectors in the euro area based on firm-leveldata.1

Beginning in 2003 and continuing into 2004, there was a substantial recovery of net income-to-sales ratios across almost all corporate sectors (see Chart B5.1).2 Profitability based on thismeasure performed strongest in more export-oriented sectors, especially the manufacturingsector, in an environment of strong global demand. In addition, the profitability of thetransportation and communications sectors also picked up significantly, the latter possiblyreflecting the efforts made in the telecommunications industry to cut back on operating costs.By contrast, the improvement in profitability of more domestically oriented sectors, such asretail trade (and in part wholesale trade), was more muted. Although this probably reflected tosome extent the sluggishness of domestic demand during this period, profitability in thesesectors has also tended to be less cyclical.

Turning to financial leverage, a pattern common to all sectors was a significant build-up of debtin the late 1990s and 2000, indicated by rising debt-to-total assets ratios (see Chart B5.2). Theaccumulation of debt was particularly strong in the transportation and communications sectors,and above average in the retail and wholesale trade sectors.3 Debt ratios for the manufacturingand construction sectors, on the other hand, stood at more moderate levels. Later on, there wasan overall stabilisation of debt-to-total assets ratios, as firms started restructuring and

Chart B5.1 Net income-to-sales rat io acrossnon-f inancial corporate sectors

(%, annual data)

Sources: Thomson Financial Datastream and ECB calculations.

-10

-8

-6

-4

-2

0

2

4

6

8

-10

-8

-6

-4

-2

0

2

4

6

8

1996 1997 1998 1999 2000 2001 2002 2003 2004

total non-financial corporations manufacturingconstruction wholesale traderetail trade transportation, communications, etc.

Chart B5.2 Debt-to-total assets ratio acrossnon-f inancial corporate sectors

(%, annual data)


0

5

10

15

20

25

30

35

40

45

0

5

10

15

20

25

30

35

40

45

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

constructionmanufacturingretail tradeservicestransportation, communications, etc.wholesale tradetotal

1 For a more detailed analysis of micro data from a corporate f inance perspective, see ECB (2005), “Developments in CorporateFinance in the Euro Area”, Monthly Bulletin, November, pp. 75-90.

2 Net income is a narrow corporate prof it indicator and is defined as the operating and financial prof it after interest expenses,taxation and extraordinary items.

3 The transportation, communications, electric, gas and sanitary services sector includes air transport; railroads; transportationservices; water transportation; motor freight services; and electric, gas and sanitary services.

49cECB



ENVIRONMENTdeleveraging their balance sheets, and in 2004 there was even a broad-based reduction. Thedeleveraging efforts were particularly notable in the communications sector. In addition,manufacturing firms, supported by the strength of cash flows, also appeared to take theopportunity to reduce their debts. In the retail sector, the debt-to-asset ratio stabilised at a highlevel in 2004, probably reflecting the relatively weak earnings performance of this sector.Likewise, the debt ratio of the wholesale sector (both with respect to the durable and non-durable goods sub-sectors) remained at a relatively high level at the end of 2004.

All in all, patterns in sectoral profit and leverage indicators show that the profitability andindebtedness of export-oriented manufacturing companies has improved substantially inrecent years. In addition, the balance sheet conditions of the communications sector alsoimproved strongly in 2003 and 2004, owing to rising profits and significant debt-reducingefforts. By contrast, profit developments in the retail trade sector, and to some extent also in thewholesale trade sector, were more muted, partly reflecting weak private consumption growthin the euro area. As a result, the debt ratios of these sectors remained high. This means that thebalance sheet conditions of some companies, particularly in the retail sector, could prove to bevulnerable to continued weakness in domestic demand, and any unexpected deterioration couldimpair the ability of these companies to honour their debt obligations. To the extent that euroarea banks tend to have large exposures to the retail sector, including to many SMEs and toother consumer-oriented companies, such a scenario might pose risks for the soundness of thebanking sector.4

4 For a further analysis of the sectoral credit risk exposures of euro area banks, see Box 7.

Throughout the first three quarters of 2005, thegrowth of corporate sector earnings more orless consistently exceeded the expectations ofmarket analysts. Looking ahead, althoughmarket analysts are expecting some slowdown

Chart 2.3 Actual and expected corporateearnings in the euro area

(Jan. 2000 - Sep. 2005, %)

Source: Thomson Financial Datastream.

actual EPS growth (yoy)long-term expected EPS growth12-month-ahead expected EPS growth

-20

-10

0

10

20

30

40

50

-20

-10

0

10

20

30

40

50

2000 2001 2002 2003 2004 2005

in the pace of corporate earnings growth overthe next 12 months, future profitability growthof larger listed corporates is still expected toremain strong (see Chart 2.3).

The efforts that corporations have made inrecent years to clean up their balance sheetshas, combined with increases in companies’financial asset holdings, caused the debt-to-financial assets ratio of the sector to declinefrom 2003 onwards. From a peak of over 80% inearly 2003, this ratio had declined to slightlyover 70% by Q1 2005 (see Chart S29). Hencethe ability of firms to repay debt by liquidatingfinancial assets, if needed, improved.

Although the recent further strengthening ofcorporate sector profitability improved theavailability of internal funds to finance firms’operations, external funding picked upsignificantly in the first eight months of 2005,reaching the highest level seen since mid-2001(see Chart 2.4).


Chart 2.4 Breakdown of the real annualrate of growth of external f inancing tonon-f inancial corporations in the euro area(Q1 2000 - Q3 2005, %)

Source: ECB.Note: The annual rate of growth is def ined as the differencebetween the actual annual growth rate and the GDP deflator.

quoted sharesdebt securities issuedMFI loans

-1 -1

0

1

2

3

4

5

6

0

1

2

3

4

5

6

2000 2001 2002 2003 2004 2005

Chart 2.5 Cost of external f inancing ofnon-f inancial corporations

(Jan. 1999 - Aug. 2005, basis points)

Sources: ECB, Thomson Financial Datastream, Merrill Lynch,Consensus Economics Forecast and ECB calculations.Note: The real cost of external f inancing is calculated as aweighted average of the cost of bank lending, the cost of debtsecurities and the cost of equity, based on their respectiveamounts outstanding and deflated by inflation expectations.The introduction of MFI interest rate statistics at thebeginning of 2003 led to a statistical break in the series.

1

2

3

4

5

6

7

8

1

2

3

4

5

6

7

8

1999 2000 2001 2002 2003 2004 2005

overall cost of financingreal long-term MFI lending rates real short-term MFI lending ratesreal cost of market-based debtreal cost of quoted equity

This pick-up may have been induced by thepersistently low costs of debt financing (seeChart 2.5). In particular, firms may have“frontloaded” their borrowing, takingadvantage of cheap terms to meet their futurefinancing needs and/or to restructure the

maturity profile of existing debt at lessexpensive terms. There have also been someindications that a further factor underlying thisrise in borrowing by firms was an increasedneed for funds to finance M&As in 2005 (seeChart 2.6, 4th panel from left). In the first three

Chart 2.6 Demand for loans and credit l ines to enterprises, and contributing factors

(Q3 2003 - Q4 2005, net %)

Source: ECB Bank Lending Survey.

-50

-40

-30

-20

-10

0

10

20

30

40

-50

-40

-30

-20

-10

0

10

20

30

40

realisedexpected

fixedinvestment

2003Q4Q3 Q1 Q2 Q2

2004 2005 2005 2005 2005 2005Q3 Q3Q4 Q1

2003Q4 Q1Q4Q3 Q2 Q2 Q2

2004Q3 Q4 Q1

2003Q3

2004Q3 Q4 Q1

2003Q3 Q4 Q1 Q2 Q2 Q3

2004 2003 2004Q3 Q3 Q4 Q1 Q1Q2 Q3 Q4 Q2 Q3Q3 Q4 Q1 Q1Q2 Q3 Q4

Factors contributing to loan demand

inventoriesand working

capital

M&A andcorporate

restructuring

debtrestructuring

51cECB



ENVIRONMENT

quarters of 2005 this amounted to €168 billion,compared with €108 billion in the same periodin the year before (i.e. an increase of around56%). Apart from debt restructuring, thestrengthening of demand for short-termfunding may have been driven, to some extent,by an increasing need for working capital (seeCharts S30 and S31).

Faster growth in borrowing by firms caused thedebt-to-GDP ratio of the sector to rise in thesecond and third quarters of 2005 (see ChartS28). While ongoing debt restructuring – mainlythrough an effective shortening of the maturity ofdebt – reduced the financing burden (interestpayments) of the corporate sector, it did not lowerthe amount of debt outstanding. The persistentlyhigh level of debt leaves many companiesvulnerable to a deterioration in balance sheetsand/or an increase in interest rates.

Even though market yield curves and the termstructure of retail bank lending rates to firmsflattened throughout the first eight months of2005, which might ordinarily have beenexpected to reduce incentives to borrow in theshort term, the increasing tendency ofcompanies to take on debt at floating and/orshort-term rate fixation (as reported in Box 4 ofthe June 2005 FSR) continued over this period(see Charts 2.7 and 2.8).

Banks seem to have been willing to take onmore credit risk up until the third quarter of2005 in order to boost profitability (seeBox 11). Prior to the third quarter, this wasreflected in a continued net easing of creditstandards on loans to enterprises (see ChartB11.1); and on riskier loans to SMEs (see Chart2.9). During the third quarter of 2005 a slightnet tightening occurred, although it is too earlyto say whether this represents a turning point in

Chart 2.7 Debt securit ies issued by euroarea non-f inancial corporations – f ixedversus f loating rate(Jan. 2003 - Aug. 2005, % per annum)

Source: ECB.

fixedfloating

-20

-10

0

10

20

30

40

-20

-10

0

10

20

30

40

2003 2004 2005

Chart 2.8 New business loans to euro areanon-f inancial corporations with short-terminterest rates and term spreads(Jan. 2003 - Aug. 2005)

Source: ECB.

floating rate and up to 1 year initial rate fixation (% of total loans, left-hand scale)differential between long-term and short-term MFI interest rates (% points, right-hand scale)

78

80

82

84

86

88

90

92

2003 2004 20050.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3

Chart 2.9 Credit standards applied to theapproval of loans to non-financial corporations –small and medium-sized versus large enterprises(Q1 2003 - Q3 2005, net percentage changes)


-20

-10

0

10

20

30

40

50

60

-20

-10

0

10

20

30

40

50

60

Q32005Q2Q1Q4Q3

2004Q2Q1Q4Q3Q2Q1

2003

largesmall and medium


credit standards. This occurred against abackground of compressed interest marginsand strong competition, as well as perceptionsof a deterioration in the economic outlook.

The perception that banks may have becomemore sensitive to taking on additional creditrisk is reflected, to some extent, in the spread ofmonetary financial institution (MFI) interestrates on loans to SMEs over comparable marketrates. By August 2005, this spread hadincreased by some 40 basis points compared toits lowest point in mid-2004 (see Chart S67).

MARKET INDICATORS OF CORPORATE SECTORFRAGILITYMarket-based indicators have continued to pointtowards improved credit risk assessments for thenon-financial corporate sector over the past sixmonths. The distribution of expected defaultfrequencies (EDF) – a market-based indicator ofthe probability of default over a 12-monthhorizon – for the non-financial corporate sectorbecame significantly more compressed at lowerlevels in September 2005 (see Chart S32).

The better assessment of non-financialcorporate sector credit risks can largely beattributed to a notable improvement in theassessment of default expectations for largeeuro area firms. While the EDFs for smallercorporations have also improved in the last sixmonths, this was far less pronounced (see ChartS33). Nevertheless, to the extent that theseforward-looking indicators can provide anindication of the future performance of loans tothe non-financial corporate sector, thecorporate credit risk outlook for banks appearsto be improving.

CORPORATE SECTOR RISKSThe strengthening of corporate sector financialpositions over recent years has beenacknowledged in rising equity prices, atightening of credit spreads (see Section 3 onthe euro area financial markets) and decliningEDFs. Moreover, the ratio of credit ratingupgrades to downgrades became balanced inlate 2004 for the first time since Q3 1998 (see

Chart 2.10 European non-f inancialcorporate sector downgrades,upgrades and balance(Q1 1995 - Q3 2005, numbers)

Source: Moody’s.

-50

-40

-30

-20

-10

0

10

20

-50

-40

-30

-20

-10

0

10

20

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

upgradesdowngradesbalance

Chart 2.10), reflecting perceptions ofimproving creditworthiness.

Looking ahead, the main risks facing the euroarea non-financial corporate sector continue tobe the risk of a prolonged period of high oilprices as well as downside risks to economicgrowth. To some extent, the rise in oil pricesappears to be reflected in market analysts’expectations regarding corporate sector profitgrowth, which seems set to deceleratesomewhat. As recent patterns of profitabilityacross industrial sectors have shown, the firmsat greatest risk should economic activity proveweaker than expected are likely to be in sectorsthat are more domestically oriented, such asSMEs (see Box 6).

A further source of risk for the non-financialcorporate sector relates to the relatively highlevels of indebtedness. The tendency of firmsin recent years to shorten the effective maturityof their borrowing could imply a significantdeterioration in corporate balance sheetconditions should short-term interest rates rise.

53cECB



ENVIRONMENTBox 6

CORPORATE EARNINGS AND SECTORAL EXPOSURE AT RISK IN THE EURO AREA

An important determinant of corporate sector creditworthiness is the (expected) profitabilityof firms. When firms’ profitability begins to improve, the availability of internal sources offinance also rises, and often this is associated with, and even anticipated by, narrowingcorporate bond spreads (see Chart B6.1). In early 2005 there were some signs of a decelerationin the rate of profit growth of stock exchange-listed firms. Since aggregate figures may hidedifferences at the sectoral level, and because banks may have different exposures to differentcorporate sectors, this Box examines corporate earnings at a sector level, making links to thesectoral exposure at risk of euro area banks.

Although operating earnings growth in theeuro area in the first three quarters of 2005was very strong, there were substantialdifferences in performance across sectors (seeTable B6.1). The early product chain sectors,such as resources (RES), basic (BI) andgeneral industries (GI), demonstratedcomparatively strong earnings growth,despite high and rising oil prices. Theearnings growth of the financial sector (FIN)was also strong, albeit consistently lower thanthat of the non-technology, media andtelecommunications (TMT) sector in 2004and 2005. Most striking is that the annualgrowth rate of reported earnings in the

Chart B6.1 Corporate earnings growth andbond spreads in the euro area

Sources: Thomson Financial Datastream and Merrill Lynch.

-20-15-10

-505

1015202530

1999 2000 2001 2002 2003 2004 2005

50

100

150

200

250

300

corporate earnings growth (left-hand scale)BBB-rated corporate bond spread (inverted right-hand scale)

RES BI GI CC NCC CS NCS UTI IT FIN TMT Non-TMT

Q1 2004 13.3 -15.8 -14.0 -5.9 0.9 -11.8 49.4 5.3 -9.8 -6.9 26.7 -5.2Q2 2004 11.5 -3.6 24.1 -6.0 0.4 -9.0 22.9 13.0 0.3 0.8 14.1 2.8Q3 2004 16.1 2.1 19.9 8.9 14.4 1.0 23.4 11.5 13.9 4.2 18.2 7.7Q4 2004 25.5 16.9 49.6 8.1 23.6 27.3 10.9 16.9 3.9 2.8 16.6 13.0Q1 2005 29.5 35.3 48.6 8.7 21.4 32.4 7.0 27.8 6.5 11.7 12.0 20.8Q2 2005 40.0 48.9 30.5 15.9 30.1 44.3 14.6 33.2 8.5 15.2 18.8 25.3Q3 2005 31.4 39.4 33.1 5.5 13.9 41.2 -2.4 23.2 5.0 14.4 8.7 19.9

Table B6.1 Reported earnings growth of stock market- l isted companies broken down bysector(% per annum, average of monthly data)

Source: Thomson Financial Datastream.Note: 2005 Q3 up to August. RES = Resources, i.e. mining, oil & gas; BI = Basic industries, i.e. chemicals, construction &building materials, forestry & paper, steel & other metals; GI = General industrials, i.e. aerospace & defence, diversif iedindustrials, electronic & electric equipment engineering & machinery; CC = Cyclical consumer goods, i.e. automobiles,household goods & textiles; NCC = Non-cyclical consumer goods, i.e. beverages, food producers & processors, health,packaging & printing, personal care & household products, pharmaceuticals, tobacco; CS = Cyclical services, i.e. distributors,general retailers, leisure, entertainment & hotels, media & photography, restaurants, pubs & breweries, support services,transport; NCS = Non-cyclical services, i.e. food & drug retailers, telecommunication services; UTI = Utilities, i.e. electricityand gas distribution; IT = Information technology, i.e. information and technology hardware, software & computer services; FIN= Financials, i.e. banks, insurance, life assurance, investment companies, real estate, speciality & other f inance; TMT =Technology, media and telecommunications; Non-TMT = Other than TMT.


cyclical consumer goods (CC), non-cyclical services (NCS) and IT sectors slowed downsignificantly in the course of 2005.

Given that earnings growth matters for the creditworthiness of firms, it is of interest to considerthe patterns of exposure at risk of euro area banks to the different sectors (see Table B6.2) Thelargest exposure at risk of euro area banks at a sectoral level in early 2005 was the consumercyclical sector, a sector which at the same time was showing rather low earnings growthfigures. Even the next two sectors to which euro area banks had large exposures in early 2005 –i.e. the financial and TMT sectors – endured a slowdown in earnings growth through 2005.Between early 2004 and early 2005, euro area banks increased their exposure at risk to thefinancial sector, whereas they reduced their exposure to the early-product chain sectors, whichsubsequently recorded very strong earnings growth figures in the course of 2005.

2.3 BALANCE SHEET CONDITIONS OF THEHOUSEHOLD SECTOR

Lending to households represents an importantshare of total bank lending in the euro area: as aproportion of the total loans outstanding ofeuro area banks, households account for almostone-third.2 This means that the condition ofhousehold sector balance sheets is importantfor the financial condition of banks. In

2 This f igure, which includes lending to individual enterprises, isbased on unconsolidated MFI data on outstanding amounts ofloans for June 2005.

BIC EUTI CAP CC NCC FIN TMT

Total exposure EUR billions 609 224 224 1,682 805 6,360 218Sectoral EDF % probability 0.25 0.08 0.50 0.46 0.23 0.07 1.13Exposure at risk EUR billions 1.52 0.18 1.12 7.74 1.85 4.45 2.47

in % all sectors 7.9 0.9 5.8 40.0 9.6 23.0 12.8

% change March 2005 – June 2004 -66.9 -87.6 -50.3 -14.5 32.1 46.3 -14.5

Table B6.2 Euro area bank exposure at r isk broken down by sector

(March 2005)

Sources: Banking Supervision Committee and ECB calculations.Note: The euro area refers to the sum of nine euro area countries with only large exposure data for Finland and no data forGreece, the Netherlands and Luxembourg. BIC = Basic industry and construction; EUTI = Energy and utilities; CAP = Capitalgoods; CC = Consumer cyclical goods; NCC = Non-cyclical consumer goods; FIN = Financials; TMT = Technology, media, andtelecommunications.

All in all, data on sectoral earnings growth, together with bank exposure at risk data for the euroarea, show that the greatest exposures of euro area banks have been to sectors that have beenenduring decelerating rates of growth in earnings (cyclical consumer goods, financial andTMT sectors). By contrast, exposures have been lower to sectors which have shown very strongearnings growth (energy and utilities, basic industry and capital goods sectors). Lookingahead, it appears that euro area banks will most notably be exposed to sectors that show lesscapacity of generating internal funds and thus a lower level of creditworthiness. Consequently,it cannot be excluded that the general decline in euro area loan loss provisions may only proveto be temporary.

addition, an important contributor to bankingsector profitability over recent years has beenlending to households for house purchase.Hence, any deceleration in mortgage lendinggrowth would impinge on banking sectorprofitability.

55cECB



ENVIRONMENTSo far, there has been little sign of deceleration,but increased competition in the mortgagelending market – spurred to some extent by thefavourable risk-weighting of mortgage loanswithin the Basel II framework – has entailedsome narrowing of euro area banks’ interestrate margins on mortgages.

In the third quarter of 2005, the pace of banklending growth to euro area householdsremained strong (see Chart S36). The bulk ofthis growth in lending was for house purchase,as banks narrowed their margins on housingloans against a background of reduced concernas to future housing market prospects, asindicated in the October 2005 ECB BankLending Survey. This narrowing of marginswas mainly due to the strength of competitionin mortgage lending, although banks didtighten their lending standards with regard toriskier mortgage loans. Another factor thatcontributed to the strength of overall lendinggrowth to households was a pick-up inconsumer lending growth. This brought thehousehold sector debt-to-GDP ratio to a newhigh of 57% in the third quarter of 2005 (seeChart S34).

Notwithstanding the rise in the euro areahousehold sector debt-to-GDP ratio, thesector’s indebtedness has remained low byinternational standards.3 From a financialstability perspective, it is not the level of debtthat matters per se, but the sustainability of agiven level of debt. Ultimately this depends onthe ability of households to service outstandingobligations out of income and, possibly, assetsin the case of adverse disturbances to income.

In terms of the ability of households to honourtheir obligations out of income, the total debtservicing burden of the household sector(repayment of the principal and interestpayments) is estimated to have remainedlargely unchanged since 2000, at around 11%of disposable income (see Chart S37). Thisratio has remained stable because the increasein household sector indebtedness was offset bythe fall in interest rates to low levels.

3 See ECB (2005), Financial Stability Review, June, Box 6.4 See Special Feature C in this issue of the FSR on “Assessing the

f inancial vulnerability of euro area households using micro-level data”.

Debt sustainability is also influenced by thefeatures of mortgage loans. In particular, otherfactors that may have lowered debt servicingburdens have been a tendency towards greaterflexibility in repayment terms and alengthening of the average maturity of loans(see Box 7). There is also evidence thatmortgage-indebted households in the euro area– i.e. those carrying the bulk of the householdsector debt – have tended to be in the highestincome categories.

Turning to the ability of households to repaydebt out of assets, indicators of householdsector solvency, such as ratios of debt to liquidfinancial assets and debt to total financialassets, have remained comfortable (see ChartS35). In other words, the overall householdsector had sufficient liquid funds and otherfinancial assets available to repay loans ifneeded. At a micro level, there is someevidence for the euro area that high-incomehouseholds, which carry most of the debt, heldmore assets and liquid assets than others overthe period 1994-2001. Moreover, over thatperiod, their ability to save has tended to behigher.4

HOUSEHOLD SECTOR RISKSThe main sources of risk for householdsector balance sheets that can have financialstability implications comprise unexpecteddisturbances to household incomes, propertyprices, and interest rates. In particular, anydeterioration in financial resources (incomeflows) or financial commitments (repaymentburdens) could imply heightened credit risksfor banks.

Compared with the June 2005 assessment, risksfacing euro area households are broadlyunchanged. However, banks responding to theOctober 2005 Bank Lending Survey reported areduced perception of risk, in terms of


expectations regarding general economicactivity and housing market prospects, in thethird quarter of 2005.

UNCERTAINTY REGARDING INCOMEEXPECTATIONS AND THE MACROECONOMYEvidence based on household-level data tendsto suggest that – at least for the period1994-2001 – changes in households’vulnerability appear to have been notablyinfluenced by developments in householdincome. This, in turn, is driven bymacroeconomic developments and is closelyrelated to developments in the labour market.Therefore, the uncertainty surrounding theoutlook for economic activity in the euro areacould have implications for employmentprospects and the disposable income of euroarea households.

Survey results from the European CommissionConsumer Survey for the second and the thirdquarters of 2005 showed a slight deteriorationin the way in which euro area householdsexpect their financial situation to develop overthe next 12 months. The main reason for thisappears to be the perception that employmentprospects will not improve significantly in the

near future (see Chart 2.11). It also cannot beruled out that the financial situation of somehouseholds, especially in the lowest-incomesegments, could be tested by the strength of oilprices.

From a financial stability viewpoint, theimpact of any deterioration in household sectorincome will ultimately depend on thedistribution of debt across different categoriesof households, which face different financialconditions. Although timely estimates are notavailable, there are some indications that thebulk of household sector debt in the euro areahas tended to be held by households in thehighest income categories, and this distributionof indebtedness tends to change slowly overtime.

RISKS TO HOUSE PRICESEuro area residential property prices remaineddynamic during late 2004 and into 2005. Inboth nominal and real terms, for the euro areaas a whole, the recent increases are of the samemagnitude as the increases during the lasthousing market peak in the early 1990s (seeChart 2.12).

Chart 2.11 Euro area households’ f inancials ituation and unemployment expectations

(Q1 1998 - Q3 2005, three-month moving averages ofpercentage balances)

Source: European Commission Consumer Survey.Note: “Balance” refers to the percentage of positive answersminus the percentage of negative answers. An increase in anegative balance indicates less pessimistic expectationsoverall.

-4048

121620242832364044

1998 1999 2000 2001 2002 2003 2004 2005-6

-4

-2

0

2

4

6

unemployment expectations over the next 12 months (left-hand scale)expectations about households’ financial situation over the next 12 months (right-hand scale)

Chart 2.12 Residential property pricegrowth rate in the euro area

(1991 - 2004, % per annum)

Sources: National data and ECB calculations.Note: Real series calculated using the euro area HICP as adeflator. Euro area series constructed from national seriescovering more than 90% of the euro area GDP.

nominalreal

-2

0

2

4

6

8

10

-2

0

2

4

6

8

10

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

57cECB



ENVIRONMENTThis period of strong mortgage lending growthto households in the euro area has in recentyears coincided with rising house prices inmany euro area countries. Lending for housepurchase grew at an annual rate of 8.1% in 2003and 10% in 2004, and available data for loansfor house purchase show a continuation of thispace of growth in 2005 (see Chart S36). At thesame time, nominal house prices increased by7.1% in 2003 and by 7.4% in 2004. For the euroarea as a whole, favourable financingconditions and, possibly, the expectation ofcapital gains appear to have contributedsignificantly to the recent strength of housingdemand, rather than the growth of householddisposable income.

The underlying dynamics of the overall euroarea index reflect mixed developments at theMember State level. This continues to indicatethat there is a strong national dimension tohousing markets. Fundamental factors such asreal economic activity, demographic changes,policy measures as well as short-run stickinessin supply have contributed to the recentdivergent price dynamics in euro area MemberStates. The indications are that house prices

Chart 2.13 Bui lding permits and residentialinvestment in the euro area

(Q1 1999 - Q2 2005)

Sources: National sources and ECB calculations.Note: Germany is excluded owing to the effect ofreunif ication on aggregate residential investment.

building permits granted (% per annum, 3-months moving average, left-hand scale)residential investment (% of GDP, right-hand scale)

-8

-4

0

4

8

12

16

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

4.5

4.6

4.7

4.8

4.9

5.0

5.1

Chart 2.14 House price-to-rent ratio for theeuro area

(1999 - 2004, index: 1999 = 100)

Sources: National sources and ECB calculations.

euro areaGermanyFranceItalySpainNetherlands

8090

100110120130140150160170180

8090100110120130140150160170180

1999 2000 2001 2002 2003 2004

continued to rise in early 2005 in several euroarea countries, especially in Spain, France andIreland. By contrast, in the Netherlands, houseprices grew at a moderate rate in early 2005.

Concerning the likely future supply ofresidential property, which is an importantfactor for future house price developments, thepicture has been somewhat mixed. On the onehand, some indicators, such as building permitsgranted and the number of house completions,show that the supply side of the housing markethas partly responded to the strength of demand(see Chart 2.13). On the other hand, data thathave become available in the six months sincethe June 2005 FSR have shown a decelerationin the growth of construction investment.

Owing to the importance of mortgage lendingas a revenue source for banks, developments inresidential property prices have become animportant factor in the assessment of risks tobanking sector stability. In this vein,concerning the valuation of house prices in theeuro area, faster growth in house prices thanrents has meant rising house price-rent ratios inseveral countries (see Chart 2.14).


To the extent that house price-rent ratios haverisen above historical values, this may notnecessarily imply an immediate and imminentrisk of downward adjustment. Adjustmenttowards intrinsic values could come eitherthrough rising rents or through falling houseprices. However, if house prices were todecline, the immediate impact on banks wouldbe felt through income generated frommortgage lending activities and through creditquality. There could also be second-roundwealth effects as households changeconsumption patterns; however, the empiricalmagnitude of these effects in the euro arearemains uncertain.

For the euro area as a whole, an uncontrolledcorrection in house prices does not appearlikely in the short run. Despite this, thecontinued strength of house prices in someMember States calls for ongoing monitoringand surveillance.

According to the results of the October 2005ECB Bank Lending Survey, banks continued tosee housing market prospects as a reason fortightening lending standards on the granting ofmortgage loans. However, they reported a netdecrease in the perception of the related risk inthe third quarter of 2005. Over the same period,households’ more optimistic view regardinghousing market developments contributed to anincrease in the net demand for housing loans.

INTEREST RATE RISKOverall, the interest rate risk facing householdsin the euro area has not changed in the sixmonths since the June 2005 FSR.

From a financial stability viewpoint, theimpact of any change in interest rates onhouseholds’ debt sustainability dependsnotably on country-specific mortgage contractfeatures, in particular the interest ratevariability regime of the outstanding loans.ECB estimates suggest that around one-third ofthe total outstanding mortgage debt in the euroarea could be exposed in the short run to achange in interest rates.5

The balance sheets of new borrowers could bemore sensitive as a greater proportion of themare indebted at variable rates. While this mightbe true for some individual countries, at theeuro area level, however, the share of newmortgage loans at floating rates and with aperiod of interest fixation of up to one year hasrecently decreased (47% in August 2005,compared with a peak of 59% in November2004). At the same time, the share of newmortgage loans with a period of interestfixation of over ten years had, by August 2005,increased to its highest level over the last twoyears, standing at 22%.

Finally, a factor mitigating the overallsensitivity of household mortgage debt tointerest rate changes might be the increaseddiffusion of variable rate products, whereby anincrease in the interest rate translates into alonger repayment period, with the monthlypayment remaining unchanged (see Box 7).However, this could be seen as merelypostponing the increase in the burden, raisingthe issue of outstanding debt sustainabilityover a longer time period.

ASSESSMENT OF HOUSEHOLD SECTOR RISKSAll in all, even though euro area householdsector indebtedness has continued to rise, itdoes not appear to constitute a threat to thestability of the euro area financial system in thenear term. Debt servicing burdens haveremained stable, and aggregate householdsector solvency remains comfortable.Moreover, indications are that the most heavilyindebted households in the euro area also tendto be those in the highest income categories.

Looking ahead, the risks facing the euro areahousehold sector as a whole have not changedsignificantly in the past six months, the mainshort-term risk being related to macroeconomicdevelopments, especially income prospects.

5 This is due to the importance in many countries of f ixed orquasi-f ixed interest rate mortgages (i.e. with long initialperiods of interest rate fixation). See ECB (2004), FinancialStability Review, December, Box 6.

59cECB



ENVIRONMENTHowever, some areas of vulnerability remain.The distribution of risks to financial stabilityemanating from the household sector arespread unevenly throughout the euro area.Differences exist across countries in terms ofdebt levels, contractual interest ratevariability, the length of loan terms, and houseprice developments. It cannot be ruled out thatat least some of the cross-country differencesin debt may be due to different equilibrium debtlevels.

Some of the countries that have experiencedsubstantial increases in house prices in recent

years have a high proportion of variable ratedebt. This may amplify the effects of anyinterest rate changes, especially for householdswith high levels of outstanding debt, lowhousing equity, low financial asset buffersand/or uncertain employment and incomeprospects. Moreover, a substantial reversal ofhouse prices may entail capital losses with animpact on household balance sheets, whichcould also indirectly affect banks. Hence, thecontinued strength of house prices in someMember States calls for ongoing monitoringand surveillance.

Box 7

THE INFLUENCE OF MORTGAGE PRODUCT INNOVATIONS ON RISKS TO HOUSEHOLD DEBTSUSTAINABILITY

In an environment of strong competition, banks in the euro area have been offering newmortgage products targeted at a larger number of borrowers. With these new products, twoprevious obstacles to borrowing have been removed. First, it is now possible in some countriesfor households to borrow higher amounts with little or no down payment, through higher loan-to-value ratios. Second, in a number of countries, other products have become available,allowing middle and lower-income borrowers to alter repayments relative to their financialresources, while borrowing larger amounts than might have been possible in the past. This hasmainly been achieved by extending the average loan maturity (up to 30-35 years in somecountries). This Box reviews the specific features of these mortgage products and theirimplications for the sustainability of household debt.1

In many euro area countries, banks are increasingly offering a variety of types of innovativemortgage products. First, “accordion” variable rate mortgages offer the option of keeping themonthly instalment constant, even in the case of a change in the interest burden, the adjustmentbeing made through an extension of the loan maturity. In the euro area, such products exist inBelgium, Italy, Spain, France and Greece. Second, mortgage products are now increasinglyoffering a wide range of flexible repayment options (such as deferred start, payment break orreduced starting payments), allowing borrowers to match their repayments to their cash flows,which can be affected by seasonal increases in expenses (for instance, a “payment holiday” canbe granted for one or two months during the summer or at the end of the year). Finally,“interest-only” or “amortisation free” mortgage loans allow the deferral of the payment of theprincipal for a given period or even until the end of the loan.

According to a recent study,2 interest-only products are now available in most euro areacountries (with the exception of Finland; no information was reported for Austria, Greece and

1 Other types of mortgage products have recently appeared, such as equity release loans, foreign currency loans and reversemortgages. However, this Box focuses on innovations that have the greatest impact on households’ monthly repayment burden.

2 See London Economics (2005), “The Costs and Benefits of Integration of EU Mortgage Markets”, Report for the EuropeanCommission, DG Internal Market and Services, August.


3 See P. J. A. van Els, W. A. van den End and M. C. J. van Rooij (2005), “Financial Behaviour of Dutch Households: Analysis of theDNB Household Survey 2003”, BIS Papers No 22 – Investigating the Relationship between the Financial and Real Economy, April.

Luxembourg). In the Netherlands survey results indicate that 41% of outstanding mortgageswere interest-only in 20033; they also tend to be more common among lower-incomehouseholds. However, it is unlikely that interest-only mortgage loans in the Netherlands aregranted to finance the total value of the property. They are rather often used in combinationwith another type of loan, or as a second mortgage, for instance to finance renovations.Moreover, Dutch banks tend to grant interest-only mortgages with rather conservative loan-to-value ratios. In Spain, most mortgage lenders now offer a wide range of products with moreflexibility in repayment schemes. They have recently started to grant mortgages under whichborrowers pay only interest for a period of one to three years. In France, loans with a deferredcapital repayment are only granted in special cases (e.g. subsidised loans and student loans).Interest-only loans, whereby the repayment occurs at the end of the loan duration, are mostlygranted to investors for buy-to-let purposes, to take advantage of particular fiscal schemes.They are often offered together with an investment product, allowing the lump sum forrepayment to be built up.

Typically, interest-only mortgage products were originally designed for wealthier households,which tended to use them as a cash management tool – investing the cash freed up during thisperiod at a higher return – and which were able to sell, if necessary, financial assets to pay offthe loan amount. They were also suited for households with irregular income, but able to makevoluntarily early principal repayments when they have more income, or for young householdsexpecting their income to rise sharply in the near future. However, for many “ordinary”borrowers, such flexible mortgage products have now become the best financing option,allowing them to overcome the financial hurdle to home ownership brought about by the recentincrease in house prices, and to adapt their repayments to the pattern of their financialresources.

However, innovative mortgage products do potentially contain certain specific risks. A longerloan duration and amortisation period entail a higher probability that the household could facedebt sustainability problems, for instance caused by a period of unemployment with a lowerincome, or loss of income altogether. With regard to interest-only loans, they might be a goodchoice for buyers intending to move or refinance – and therefore repay the principal – before orat the end of the interest-only period. However, after the initial amortisation-free period,borrowers could face a sudden, sharp increase in their financial burden for which they might beunprepared. Moreover, should house prices decline, there is a higher risk that householdswould be left with low or even negative net housing equity, the outstanding balance of the loanexceeding the value of their houses. Finally, the total amount of interest paid will be higherover the term of the loan.

61cECB



ENVIRONMENTIn the euro area, quantitative information on these new mortgage products is scarce, making itdifficult to assess the overall financial stability implications. However, some lessons can bedrawn from recent developments in the US, where the growing popularity of interest-onlymortgages has recently raised financial stability concerns.4 Interest-only mortgages (deferringprincipal payments for a period of three to ten years) are now being offered by most lenders,and represented a third of home purchase loans originated in 20045, up from 5% in 2003.However, the amortisation-free period is substantially longer than in the euro area (up to halfthe total duration of the loan), potentially resulting in a high increase in the monthly paymentsat the end of this initial period (anecdotal evidence suggests that the monthly payment couldjump by 50%, even in the absence of any interest rate rise).

Available products in the US also include option adjustable rate mortgages (ARMs), or flexibleARMs, allowing the borrower to choose a repayment scheme whereby payments in the initialperiod (five to ten years) might cover in extreme cases only a part of the interest payment, theremainder being added to the outstanding loan balance to be repaid later. There are alsoconcerns that these higher-risk ARMs are increasingly being offered to riskier borrowers, whomay face greater difficulties adjusting to the rise in their monthly payments at the end of theinitial period. However, at present such products, which could potentially result in a“negative” amortisation of the loan (meaning that the outstanding balance increases over timeinstead of decreasing, as a result of accumulated deferred interest payments), do not appear tobe available in the euro area.

From a financial stability viewpoint, while the innovative mortgage products that arebecoming increasingly available in the euro area allow households to keep their monthly debtservicing burdens at reasonable levels in the short run, longer-term risks could be increasing,especially as the ability of households to make large principal repayments after a considerableperiod of time is largely untested. This would call for closer monitoring of how the nature ofrisk-sharing in mortgage lending is being altered by product innovation.

4 See Federal Reserve Board, Monetary Policy Report submitted to Congress on 20 July 2005, which states that “Recently there hasbeen increased use of potentially riskier types of mortgages, including adjustable-rate and interest-only loans, which could posechallenges to both lenders and borrowers.”

5 According to data from the real estate information firm Loan Performance (see for instance the annual report on “The State of theNation’s Housing 2005”, issued by the Joint Center for Housing Studies at Harvard University). These data refer to loans packagedfor resale as mortgage-backed securities, and thus do not cover the entire market.

63ECB


I I I T H E E URO A R E A F I N ANC I A L S Y S T EM3 EURO AREA FINANCIAL MARKETS

Since the June 2005 FSR, conditions in the euroarea money market have remained favourable,as perceptions of counterparty credit riskswere rather low and participants couldsmoothly manage their liquidity needs. Despitean aggressive global hunt for yield, yields inthe euro area bond markets appear to be less ofa conundrum than in the US. However, anysignificant upturn in long-term bond yields,either possibly induced by a reappraisal ofrisks or transmitted from the US bond markets,remains an important source of risk for thefunctioning of the euro area capital markets. Inaddition, a rise in euro area short-term risk-free interest rates would be unlikely to leavethe bond markets, especially the corporatesegments, unaffected. In turn, upside risks torisk-free interest rates and the credit riskpremium could also lead to a revaluation ofpricing in euro area equity markets, especiallyin connection with a disorderly unwinding ofglobal imbalances, and given that the earningscycle is probably close to its peak.

3.1 KEY DEVELOPMENTS IN MONEY MARKETS

MONETARY POLICY RATES REMAIN UNCHANGEDIN THE EURO AREAFrom a financial stability viewpoint,conditions in the euro area money market areimportant for at least two reasons. First, theECB implements its monetary policy in thismarket, and effective implementation requiresthe market to function smoothly. Second, banksusually secure their liquidity needs in thismarket. Because of this, the smoothfunctioning of the money market can contributeto the stability of the banking system as awhole.

There has been no change in the level ofmonetary policy interest rates in the euro areasince June 2003, with the minimum bid rate forthe main refinancing operations remaining at2% since then. At the time of the June 2005FSR, market participants had been expectingthat the ECB would raise its interest rates by

25 basis points in the second quarter of 2006.However, supported by some recent economicgrowth data and upward revisions in theforecasted inflation rates, market participants’expectations of a rate hike by the ECB haverecently been brought forward. By earlyNovember 2005, expectations of a 25 basispoint interest rate hike by the ECB, as derivedfrom money market derivatives prices, hadshifted to the first months of 2006.

GENERAL MONEY MARKET CONDITIONS REMAINFAVOURABLEFrom a financial stability viewpoint, generalmoney market conditions in the euro areamoney market remained favourable for thefollowing three reasons.

First, perceptions of counterparty credit risksin euro area money markets remained ratherlow. These perceptions can be revealed inpatterns of interest rate spreads betweenuncollateralised interbank money market ratesand collateralised repo rates. Fears that Fordand GM would be downgraded by credit ratingagencies, followed by the actual event,appeared to have some, albeit limited andshort-lived, impact on very short-term spreadsin the early months of 2005 (see Chart S38).Overall, these spreads have changed little andhave remained low across all short-termmaturities over the past six months.

Second, the secured money market segment1

continued to grow relative to the unsecuredsegment, and is now the largest money marketsegment in the euro area, as the 2004 ECB EuroMoney Market Survey revealed (see Box 8).This is a positive development from a financialstability viewpoint, as it suggests thatmoney market counterparties are showingan increasing preference for limiting theircounterparty risks.

Third, liquidity conditions in the euro areamoney market remained favourable. Liquidityconditions across different segments and

1 Often referred to as the repo market segment.


maturities of the money market can beevaluated by the monitoring of patterns in bid-ask spreads (see Chart S39). Already low inearly May 2005, bid-ask spreads at the one andthree-month maturities for EONIA swap rateshave narrowed further over the past six months.While the further narrowing of these spreadscan be seen as a sign of high market liquidity,suggesting that, on aggregate, marketparticipants have faced little difficulty inaccessing short-term funding, it also reflectsthe increasing use of electronic tradingplatforms. As discussed in the June 2005 FSR,it cannot be excluded that, by eroding buffersfor market movements, very tight bid-offerspreads could adversely affect risk-returntrade-offs in market-making activity.2

3.2 KEY DEVELOPMENTS IN CAPITALMARKETS

GOVERNMENT BOND MARKETSIn the euro area fixed income markets, afterMay 2005, long-term government bond yields –which set a benchmark for the level of longer-term risk-free interest rates – reached historicallows of 3.1% in September 2005 against abackground of high oil prices, and in theimmediate aftermath of Hurricane Katrina.

Thereafter, long-term government bond yieldsedged up to around 3.5% in early November2005.

Underlying the drop in nominal euro area long-term bond yields in the course of 2005 has beena significant decline in real long-term interestrates to very low levels. Rising oil prices seemto have been one of the important factorsaccounting for the drop in long-term real rates(see Chart 3.1). The rise in oil prices appearedto lead to concerns among market participantsabout prospects for economic growth ratherthan inflation. Apart from the negative impacton real yields through lower euro area growthexpectations, the rise in oil prices may alsohave led to a strengthening of demand for long-term bonds by oil-exporting countries, throughinvesting their oil revenues to some extent ineuro-denominated bonds.

By contrast with patterns in the US marketyield curve, which almost became inverted inlate 2005, the euro area yield curve remainedcomparatively steep (see Chart 3.2). As aconsequence, euro area yield curve carry tradesbecame relatively attractive, especially when

Chart 3.1 Euro area 2015 real bond yie ldand oi l pr ice

(Jan. 2005 - Nov. 2005)

Sources: BIS, Bloomberg, Reuters and ECB calculations.

25

30

35

40

45

50

55

60

65Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct. Nov.

2015 inflation-indexed linked bond (%, left-hand scale)oil price (EUR per barrel, inverted right-hand scale)

1.7

1.6

1.5

1.4

1.3

1.2

1.1

1.0

0.9

2005

Chart 3.2 Euro area and US yie ld curves

(%)

Sources: Reuters and ECB calculations.Note: The yield curves are estimated using an extension ofthe Nelson-Siegel method applied to zero-coupon swap rates.

1 M 1 Y 2 Y 3 Y 4 Y 5 Y 6 Y 7 Y 8 Y 9 Y 10 Y

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.0

5.5

5.0

4.5

4.0

3.5

3.0

2.5

2.0

early November 2005 euro areaearly November 2005 US

early May 2005 euro areaearly May 2005 US

2 See ECB (2005), Financial Stability Review, June, Box 7.

65cECB


III THE EURO AREAFINANCIAL

SYSTEMinvestors did not exclude the possibility ofdownward pressure on the US dollar stemmingfrom wide external imbalances (see Section 1).The steepness of the euro area yield curve has

also served to support the interest rate marginsof euro area banks, as banks traditionally fundlong-term assets (e.g. loans) with short-termliabilities (e.g. deposits).

Box 8

STRUCTURAL TRENDS IN THE EURO MONEY MARKET

The fifth study on the structure and functioning of the euro money market was conducted inearly 2005 by the ECB together with the 15 national central banks (NCBs) that were membersof the European System of Central Banks (ESCB) before 1 May 2004.1 The study was based onturnover data collected from banks covering the second quarters of 2003 and 2004. This Boxreports on some of the main findings of this study, and draws attention to the following threemain findings. First, although changes in the aggregate turnover of the euro money market werelimited between the second quarters of 2003 and 2004, there were some notable compositionalchanges, especially in terms of growing activity in secured relative to unsecured market.Second, overall activity in the euro money market became less concentrated, although largedifferences have remained across market segments. Third, the growing use of electronicplatforms to make transactions in many market segments produced a further narrowing of bid-offer spreads.

Following the rise that took place between 2000 and 2003, aggregated turnover in the euromoney market appeared to stabilise in Q2 2004 (see Chart B8.1). This was due to offsettingpatterns in different segments of the market. Although activity fell in the overnight index swap(OIS) market as well as in the cross-currency and foreign exchange swaps markets, there was arise in turnover in the unsecured, secured, other interest rate swap (IRS), forward rateagreement (FRA) and short-term securities segments (see Chart B8.2). Furthermore, as in Q22003, the secured segment remained the largest money market segment in Q2 2004, accountingfor around 36% of total market turnover.

The increasing share of secured transactions can be seen positively from a financial stabilitypoint of view, as it shows that market participants have a preference for limiting their creditrisk exposures. A further development related to the increase in the secured market segmentwas a substantial increase in tri-party repo activity.2 This also reduced the counterparty andoperational risks related to settlement.

The degree of concentration in money market activities can provide an indication of both themarket’s dependency on individual institutions and the risks for market functioning if asignificant counterparty were forced to exit. There are indications that the overall level ofactivity in the euro money market has become less concentrated in recent years. Nevertheless,degrees of concentration vary widely across different market segments. The least concentratedsegment of the money market in Q2 2004 was the unsecured segment, where the ten most active

1 See ECB (2005), Euro Money Market Study 2004, May. The study is based on data received from a sample of credit institutions,implying that the f indings must be interpreted with caution, as they are not necessarily representative of the euro money market asa whole.

2 A tri-party repo is a repo that involves a third party, commonly a custodian bank, acting as an agent to exchange cash and collateralfor one or both counterparties.


institutions accounted for around 35% of the total turnover. However, some segments haveremained highly concentrated. For instance, the FRA, other IRS and cross-currency swapsegments remained highly concentrated: the ten most active institutions in each of thesesegments accounted for around 70% of total turnover (see Table B8.1).

Concerning the integration of the euro money market, the most relevant change in thegeographical counterparty structure in Q2 2004 was the loss of predominance of transactionswith national counterparties for short-term securities. Indeed, in the short-term securities (andcross-currency swap) segments, cross-border transactions with other euro area counterpartiesbecame the highest among all market segments (see Chart B8.3). From a financial stabilityviewpoint this development is important as it reduces country-specific risks by spreading risksmore widely. By contrast, however, the share of transactions with national counterpartiesremained relatively high in the secured market segment, indicating that the integration ofnational repo markets across the euro area continued to proceed at a slower pace.

OIS other IRS FRAs FX swaps cross-currencyswaps

top 5 banks 42 62 57 38 52top 10 banks 62 79 78 64 75

Table B8.1 The share of the f ive and ten largest banks in total act iv ity in OTC derivativesmarket(%, Q2 2004)

Source: ECB.

Chart B8.2 Average dai ly turnover by moneymarket segment

Source: ECB.Note: The Q2 2000 unsecured volume is taken as the base. Nodata on FRA turnover were available in 2000.

Chart B8.1 Aggregated turnover in the euroarea money market

Source: ECB.Note: The Q2 2000 volume is taken as a base. No data onFRA turnover were available in 2000.

short-term securitiesFRAcross-currency swapsother IRSFX swapsOISsecuredunsecured

0

25

50

75

100

125

150

175

200

0

25

50

75

100

125

150

175

200

Q2 2000 Q2 2001 Q2 2002 Q2 2003 Q2 2004

Q2 2000Q2 2001Q2 2002Q2 2003Q2 2004

0

25

50

75

100

125

150

175

200

0

25

50

75

100

125

150

175

200

un-secured

secured OIS FXswaps

otherIRS

cross-currencyswaps

FRA short-termsecurities

67cECB



SYSTEM

3 See, for instance, ECB (2005), Financial Stability Review, June.4 The EONIA Swap Indices were calculated for the f irst time on 20 June 2005.

Finally, the following two additional structural developments were observed in the euro moneymarket. First, electronic trading continued to grow in Q2 2004 in most market segments (e.g. inthe secured, IRS and foreign exchange swap segments), and especially in the secured marketsegment (see Chart B8.4). While electronic trading accounted for a very large share of totalactivity in the secured markets, it has remained rather small in most of the other over-the-counter (OTC) derivatives markets.

Overall, the increasing share of electronic trading in some segments of the euro money marketcan be seen as a positive development from the market’s viewpoint, as it can enhance the pricediscovery process and liquidity, reduce operational risks, and lower the cost of trading.However, in the case of abnormal market conditions, the role of electronic trading platforms,especially those of quote-driven systems, can have a destabilising effect on the functioning ofmarkets and can for example lead to sudden withdrawals of liquidity from the markets.3

Second, the creation of the EONIA Swap Index4 by EURIBOR-ACI will probably furtherstimulate the development and enhancement of the overnight swap market segment, since itshould provide a new benchmark for derivatives markets. By improving the choice ofinstruments available to market participants for hedging against, or speculating on interest raterisks, the development of derivatives products in the euro money markets should contributepositively to financial stability by facilitating the dispersion of risks.

Chart B8.4 Trading structure by moneymarket segment

(%)

Source: ECB. Source: ECB.

Chart B8.3 Geographical counterpartystructure by money market segment

(%)

via electronic broker via voice broker direct

un-secured

secured OISFXswaps

IRS cross-currency

swaps

FRA short-term

securities

0

20

40

60

80

100

0

20

40

60

80

100

2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004

otherseuro areanational

un-secured

secured OISFXswaps

IRS cross-currency

swaps

FRA short-term

securities

0

20

40

60

80

100

0

20

40

60

80

100

2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004 2003 2004


Other factors that appear to account for the verylow level of real long-term interest rates in theeuro area include higher demand for assets withlong durations from institutional investors (e.g.pension funds and life insurance corporations)eager to close balance sheet mismatches,motivated in large part by regulatory changesand in anticipation of proposed legislation (seeChart 3.3).

Another structural development which mightcontribute to a low level of bond yields is theincreasing savings activity of the baby boomgeneration and the ageing population ingeneral.3

A further factor that may have contributed tothe lowering of euro area bond yields isgrowing exchange rate diversification awayfrom the US dollar on the part of Asian andother central banks. Some studies suggest thatthis tendency could increase if the US dollarwere to depreciate against the euro in the periodahead.4

Concerning the risks in euro area bond markets,the level of yields appears to be less of aconundrum than in the US, given the steepereuro area yield curve and lower (potential)economic growth expectations. As a corollary,the risk of an unexpected and significant rise ineuro area bond yields would appear to be lowerthan in the US. Nevertheless, it seems unlikelythat a significant upturn in long-term bondyields in the US would leave euro area bondmarkets unaffected.

Indicators of the balance of risks to long-termbond yields in the period ahead as perceivedby market participants pointed throughout2005 to continued concerns about thepossibility of a sudden rise in long-term bondyields. The option-implied skewness – ameasure of the degree of asymmetry in theprobability distribution of likely outcomes –remained significantly positive in late 2005(see Chart S40).

Chart 3.3 Net purchases of long-term bondsand equit ies by euro area insurancecorporations and pension funds(Q1 1999 - Q1 2005, EUR billions, one-year movingcumulative purchases)

Source: ECB.

-20

0

20

40

60

80

100

120

140

160

-20

0

20

40

60

80

100

120

140

160

1999 2000 2001 2002 2003 2004 2005

long-term bond net purchasesquoted shares net purchases

3 For a more in-depth analysis of both structural developments,see “Ageing and Pension System Reform: Implications forFinancial Markets and Economic Policies”, report prepared atthe request of the Deputies of the Group of Ten, October 2005;and A. Maddaloni, A. Musso, P. Rother, M. Ward-Warmedingerand T. Westermann (2005), “Macroeconomic Implicationsof Demographic Developments in the Euro Area”, ECBOccasional Paper, forthcoming.

4 See M. D. Chinn and J. A. Frankel (2005), “Will the EuroEventually Surpass the Dollar as Leading InternationalReserve Currency?”, NBER Working Paper, No W11510,August.

5 See ECB (2005), Monthly Bulletin, June, Box 3.

CORPORATE BOND MARKETSCorporate bond spreads in the euro area at thelower end of the rating spectrum reachedhistorical lows in early 2005, but widenedsignificantly later on (see Charts S47 and S48).While first the fear of and ultimately the actualdowngrading of Ford and GM at the start of2005 played some role in this (see Box 9), thestrong reversal in the course of 2005 can also beattributed to changing fundamentals of euroarea corporations, other firm-specific news,and factors related to market dynamics.5 Forinstance, there have been indications that thecorporate earnings cycle may have turned inearly 2005 (see Box 6). In addition, the pace ofcorporate balance sheet repair has slowed down(see Section 2).

69cECB



SYSTEM

Box 9

DEVELOPMENTS IN THE EUROPEAN CREDIT DERIVATIVES MARKETS

European credit derivatives markets have, like their US counterparts, experienced rapidgrowth in the past few years. If history is a guide, such rapid growth is often accompanied by anincreased potential for instability should conditions take a turn for the worse. This Boxdiscusses the financial stability implications of recent events in these markets.

In May 2005 the credit ratings of GM and Ford, both global car makers and major issuers ofcorporate debt, were downgraded by all three major credit rating agencies: Standard and Poor’sand Fitch lowered their ratings to speculative grade, while Moody’s cut its ratings to the lowestinvestment grade before also classifying them as non-investment grade in August 2005. GM’sratings were downgraded still further following the filing for bankruptcy of its major partssupplier and former subsidiary Delphi Corp. in October 2005. At the time of the firstdowngrades, GM and Ford had global debt outstanding of USD 453.1 billion. According toLehman Brothers, one of the leading providers of bond indices, outstanding bonds from GMand Ford eligible for European high yield index inclusion totalled €12.5 billion and €8.2 billionrespectively, representing 27.7% of the new high-yield European market.

The May 2005 downgrades led to a very sharp, although only temporary, widening of yieldspreads in the credit markets. The sharp spread widening in the cash market was reversedrelatively quickly, and by June 2005, spreads had already retraced part of their widening andhad come back to their April levels, narrowing still further during July and August 2005. Bothissuers actually returned to the market in July, issuing new debt at yield levels not much higherthan before the downgrades. The smooth performance of the cash markets can be attributed to

Chart 3.4 Euro area BBB-rated corporatebond spread and the real short-terminterest rate(Jan. 2001 - Oct. 2005)

Sources: Bloomberg, Eurostat and Reuters.

euro area BBB-rated corporate bond spread (basis points, left-hand scale)euro area real three-month money market interest rate (%, right-hand scale)

50

100

150

200

250

300

350

400

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

2001 2002 2003 2004 2005

To some extent, the widening of spreads oflower quality issuers may also partly reflectmore prudent pricing of corporate default risk,

6 See G. de Bondt (2004), “The Balance Sheet Channel ofMonetary Policy: First Empirical Evidence for the Euro AreaCorporate Bond Market”, International Journal of Finance andEconomics, 9, 3, pp. 219-28.

following a prolonged period where corporatebond markets benefited from an aggressivehunt for yield in an environment of ample(global) liquidity and low real interest rates.Patterns in EDFs, where little upturn was seen,differed from those in corporate bond marketsduring 2005. EDFs are, however, projected tobe higher in 2006 than in 2005.

In the period ahead, a possible trigger for acontinued widening of spreads could be a risein the real short-term risk-free interest rate,since past experience shows a positiverelationship between the BBB-rated corporatebond spread and the real short-term risk-freeinterest rate in the euro area (see Chart 3.4).6


Source: Thomson Financial Datastream.Note: Spread between seven to ten-year yield to maturity andeuro area seven to ten-year government bond yield.

Source: Bloomberg.

Chart B9.1 iTraxx f ive-year CDS spreads

(basis points)

Chart B9.2 Corporate bond spreads in theeuro area

(basis points)

crossoverautoshigh volatilityinvestment grade

Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct.2005

0

50

100

150

200

250

300

350

400

0

50

100

150

200

250

300

350

400

BBBAAAAAA

Jan. Feb. Mar. Apr. May June July Aug. Sep. Oct.2005

0

20

40

60

80

100

120

140

160

0

20

40

60

80

100

120

140

160

several factors, namely: the anticipation of the downgrades, reflected in the spread widening ofthe two issuers and the markets in general since at least mid-March1 (see Charts B9.1 andB9.2); a more flexible management of portfolios by fixed income managers, using morecustomised benchmarks and increased tracking error possibility adopted after the disorderlymarket action of WorldCom’s downgrade in 2002; and the continuing strength of creditfundamentals and corporate earnings, with European high-yield default rates remaining verylow for several years. The effects of the October 2005 GM downgrade, triggered by the DelphiCorp. filing for bankruptcy, on its corporate debt prices may turn out to be longer-lasting, andthe outcome will depend on additional factors (e.g. negotiations with trade unions and futuredevelopments in car sales).

The rapid development of credit derivatives is another reason for the relatively smoothbehaviour of the corporate bond markets. Cash bond investors can effectively unwind theirexposures to individual bond issuers or to entire sectors through the use of credit derivatives.This gives investors the possibility to withstand the immediate impact of possible downgradesand would make it less urgent to liquidate the affected issuers’ bond holdings. On the otherhand, investors involved in trading-oriented strategies typically prefer to stay away from thecash bond market as the credit derivatives markets offer them greater flexibility and liquidity.The resilience of the cash bond markets to adverse market events has thus been strengthenedrelative to the situation before the emergence of the credit derivatives markets.

While it seems that credit derivatives markets have to some extent sheltered the cash marketfrom a rise in volatility after the credit event, some segments of the credit derivatives marketshave themselves experienced significant market upheaval. The launch in June 2004 of a new

1 Further evidence of the market’s anticipation can be found in the increased use of the two issuers’ credit default swap (CDS) contractterminations, which could be observed even before the downgrades were announced, and continued at a strong pace afterwards. Bycancelling “redundant” contracts on dealers’ books without significantly changing the market risk prof ile of their positions, theterminations reduce not only their exposure to the issuers but also legal and operational risks. For a more detailed description of thederivatives contract terminations, see ECB (2005), Financial Stability Review, June, Box 17.

71cECB



SYSTEMsingle family of credit default swap (CDS) indices2 – iTraxx in Europe and Asia and CDX inNorth America – has facilitated the development of a liquid secondary market for standardisedindex tranches that allows investors to express a view on spread direction and defaultcorrelation. In their search for high returns, trading-oriented and leveraged investorsconcentrated on buying the high yielding iTraxx equity tranche (i.e. selling credit protection bybetting on a low level of defaults). Assuming that credit spreads on different tranches willcontinue to move in parallel, as they had done until spring 2005, many investors delta-hedgedagainst spread widening by selling the more senior and lower yielding (the so-calledmezzanine) tranches (which are less exposed to default risk, but more exposed to spread risk).Such a hedge is in principle neutral to the parallel move in credit market spreads that occurswhen default correlations in the underlying asset portfolio remain broadly constant. However,the GM and Ford downgrades have increased the idiosyncratic risk of some names andindustries within the underlying portfolio. This led to a dispersion and widening of spreadswithin the equity tranche, causing prices to fall. As a result, some investors were forced tounwind their exposures due to mark-to-market losses, and the price of the equity tranche fellfurther. On the other hand, the spreads of the mezzanine tranches narrowed, and prices rose;investors thus lost money on both legs of the position.

It was notable that the market for synthetic CDOs – which consists of large pools of CDSs –remained largely unaffected, as these instruments are mainly held to maturity by investors whousually do not follow short-term trading strategies. The low concentration of single names inthe underlying portfolios – which is the result of the lessons from previous events such asParmalat, where exposures reached up to 6% in a number of CDOs – meant that the vastmajority of the CDO tranches that included the two carmakers remained unaffected by therating changes. With higher portfolio diversification, the CDO market is now more able to copewith idiosyncratic shocks. As a consequence, CDO primary market volumes remained strong,indicating sustained interest by investors, as the overall fundamental situation has notchanged. This effect was even more pronounced after the downgrades in October 2005, asrating agencies reported that Delphi was referenced in more than a third of synthetic CDOs, andGM in even more. Nevertheless, the immediate impact on the ratings of CDOs was rather mild,due both to diversification effects and gradual adjustments of the tranches’ ratings over time(Delphi’s ratings, for example, had fallen from investment grade at the end of 2004 to defaultby October 2005).

From a financial stability point of view, the main lesson to be drawn from the incident is thatrecent structural innovations in credit risk transfer markets have extended linkages betweenCDOs, corporate bonds and credit derivatives markets, and have thereby altered pricingdynamics. On one hand, the evolution of credit derivatives has allowed a smoother handling ofprice adjustments in the underlying cash market and has helped to diminish the market impactof mechanical bond index changes. The iTraxx credit indices have also proven their value as ahedging instrument during stressed market conditions. High levels of activity and contractingbid-offer spreads even under market stress have shown that CDS indices are now traded inEurope with a sustained level of liquidity. On the other hand, the tensions in the index tranche

2 For a more detailed description of tradable CDS indices, see ECB (2005), Financial Stability Review, June, Box D.2. These indiceshave also made pricing more transparent, since CDS indices provide a market estimation of default correlation. As a consequence,trading-oriented investors, mainly hedge funds, which had previously concentrated on single-name CDS and avoided portfoliocredit derivatives because of their unsatisfactory liquidity and price transparency, have rapidly entered the standardised market forCDS indices and index tranches.


EQUITY MARKETSEuro area stock prices continued to rise afterMay 2005, reaching their highest levels sinceearly 2002 (see Chart S41). In the second halfof 2005, the small and mid-cap segments of theeuro area stock markets reached all-time highs(see Chart 3.5), whereas large caps remainedabout 30% below the peak of March 2000.

The factors supporting the stock marketsincluded continued low risk-free interest rates,double-digit growth in earnings (see Box 6),and low stock market volatility (seeChart S43). Moreover, the earnings estimatesmade both by brokers and companies wereconsistently revised upwards on a net basisbetween May and October 2005 (see Chart 3.6).

A further factor that might have explained theoutperformance of the euro area stock market

vis-à-vis the US was the depreciation of theeuro against the US dollar (see Chart 3.7).

Euro area stocks were popular among fundmanagers after May 2005, who, according tosurveys, continued to overweight euro areaequities.7 This overweighting relates to the factthat surveyed fund managers viewed euro areaequities as being intrinsically undervalued.The plethora of flows into hedge funds andlimited non-equity investment opportunitiescould also have prompted some hedge funds toshift more money into euro area equityholdings, as some market participants havesuggested.

7 See Merrill Lynch (2005), “Anything but US equities”, GlobalFund Manager Survey, October.

market showed quite clearly that while the increased participation of hedge funds in all majorsegments of the credit markets adds liquidity, it also raises the potential that a liquidity squeezeand price dislocation could spread across multiple, interconnected credit markets. Since hedgefunds’ investments in credit markets tend to be highly leveraged, their potential impact onmarkets can be much greater than the notional size of these investments.

Chart 3.5 Dow Jones EURO STOXX totalreturn index

(Jan. 1992 - Nov. 2005, index: Jan. 1992 = 100)

Source: Dow Jones EURO STOXX.

largemidsmall

0

100

200

300

400

500

600

0

100

200

300

400

500

600

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Chart 3.6 Dow Jones EURO STOXX earningsrevis ions rat io

(Jan. 2001 - Oct. 2005, %)

Source: Thomson Financial Datastream.Note: The revisions ratio is the difference between thenumbers of earnings estimates revised upwards anddownwards by respectively companies and brokers as aproportion of the total number of earnings estimates.

by companies (left-hand scale) by brokers (right-hand scale)

-50

-40

-30

-20

-10

0

10

20

30

2000 2001 2002 2003 2004 2005-20

-15

-10

-5

0

5

10

15

20

73cECB



SYSTEM

Notwithstanding the fact that institutionalinvestors considered euro area stock prices tobe undervalued, some stock market valuationmetrics such as the price-earnings (P/E) ratiobased on ten-year trailing earnings haveremained rather high (see Chart S42).

High valuations in euro area stock markets mayalso provide some explanation for thebuoyancy of SPO activity in the euro area in2005 (see Chart S46). The total value of annualSPO deals came very close to the peaksobserved in mid-2000 and mid-2001. At thesame time, IPO activity remained substantiallylower at levels earlier observed in end-2001and early 1999.

Looking at the risks to euro area equitymarkets, stock market uncertainty derived fromthe distribution of options prices increased inthe second half of 2005, showing both strongerupward and downward risks (see Chart S44).Investors perceived in October 2005 a higherlikelihood of stronger increases or decreases ineuro area equity prices. A similar pictureemerges from implied stock market volatility,which crept slightly up in October 2005 (seeChart S43).

However, it cannot be excluded that some ofthe factors that have contributed to rising euro

8 For empirical evidence, see G. de Bondt (2005), “Does theCredit Risk Premium Lead the Stock Market?”, AppliedFinancial Economics Letters, 1, 5, pp. 263-68.

Chart 3.7 Relat ive performance betweeneuro area and US stock prices and EUR/USDexchange rate(Jan. 1999 - Oct. 2005)


USD/EUR (left-hand scale)S&P500 vis-à-vis DJ EURO STOXX (natural logarithm, right-hand scale)

1999 2000 2001 2002 2003 2004 2005

1.40

1.30

1.20

1.10

1.00

0.90

0.80

0,7

0,6

0,5

area stock prices in 2005 may in the end peterout. For instance, the pace of corporateearnings growth has shown signs ofdeceleration, and analysts’ expectations pointtowards a further slowdown. Moreover, asexpectations adjust, positive earnings surprisescannot be expected to provide ongoing supportto the market. At the same time, some upsiderisks remain for long-term risk-free interestrates which, if they were to crystallise, andespecially in connection with a disorderlyunwinding of global imbalances, could lead toa reappraisal of pricing in equity markets. Inaddition, if corporate bond spreads were towiden significantly, this could trigger a turn inthe euro area equity markets.8


4 THE EURO AREA BANKING SECTOR

Information that has become available sincethe June 2005 FSR confirms that euro areabanks’ profitability improved further,continuing the positive trend that began in2003. Profitability also improved among thosebanking sectors that had previously reportedweaker results, as well as in many MemberStates where economic growth rates wereless favourable. However, although improvedfinancial results may have made banks moreresilient to vulnerability, in the near term,downside risks remain in the medium to longterm which originate from sources bothinternal and external to the banking sector.

4.1 FINANCIAL CONDITIONS IN THE BANKINGSECTOR

BANKS’ PROFITABILITY IMPROVEDACROSS THE BOARDConsolidated data1 for 2004 show that theprofitability of euro area banks improved further(see Chart 4.1), consolidating the recovery thatbegan in the previous year. Behind thestrengthening of banks’ profitability, reductionsin the flow of provisions and growth in lending tohouseholds – mostly for housing purposes –continued to feature prominently.

ROE for euro area domestic banks, the mainindicator signalling increased profitability,stood at 10.54% at end-2004, increasing byalmost three percentage points from 2003 (seeTable S5). The developments in return on assets(ROA) closely followed those for ROE. Inaddition to the increase in average profitability,the distribution of ROE for euro area banksshifted to the right (see Chart 4.2). This indicatesthat both tails of the distribution (i.e. the worstand best performing banks) were more profitablein 2004 than 2003. The ROE of foreign banks,which includes both euro area and non-euroarea-owned foreign banks, stood at 10.46% atend-2004.2

Available data for the first half of 2005 for a setof large euro area banks indicate that the trendobserved in the 2004 consolidated banking datacontinued into 2005 (see Box 10 and Table S9).It is important to note, however, that apartfrom a difference in the sample between theconsolidated banking data and the data usedin Box 10, any direct comparison between

Chart 4.2 Frequency distr ibution of thereturn on equity for euro area banks

(2002 - 2004, after tax and extraordinary items,% of total assets)

Source: Banking Supervision Committee.

200220032004

0

5

10

15

20

25

30

35

40

0

5

10

15

20

25

30

35

40

< 0% 0-5% 5-10% 10-15% 15-20% > 20%


Chart 4.1 Prof itabi l ity and cost-to- incomeratios of euro area banks

(2002 - 2004, %)

0

2

4

6

8

10

12

2002 2003 20040

10

20

30

40

50

60

70

cost-to-income ratio (right-hand scale)ROE (left-hand scale)

1 See ECB (2005), EU Banking Sector Stability, October,Statistical Annex, Box 1 for a description of the consolidatedbanking data.

2 See Special Feature D in this Review, “What determines euroarea bank prof itability”, for an examination of the empiricalimportance of various factors on euro area banks’ financialperformance over a longer period of time.

75cECB



SYSTEMBox 10

FINANCIAL CONDITIONS OF LARGE EURO AREA BANKS

Problems in individual large euro area financial institutions could potentially spill over to otherparts of the euro area financial system. For this reason, it is particularly important to monitorclosely developments in these institutions. This Box complements the analysis in the main text byreviewing the recent financial results of a sample of large euro area banks. Because of the varyingdates of implementation of IFRS by European banks, some of the set of euro area banks used inprevious editions of this Review have begun to compile their financial statements under IFRS,whereas others have continued reporting under local or US GAAP (Generally AcceptedAccounting Principles). During the implementation period, this complicates to some extent theanalysis of financial statements from a financial stability perspective, given that IFRS and non-IFRS accounts are not directly comparable. Consequently, in order to ensure a consistentanalysis, the large banks are analysed in two sub-groups depending on whether they reported their2005 accounts under IFRS or non-IFRS.1

IFRS-reporting banksAs only a limited amount of historical data exists for comparison, any inferences on performancemust be treated with a high degree of caution because the restated accounts for end-2004 or mid-2005 are not audited, and because institutions vary considerably in the degree to which they haveimplemented the accounting standards IAS 32, IAS 39 and IFRS 4 in their pro forma 2004 figures.Furthermore, given that 2005 is viewed by banks and rating agencies as a transitional year foraccounting purposes, there is uncertainty about the extent to which the development of banks’financial results reflects either underlying circumstances or accounting changes. For example,profitability, provisions, and the overall size of banks balance sheets may be affected. The fullimpact will probably only become evident in 2006. In the short term, there may be importantcountry-specific effects related to IFRS implementation that may affect the interpretation of theresults.

Profitability for the largest euro area banks reporting under International Accounting Standards(IAS), as measured by the weighted average ROE, increased from 13.6% in 2004 to 20.8% in thefirst six months of 2005 (see Chart B10.1). While it is difficult to assess the degree to which thechange in accounting standards affected this performance, this does suggest that the underlyingperformance of most large euro area banks has remained solid. In fact, all of the institutions in thesample reporting results for the period up to the end of June 2005 posted an increase in ROE.

For banks reporting under IFRS, net interest income fell from 1.55% of total assets in 2004 to0.93% at the end of Q2 2005. As with those banks that did not report financial results underIFRS, this reflected increased competition as well as the need to fund new loan growth frommore expensive sources. By contrast, non-interest income such as fee and commission incomerose for most institutions in the first half of 2005.

Provisions declined from 0.16% of total assets at the end of 2004 to just over half that figure at0.08% of total assets at the end of H1 2005. Costs also declined, with the weighted average

1 For a detailed overview of IFRS, see Special Feature E in this Review entitled “Main effects from the new accounting framework onbanks”.


cost-to-income ratio falling from 63.52% in 2004 to 61.44% in June 2005. The cost-to-incomeratio not only decreased in the best performing group of institutions (first quartile for the cost-to-income indicator), but also for those performing worse than the weighted average (the thirdquartile), perhaps signifying that cost control has been firmly established.

Increased profitability has also led to the continued strengthening of capital ratios. The Tier 1ratio increased over the period, rising from 7.53% in 2004 to 8.24% in 2005, marking anincrease in the buffers available to banks to cushion against unexpected losses (seeChart B10.2). The overall regulatory solvency ratio also increased over the period from11.36% in 2004 to 11.92% at the end of the second quarter of 2005.

Non-IFRS-reporting banksOn the whole, the financial positions of large euro area banks not reporting under IFRS continuedto improve in the first half of 2005, consolidating the performance of the previous two years. Thiswas mainly driven by reasonable performance from non-interest income sources, and benefitedfrom the reduction in charges for credit losses. However, generating sustainable revenue in somedomestic markets remains challenging for certain institutions. Profitability, as measured by theweighted average ROE, continued to improve from about 4.5% in 2004 to just over 15.3% inQ2 2005 (see Chart B10.3). Furthermore, the weaker institutions also managed to improve theirperformance, with banks in the bottom quarter of the distribution increasing profitability from-2.15% in 2004 to 14.50% in the first six months of 2005.

Banks’ net interest income as a percentage of total assets continued to fall, dropping from aweighted average of 0.61% for 2004 as a whole to 0.55% in the first half of 2005. The continuedeffect of low nominal interest rates, increased competition in certain retail segments, and lowloan demand in some of these banks’ domestic markets has led to sustained pressure onmargins. Most institutions also reported increased non-interest income, especially fee andcommission income.

Improved credit risk conditions led to a moderate decline in provisions from a weightedaverage of 0.09% of total assets in 2004 to 0.08% of total assets at the end of Q2 2005. This

Chart B10.2 Frequency distr ibution of Tier 1ratio for large euro area banks ( IFRS)

Sources: Published accounts of individual banks and ECB calculations.Note: Data for the f irst half of (H1) 2005 are annualised and unaudited and are based on 14 banks from a sample of 15.

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

< 0% 0-5% 5-10% 10-15% 15-20% > 20%

2004H1 2005

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

< 6% 6-7% 7-8% 8-9% 9-10% > 10%

2004H1 2005

Chart B10.1 Frequency distr ibution ofreturn on equity for large euro area banks( IFRS)

77cECB



SYSTEM

results for euro area banks in 2004 and 2005 isfurther complicated by the application of newreporting requirements in 2005, following theadoption of International Financial ReportingStandards (IFRS).

OPERATING INCOME FELL MARGINALLYAccording to the consolidated banking data,the total income of euro area banks fell

Chart B10.3 Frequency distr ibution ofreturn on equity for large euro area banks(non-IFRS)

Chart B10.4 Frequency distr ibution of Tier 1ratio for large euro area banks (non-IFRS)

Sources: Published accounts of individual banks and ECB calculations.Note: Data for the f irst half of (H1) 2005 are annualised and unaudited and are based on three banks from a sample of six.

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

< 0% 0-5% 5-10% 10-15%

2001 2002 2003 2004 H1 2005

2001 2002 2003 2004 H1 2005

0

10

20

30

40

50

60

70

0

10

20

30

40

50

60

70

< 6% 6-7% 7-8% 8-9% 9-10% > 10%

level of provisioning is particularly low compared to historical norms. While some individualinstitutions have indicated that provisions may increase slightly during the second half of theyear, it cannot be ruled out that the adequacy of some banks’ provisioning could be tested ifcredit conditions were to deteriorate unexpectedly.

Cost control measures continued to be implemented by most banks in the first half of 2005 inorder to sustain profitability. The weighted average cost-to-income ratio decreased from75.05% in 2004 to 71.93% in the first half of 2005. The weighted average Tier 1 ratio increasedfrom 8.06% in 2004 to 8.28% during the first half of 2005, with the weakest performinginstitutions also managing to increase their Tier 1 ratios (see Chart B10.4). Overall solvencyratios improved as well (see Table S9).

A notable development was that most indicators of financial conditions for both sets of banksimproved regardless of the accounting method followed. If this trend continues for theremainder of 2005, the resilience of euro area banks to adverse shocks should improve further.However, for some institutions without strong underlying performance in their most importantmarkets, it remains to be seen how durable the current return to profitability will be, given thereliance of these institutions on certain types of non-interest income sources, in tandem withcost-cutting to boost overall profitability.

marginally, as a percentage of total assets,between 2003 and 2004, while net interestincome decreased slightly as a share of totalincome (see Table S5).

Turning to the main sources of banks’ income,the low interest rate environment supported arise in lending to households for housingpurposes at an annual rate of 7.9% in December


2004, a rate similar to what was experienced in2003 on the basis of MFI data. This growth ratefurther accelerated to 10.2%, on an annualbasis, according to preliminary data for the firstsix months of 2005.

There is also some evidence of a new lendingpattern developing for euro area banksconcerning lending to non-financialcorporates, especially the smaller ones.Borrowing by non-financial corporates wassluggish in 2002 and 2003, but started to pickup in 2004 and early 2005. Consequently, loansfrom euro area MFIs became an increasinglyimportant source of funds for non-financialcorporates in 2004. The rate of corporateborrowing growth and the share of MFI lendingin it is, however, still far from the peaksrecorded in the period 1999-2000.

Nonetheless, if prolonged, the increase incorporate borrowing could provide banks witha desirable differentiation in sources ofincome. In fact, in the persistently low interestrate environment of the last few years, bankshave probably faced constraints in supportingtheir operating income in mature markets,which may have led them to increase theirexposures to potentially more risky assets orgeographical regions. In this light, a recoveryin borrowing from the corporate sector wouldbe beneficial. This development could alsoreduce banks’ dependence on lending tohouseholds for housing purposes, especiallyfor banks in those countries where house priceshave increased substantially over the last fewyears.

Notwithstanding the potential recovery inborrowing by corporates, evidence of asustained improvement in this source ofincome for banks is still elusive. Corporatesappear reluctant to increase their borrowing forlong periods of time. The largest part of theincrease in borrowing volumes is for maturitiesof no more than five years, irrespective of thepotential benefits of locking in generally lowinterest rates at present (see Chart S54).

The aggregated figures for domestic banks inthe consolidated banking data show that therewas a very small decline in lending in banks’balance sheets as a share of total assets in theeuro area (see Table S6). Part of this fall isrelated to the more rapid increase in euro areabanks’ assets. The group of large banksrecorded the most noteworthy reduction inloans as a share of total assets. It cannot beexcluded that the development in lending bylarge banks may indicate a broad-basedcontainment of lending by banks, in relation totheir accumulation of total assets.

Apart from the developments in terms of lendingvolumes, information on loan pricing shows thatlending margins, based on MFI data, fluctuatedrather sharply in the first half of 2004 beforestabilising in the second half of 2004 and into thefirst half of 2005 (see Chart S57). Banks appearto be experiencing strong competition in theirlending activities, especially to households, andthus face significant constraints in increasingtheir lending margins (see also Box 11).Therefore, the tentative increase in margins mayoverestimate the potential for increased interestincome for euro area banks going forward.

Deposit margins have evolved less positively.As banks have increased their lending, they havehad to compete for depositors’ funds; however,with the household sector increasingly burdenedby housing debt, deposits may not have beenwidely available to banks. The decliningmargins for most of 2004 and the first half of2005 indicate that euro area MFIs were facedwith a scarcity of funding from deposits (seeChart S58). The information from banks’balance sheets points in the same direction: as ashare of total assets, deposits remained basicallyunchanged in the euro area (see Table S6).

The gradual flattening of yield curves in theeuro area could impair banks’ efforts toimprove their net interest income goingforward. A flatter yield curve reduces therevenues that banks derive from traditionalmaturity transformation business. Apart from

79cECB



SYSTEMlower revenues, the continuing increase inlending by banks, as also indicated by thesustained demand for consumer and housingloans and the increasing demand for corporateloans in the latest euro area Bank LendingSurvey (October 2005, see Box 11), coupledwith strong competition among banks, couldnegatively affect the quality of banks’ loanportfolios.

Net non-interest income as a share of totalassets remained broadly unchanged (see TableS5). Among the components of non-interestincome, fees and commissions increased in

2004 both in terms of total assets and of totalincome, and made a positive contribution to thenet non-interest income of euro area banks. Bycontrast, profits from securities and foreignexchange trading (henceforth “trading andforex results”), fell for all bank groups in theeuro area, both as a share of total assets andtotal income. While fees and commissions mayalso be earned by banks on their moretraditional lending activities, trading and forexresults are more directly correlated withsecurities market conditions, where the lowvolatility environment may provide limitedopportunities for trading activity.

Box 11

ASSESSING FINANCIAL STABILITY IMPLICATIONS OF RECENT FINDINGS FROM THE ECB BANKLENDING SURVEY

The ECB’s Bank Lending Survey (BLS) provides timely qualitative evidence of the lendingpolicies of the euro area banking sector, and should be a useful tool for detecting turning pointsin the credit cycle and potential credit crunches facing euro area households and firms. ThisBox examines recent developments in banks’ credit standards – and the underlyingdeterminants – on the approval of loans to households and loans since early 2005, as reported inthe October 2005 BLS.

Chart B11.1 Changes in credit standards appl ied to the approval of loans or credit l ines toenterprises


-40

-20

0

20

40

60

-40

-20

0

20

40

60

realisedexpected

FACTORS CONTRIBUTING TO TIGHTENING CREDIT STANDARDS

costs related tobank’s capital

position

competitionfrom other

banks

industry or firm-specific outlook

expectationsregarding general

economic activity

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q3Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q32003 2004 2005 2003 2004 2005 2003 2004 2005 2003 2004 2005 2003 2004 2005


Chart B11.2 Changes in credit standards appl ied to the approval of loans or credit l ines tohouseholds for house purchase


realisedexpected

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q42003 2004 2005

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q32003 2004 2005

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q32003 2004 2005

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q32003 2004 2005

-20

-10

0

10

20

30

-20

-10

0

10

20

30

housing marketprospects

expectationsregarding general

economicactivity

FACTORS CONTRIBUTING TO TIGHTENING CREDIT STANDARDScompetition

from other banks

According to the BLS, in the third quarter of 2005 banks reported more or less unchanged(compared with the previous quarter) credit standards on the approval of loans to enterprises.The slight net tightening (+2%) in the third quarter of 2005 occurred following fiveconsecutive quarters of net easing of credit standards towards the non-financial corporatesector. While it is still too early to tell whether this change constitutes a turnaround in theeasing cycle of credit standards on corporate loans, it may partly reflect a strong increase inperceived loan demand (as reported in the October 2005 BLS), enabling banks to attractborrowers without having to ease credit standards. Broadly unchanged underlying factors(such as the industry or firm-specific outlook, and expectations regarding general economicactivity) compared with the previous quarter contributed to ending the net easing of creditstandards (see Chart B11.1). The increase in corporate loan demand may also have mitigatedthe effects of competition from other banks on credit standards applied to loans. With regard tothe terms and conditions by which credit standards were applied, banks reported that marginson average loans were lowered (although less than in previous quarters), while non-interestrate charges and margins on riskier loans tended to support a net tightening of credit standards,suggesting that banks were becoming more discriminating in their pricing of risks.

With regard to the approval of loans to households for house purchase, banks reported a neteasing of credit standards in the third quarter of 2005. This was in line with developments inprevious quarters, except for the second quarter of 2005 when a slight net tightening wasreported. The net easing seemed to reflect, in particular, reduced concerns regarding housingmarket prospects as well as a slight improvement in expectations concerning general economicactivity (see Chart B11.2). Moreover, competition from other banks continued to contribute tothe net easing of credit standards. The tightening of credit standards on loans for housepurchase was mainly carried out through the margin on riskier loans, while margins on averageloans as well as less stringent loan-to-value ratios contributed to the net easing.

81cECB



SYSTEM

In the third quarter of 2005, euro area banks reported broadly unchanged credit standards on theapproval of loans for consumer credit and other loans to households, following four quarters ofnet easing (see Chart B11.3). The factors behind the applied credit standards were more or lessunchanged from the previous quarter, with competition from other banks (and from non-banks)contributing to a net easing, while factors such as borrowers’ creditworthiness, the risk ofcollateral demanded and expectations of the general economic activity continued to pulltowards a net tightening. With regard to terms and conditions, the net easing of credit standardson consumer credit and other loans to households was carried out, in particular, through lowermargins on average loans; margins on riskier loans were not eased, however.

Overall, as in previous quarters, in the second and third quarters of 2005 euro area bankslargely continued to ease or to keep credit standards on loans to the non-financial private sectorbroadly unchanged. The net easing was mainly driven by strong competition from other banks,but took place against a background of deteriorating expectations regarding economic activity.This suggests that banks may have taken on more risk in order to gain market share and boostprofitability. In the most recent quarter, however, there are some signs that banks have reactedto the still moderate economic growth prospects and increasing loan demand by ending the neteasing of credit standards. Moreover, the perceived increase in risk-taking in recent quartersseems to have been reflected in more differentiated pricing of loans. All in all, the questionwhether these developments could entail future difficulties, such as a deterioration in thequality of banks’ credit portfolios, depends on the extent to which the accumulated risksmaterialise in the period ahead.

Chart B11.3 Changes in credit standards appl ied to the approval of loans or credit l ines forconsumer credit and other loans to households


realisedexpected

Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q3Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q32003 2004 2005 2003 2004 2005 2003 2004 2005 2003 2004 2005 2003 2004 2005

-20 -20

-10

0 0

10

20

30

-10

10

20

30

creditworthiness ofconsumers

FACTORS CONTRIBUTING TO TIGHTENING CREDIT STANDARDSrisk on

collateral demanded expectations

regarding generaleconomic activity

competition fromother banks


COST EFFICIENCY IMPROVEDCost containment, which contributedpositively to banks’ profitability in 2002 and2003, continued at a slower pace in 2004 (seeChart 4.1). Nevertheless, total costs, as a shareof total assets, still declined for all bankcategories (see Table S5).

Developments varied among the components oftotal costs. For example, the share of staff costsdeclined only marginally, and this left banksrelying on administrative and other costs toreduce their total costs. This may imply limitedscope for future cost-efficiency gains. Asbanks in the euro area have relied rather heavilyon cost-cutting to support profitability over thelast few years, the exhaustion of this source ofprofitability may exert pressure on results inthe future.

PROVISION FLOWS DECLINED FURTHER, BUTTHE COVERAGE RATIO INCREASEDGeneral economic conditions in the euro areahave supported banks’ activities: insolvenciesin the household and corporate sectors declinedfurther, any deterioration in banks’ assetquality was marginal, loan losses declined, andsignificant write-offs or write-downs of creditoverdue in the previous years in at least a fewMember States reduced the impact of creditrisk on the banks’ balance sheets. Against thebackground of such a benign credit environment,the aggregate flow of provisions declinedfurther, and as a share of total assets the stockof provisions also fell between 2003 and 2004(see Table S5 for the flow of provisions andTable S7 for the stock). Developments variedacross countries, however, as in at least onelarge country the level of provisions fell as anormalisation in the aftermath of the highprovisioning activity observed in 2002, wheninsolvency figures had risen sharply.

An additional factor expected to affect the flowof provisions in 2005 is the move of stockmarket-listed banks (and, in some countries, allbanks) to IFRS.3 While it is still premature tomake a precise assessment of the impact of thelevel of provisions, not least because the

impact is likely to differ across the euro areadepending on the accounting regime adopteduntil 2004, two broad indications cannevertheless already be provided: (1)provisioning requirements under IFRS may bemore restrictive, as IFRS require objectiveevidence of impairment that is supported bycash flows before an impairment provision canbe created; and (2) the ‘Funds for GeneralBanking Risks’, or general provisions, nolonger exist. Given that the credit cycle is at apotentially delicate juncture, the impact ofthese accounting changes on provisions callsfor careful monitoring in 2005 to ensure thatcapital buffers remain at a prudent level.

Indeed, beyond the short-term horizon, theoverall low levels of provisions could bequestioned. While the low level may be justifiedat present by the benign credit risk environment,a less benign environment would make thesituation more difficult for banks. Possibletriggers for a deterioration include a rise inshort and long-term interest rates, or a suddenworsening in the macroeconomic environment.In at least a few Member States, there is aperception that credit quality, particularly in thehousehold sector, could weaken over themedium term, and it cannot be excluded that theability of the present stock of provisions toprovide a sufficient buffer may be tested.4

In contrast to the observed changes inprovisioning as a share of total assets, changesin asset quality and the coverage ratio, i.e. theratio of provisioning stocks over total non-performing and doubtful assets, paint a morepositive picture (see Chart 4.3). For the group ofdomestic banks, the share of non-performingand doubtful assets as a percentage of ownfunds, gross and net of provisioning, decreased

3 See Special Feature E, “Main effects of the new accountingframework on banks”, in this Review for a more thoroughassessment of the implications of the new accountingstandards.

4 In most euro area countries, banks are in general not allowed toprovision in anticipation of the credit cycle, and provisioninglevels are mostly based on historical asset impairment.Nonetheless, recent low levels of provisions have reducedbanks’ buffers in the event of a downturn in the credit cycle.

83cECB



SYSTEM

between 2003 and 2004. The largest reductionwas in gross terms (see Table S7).

The positive developments in coverage ratiosamong most groups of banks should beconsidered with caution, as the picture of thequality of the assets portfolio of banks isbackward-looking and is in general highlysensitive to changes in the phase of the creditcycle. A sudden deterioration in credit qualitywould cause a drop in coverage ratios. Dueto the possibility, albeit remote, of suchcircumstances, the absolute amount of the flowand the stock of provisions would be the moremeaningful indicator.

SOLVENCY BROADLY UNCHANGEDBanks’ capital adequacy in the euro arearemained broadly unchanged in 2004, as theoverall solvency ratio slightly fell and theTier 1 ratio remained broadly constant (seeTable S8). These developments are broadlyconsistent with the indicators for large banksbased on published accounting data for the firsthalf of 2005 (see Box 10). Large banks have asubstantial impact on the aggregate indicators,and therefore the share of the total bankingsector with an overall solvency ratio of lessthan 9% increased.

Looking at the information that is collected atcountry rather than bank level, the dispersion inthe overall solvency ratio increased in 2004.This would suggest that, together with thedownward shift in the average level, somepockets of fragility may be developing,although it must be borne in mind that solvencyfigures remain comfortably above theregulatory minimums (see Chart 4.4).

The share of risk-weighted assets of total risk-adjusted assets declined for the group ofdomestic banks (see Table S8), owing to areduction in risk-adjusted trading book assets.On the other hand, risk-adjusted off-balancesheet assets increased (see Table S8).

LIQUIDITY DEVELOPMENTS MIXEDDevelopments in euro area banks’ liquiditywere mixed in 2004. On the assets side,liquidity increased on the basis of the indicatorcovering assets with the lowest maturity, whileit declined in relation to the broadest liquiditymeasure, which includes assets with longermaturities (see Table S6 and Chart 4.5).5

Chart 4.3 Euro area banks’ total provis ions

(2002 - 2004, % of non-performing and doubtful assets)


200220032004

0

20

40

60

80

100

120

0

20

40

60

80

100

120

domestic banks foreign banks

Chart 4.4 Frequency distr ibution of overal lsolvency ratios for euro area banks

(2002 - 2004, % of risk-weighted assets)


200220032004

05

101520253035404550

0

5101520253035404550

< 7% 7-8% 8-9% 9-10% 10-11% 11-13% > 13%

5 The narrowest liquidity indicator, Liquidity asset ratio 1 in thetables, covers only cash and short-term government debt, as ashare of total assets; the intermediate indicator, Liquidity assetratio 2, also includes loans to credit institutions; and the widerliquidity indicator, Liquidity asset ratio 3, includes in additionalso longer-term debt securities issued by public bodies.


Against this background, the expectation of anupward shift in long-term interest rates mayhave encouraged banks to keep most of theirfunds in assets with the shortest maturities,such as cash and government debt.

On the liabilities side, banks decreased theirreliance on the interbank market, as theamounts owed to credit institutions as a shareof total assets fell (see Table S6). Banks partlycompensated for this by stepping up the use ofmarket instruments such as debt certificatesand subordinated liabilities. The mix ofincreased market funding and decreaseddeposits in some euro area Member Statescould indicate a tightening in the price ofliquidity of banks in those countries. Given thatmarket funding is more expensive for banks,such developments may have a negative impacton banks’ profitability in the near future.

4.2 RISKS FACING THE BANKING SECTOR

Following the review of the financialconditions of the euro area banking sector, thissub-section considers the risks facing the euroarea banking sector in the period ahead. Itshould be emphasised that these risks arenot highlighted with the aim of identifyingthe most probable outcome, but rather with

the intention of highlighting potential andplausible sources of downside risks with regardto the likeliest outcome.

Given the favourable developments in thefinancial performance of the euro area bankingsector, including many of the weakest performingbanks, the baseline assessment of the outlook forbanks is broadly positive. This assessment is alsosupported by the fact that in many euro areacountries, improved banking sector performancetook place against a backdrop of subduedeconomic growth. Despite this rather benignoverall assessment, certain vulnerabilities can beidentified, some internal and others external.Box 12 reports on the results of the survey of themain risks going forward as identified by thebanks themselves in spring 2005.

Starting with sources of risk and vulnerabilitiesinternal to the banking sector, as discussed insub-section 4.1, the improvement in bankingsector profitability was supported by asubstantial reduction in the number of corporatesector defaults and non-performing loansbetween 2002 and the first half of 2005, and acorresponding fall in provisioning for loanlosses by banks in almost all countries. Thereare some signs, however, that the cuts inprovisioning may have advanced rather far as, in

Chart 4.5 Euro area banks’ l iquid asset ratios

(2002 – 2004, % of total assets)

Source: Banking Supervision Committee.Note: Liquid asset ratio 1 includes cash and short-term government debt, while Ratio 3 in addition includes loans to creditinstitutions and debt securities issued by public bodies.

200220032004

0.0

0.5

1.0

1.5

2.0

2.5

3.0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

domestic banks’liquid asset ratio 1

foreign banks’liquid asset ratio 1

domestic banks’liquid asset ratio 3

foreign banks’liquid asset ratio 3

0

5

10

15

20

25

30

35

40

0

5

10

15

20

25

30

35

40

85cECB



SYSTEMan increasing number of cases, banks havereported that provisions had reached all-timelow levels. Going forward, the adequacy of lowlevels of provisioning will therefore rely on theviability of expectations that the credit quality ofbank loan portfolios will continue to developfavourably, which in some countries could bequestioned in the current phase of the creditcycle.

Although abating in intensity compared withprevious years, banks’ efforts to cut costs, as ashare of total assets, continued to reflect tightcompetition and, in some countries, thedifficulty of generating income from the corematurity transformation business in a lowinterest rate environment. Since provisioningand costs both have lower limits, in order tosustain profitability banks might come underincreasing pressure to resort to non-interestincome, thereby possibly taking on greaterrisk. Furthermore, as discussed elsewhere inthis Review, the uncertainty surrounding theadoption of the new IFRS accounting standardsmay contain transitory risks to banks insofar astheir practices of reporting provisioning need tobe amended.

The principal sources of external risks andvulnerabilities in the operating environment ofbanks continued to be high oil prices and thepersistence of wide global imbalances. High oil

prices, should these persist, could adverselyaffect the ability of non-financial sectors tohonour their liabilities to banks. A disorderlyunwinding of global imbalances could, ifthis were to materialise, result in heightenedexchange rate and bond market volatility.While the direct exposures of euro area banksto the oil market and to foreign exchange risksseem rather limited, the risks posed byexchange rates could prove to be correlatedwith the credit risks of non-financial sectors –especially firms in the export sector and theirsuppliers such as SMEs.

The protracted period of low short-terminterest rates and ample global liquidity has –via the global quest for yield – helped compressspreads across the credit spectrum toexceptionally low levels. The tightness ofcredit spreads could represent a greater thannormal market risk for banks, and may alsohave affected their pricing of risks in thecontext of corporate lending.

To assess the possible impacts of these risks,this sub-section first provides an assessment ofEU banks’ credit risks originating from theirlending to the household and corporate sectors.It then assesses the market risks from interestrate, exchange rate and other market exposures,as well as banks’ exposures to emergingmarkets and hedge funds.

Box 12

SURVEY ON MAJOR EU BANKS’ PERCEPTION OF RISKS IN THE YEAR AHEAD

This Box summarises the answers from a survey of banks’ perceptions of main risks for the yearahead. The survey was undertaken by the Banking Supervision Committee, with the assistanceof the Working Group on Developments in Banking (WGBD), in February and March 2005,covering all EU Member States. It was similar to the one held in 2004 (see ECB (2004),Report on EU Banking Structures, November). It was completed by 99 banks, 43 of which werefrom euro area Member States, and 56 from non-euro area Member States. A maximum of fivebanks per country participated in the survey. The respondent banks were asked to state whatthey perceived as major risks and to distinguish between macroeconomic, financial market,sectoral, strategic and regulatory risks. They were then asked to assign different scores to thesefive broad categories on a scale from 1 (very low) to 5 (most important), and to elaborate inmore detail what they felt constituted those risks.


Charts B12.1 and B12.2 present the distribution of the banks’ scores on the different broadrisks, each as a percentage of the replies rated 1 to 5, with a further disaggregation between theeuro area and non-euro area. A higher percentage of replies with score 5 for a specific risk, forexample, indicates that relatively more banks see this risk as a very important one. It should benoted that the risks reported here are directly attributable to the respondent banks, and do notnecessarily reflect the ranking of risks by supervisory authorities or central banks (as reflectedelsewhere in this Review).

The results clearly indicate that, similar to last year, euro area banks regard macroeconomicrisks as the major challenge in the period under review (i.e. until Q1 2006), followed bypossible risks stemming from the financial markets. For non-euro area banks, the balance ofrisks is more evenly spread across different sources of risk. This possibly reflects relativelymore favourable macroeconomic and financial market conditions in the latter countries, whichallows banks to pay more attention to other sources of risk.

Further detail was gained by a decomposition of the broad sources of risk into specific items,for which banks were also asked to assess the “expected negative impact” on profits, rangingfrom less than one month to one quarter or a full year’s profit. The discussion below focuses oneuro area banks, although Table B12.1 also displays evidence for non-euro area banks, whichpresents a fairly similar picture. The table shows that 86% of the respondent euro area bankssee downside credit risks stemming from a deterioration in borrower quality as a majorchallenge to profits, of which 61% regard this as having a potentially major impact on profits.Another major macro challenge for euro area banks is weak capital spending, which wouldreduce the lending volume to the corporate sector. On the other hand, banks do not considerlower consumer spending and borrowing to be a major issue. A possible house price bubble,however, does worry banks from almost all euro area countries. At the time of the survey, oilprices had not risen to the levels currently experienced, and banks regarded oil prices therefore

Chart B12.1 Sources of r isk for euro areabanks

(2005, % bank answers per score)

Chart B12.2 Sources of r isk for non-euroarea banks

(2005, % bank answers per score)

Sources: Banking Supervision Committee and ECB calculations.Note: Data not available for Irish banks.

0102030405060708090

100

0102030405060708090100

1lowest

importance

2 3medium

4 5highest

regulatorystrategicsectoralmarketsmacro

1lowest

importance

2 3medium

4 5highest

regulatorystrategicsectoralmarketsmacro

0102030405060708090

100

0102030405060708090100

87cECB



SYSTEM

MacroeconomicCredit risk-deteriorationborrower quality 86 61 73 24 35 68 41 47 29 26Weak capital spendingcorporates 60 40 54 46 80 45 9 72 16 40House price bubble 42 61 44 56 46 57 23 31 69 69Sustained high or rising oilprices 30 31 23 77 80 32 9 33 61 80Credit risk-overheating-high indebtedness 26 18 73 18 0 59 29 36 61 69Global imbalances 23 50 20 80 100 16 2 67 33 0Financial marketsMajor downturn ofequity market 67 36 79 3 0 38 7 71 24 50Decrease in interest rates 49 40 38 14 0 59 29 33 58 50Increase in interest rates 40 78 41 59 50 57 25 34 59 57Exchange rate developments 37 27 63 31 43 54 13 53 33 14Credit event 35 29 60 33 38 39 14 73 14 0Flattening of the yield curve 33 69 71 0 0 36 20 35 55 55SectoralCompetitive pressure/overcapacity 88 41 42 55 39 96 55 15 83 90Emergence of new players 63 20 37 44 36 63 25 43 46 57M&A and consolidation,need for size 47 25 30 65 67 50 11 43 50 67StrategicNeed to increaseperformance/efficiency 77 45 39 55 32 82 45 24 70 72Expansion in new markets 40 41 41 59 100 34 7 42 58 50Need to increase size 37 19 69 31 10 43 18 46 50 50Operational risks 37 25 63 31 43 32 13 50 39 29Risks from financialinnovations 35 13 20 80 50 34 7 37 58 75RegulatoryInternational accountingstandards 93 40 20 70 88 70 30 26 67 76Basel II 72 42 29 71 56 86 29 19 77 81Corporate governance orreporting burden 49 24 5 95 100 36 7 25 75 75New EU Directives 44 50 21 79 80 46 18 31 62 70Increasing importance ofrating agencies 33 36 36 64 100 20 4 36 64 100

Table B12.1 Respondent banks’ main r isks stemming from f ive broad classes of r isk

(2005, %)

Sources: Banking Supervision Committee and ECB calculations.Note: “All” refers to the number of respondents mentioning a certain issue (as a % of the number of respondents). “Major”denotes the % of banks that identif ied the issue as having a major impact on prof its (equal to quarterly prof it or more).“Trend” compares the importance of the issue with one year ago, and gives the percentage of banks that identif ied aconstant or upward movement in its importance. “Major/upward trend” refers to the % of banks that identif ied the issue ashaving a potentially major impact on prof its and more impor tant than one year ago (as a % of banks that expected a majorimpact).

euro area banks non-euro area banks

trend trend

all major constant upward major/ all major constant upward major/upward upward


THE OUTLOOK FOR CREDIT RISK IS FAVOURABLEBUT DOWNSIDE RISKS EXIST

Banks’ exposure to the household sectorcontinues to increaseA significant share of the lending portfolios ofeuro area banks is composed of loans extendedto households, with non-consolidated datasuggesting that lending to householdsrepresents roughly 36% of the total lendingstock in the euro area. Against the backgroundthat lending to the household sector has, forseveral years, been the most rapidly growingline of business for banks, continued growth inexposures to the household sector counts as apotentially increasing source of credit risk,despite the generally higher quality of creditgranted to households relative to the corporatesector.

Reflecting favourable financing conditions,households continued to accumulate more debtin the first half of 2005. As household debtratios rise, the vulnerability of households toany negative income or interest rate shocksshould also increase.6 All else being equal, thismeans that adverse disturbances could have agreater impact on the non-performing loanrates of banks than in the past.

Most of the rise in household indebtednessreflects strong mortgage financing growth.Owing to the importance of this revenue sourcefor banks, developments in residential propertyprices have become an important factor in the

as having a more remote and indirect impact on their profits, although the trend has been risingcompared with 2004.

As for financial markets, around two-thirds of euro area banks factored in a possible downturnin equity markets later on in 2005. However, this seems to be a matter of normal caution, sincein about 80% of the cases the issue has not become more pronounced compared with one yearago. Interest rate changes were a matter of concern for most euro area banks, although thedirection in which banks expected these to influence their performance proved to be quitevaried. Some banks expected to be mainly affected by rising rates, others by falling rates, andstill others were concerned by a flattening of the yield curve or increases in interest ratevolatility.

In the category of industry sector risks, almost 90% of euro area banks said they had beenaffected by tighter competitive conditions, which has been reflected in margin pressure in bothdeposit-taking and lending. In addition, around 60% of the surveyed banks were expectingcompetition from new market players (e.g. in consumer finance) to increase.

Although performance improved in 2004, many banks seemingly still see a need to take furthermeasures to improve their situation. Otherwise, most banks seem to be relatively confidentabout their strategic choices.

Finally, in line with some other banking industry surveys, the results showed that for mostbanks the importance of regulatory and accounting issues has increased substantiallycompared with one year ago. Around 70% of euro area banks regarded the new regulatoryframework (Basel II) as a challenge, while more than 90% mentioned that they expected tonotice the effects of the new accounting standards.

6 See Special Feature C, “Assessing the f inancial vulnerability ofeuro area households using micro-level data”, in this Review,which highlights some of the characteristics of indebtedhouseholds in the euro area.

89cECB



SYSTEMassessment of risks to banking sector stability,warranting close monitoring of the housingmarket.

Turning to banks’ credit conditions, despite therather substantial increase in loan-to-value(LTV) ratios in the flow of new mortgages inmany euro area Member States, the averageLTV ratios of the mortgage lending stock in theeuro area are still rather conservative. Thissuggests that a fall in house prices, should thishappen, would need to be fairly large beforecredit quality in the euro area is more seriouslyeroded. However, declining house prices areoften associated with rising unemploymentcaused by a macroeconomic shock. In such ascenario, the risk of default would rise intandem with the erosion of the value ofcollateral (the residential property), therebyreducing credit quality more directly.

Available information suggests that the currenthousehold sector financial buffers should helphouseholds to absorb reasonably large adversemacroeconomic shocks. The ratio of debt tofinancial assets stabilised in 2003 at around30% (see Chart S35). Moreover, the estimatedtotal debt servicing burden of the householdsector (repayment of the principal, plus interestpayments) has remained broadly stable since2000, at around 12% of disposable income (seeChart S37).

However, the overall impact of a change ininterest rates on households’ debtsustainability will notably depend on country-specific features of mortgage contracts, inparticular the interest rate variability regime ofthe outstanding loans. In particular, the impacton households could be more substantial incountries where variable rates are morecommon. Furthermore, new borrowers wouldbe more sensitive to changes, as a greaterproportion of them are indebted at variablerates, and as they have not yet amortised a highamount of their principal.

The mortgage lending market has in recentyears been characterised by intensified

competition, to some degree spurred by thefavourable risk-weighting of mortgage loanswithin the Basel II framework. As mentionedabove, in most euro area countries there hasbeen a general tendency over the last few yearstowards granting higher LTV ratios, with banksin an increasing number of countries offeringloans up to 100% of the value of the house.Since this has often gone hand in hand with anextension of the loan maturity, up to 30-35years in some countries, the typical monthlyinstalment rate has been kept broadlyunchanged (usually granted up to 30% of thehousehold’s monthly income). While themonthly debt servicing burden has notincreased dramatically, the likelihood thatborrowers could struggle to meet paymentdifficulties over longer repayment periods withslower loan amortisation has become higher.

As the ECB Bank Lending Survey shows (seeBox 11), banks in the euro area slightlytightened their credit standards in the secondquarter of 2005 in response to growingperception of risk, particularly in terms ofworsening housing market prospects.However, the banks had once more eased theircredit standards by the following quarter,reversing the tightening on account of reducedconcerns regarding housing market prospects,coupled with a slight improvement inexpectations concerning general economicactivity. Looking forward, it remains unclearwhether banks will continue to ease their creditstandards or whether they could engage in areassessment of the housing market.Nonetheless, banks’ risk managementtechniques have improved, allowing them tocope more effectively with complexinstruments used to transfer credit risks to thesecondary market. This could in part havecontributed to a lower share of non-performingloans in total loans to households.

Regarding the developments in unsecuredconsumer credit, outstanding amountscontinued to rise at a brisk pace in somecountries, compared with more moderateoverall picture. The growth rate of consumer


credit continued on a modest upward trend untilJuly 2005, with the stock of consumer loans andother credit as a proportion of total householdloans standing at 13.5%. Intense competitionamong banks has contributed to a slight neteasing of banks’ credit standards for consumercredit, although this has been recentlymitigated by an increased perception of risk ondemanded collateral and on consumercreditworthiness. Since households infinancial distress tend to default on consumercredit before defaulting on mortgages, paymentarrears are more likely to show up first in theformer category of loans. However, there arecurrently no indications of a systemic increasein payment disturbances and worseningconsumer credit quality in the euro area.

Corporate sector credit risks have easedThe information that has become availablesince the June 2005 FSR suggests that thequality of euro area banks’ corporate loanportfolios may have started to improve. Themain factors underlying this were the strengthof corporate profitability coupled withfavourable financing conditions. Bankingsectors in most countries seem to havebenefited from a benign credit environment,although marked differences across countriesstill continue to exist. In the foreseeable future,downside risks to corporate credit quality mayarise from a slower than expected economicrecovery in some Member States owing to weakdomestic demand and persistently high oilprices, or from a possible turnaround in thecredit cycle.

The possibility that sluggish corporate creditdemand could hurt banks’ profitability seemsto have gradually eased in the euro areaMember States. Supporting this assessment,bank lending to non-financial corporationspicked up considerably in several countriesbetween mid-2004 and mid-2005. The annualgrowth of short-term MFI loans to non-financial corporations has reached its highestrate since late 2001 (see Chart S30). Increasingfinancing needs for inventories, workingcapital and M&A activity as well as for

leveraged buyout activity may havecontributed to accelerating credit growth. Tothis end, syndicated lending activity in theeuro area has increased, also reflectingsignificant growth in the leveraged segment ofthe market. According to market participants,the activity in syndicated lending is expectedto evolve dynamically (see Box 13). On theother hand, weak fixed investment activityand the increased availability of internalfunding sources owing to improved corporateprofitability continued to have a dampeningeffect on corporate credit demand.

As evidence that supply-side factors mighthave also played a role in stronger loan growth,banks continued to ease their credit standardsfor loans to enterprises in the first half of 2005,and only slightly tightened their creditstandards in the third quarter (see Chart 4.6).The October ECB Bank Lending Survey (seeBox 11) indicates that for the banks in the euroarea, the relaxation of credit standards hasmainly reflected intensified competition, whileit is probably too soon to tell whether thechange in the third quarter constitutes aturnaround in the easing cycle. This changemay in fact at least partly reflect a strongincrease in perceived loan demand, which has

Chart 4.6 Annual growth of euro area MFIloans to non-f inancial corporations andchanges in credit standards(Q1 2003 - Q3 2005)

Sources: ECB and ECB Bank Lending Survey.Note: The net percentages refer to the difference between thesum of percentages for “eased somewhat” and “easedconsiderably” and the sum of percentages for “tightenedsomewhat” and “tightened considerably”.

1

2

3

4

5

6

7

8

2003 2004 2005-50

-40

-30

-20

-10

0

10

20

loan growth (% per annum, left-hand scale)net easing of credit standards (%, right-hand scale)

91cECB



SYSTEMallowed banks to attract borrowers withouthaving to ease credit standards. Additionalinformation on the lending conditions to thecorporate sector can be gathered from recentdevelopments in the syndicated lendingbusiness. There, margins on leveraged loansmight not be appropriately differentiatingacross the risk profile of EU borrowers, thuscreating concerns about possible ongoingmispricing (see Box 13). In the event of slowerthan expected economic growth, it cannot beexcluded that the loosening of credit standards,combined with the low level of provisioning inseveral countries, could leave some banks witha rather thin cushion against deterioratingcredit quality, especially for those banks withlower than average profitability results.

Compared with the assessment in the June 2005FSR, there are indications that risk perceptionsof banks concerning SMEs might have startedto improve in some countries. Country-levelinformation suggests that the increase in thenumber of corporate insolvencies, which hasbeen mainly driven by developments in theSME sector, may have slowed down in the euroarea as a whole and even decreased in somecountries in the first half of 2005. This positivedevelopment, however, should be treated withcaution as in those countries where the numberof insolvencies fell, the decline took place fromhistorically high levels.

Banks in the euro area have reported that therecent easing of credit standards was alsoextended to SMEs, suggesting an increasedwillingness on the part of banks to lend tosmaller companies. Notwithstanding thetentative evidence on the improved riskperception of banks concerning SMEs in someeuro area countries, the loosening of creditstandards might also have reflected a decreasein banks’ risk aversion in an environment ofnarrow interest margins. Developments in thepricing of small loans by banks were mixed inthe first eight months of 2005. While lendingmargins declined in the first couple of monthsof the year, they have since widened in thesecond quarter of 2005 to stabilise in July and

August 2005 at levels recorded at the end of2004 (see Chart S67). In sum, the developmentof margins suggests that banks’ perception ofSMEs’ credit risk fluctuated throughout thefirst eight months of 2005.

Notwithstanding accelerating credit growth tonon-financial corporations since mid-2004,euro area banks’ aggregate exposures at riskare likely to have declined across mostindustries owing to a continued decline in themedian EDFs (see Chart 4.7). Lending to thosesub-sectors which are vulnerable to furtherincreases in oil prices is, however, stillperceived as relatively risky, and could resultin higher credit losses for banks with sizeableexposures to these industries.

Regarding banks’ exposure to the commercialreal estate sector, significant differencescontinue to exist across euro area countries.On many occasions, loans related to thecommercial real estate sector do not accountfor a significant part of the loan portfolio. Onthe other hand, experiences of those countrieswhere banks have sizeable exposures havebeen diverse over the past year. In at least onelarge country, falling rents and commercial real

Chart 4.7 Expected default frequency fordi f ferent euro area industr ia l sectors

(medians)

Source: Moody’s KMV.Note: The sectors are as follows: basic materials andconstruction (BaC), capital goods (Cap), consumer cyclicals(CCy), non-cyclicals (CNC), energy and utilities (EnU),f inancial (Fin), and technology and telecommunications(TMT).

March 2005June 2005September 2005

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

BaC Cap CCy CNC EnU Fin TMT


estate prices continue to pose risks for creditsextended by banks to this sub-sector. In somecountries, banks have reported that riskperceptions may have improved somewhatowing to a recovery in rental growth orrelatively high pre-sale rates for offices.

Finally, developments in the market indicatorsof corporate credit risk have been mixed sincethe publication of the June 2005 FSR. After atightening in early 2005, spreads on bondsissued by euro area non-financial corporationsrose sharply in April and May, triggered by the

credit event affecting the US automobilemanufacturers. Corporate bond spreads havemoderated somewhat since then, but remainslightly higher than in early 2005. EDFs oflarge euro area firms continued to decline in thefirst three months of the year, although thisimprovement appears to have stopped in thesecond quarter. Overall, there are as yet noindications that the recent fluctuation incorporate bond spreads reflects a permanentworsening in perceptions of corporate creditrisk on the part of market participants.

Box 13

THE SYNDICATED LOAN MARKET IN THE EURO AREA MATURES INTO A DISTINCT ASSET CLASS

The global syndicated loan market has grown significantly over the last decade, with the totalamount of gross issuance more than tripling between 1994 and 2004 (see Chart B13.1). Theshare of euro area borrowing rose from close to negligible levels just a few years earlier toreach 25% of global lending in 2005 (from January to October).

On the primary syndicated loan market – where loans are originated – loans can be broadlydistinguished into investment-grade and leveraged according to the credit quality of theborrower. Spreads on investment-grade loans – which still account for the largest share ofgross signings – are at record lows, reflecting high levels of liquidity and strong competition atthis end of the rating scale (see Chart B13.2). Under these circumstances, conditions forborrowers have improved: maturities of extended loans have increased from 4.1 years in 1999to over 6 years in 2004, and a less frequent use of covenants was reported. At the same time, the

Chart B13.1 Annual volumes of syndicatedloans

(USD billions)

Source: Dealogic (Loanware).

0

500

1,000

1,500

2,000

2,500

3,000

0

500

1,000

1,500

2,000

2,500

3,000

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Oct.2005

volume euro areavolume non-euro area

Chart B13.2 Euro area investment andleverage grade volumes

Source: Dealogic (Loanware).

0

50

100

150

200

250

300

350

0

50

100

150

200

250

300

350

1999 2000 2001 2002 2003 2004 Oct.2005

investment-grade volume (EUR billions) leveraged-grade volume (EUR billions)investment average margin (basis points) leveraged average margin (basis points)

93cECB



SYSTEM

1 Second lien loans are usually secured by a second-ranking charge on assets backing senior secured debt, and are typicallysubordinated in their rights to receive principal and interest payments from the borrower to the rights of the holders of seniorsecured debt.

2 PIK instruments pay interest in the form of additional loans (or bonds) instead of cash, thereby increasing the principal, and areclassified as first-loss tranche but with preference in liquidation to equity securities.

3 In November 2004 Standard & Poor’s launched the European Leveraged Loan Index (ELLI), tracking back to December 2002.

Chart B13.3 Secondary trading volume inEurope

(EUR billions)

Source: Loan Market Association.

combination of an erosion in pricing and thepotential weakening of lenders’ positions inthe loan documentation have given rise toconcerns regarding possible mispricing ofcredit risk.

As a result of the hunt for yield, the supply ofcapital for the leveraged loan segment hasgrown significantly, contributing to strongercompetition in the primary market. Unlike inthe investment-grade segment, the margins inthe leveraged loan market have slightlyincreased (see Chart B13.2), althoughaccording to recent market views, thegrowing competition is exerting downwardpressure, and pricing in the euro area may notreflect credit risk differentials, at least when

compared to similarly rated credit in the US. The broadening of the investor base in theleveraged loan market – which now includes insurance corporations, hedge funds andspecialised CDO managers – has promoted product innovations which have in turn led to theemergence of new structured loan tranches with various levels of subordination, such as secondlien loans1 or payment-in-kind (PIK) instruments.2

The growth of the European secondary market, where syndicated loans are traded, reflects thebroadening of the investor base, and was supported by more uniform market practices andstandardised documentation produced by the Loan Market Association; however, at around 6%of the primary market, it remains still small. The secondary market is differentiated into par/near par, leveraged and distressed segments depending on the price to par at which the loan istraded. The secondary leveraged trading volume, non-existent at the launch of the euro, hassince rapidly developed into the most important secondary market segment, accounting forabout 50% of the trading in the first half of 2005 (see Chart B13.3).

Three recent developments in the leveraged loan market provide a further indication of theestablishment of the syndicated loan market as a distinct asset class. Firstly, syndicated loans areincreasingly assigned ratings by the rating agencies, and a new set of recovery ratings (estimatingthe likely recovery of the principal in the event of default) has been introduced. Secondly, the firstindex for leveraged loans that provides a benchmark against which investors can assess theperformance of their investments was launched3, and is likely to enhance transparency andliquidity in the market still further. Thirdly, the arrival of CDS referencing European leveragedloans may attract additional investors as it will enable them to manage actively their exposure.

These developments, coupled with considerable growth in both the primary and the secondarymarket as well as significant changes in the product structure, have contributed to turn

0

5

10

15

20

25

30

35

0

5

10

15

20

25

30

35

1999 2000 2001 2002 2003 2004 H12005

par/near pardistressedleveraged


leveraged loans into a debt product that allows financial market participants to tailor theircredit risk exposures. However, although these developments are indeed contributing toensuring a better allocation of risks in financial markets, they may also give rise to concerns onfinancial stability grounds since, according to some market participants, credit risk may not beadequately reflected in the pricing, which is evident in the small or non-existent differences inprices for different layers of credit quality. At present, pricing is mainly driven by strongcompetition in the primary market and reflects to a lesser extent credit risk differentials.Pricing conditions that may currently appear to be appropriate could prove inadequate in thecase of a turn in the credit cycle. A rise in global interest rates and an increase in marketvolatility, coupled with a significant increase in corporate sector leverage, may well contributeto such a turn. An improvement in the pricing of syndicated credits is however on the horizon.The recent launch of an official index for leveraged loans is expected to lead to better creditdifferentiation in the pricing. Furthermore, the impact of Basel II may lessen the prevailingimportance of the current supply-demand dynamics and could also promote more efficient loanpricing with differentiation according to the risk profile of the borrowers.

MARKET-RELATED RISKS

Interest rate risks not materially changedBanks are exposed to interest rate risks eitherdirectly, via interest rate-sensitive positions intheir trading and banking books, or indirectlythrough credit risk. Interest rate risk scenariosmay include, on one hand, a sudden upturn inlong-term interest rates or, on the other hand,a protracted low interest rate environmentassociated with a flattening of the yield curve. Itshould be emphasised, however, that while thegradual increase of long-term interest ratesremains the main scenario, a more protractedperiod of low long-term yields cannot beexcluded. Such a scenario would imply risks tobanks mainly due to reduced earnings from thecore maturity transformation business.

The relative size of capital requirementsprovides estimates for banks’ direct exposures tointerest rate risk. Information available oninterest rate VaR for a sample of 15 large euroarea banks indicates that several large banksincreased their exposure to interest rate risk in2004, whereas others reduced their exposure to asignificant extent (see Chart 4.8). Based oncountry-level information and financial resultsof large banks, some banks might have furtherincreased their risk-taking in the first half of2005. It is important to stress, however, that the

VaR values remain low relative to banks’ capitalor income.

Interest-sensitive positions in the bankingbook are another source of potential directlosses stemming from adverse changes ininterest rates. However, given the lack ofharmonised information across both banks andMember States, it is difficult to gauge therelative importance of this type of riskexposure across the euro area banking sector. Itcan, nevertheless, be expected that banks incountries where fixed income portfolios

Chart 4.8 Interest rate VaR for selectedlarge euro area banks

(2003 - 2004, % of Tier 1 capital)

Source: Financial disclosures of banks.

20032004

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

bank1

bank2

bank3

bank4

bank5

bank6

bank7

bank8

bank9

bank10

bank11

bank12

bank13

bank14

bank15

95cECB



SYSTEMaccount for a significant part of the bankingbook may be particularly vulnerable to interestrate increases.

The indirect effects of potential interest rateincreases that could materialise via credit riskmay be more important for banks than the riskof direct losses.7 In this context, the risk ofincreased correlation between interest ratesand credit risk may also be considered as animportant source of risk for banks. A possiblesudden widening of credit spreads triggered byan increase in risk aversion or a significantnegative credit event could affect banksthrough a rise in credit risk in relation to certainasset classes (such as corporate and emergingmarket-related loans), as well as through anincrease in trading book losses.

To conclude, there are some indications thatbanks may have increased their trading book-related exposures over the past year, althoughthe relative size of these exposures appears tohave remained modest. Nevertheless, indirecteffects implied by certain interest rate riskscenarios may be more significant, although ondifferent time horizons.

In the short term, an unanticipated large increasein long-term interest rates could have aconsiderable short-term impact on banks’profitability by prompting a deterioration incredit quality and lower credit demand. Tomaintain profitability, these losses would haveto be balanced by increased income frommaturity transformation on new loan businesses.

A prolonged period of low long-term interestrates, on the other hand, would have a negativeimpact on banks’ net interest income and couldencourage banks to seek revenues from riskieractivities such as lending to emerging markets.Moreover, the low interest rate environmentcould contribute to a further build-up offinancial imbalances, for example through arise in private sector indebtedness, which mayhave an impact on banks’ credit risk, but with asignificant time-lag.

Direct exchange rate exposuresremain containedBanks’ foreign exchange rate exposures can beclassified into two types: direct and indirect.Country-level information available for thefirst half of 2005 suggests that no significantchange in the direct exposure of banks toexchange rate risk has taken place. Banks’ openforeign exchange positions have in generalremained low, both in absolute terms and as ashare of regulatory own funds. Consolidateddata for 2004 show that the share of euro areabanks’ trading book own funds requirementsfor foreign exchange risk remained broadlyunchanged from 2003 (see Table S8). Theseexposures are generally lower than therequirements for debt or equity instruments.

Moreover, there is little evidence of currencymismatches between the asset and liabilitysides of banks’ balance sheets on an overallbasis. Nevertheless, some risks may exist forbanks with branches or subsidiaries in fastergrowing retail markets. These institutions maybe exposed to currency risk via unhedgedcurrency mismatches as well as to thetranslation of currency changes on the balancesheets of subsidiaries recorded in othercurrencies than their domestic currency.

Banks can also be indirectly exposed toforeign exchange risk through changes inthe competitiveness of non-financial firmsor defaults of loans denominated in foreigncurrencies. The former type of indirect effectswould have an impact not only on banks withopen positions in foreign currencies, but ingeneral on all banks whose borrowers’ creditquality could suffer because of a loss ofcompetitiveness. The latter type of indirecteffects could materialise via repaymentdifficulties triggered by adverse currencymovements in those countries where substantiallending in foreign currencies, both tohouseholds and corporates, has been encouragedby favourable interest rate differentials. So farthere has been little sign that the fluctuations in



the euro exchange rates have translated intobalance sheet problems for firms. However, it ispossible that large exchange rate movementsrelated to a correction in global imbalances andthe associated adjustments in consumption indeficit countries could pose problems for euroarea banks.

Equity market exposures increasedConsolidated data for 2004 show that theexpansion of trading activity contributed to aslight increase in euro area banks’ exposures toequity market risk as measured by the increasein the share of the banks’ trading book ownfunds requirements for this risk category (seeTable S8).

Since the publication of the June 2005 FSR,banks’ income is likely to have increased bothin terms of fees and commissions and in termsof trading activity, given that equity markets inalmost all euro area countries recorded furthergains between June and early November 2005.In some cases these revenue sources could havemore than offset lower income from the corematurity transformation business.

Exposures to hedge funds warrant monitoringSince early 2003, inflows into hedge fundshave grown at a brisk pace. This developmentcontinued throughout 2004 and into the firsthalf of 2005, although with some moderationin the second quarter of 2005 due to poorerhedge fund return performance. In particular,certain credit strategies suffered in March-April 2005, owing to market developmentsassociated with the downgrades of GM andFord to sub-investment-grade (see Chart 4.9).As a result, some banks may have faced higherrisks on their trading and lending exposures tohedge funds during this period.

Banks’ agreements with hedge funds usuallyinclude triggers of net asset value (NAV)decline, which when breached would allowbanks to terminate all transactions and seize thecollateral held. NAV can fall for two mainreasons: negative performance, or investorredemptions. Commonly, 15-20%, 25-30% and35-45% thresholds are applied for NAVdeclines during one, three and 12 monthsrespectively on a rolling basis. In the first halfof 2005, the share of hedge funds breaching

Chart 4.10 Share of hedge funds breachingtr iggers of net asset value (NAV) decl ine

(Jan. 1994 - June 2005, % of total NAV)

Sources: Lipper TASS database and ECB calculations.Note: Excluding funds of hedge funds. NAV is equivalent tocapital under management, i.e. its declines includeredemptions. If several assumed NAV decline triggers werebreached, then the fund in question was included into onlyone group with the shortest rolling period.

-40% on a rolling 12-month basis-25% on a rolling 3-month basis-15% on a monthly basis

0

5

10

15

20

25

0

5

10

15

20

25

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Chart 4.9 Distr ibution of hedge fundreturns

(Jan. 2004 - June 2005, %, net of all fees)

Sources: Lipper TASS database and ECB calculations.Note: Excluding funds of hedge funds.

-8Jan. Apr.

2004July Oct. Jan. Apr.

2005July

-6

-4

-2

0

2

0

20

40

60

median - 25th percentile25th - 10th percentile10th percentile - minimum

median25th percentile10th percentile

% of capital by return ranges, right-hand scale

Distribution of returns,left-hand scale

97cECB



SYSTEMassumed triggers was higher than in 2004 (seeChart 4.10). Barring weak performance orhigher levels of investor withdrawals, this factcould also indirectly signal a more active use ofleverage among hedge funds globally.

An additional issue is the potential operationalrisk related to the increased involvement ofhedge funds in the markets of complexstructured credit products. The tendency ofhedge funds to take a more short-term view onthese products, as opposed to traditionalinvestors who often hold the instruments tomaturity, has increased the number of settlementprocedures required to verify the holder of theasset after a trade has taken place. It cannotbe excluded that in more stressed marketconditions, when a large number of investorsmay be willing to exit similar trading positions

simultaneously, accumulated settlement backlogscould worsen the distortions to the functioningof the markets that are generated by anunwinding of “crowded trades”.

For banks, the greatest challenge remains theinteraction of market, credit and illiquidityrisks when larger price movements orunexpected changes in correlations force hedgefunds to liquidate their leveraged and possiblyundiversified positions. In such cases,leveraged market risk faced by hedge fundsleads to credit risk for banks, which could befurther exacerbated by the drying up ofliquidity in affected markets. Moreover, the so-called crowding of hedge fund trades couldfurther magnify the risks for banks, especiallyif banks’ proprietary trading desks use similarhedge fund-like strategies (see Box 14).

Box 14

LARGE EU BANKS’ EXPOSURES TO HEDGE FUNDS

Over the last couple of years, the hedge fund industry has expanded rapidly. Because of theimportant role that hedge funds play as participants in financial markets and as counterpartiesto financial institutions, especially banks, it has become increasingly important to monitortheir activities and to assess the implications for financial stability. Against this backdrop, theESCB Banking Supervision Committee decided to investigate the links between EU banks andhedge funds. This Box reports on the main findings of this survey.1

The survey excluded subsidiaries and branches of non-EU banks, some of which, primarily USones, were leading global financing and trading counterparties of hedge funds. More than 40EU banks from 14 countries provided qualitative comments and sometimes quantitative dataon their connections with hedge funds. Based on the provided coverage information, 35surveyed large banks (including 11 smaller banks with mainly investment exposures) as agroup constituted around 1%, 55% and 38% of respectively the total number, consolidatedassets and Tier 1 capital of all eligible banking groups in these countries. Some quantitativedata was supplied by 22 large banks from seven EU Member States.

Regarding banks’ financing exposures, at the end of 2004, for the 14 large banks from sixcountries (AT, DE, ES, FR, NL and SE), the absolute amount of cash lending to hedge fundscollateralised with securities (e.g. via reverse repurchase agreements) totalled almost €100billion, and large banks from two countries clearly dominated in the sample (see Chart B14.1).For the smaller sample of five banks from four countries, which also provided 2003 data,lending increased 1.5 times in 2004. In general, banks extended either no or only negligible

1 ECB (2005), Large EU banks’ exposures to hedge funds, November.


Chart B14.1 Cash lending to hedge fundscol lateral ised with securit ies

(end-2004, % of Tier 1 capital and assets)

Chart B14.2 Investments in hedge funds

(end-2004, % of Tier 1 capital)

Source: Banking Supervision Committee.Note: Based on a sample of 14 large banks from six EUMember States.

Source: Banking Supervision Committee.Note: Based on a sample of 16 large banks from six EUMember States.

1.5%

50.2%

0 2 4 6 8 10

TotalCountry F Country E Country DCountry C Country B Country A

TotalCountry F Country E Country D Country C Country B Country A

500 100 150 200 250

% of Tier 1 capital (higher axis)

% of assets (lower axis)

0

1

2

3

4

5

6

7

8

9

0

1

2

3

4

5

6

7

8

9

totalcountry F

country E

country D

country C

country B

country A

managed within a groupmanaged by unconnected management firms

1%1%2%

8%

4%

2%

3%

4%

2%

amounts of unsecured lending, and many banks had policies completely forbidding unsecuredcredit exposures to hedge funds. A number of banks indicated that lending spreads had declinedover 2004, especially for lending to larger hedge funds, as competition in this segment was themost intense.

In many EU Member States investments in (funds of) hedge funds were the major andsometimes the only form of direct links with the hedge fund industry. Banks saw suchinvestments as a way of gaining attractive risk-adjusted returns and improving thediversification of their investment portfolios. At the end of 2004, the total amount ofinvestments in hedge funds by 16 large banks from six countries (AT, DE, ES, FR, NL and SE)exceeded €9.4 billion, although most of these investments were made by large banks in twocountries (see Chart B14.2). In 2004, total investments by the smaller sample of five banksfrom four countries that also provided 2003 data increased by 52%, and allocations tounconnected hedge funds grew more rapidly.

Regarding trading exposures, for five large banks from three countries (DE, FR and SE) theestimated gross market value of OTC contracts outstanding with hedge funds in derivativesmade up 2.7 % of all outstanding banks’ OTC contracts in derivatives at the end of 2004. In thecase of OTC interest rate derivatives, the share was 2.4%. Based on banks’ comments and somequantitative evidence, it seems as if hedge funds were not key banks’ counterparties in creditrisk transfer markets and probably, on aggregate, were net credit protection buyers from banks.

Finally, on banks’ income exposures, according to the quantitative data from nine large banksfrom four countries (AT, FR, NL and SE), banks earned nearly €0.8 billion from hedge funds in2004. However, the share of net income derived from hedge funds was not high in relation tototal net income and its sub-components, although proportions were higher for net tradingcommissions (see Chart B14.3). Across countries, net trading commissions made up the largestshare of total net income derived from hedge funds. Moreover, for the smaller sample of fourbanks from three EU Member States that also provided 2003 data, the growth of total netincome and its sub-components derived from hedge funds was much faster than the net income

99cECB



SYSTEM

growth from all activities in 2004. This positive contribution may further intensify banks’efforts to foster hedge fund-related business and to attract more hedge fund clients, most likelyputting further pressure on applied price and non-price credit terms.

With respect to risk management practices, most banks that extensively dealt with hedge fundshad specific guidelines for this interaction as well as advanced risk management systems, orwere in the process of further enhancing them. Surveyed banks generally had stringentrequirements for exposures to hedge funds, with a strong emphasis on collateralisation.Nearly all cash-lending exposures to hedge funds were collateralised. Moreover, many bankswith higher financing and trading exposures used sophisticated potential future creditexposure (PFE) measures to calibrate the expected downside risks of their hedge fundexposures that arise from the interaction of market, credit and illiquidity risks. Most banks alsoreported that they used stress tests to evaluate the potential effects of volatile or illiquidmarkets on their exposures. Regarding recent developments, banks did not see any systematicincrease in risk-taking, as leverage levels across hedge fund clients seemed to be moderate andlower than in 1998, even though funds of hedge funds were reported to be increasing leverage.It has to be noted, however, that banks generally did not have any information on off-balancesheet leverage arising from trading in derivatives.

The survey also highlighted several areas with scope for further improvement that couldbecome a cause of concern, particularly if the current rather benign market conditions were tochange abruptly. These are:

(i) counterparty discipline, as applied by banks, was found to be under pressure owing tohighly competitive market conditions. Hedge funds, particularly the larger ones, weresuccessful in negotiating less rigorous credit terms, including, for example, lowerlending spreads, higher NAV decline triggers or trading on variation margin only;

Chart B14.4 Scope of stress test ing

(% of banks, answers not mutually exclusive)

Source: Banking Supervision Committee.Note: Based on a sample of 20 large banks from seven EUMember States.

10%

70%

15%

30%

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

no stress tests individual hedge fundexposures

exposures to groups of

hedge funds(strategies)

all exposures to hedge

funds

Chart B14.3 Share of net income derivedfrom hedge funds

(2004)

Source: Banking Supervision Committee.Note: Based on a sample of eight large banks from four EUMember States.

3%

6%

3%

1%0

3

6

9

12

0

3

6

9

12

total net income

net trading commission

net other non-interest

income

net interest income

totalcountry Acountry Bcountry Ccountry D


Emerging market exposures increased furtherAs discussed in detail in sub-section 1.1,against a background of increasing commodityprices and a low interest rate environment, theeconomic performance in many emergingmarkets was rather favourable in the first halfof 2005. Consequently, available data suggestthat euro area banks further increased theirexposures to these countries (see Charts S59and S60 and Table S4).

Banks have particularly increased their alreadylarge exposures to Latin American countrieswhere most large economies, with theexception of Argentina, have seen furtherinflows. Exposures have risen particularly

strongly in Mexico and in Brazil, owing to therather benign economic performance of thesetwo countries which could have eased earlierconcerns about risks arising from investmentsin the local bond markets. Completednegotiations on the restructuring of thedefaulted debt in Argentina could havemitigated one source of uncertainty in themarket. Nonetheless, international investors,including some euro area banks, were forced toaccept a significant discount on their claims ondebt issued by the Republic of Argentina.

Exposures of euro area banks to selectedmarkets in Asia also either stabilised orincreased further.

(ii) most stress tests applied by banks, particularly the regular ones, included only historicalscenarios and often were applied to individual hedge funds only (see Chart B14.4). Inaddition, the stress testing of collateral was less common and offers further scope forimprovement;

(iii) aggregation by banks of their exposures to hedge funds across the entire financial groupand/or different business areas/geographical regions was sometimes seen as problematic;

(iv) hedge fund disclosures and information on leverage were, despite some progress, laggedand not always adequate. In many cases hedge funds still provided banks with relativelycrude measures of leverage, although an increasing number of hedge funds weresupplying more advanced risk-based measures of leverage;

(v) banks’ descriptions of their risk management practices also raised questions whetherbanks were sufficiently taking into account and/or had enough timely information on thewhole portfolio structure of hedge funds, particularly on the larger ones with financingand trading relationships with several counterparties.

All in all, direct exposures to hedge funds of the large EU banks surveyed varied acrosscountries and generally were not substantial in relation to banks’ balance sheets and totalrevenue. However, even the limited data provided indicated that exposures were growingrapidly, although in most EU Member States these have remained negligible and/or mainly inthe form of investments. It is very likely that the absolute and relative size of exposures tohedge funds will increase further in line with the continuing expansion of the whole hedge fundindustry, and in particular its European segment. Most of the recommendations that were raisedafter the near-default of LTCM remain relevant, and banks should further continue tostrengthen the risk management of their exposures to hedge funds. Moreover, banks shouldresist market pressures to lower credit standards applied and should continue to insist on moretransparency from hedge funds.

101cECB



SYSTEM4.3 SHOCK ABSORPTION CAPACITY OF THEBANKING SECTOR ON THE BASIS OFMARKET INDICATORS

MARKET INDICTORS CONTINUE TO SUGGEST APOSITIVE SHORT-TERM OUTLOOKThroughout much of 2004 and 2005, financialmarket indicators suggested that marketparticipants were still optimistic and confidentabout the robustness of euro area banks and thebanking industry’s future earnings prospects.Therefore, for the time being, the markets seemrather convinced that the euro area banks’ shockabsorption capacity remains comfortable,despite the continued decline in loan lossprovisioning reported in 2004.

The ratio of banks’ share prices relative to thegeneral stock market index rose overallbetween late 2003 and 2005, reaching newheights in early November after a temporarydecline in the second quarter of 2005. Despite aslight decrease of late, the bank stock marketindex remains well above the market index(see Chart 4.11). This suggests that market

participants may have become more optimisticabout the future earnings prospects of banks,and/or that they perceived the risks facing thebanking industry to have declined. On theperformance side, euro area banks’profitability improved significantly in 2004and in the first half of 2005 (see also sub-section 4.1), creating expectations of furthergrowth. On the vulnerabilities side, euro areabanks’ credit risk outlook also has tentativelyimproved (see sub-section 4.2). In particular,the substantial reduction in the number ofcorporate sector defaults and non-performingloans is legitimising market participants’rather benign perception of the risks lyingahead.

Despite the generally favourable developmentsat the level of the euro area banks’ stock index,uncertainty – as captured by increased standarddeviation of the RND function derived fromoptions prices – has increased since the firstquarter of 2005 (see Chart 4.12). This couldreflect uncertainty regarding the generalmacroeconomic and financial environment,

Chart 4.12 Risk-neutral probabi l ity densityfunction on the Dow Jones EURO STOXX bankindex

Sources: Bloomberg and ECB calculations.Note: The methodology for estimating the risk-neutralprobability density functions is explained in Box 13 of theJune 2005 FSR. The full range of strike prices needed tocalculate the function can generate a measure of standarddeviation that can from time to time slightly deviate from theindex implied volatility f igures that are also provided byBloomberg and reported in Charts S66 and B15.1.

15 March 2004 15 March 2005 4 November 2005

0.00

0.01

0.02

0.03

0.04

0.05

0.00

0.01

0.02

0.03

0.04

0.05

200 220 240 260 280 300 320 340 360 380 400

value of the Dow Jones EURO STOXX bank index


Chart 4.11 Ratio of the Dow Jones EUROSTOXX bank index to the overal l marketindex for the euro area(Jan. 1999 - Nov. 2005, index: Jan. 1999 = 100)


70

80

90

100

110

120

70

80

90

100

110

120

1999 2000 2001 2002 2003 2004 2005


and the potential impact on banks’ futureprofitability. Nevertheless, possibly reflectinga series of positive financial results by severallarge banks published in late October and inearly November, the function remained skewedto the right, suggesting that market participantscontinued to assign a higher probability tonear-term increases rather than to decreases inthe stock index.

Supporting the hypothesis of positive futureearnings prospects, in the third quarter of 2005the earnings-per-share forecasts for largebanks in the euro area showed a continued verypositive outlook. (see Chart 4.13).

The fact that the euro area banks’ earningsgrew rather steadily in 2004 and untilearly November 2005, while stock pricestemporarily dropped in the second quarter, isreflected by the V-shaped development in theprice-earnings ratio (see Chart S65). Growth inbanks’ earnings in this period was still mainlydriven by improvements in cost efficiency,lower provisions for loan losses, and afurther increase in lending to households, ashighlighted in sub-section 4.1.

Turning to market-based assessments of theresilience of the euro area banking sector, themedian distance-to-default (DD) for the set oflarge euro area banks fluctuated throughout thefirst ten months of 2005 between verycomfortable levels (see Chart S62).8 For banksin the lowest percentile, there was a notableimprovement in the DD in early 2005, followedby a drop in the second quarter. The temporaryincrease in default risks most likely reflectedthe turbulence in credit markets caused by theGM/Ford downgrades in April 2005. Overall,the behaviour of this indicator is in line with theimprovement in euro area banks’ performanceand the overall balance of risks facing thebanking sector discussed in the other parts ofthis Review. According to this information, theeuro area banking sector would, therefore, be ina comfortable position to honour its debtobligations, and continues to move away fromthe possibility of future financial distress.

CDS spreads on the debt of euro area financialinstitutions steadily declined for most of 2004.However, these spreads widened in the secondquarter of 2005 (see Chart 4.14), with thisdeterioration apparently being driven by thetwo large corporate downgrades. Investors’concerns about possible exposures of largeeuro area banks to the troubled automobilesector proved to be limited, although by nomeans negligible, as witnessed by the wideningof euro area banks’ subordinated debt spreadsto over 50 basis points. These spreadssubsequently reverted back towards the lowlevels seen at the beginning of 2005.

Spreads in the offers to buy and sell protectionon European non-financial institutions’ debtsuffered a similar widening in the secondquarter. Although the widening of thesespreads was also rather short-lived, spreadsresumed at levels slightly higher than thoseseen in the first quarter of 2005. Throughout2004 and 2005 spreads on both senior andsubordinated debt issued by financialinstitutions remained systematically lowerthan spreads on non-financial institutions’debt, which reflects confidence among

Chart 4.13 Large euro area banks’ earningsper share (EPS) and 12-month-aheadforecasts(Q1 1999 - Q3 2006, weighted average, %)

Sources: Thomson Financial Datastream, I/B/E/S and ECBcalculations.

Q3 2005Q1 2005

1

2

3

4

1

2

3

4

1999 2000 2001 2002 2003 2004 2005 2006

8 The DD measures the distance between the market value of af irm’s (a bank’s) assets and the point at which it is insolvent. Formore details, see ECB (2005), Financial Stability Review, June,Box 14.

103cECB



SYSTEM

investors regarding the resilience androbustness of the euro area financial sector.

A comparison between the message thatemerges from the forward-looking financialmarket indicators, on one hand, and the viewthat is provided by the assessment of risks inthe previous sub-section, on the other hand,points towards a common baseline scenario ofcontinuing euro area banking sector stability.

However, while the market indicators analysedin this section necessarily focus on the mostlikely outcome, risk assessment can providesome hints about the distribution of risksaround this baseline scenario. On the negativeside, there is uncertainty regarding the near-term macroeconomic outlook, particularly dueto the very high level of oil prices, continuingimbalances in the global financial system, aswell as the increasing indebtedness of thehousehold sectors in most euro area countries.Given the above, the vulnerabilities arisingfrom the low level of provisioning by euro areabanks could materialise in a scenario wherebycredit conditions deteriorate more markedly.On the positive side, high levels of profitabilitycould have increased banks’ resilience in theshort term. In addition, banks in the euro areahave improved their risk management, which islikely, at least partially, to shelter them fromsevere losses in the event of worsening creditquality.

Finally, looking at the developments in thevolatility of euro area banks’ stock prices, thispositive assessment could extend to themarkets’ current perception regarding the riskof a cascading failure in the euro area bankingsystem (see Box 15).

Chart 4.14 European f inancial and non-f inancial inst itutions’ credit default swaps

(May. 2002 - Nov. 2005, basis points, five-year maturity)

Source: JP Morgan Chase and Co.Note: European f inancial institutions and non-f inancialinstitutions correspond to JP Morgan Chase & Co.def initions.

financial institutions subordinated debtfinancial institutions senior debtnon-financial institutions

020406080

100120140160180200

020406080100120140160180200

2002 2003 2004 2005

Box 15

A DECOMPOSITION OF EURO AREA BANK STOCK VOLATILITY

Since the recovery of the euro area stock markets started in March 2003, the Dow Jones EUROSTOXX banking sector index has performed strongly, increasing by roughly 80% up toNovember 2005.1 Moreover, the upturn in euro area bank’s stock prices has been accompaniedby a declining trend in various measures of stock market uncertainty. In this respect, bothrealised volatility and forward-looking measures such as implied volatility on options ofthe bank index have declined in tandem to relatively low levels by historical standards (seeChart B15.1). This development of lowered uncertainty would imply that the risks facing theEuropean banking sector are currently assessed as rather benign. By using data on individualbank stocks, this Box decomposes the decline in overall bank index volatility into two separateparts: the first measuring the contribution from single stock variances; and the secondreflecting the covariation between stocks making up the index. Given that the degree of

1 The Dow Jones EURO STOXX 50 has increased by slightly less than 50% over the same time period.


covariation among individual banks’ stock returns is important for assessing the risk ofcommon vulnerability to similar shocks, and that this covariation from time to time may bedominated by some subset of banks (possibly changing over time), this decomposition mayalso provide some tentative hints about changes in the “systemic risk” embedded in financialmarket prices.2

In order to extract time-varying measures of the volatility of individual stocks and thecorrelation structure between them, a multivariate GARCH model of daily returns for 38 banksincluded in the Dow Jones EURO STOXX banking sector index is employed.3 These estimatesare then used – together with the individual weights of the banks represented in the index – todecompose the overall index volatility into both the individual variance part and the covariancepart. The standard expression describing the variance of a portfolio of assets is used for thedecomposition:

Portfolioσ ∑=2 , (1)

where �2Portfolio is the total variance of the portfolio and wi represents the weight of the i:th

individual stock. Chart B15.1 shows an annualised version of the calculated portfolio volatilitytogether with the evolution of the implied volatility for the banking index and the realisedindex volatility. Given the rather small and unsystematic difference between the threevolatility measures, the multivariate volatility model that underlies variance expression (1)may be regarded as providing a fairly sound illustration of volatility developments over thesample under consideration.4

One simple way of gauging information regarding euro area banking sector risk is to considerthe evolution of components of expression (1), which sums elements of the co-movementbetween stocks. For the purposes of financial stability analysis, risk assessment typicallydepends on these covariance patterns, even if overall volatility is low. However, not all banksrepresented in the index are necessarily equally important in terms of their contribution to thetotal covariance component. Taking this consideration into account, the variancedecomposition proposed here might shed some light on the impact of some subset of banks ontotal volatility.

The analysis shows that the contribution from the covariance part in (1) has remained more orless unchanged over the last few years for the index in total, fluctuating at around 85% of totalportfolio variance.5 The stable evolution of this proportion confirms that the recent decline inthe volatility of the banking index has been driven by a reduction in the covariation betweenindividual bank stocks.

2 Bank stocks are assumed to be efficiently priced in that they reflect all publicly available information, both in terms of individualbanks’ balance sheet risks and the relationships between different banks’ risks.

3 Ten banks, accounting for less than 9% of the total bank index, have been excluded from the calculations owing to data limitations.Data from 1 June 2000 to 1 November 2005 have been used to estimate a Dynamic Conditional Correlation model with GARCH(1,1)margins. See R. Engle (2002), “Dynamic Conditional Correlation – A Simple Class of Multivariate Generalized AutoregressiveConditional Hetroskedasticity Models”, Journal of Business and Economic Statistics, Vol. 20, No. 3, July.

4 Comparing realised volatility and volatility measures extracted from options prices may be somewhat misleading, as the former isbackward-looking while the latter is forward-looking. However, as market participants tend to use realised volatility when formingtheir expectations about future volatility, the two measures usually exhibit similar movements over time.

5 This f inding is not surprising, given that the bulk of total risk in a diversif ied portfolio should be made up of the covariance betweenthe individual stocks.

ijj

n

i

n

ji ww σ∑∑

i = 1 j = 1

105cECB



SYSTEM

From a financial stability perspective, increased co-movement between large institutionsmight be perceived as particularly important. An estimate of the variance contribution from asubset of the ten banks with the highest weights in the index – comprising about 60% of thetotal market capitalisation – shows that the level of co-movement actually declined betweenMarch 2003 and November 2005. Chart B15.2 shows that the contribution from these teninstitutions alone in terms of covariation was reduced by approximately 10% of totalcovariance, thereby suggesting a slight reduction in “systemic risk” as measured by thisparticular indicator. Thus, the change in the covariance structure among euro area bankingstocks, together with the observed reduction in implied volatility, lends some support to theview that risks are currently assessed as being manageable for the banking sector as a whole.

Chart B15.2 Covariance contribution of theten largest banks included in the Dow JonesEURO STOXX bank index(% of total index covariance)

Source: ECB calculations.

60

65

70

75

80

85

2001 2002 2003 2004 200560

65

70

75

80

85

large banks’ contribution to total covariance30-day moving average

Chart B15.1 Real ised, impl ied and model-based volat i l i ty on the Dow Jones EUROSTOXX bank index(annualised standard deviation of daily returns, %)


0

10

20

30

40

50

60

70

2001 2002 2003 2004 20050

10

20

30

40

50

60

70

realised volatilityimplied volatilitymodel volatility

THE UPWARD MIGRATION IN CREDIT RATINGSHAS SLOWED DOWNThe generally positive assessment of theresilience of the euro area banking sector,although qualified by downside risks, isbroadly shared by the three major ratingagencies, which report that the average creditquality of EU (debt) issuers improved further in2004 and in the first two quarters of 2005,albeit at a decelerating rate. However, for theset of western European banks, Moody’sreports a significant decrease in the upgrade-to-downgrade ratio from 4 in 2004 to 0.5 in thefirst half of 2005. This indicates thatthroughout the first six months of 2005, thenumber of downgrades was twice the number ofupgrades (see Chart 4.15 where the dotted linedenotes the difference between upgrades anddowngrades). Moderate economic growth and

increasing competition were seen as the mainfactors behind the slowdown in the upwardmigration rate. The still positive assessment ofthe credit quality of banks was, on the otherhand, supported by the strength of banks’earnings growth in 2004 and the first half of2005, as well as some signs of acceleration inthe pace of cross-border consolidation in theEU banking sector.

The ratings of the largest euro area banksremained stable during the second quarter of2005. However, there seems to remain littleupside potential for the long-term ratings ofthese banks, as the majority are now rated in the“AA” range. Most upgrades affected banksrated within the “A” to “B+” range, while therewere no upgrades of banks with a rating of “A”or higher.


4.4 OVERALL ASSESSMENT

The financial conditions of euro area banksimproved in 2004 and the first half of 2005,continuing the positive developments recordedin 2003. Profitability increased for the fullspectrum of banks, including those nationalbanking sectors that had substantiallyunderperformed in previous years. Thisdevelopment was associated with sustainedgrowth in lending to households, mostly forhousing purposes, and an incipient recovery inlending to the corporate sector, including toSMEs. However, banks also experiencedincreased competition from their peers ingranting loans, especially to households. Thisfact contributed to a continuing decline in theshare of net interest income in total income andstill rather narrow lending margins.

While euro area banks at least partiallycontinued to exploit the benefits toprofitability that can be reaped from cost-cutting, their profitability was also affected bythe further decline in provisions. Given thecontemporaneous improvement in assetquality, however, the coverage ratio hasincreased.

Overall, considering the outlook for euro areabanks in the foreseeable future, broadlypositive short-term indications of banks’conditions suggest that, notwithstanding somedifferences across countries, the euro areabanking sector has benefited from generallybenign credit and liquidity conditions. On theother hand, the banks in the euro area haveachieved improved financial results in anenvironment of sluggish economic growth inmany of the largest Member States. On thewhole, the assessment suggests that the euroarea banking sector is rather robust. However,from a longer-term perspective, certaininternal and external risks to the banking sectorcould emerge.

Concerning internal risks, banks have facedmounting difficulties in increasing ormaintaining their interest income and interestrate margins in mature markets, given the lowinterest rate environment and strongcompetition from their peers to securesustained lending growth. As a result, banksmay have started to loosen their creditstandards, possibly increasing their futureexposure to credit risk, or might have begun tosearch for alternative, possibly riskier, sourcesof income. A second internal risk is representedby the historically very low levels of provisionsat present. Although the coverage ratio hasincreased, low levels of provisioning have leftbanks exposed to potentially high costs in theevent of a deterioration in credit quality.Moreover, banks could potentially havealready largely exhausted the boost toprofitability that can be derived from cost-cutting or cost containment, thus reducing thenumber of sources they may resort to in order tomaintain satisfactory profitability levels in theevent of a deterioration in the creditenvironment.

The main external risks include increasinglyhigh oil prices and persistently wide globalimbalances. Although the direct exposures ofeuro area banks to these risks are likely to belimited, the risks posed by exchange rates

Chart 4.15 Western European bankingsector rating downgrades, upgrades andbalance(1996 - Q3 2005, number)

Source: Moody’s.Note: “Western European” corresponds to Moody’s owndef inition.

upgradesdowngradesbalance

-25-20-15-10

-505

10152025

-25-20-15-10-50510152025

1996 1997 1998 1999 2000 2001 2002 2003 2004 Q32005

107cECB



SYSTEMcould prove to be correlated with the creditrisks of non-financial sectors – especially firmsin the export sector and their suppliers such asSMEs.

The interest rate risks faced by banks reflect thesomewhat different impact that these externalvulnerabilities could exert, if they were tomaterialise. While the main scenario remainsone of gradually increasing long-term interestrates, the likelihood of protracted low long-term interest rates and a flat yield curveenvironment might have increased, potentiallyexposing banks to further difficulties ingenerating interest income from their corebusiness. The current tightness of creditspreads could represent greater than normalmarket risk for banks, and may also haveaffected their pricing of risks in the context ofcorporate lending. Owing to the incipientrecovery in borrowing from the non-financialcorporate sector, which, per se, is a broadlypositive development for banks’ incomedifferentiation, banks have also become morevulnerable to potential mispricing of creditrisk.

Market indicators present a broadly positiveforward-looking assessment of banks’ risk andprofitability outlook. However, most marketindicators tend to focus on the most likelyoutcome, whereas an overall risk assessmentshould emphasise the most disruptive,although still plausible, outcome in terms ofdistribution of risks around this baselinescenario. The indicators that capture thedistribution show somewhat increaseduncertainty about future developments inbanks’ profits, possibly reflecting uncertaintyregarding the macroeconomic and financialenvironment.

All in all, the current financial condition ofeuro area banks is broadly satisfactory, and nospecific items can be identified that wouldtrigger the emergence of the more negativescenario. However, despite the overall positiveassessment in the near future, the combination ofincreased loan growth and lower provisioningmay leave banks more exposed to credit risk andto a sudden deterioration in the currently benigncredit and liquidity environment.


5 OTHER EURO AREA FINANCIALINSTITUTIONS

Profitability in the euro area insurance industryimproved further in 2004. The strengthening offinancial positions was mainly driven by strongunderwriting results. Investment income,however, remained subdued in an environmentof very low interest rates. Capital basesimproved in both the non-life and reinsuranceindustries in 2004, whereas solvency positionsin the life insurance industry remainedunchanged. By late 2005, the outlook for theinsurance sector seemed positive, with the riskfacing the industry appearing rather contained.

5.1 FINANCIAL CONDITIONS IN THEINSURANCE SECTOR1

FURTHER IMPROVEMENT IN PROFITABILITY IN2004In the life insurance sector, the average ROEin 2004 stood at 10.9%, up from 9.9% in 2003(see Chart 5.1).2 Non-life insurers enjoyed asignificant improvement in profitability, withROE reaching 12.2% in 2004, compared with8.6% in 2003. By contrast, the reinsuranceindustry saw a significant decline in profits in2004 with the average ROE dropping from14.8% in 2003 to 9.7% in 2004.

THE NON-LIFE INSURANCE INDUSTRYThe strengthening of non-life insurers’ balancesheets in 2004 was entirely due to animprovement in the profitability of corebusiness. Underwriting results, which turnedpositive in 2003, continued to improve in 2004,owing to strict pricing discipline on the part ofeuro area non-life insurers. Although therewere some signs that premium prices hadpassed their cyclical peak, they declined onlyslightly in 2004 and remained above technicallevels, reflecting adequate pricing of risks.

Growth in premium written, which edged upfrom 6.9% in 2003 to 7.2% in 2004, contributedto the improvement in core business profitabilityas measured by the combined ratio.3 In 2003,given positive underwriting results, the combinedratio of the euro area non-life sector droppedbelow 100%, and in 2004 it declined still further(see Chart 5.2). The improvement was mainlydue to tighter terms and conditions, whichcontinued to limit the magnitude of claims to bepaid. Indeed, the loss ratio, which measures themagnitude of incurred losses in the current year,decreased from 74.3% in 2003 to 72.3% in 2004,whereas the expense ratio stabilised in 2004at 25.2%, after three years of improved costcontrol.

Chart 5.1 Return on equity in the euro areal i fe , non-l i fe and reinsurance sectors

(2002 - 2004, %)

Source: Bureau van Dijk (ISIS).

200220032004

0

5

10

15

20

25

0

5

10

15

20

25

life non-life reinsurance

1 The assessment of the f inancial condition of the euro areainsurance sector is based on unconsolidated accounts. Thisallows for the disentangling of large insurance f irms engagedin life, non-life and reinsurance activities. The data source forbalance sheet and income statement data was Bureau Van Dijk(ISIS database). The sample of f irms is composed of 213 lifeinsurers, 292 non-life insurers and 25 reinsurance companies.

2 ROE is calculated as the ratio of prof its after taxes andextraordinary income to capital and, when available, non-distributable reserves, claims of equalisation of non-lifeshareholders’ funds and of other reserves, and profits andlosses. The average ROE is weighted by the net premium earnedby non-life insurers and by the net premium written for life andreinsurance companies.

3 The components of the combined ratio allow the sources ofprofitability to be highlighted – cost-cutting and/or lossreductions. It is calculated as the sum of the loss ratio, whichmeasures the magnitude of incurred losses for the current year(net losses and loss adjustment expenses/net premium earned),and the expense ratio, which provides information aboutexpense control (underwriting and administrative expenses/netpremium written). Typically, a combined ratio of more than100% represents an underwriting loss for the non-life insurer.

109cECB



SYSTEM

By contrast, net investment income in the non-life sector dropped by 3.4% in 2004, following adecline of nearly 3.0% in 2003. This was mainlydue to a combination of declining euro arealong-term interest rates and increasing bondholding in insurers’ balance sheets throughout2004. Bonds that are held in a “buy and hold”strategy are currently not marked-to-market inmost of the countries of the euro area. Buy andhold bonds are valued at their amortised cost.Any changes in the bond valuation are reportedas unrealised gains and are therefore notreported in the income statement. Thisaccounting practice will be generalised in 2005with the implementation of the InternationalFinancial Reporting Standards (IFRS). Hence,the decline in long-term interest rates in 2004has simply led to insurers receiving loweredinterest rate payments. As the new bonds boughttypically had lower yields than those which hadmatured in 2004 and as the amount of thesebonds bought largely surpassed the amount ofbonds maturing in 2004, the overall impact oninvestment income and thus on profitability wasnegative.

THE LIFE INSURANCE INDUSTRYNotwithstanding the persistently low interestrate environment and the large outstanding stockof life insurance products with guaranteedreturns, profitability in the life insurance

industry improved in 2004. This was mainly dueto strong growth in sales of unit-linked products,lower bonus allocation, cost savings and thereduction of guaranteed rates of return on newlyissued savings products. Furthermore, ongoingreforms of public sector pension schemes in theeuro area favoured private saving in lifeinsurance products. Legal changes and taxadvantages that were introduced in France, Italyand Germany during the year also promotedgrowth in the demand for pension products.Although the environment of persistently lowinterest rates reduced the attractiveness ofpolicies with guaranteed returns, this waspartially compensated by the strength of demandfor unit-linked products, the sales of which grewby 21.7% in 2004. This was, however, lowerthan the corresponding figure for 2003 of 31.3%.Net premium written by the life insuranceindustry increased by 7.4% in 2004, comparedwith 11.1% in 2003. As margins on unit-linkedproducts are currently often higher than those ontraditional guaranteed return policies, there isroom for further improvement in profitability inthe period ahead if the demand for theseproducts remains buoyant. The investmentincome of the sector only grew by 2.4% in 2004,compared with 7.1% the previous year. Thispoor performance was mainly due to thedecrease in long-term interest rates in the euroarea during the second half of 2004.

Chart 5.2 Expense, loss and combined ratiosof the euro area non-l i fe insurance industry

(2000 - 2004, %)


expense ratio (left-hand scale) loss ratio (left-hand scale)combined ratio (right-hand scale)

0

10

20

30

40

50

60

70

80

2000 2001 2002 2003 200492

94

96

98

100

102

104

106

108

Chart 5.3 Frequency distr ibution of returnon equity of euro area l i fe insurancecompanies(2002 - 2004, %)


200220032004

0

5

10

15

20

25

30

35

0

5

10

15

20

25

30

35

<0 0-5% 5-10% 10-15% 15-20% >20%


The strengthening of profitability was nothomogeneous among life insurers, and theweakest companies in particular did notsucceed in increasing their earnings. At about10%, the share of firms with negative ROE didnot change in 2004 compared with 2003 (seeChart 5.3).

THE REINSURANCE INDUSTRYThe deterioration of profitability in the euroarea reinsurance sector in 2004 was broad-based. Net investment income dropped by54.6%, following an increase of 143.3% in2003. At the same time, net premium writtendeclined by 7.4% in 2004, compared with zerogrowth in 2003.4 Several factors account forthe weakness of the profitability of reinsurersin 2004. In an environment characterised bysignificant balance sheet improvement, non-life primary insurers limited their cededexposures to the reinsurance industry.Reflecting this, the retention ratio – a yardstickwhich measures the amount of risk that non-lifeprimary insurers retain on their own account –increased in 2004 to 80.6%, up from 79.1% inthe previous year.5 This reduction in risktransfer to the reinsurance sector weighednegatively on the growth in premium written.Furthermore, it cannot be excluded that thedecrease in premium written by euro area

reinsurers may also have reflected some slightloss of market share, as premium pricesdeclined only slightly in 2004.

The decline in premium written contributed to aslight increase in the combined ratio of the euroarea non-life reinsurance sector from 101.0%in 2003 to 101.5% in 2004.6 Increases inunderwriting and administrative expenses alsocontributed to the slight deterioration in thisratio. By contrast, the loss ratio showed someimprovement (see Chart 5.4).

STRENGTHENING OF CAPITAL POSITIONSIN 2004 EXCEPT IN THE LIFE INSURANCEINDUSTRYThe capital positions of euro area non-life andreinsurance companies improved in 2004. Thesolvency positions of insurers may be broadlymeasured by the ratio of surplus to net premiumwritten.7 In the non-life industry, this ratio roseto 25% in 2004, up from 24.4% in 2003 (seeChart 5.5). This strengthening of capitalpositions resulted in part from the fresh capitalbrought by new entrants into the non-lifesector, motivated by the rather strong levels ofprofitability. In particular, firms involved in

Chart 5.4 Expense, loss and combined ratiosof the euro area non-l i fe reinsuranceindustry(2000 - 2004, %)


expense ratio (left-hand scale) loss ratio (left-hand scale)combined ratio (right-hand scale)

0

20

40

60

80

100

2000 2001 2002 2003 2004100

105

110

115

120

125

4 To a certain extent, the rather high year-to-year volatility of theaggregated income statements of euro area reinsurers is areflection of very high levels of business concentration. In2004, the f ive and ten largest reinsurers accounted for 79.3%and 91.1% of the total premium written in 2004 respectively, anincrease from 2000, when the corresponding f igures were75.1% and 86.9% respectively. This high concentrationimplies that any change in the strategy followed by one of thelarger companies will make a significant contribution to theaggregated data for the euro area.

5 The retention ratio is measured as the amount of net premiumwritten over gross premium written.

6 In the euro area, the non-life reinsurance business constitutesthe bulk of reinsurance activity. Of the 29 reinsurancecompanies in the euro area with available f inancial accounts for2004, only three were specialised in the life business, and theyrepresented slightly less than 3.5% of the total premium writtenin 2004.

7 The capital position of non-life insurers and reinsurancecompanies is calculated as surplus over net premiums written;while that of life insurers as surplus over technical reserves.Whether the capital position of life companies is calculated assurplus over total assets or as surplus over total assets lesslinked products, the assessment remains identical with noimprovement in the solvency position. The ratios for life andnon-life/reinsurance companies are not comparable due tosector-specif ic accounting regulations regarding, for example,equalisation reserves.

111cECB



SYSTEMlife insurance turned to non-life activitiesin order to diversify their risk. A furtherimprovement was also seen in the reinsurancesector in 2004, with the capital positionreaching 26.1%, up from 23.8% in 2003.Despite rather subdued profits, the capital baseof the reinsurance industry mainly grewthrough earnings in 2004 as companies limitedthe amount of dividends they distributed totheir shareholders.

In the life insurance industry, there was a slightdeterioration of solvency position from 7.5%in 2003 to 7.2% in 2004. Recognition ofincreasing policyholder life expectancy onannuity business led many euro area lifeinsurers to revise their mortality assumptionsthat had previously underestimated so-calledlongevity risk (see Box 16). Hence, these firmscontinued to significantly strengthen their

Box 16

AGEING POPULATION AND LONGEVITY RISK

As the so-called baby boom generations – i.e. those born between the mid-1940s and the mid-1960s – approach retirement, policymakers in many developed countries have becomeincreasingly concerned about retirement funding and retirement income security. Withongoing pension reforms reducing the generosity of funding from public sources, moreemphasis is being placed on private saving. However, the inherent uncertainty about the lengthof human life complicates any decision regarding saving for retirement. In particular, there is arisk that individuals may outlive their resources and could be forced to reduce their livingstandards quite substantially when they reach a more advanced age, or even risk falling into apoverty trap. Longevity risk,1 which materialises when expectations regarding lifespan are notmet, has two components.2 Individual longevity risk is the risk that a person will die either priorto or after the average lifespan of his/her cohort. It can theoretically be diversified away bypooling risks in private annuity markets, where those who live longer than the average maybenefit from the contributions of those who die earlier. Collective longevity risk concerns therisk of underestimating the average expected longevity. This risk poses more challenges thanindividual longevity risk because it cannot be shared within members of the same cohort bywriting a large number of life policies. This Box discusses some of the challenges raised bycollective longevity risk, for which no simple hedge may be found.

Governments, pension funds and to a less extent life insurance companies used to bearcollective longevity risk. Due to the partial disengagement of governments from pension

1 Longevity risk concerns the upper end of the age distribution of the population. It differs from mortality risk, which is driven byshort-term extreme events such as flu epidemics.

2 See M. King (2004), “What Fates Impose: Facing up to Uncertainty”, the Eighth British Academy Annual Lecture.

Chart 5.5 Capital posit ions in the euro areal i fe , non-l i fe and reinsurance sectors

(2002 - 2004, %)


lifenon-lifereinsurance

0

5

10

15

20

30

25

0

5

10

15

20

30

25

2002 2003 2004

reserves, thus preventing a significant build-upof capital in 2004.


provision and the gradual change from defined benefit pension plans to defined contributionschemes, not only individual but also collective longevity risk has been increasinglytransferred to the household sector. At present in the euro area, the risk of outliving resources israther limited, as the bulk of pension income continues to originate from public sources and/oroccupational defined benefit schemes. Nevertheless, further reforms are likely to shift riskstowards households. Hence, there is the risk that households at retirement may find it difficultto convert accumulated wealth into a guaranteed stream of income until death. Liquid andefficient annuity markets could dampen this risk and eliminate the individual longevity riskthat is now increasingly borne by households. However, adverse selection problems anddifficulties faced by life insurers and pension funds in hedging collective longevity riskcurrently weigh on the development of such markets.

An adverse selection problem arises with the provision of individual annuities, because thosewho live longer than the average expected life span will tend to buy more annuities on avoluntary basis than others. Consequently, this raises the price of annuities, thus reducing theincentives of those new potential annuitants with shorter life expectancy to enter the market.Hence, only a very small proportion of saving is currently invested in annuities. In order tocircumvent the adverse selection problem that is associated with the individual provision ofannuities, collective schemes could be made compulsory. As a result, policy prices shouldconverge to their fair actuarial prices as mortality tables of the whole population may replacethose of annuitants currently used by annuity writers. However, any compulsory scheme wouldinvolve some problems of redistribution so that as a first step, annuities could simply be set asthe default option in defined contribution plans instead of the current practice of lump sumwithdrawals.

Regarding problems associated with hedging exposures to collective longevity risk, no simplesolutions exist. This risk is currently concentrated in corporate defined benefit pension fundsand in life insurance companies’ balance sheets. Ideally, these institutional investors maydesire to hold assets whose return is proportional to the average longevity of their annuites inorder as a hedge. Such hedging instruments do not exist yet and the absence of adequatehedging has already led to significant deficits in the reserves of pension fund balance sheets,and the problem has been exacerbated by the low level of interest rates. Indeed, when interestrates are low, any unexpected improvement in lifespan results in a significantly larger increasein reserves than the rise needed when interest rates are high.3

While the design of effective hedges for longevity risk would appear to be important, there arepractical challenges in that there are many more potential buyers of longevity protection thanthere are sellers. Indeed, the appetite of reinsurers to take further longevity risk is very limited.Owing to the legal scrutiny of financial/finite risk reinsurance contracts, reinsurers have becomemore reluctant to provide insurance against longevity risk than in the past.4 Furthermore, M&Ashave reduced the number of companies operating in this sector. Attracted by increasing premiumprices, many reinsurers have also shifted resources to the non-life sector.

3 A 10% improvement in longevity by leads to an increase by 5.4% of the net present value of the immediate annuity – an immediateannuity being a regular income payable throughout life, which is usually secured in exchange for a lump sum – to meet an annualpayment of 10,000 euro over 25 years, based on a 3% interest rate. With interest rates equal to 5% and 10% respectively, this f igurewould fall to 4.2% and 2.1%. Hence, life insurance companies and pension funds are concerned about interest rate/longevitycorrelation risk.

4 See Fitch Ratings (2005), “Reeling in the Years: VIF securitisation”, Special Report on Insurance, Europe, June.

113cECB



SYSTEM

Despite the favourable funding conditions thatprevailed in 2004, life insurers in the euro areadid not exploit external sources of funding tostrengthen their capital position, as their accessto financial markets remained limited. Instead,life insurers mainly followed a strategy of riskreduction in order to save capital. The risk ofinvestment portfolios was reduced through areallocation of asset portfolios in favour ofbonds. Companies also continued to use financial/finite risk reinsurance to mitigate risks and toreduce solvency requirements.8 Indeed, theretention ratio did not change between 2003 and2004, remaining at 96.4%. Finally, those lifeinsurers which managed to improve slightly theircapital positions achieved this through the retentionof profits by limiting dividend distributions.

Alternative risk management options involvingsecuritisation are currently available to euro

Given the lack of longevity insurance capacity within the reinsurance sector, the capitalmarkets could provide innovative solutions to hedging longevity risk. In November 2004, theEuropean Investment Bank/BNP Paribas announced their intention to issue a 25-year survivorbond, also called a longevity bond, where coupon payments would be linked to the proportionof the UK male population who were aged 65 in 2003 and who are still alive at the coupon date.The longevity risk in this operation will be born by the Bermuda-based Partner Re through areinsurance contract. However, Partner Re has apparently made it clear that it has little appetitefor additional deals. Since an active and liquid market for longevity bonds requires not onlybuyers but also sellers, there might be a role for governments to substitute for reinsurers. Byassuming collective longevity risk, either as an issuer of longevity bonds or of long-datedannuities, governments would support the development of liquid annuity markets.5

Regarding the consequences for financial stability, the lack of transparency about longevityassumptions used for the calculation of reserves may expose shareholders of companies withcorporate defined benefit plans to risks of significant declines in distributed profits and in stockprices following revelations of unrealistic assumptions in mortality rates. In the UK, as from23 September 2005, new actuarial valuations will apply which make more prudent assumptionsregarding longevity. The implementation of the new Statutory Funding Objectives requires theelimination of any potential deficits in UK defined benefit pension funds over a predeterminedperiod that is likely to be set to ten years. Owing to the relatively short period of reserverebuilding, most employers with defined benefit schemes incur a significant risk of financialdistress, even in some cases a risk of insolvency. The choice of the discount rate – which is crucialin the assessment of underfunding – is still being debated (see Box 17).

area life insurance companies with the aimof managing their capital more efficiently.For instance, the availability of mortalitybonds or, more recently, the securitisation ofexpected future profits from blocks ofinsurance business – so-called value of in-forcesecuritisation9 – has helped risk managementby life insurers located outside the euro area,albeit so far not in the euro area itself.10 For thefirst time in the euro area, in August 2005 small

5 The UK Debt Management Off ice conducted a consultation about the relevance of issuing annuity-type gilts. Most respondents –except policy advisers in the pension industry and trustees – did not support the project, pointing out concerns about the potentialilliquidity of such instruments, in that annuities could be buy-and-hold instruments. Hence, no issuance of annuities will occur inthe near future. See http://www.dmo.gov.uk/gilts/public/consdoc/cons160305.pdf.

8 These contracts involve a certain level of risk transfer fromprimary insurers to reinsurers, but with a limited exposure.

9 Value of in-force is the discounted value of the reserve releasesexpected to come through in future years based on a set ofassumptions. It can therefore be thought of as a quantificationof the degree of prudence in the reserves.

10 In the UK, both Barclays Life and Friends Provident haverecently securitised the value of in-force business contained inclosed and open blocks of life policies in order to raiseeff iciently regulatory capital. Swiss Re also completed a USD245 million securitisation of in-force life insurance policies inJanuary 2005.


European insurance companies called upcapital through the issuance of CDOs on theirsubordinated debt.

5.2 RISKS FACING THE INSURANCE SECTOR

Despite the ongoing improvement in thefinancial condition of the euro area insurancesector, some risks do lie ahead. Among theexogenous sources of risks and vulnerabilities,the insurance sector faces market risksassociated with the low level of long-terminterest rates. There are also some risks withinthe sector itself.

Among the external sources of risk facing theeuro area insurance sector, the most importantappears to be uncertainty surrounding theexpected future path of long-term interest rates(see Box 17). As market interest rates remainbelow the average guaranteed returns in someeuro area countries, profitability will remainsubdued. In the life insurance industry, anyincrease in long-term bond yields is likely toease pressures on asset-liability mismatches inbalance sheets. An improvement in solvencypositions may be expected as long as the same

accounting rules are used to value both assetsand liabilities – which is currently not the casein all euro area countries. This means that forthose countries which apply the new accountingstandards only on the asset side, the rise in long-term interest rates may prove detrimental totheir solvency position. Indeed, this woulddecrease the present value of the assets, whileliabilities would not be marked-to-market. Onthe positive side, rising interest rates mightmitigate longevity risk, which is correlated withthe level of interest rates (see Box 16).Longevity risk arises primarily from anincorrect assessment concerning people’slongevity, and is exacerbated by the impact oflow interest rates on the present value of futurepayments. The current flattening of the euroarea yield curve does not pose a challenge forlife insurers. Only an inverted yield curve withlong-term interest rates that are lower thanguaranteed rates would constitute a risk. Thefirms would then have the incentive to invest inthe short term part of the yield curve, and bydoing so would worsen the negative durationgap of their balance sheet. The sensitivity tointerest rate changes may then be amplified as aresult.

Box 17

LOW INTEREST RATES AND BALANCE SHEET VULNERABILITIES OF LIFE INSURANCE COMPANIESAND PENSION FUNDS

Owing to population ageing, the size of pension funds and insurance companies’ balance sheetshas been growing rapidly. In the euro area, the total assets-to-GDP ratio of these institutionalinvestors reached 58.5% in 2004, up from 51.3% in 2002 (see Chart B17.1).The decline in long-term interest rates since the 1990s and their persistently low levels have weakened the balancesheets of these institutions.1 The important share of bond holdings in their investmentportfolios has weighed significantly on profitability over recent years, which has remainedsubdued (see Chart B17.2). However, the main negative impact of low yields has been on theassessment of liabilities and therefore on companies’ net debt. In those countries where bondyields influence the choice of the discount rate used for reserves funding calculations, the

1 The non-life insurance sector does not provide a mixture of long-term saving and insurance in the same manner as life insurancecompanies and pension funds. It essentially faces insurance risks arising from highly uncertain flows of claims. Because of theirspecif ic risk and their shorter liability duration compared to the life insurance industry, these portfolios typically include a higherproportion of equities and a significant proportion of short-term assets with low price volatility. They are therefore less affected bythe low level of long-term interest rates.

115cECB



SYSTEM

Source: ECB.

Chart B17.2 Assets’ composit ion of euroarea insurance companies and pension funds

(%)

0

5

10

15

20

25

30

35

40

45

0

5

10

15

20

25

30

35

40

45

1998 1999 2000 2001 2002 2003 2004 2005

bondsequitymutual fundsloans

Chart B17.1 Assets-to-GDP ratios of euroarea insurance companies and pension funds

(%)

Source: ECB.

05

101520253035404550

05101520253035404550

2002 2003 2004

insurance corporationspension funds

lower the market rates, the higher the present value of liabilities. Hence, any fall in long-terminterest rates may lead to a significant funding gap in balance sheets, given the typically longeraverage duration of liabilities than of assets. This Box discusses some of the financial stabilityissues raised by the impact of the low level of long-term interest rates on life insurers’ andpension funds’ balance sheets.

In defined benefit pension funds, the long bull market for equities in the 1990s allowedcontributions to be scaled back and even to be eliminated for several years. The resultingoverfunding was furthermore amplified in some jurisdictions that had allowed constant orabove-market discount rates to be used in the valuation of liabilities at a time of declining bondyields. The strong stock market performance in the 1990s also permitted life insurancecompanies, which sold traditional policies with high guaranteed returns, to record strongprofits, even though profit margins were progressively eroded at the same time. Indeed, thedifference between the yields earned on bond holdings on the assets side and the guaranteedrate to be paid to policyholders on the liabilities side continuously decreased, even becomingnegative in some cases. For life insurance companies and pension funds alike, as annuityproviders, the low yield context furthermore magnified the risks of underfunding owing to theincrease in longevity beyond earlier actuarial projections. Indeed, when interest rates are low,any underestimation of longevity translates into a much larger funding gap, because thechanges in the present value of liabilities are of a greater magnitude (see Box 16).

This phase of strong profitability in the pension and insurance industry came to an abrupt endwith the bursting of the stock market bubble. The bear equity market from March 2000 to March2003 and the sharp fall in the level of interest rates started to put pressure not only onprofitability, but also on solvency positions. On the profitability side, both life insurancecompanies and pension funds have been affected by persistently low and declining interestrates, as they typically hold a high proportion of bonds in their assets, the bulk of which are helduntil maturity and are therefore not marked-to-market in most euro area jurisdictions.Regarding the solvency assessment, one important aspect is related to the asset-liability match.In cases of a perfect match between asset and liability cash flows at any future date, the choice


of the discount rate would not matter.However, given the current balance sheetmismatches, the choice of discount rates inassessing solvency is extremely important.To manage their balance sheet risks, bothpension funds and insurance companies useasset liability management techniques(ALM), whereby the long-term balancebetween assets and liabilities is maintainedthrough the choice of an asset portfolio withsimilar return, risk, duration and convexitycharacteristics to liabilities.2 Owing to thelimited availability of long-dated bonds forlife insurers and long index-linked bonds forpension funds, the duration of the liabilityremains higher than that of the assets. Hence,as the asset liability matching is not perfect,their balance sheet is usually not immunised against interest rate changes.

The decline in interest rates over the last decade widened this negative duration gap becauseof a larger increase in the present value of the liabilities than that of the assets. As a result, thesensitivity of balance sheets and especially of capital bases to interest rate risk has increased.In the pension fund industry, greater use has been made of market-related discount ratesfollowing the collapse of the equity market, which has boosted the present value of pensionliabilities. As a result, large funding gaps have suddenly been reported in pension funds’balance sheets.3 In the life insurance industry, the discount rate used to assess technicalreserves depends on the cost of capital. Therefore, it is likely to reflect, to a certain extent, theevolution of market interest rates, although the new accounting rules for the valuation ofliabilities are only scheduled for 2007. As a result, lower interest rates may lead to adeterioration of solvency positions of life insurers.

To restore capital bases or reduce funding gaps, pension funds and life insurance companies haveundertaken several measures. These measures have also been favoured by the implementation ofthe new accounting standards and the Solvency II project whose introduction is currentlyscheduled by 2010. Among the risk mitigation actions, there has been a significant increase inboth the share of bonds in total assets and the cutting back of equity, especially in the lifeinsurance sector. Regarding pension funds, the magnitude of such risk rebalancing has been moremodest, owing to the nature of their liabilities.4 To deal with the problem of underfunding,

2 Modif ied duration is a yardstick of the sensitivity of a bond portfolio’s value to a small change in interest rates. This relation istypically not proportional, and convexity measures this aspect of the price-yield relationship.

3 In accordance with IAS 19, pension funds may now be required to use a high-quality corporate bond yield – typically AA orequivalent – in some jurisdictions However, for the majority of European companies, whose average is triple B, using AA yield-based discount rate may lead to an overestimation of the true corporate pension def icit. In other countries, such as Germany, wherethe discount rate is f ixed by the authorities and rarely changed, potential concerns in terms of underfunding may arise with thefuture implementation of new accounting standards in the pension fund industry.

4 The liabilities of life insurers have historically tended to be defined in nominal terms, owing to the offered guaranteed return that isf ixed in money terms. Liabilities of pension funds, on the other hand, are denominated in real terms, as these grow in tandem withwage increases. Indeed, the replacement ratio is typically indexed on f inal earnings in def ined benef it schemes. Defined benefitpension funds are therefore used to hold real assets such as property and equities to match liabilities of a higher proportion than thatobserved in life insurers’ typical investment portfolios.

Chart B17.3 Linked products outstandingand as a share of total assets of euro areal i fe insurance companies


0

20

40

60

80

100

120

140

160

180

2000 2001 2002 2003 200410

11

12

13

14

15

16

17

18

linked products outstanding (EUR billions, left-hand scale)share of linked assets in total assets (%, right-hand scale)

117cECB



SYSTEM

Turning to the sources of risks andvulnerabilities within the industry, animportant risk for the non-life and reinsurancesector is the traditional cyclicality of thebusiness. As there are indications that premiumprices may have peaked in both sub-sectors, theindustry may face the risk that premium pricescould decline below levels that would ensureadequate risk pricing in underwriting.However, there are several factors that mightmitigate this risk. If long-term yields on euroarea bonds remain low, this will weigh oninvestment income, thereby mitigating the riskof a sharp fall in premium prices in the non-lifeand reinsurance sectors. In addition, most euroarea reinsurers have announced that they willfavour meeting ROE targets in the period aheadrather than seeking to expand their marketshares by lowering premium prices. Liabilitymanagement is also likely to come undertighter control in the run-up to Solvency II.Finally, Hurricane Katrina may, as a uniqueevent, result in significant losses, and couldtherefore dampen downward pressures onreinsurance prices. With current insured lossestimates in the USD 40-60 billion range, 2005may prove the costliest year yet for thereinsurance sector as a whole.11 Regarding theeuro area reinsurance industry the lossesincurred appear rather more contained atslightly more than USD 1 billion.

A risk facing the reinsurance sector is related tothe so-called finite risk reinsurance business.Finite risk reinsurance may be broadly definedas reinsurance agreements that suppose acertain level of risk transfer from the primaryinsurers to the reinsurer, but which limit thereinsurer’s risk exposure to a maximumamount.12 This allows companies to smoothentheir earnings by transferring current lossesinto future periods. This may sustain the firm’scapital base for long periods of time, but couldalso distort financial statements. In October2004, the office of the New York attorney-general began a probe into allegations thatfinite risk reinsurance may have been used tomanipulate financial accounts. The actionsundertaken by prosecutors should determinewhether the required risk transfer levels havebeen properly respected and whether certaintransactions were engineered with the aim offalsifying financial accounts. From a financialstability viewpoint, an improvement in the

11 However, the amounts of losses are still insuff icient to triggerhurricane-related catastrophe bonds issued by insurers toprotect themselves against capital depletion arising from peakrisks. A signif icant part of the current estimated losses – USD15-25 billion – is related to flooding in New Orleans. Floodinsurance is provided by the National Flood Insurance Programand will therefore not be incurred by the insurance industry.

12 From an accounting viewpoint, if the degree of risk transfer isinsuff icient, then the transaction is considered as a f inancingmechanism and is booked as a loan or liability instead of as anasset.

pension funds have adopted a variety of solutions, such as requiring higher contributions,reducing benefits, or ultimately accelerating the move towards defined contribution pensionschemes or hybrid plans, thereby transferring (at least partially) financial and longevity risks toemployees. In the life insurance industry, the persistently low yield environment has promptedthe reduction of guaranteed return on traditional saving products. This has eased pressures onprofitability, albeit to a small extent as this reduction only applies to new contracts sold; indeedthe bulk of traditional products have continued to offer higher guaranteed rates. The low-yieldenvironment has also made traditional saving products with reduced guaranteed returns lessattractive. This has benefited a new type of instrument, the so-called unit or index-linked product(see Chart B17.3). Such products are typically indexed to stock indexes, and the investment riskis borne by the policyholders. In both cases, the low-yield environment has contributed to thecontinuous and increasing transfer of risk from the balance sheets of life insurance companiesand pension funds to the household sector, although an ageing euro area population, together withlower expected returns from stock markets compared with the 1990s, remains the main factor inthe straining of balance sheets of defined benefit pension schemes.


transparency of financial statements resultingfrom these investigations should be positive.However, it may weigh on the profitability ofthe reinsurance sector in the period ahead andreduce the flexibility of primary insurers tomanage their capital efficiently.

There are some risks related to the run-up to theimplementation of Solvency II that areconnected to the consequences of possiblechanges in the financing sources of insurers.On one hand, attempts at reducing doubleleverage in balance sheets between holdinggroups and their subsidiaries may weigh onROE in the period ahead.13 On the other hand,in order to face the new regulatory environmentand, to a certain extent, to circumvent thesedownward pressures on ROE, insurers maychoose to rely heavily on subordinated debtissuance. Indeed, it seems likely that the newSolvency II regulation will allow moresubordinated debt to be accounted for asregulatory capital than is currently permittedby most countries of the euro area.14 StatedROE may be boosted by subordinated debtissuance owing to increased financial leverage.The potential negative impact on fundingspreads could impinge on the capital funding ofthe most poorly capitalised companies, i.e.those with only limited or no access at all toequity markets. This could prove problematicin an environment of relative shortage of risk-finite reinsurance and of widening spreads.

MARKET-BASED INDICATORS OF THE INSURANCESECTOR’S SHOCK ABSORPTION CAPACITYAssessment of the euro area insurance outlookthrough market-based indicators does notprovide a homogeneous picture. On one hand,the strong performance of the largest listedinsurance companies since May 2005 hastended to reflect rising investor confidence inthe capacity of the sector to show animprovement in earnings in the period ahead.Expectations of improving financial conditionswere surrounded by a very low degree ofuncertainty (see Chart 5.6).

While equity market-based indicators andexpected default frequencies have shownperceptions of declining risk in the insurancesector, patterns in subordinated debt spreads,which widened after March 2005, point towardsthe perception of rising risk (see Chart S68). TheRND became somewhat flatter in earlyNovember compared to May, reflecting a higherdispersion of expectations among marketparticipants regarding the future value of theinsurance stock index. By early November, thelikelihood of an abnormally large decline waspriced by market participants as being smallerthan that of a large increase (see Chart 5.7).

13 The new regulation – planned to be introduced in 2010 – isexpected to apply at both the holding group and subsidiarylevels. It should therefore reduce the incentive for doubleleverage, a situation whereby the holding company issuessenior debt and transfers part of the proceeds to subsidiaries inthe form of subordinated debt. As double leverage is signif icantin the insurance sector, any reduction may tend to put downwardpressure on ROE.

14 Current regulations often allow subordinated debt to accountfor up 50% of the required minimum margin. However, as mostinsurance companies operate with higher levels of capital thanthe minimum regulatory requirement, subordinated debtcurrently accounted de facto for a smaller part of capital – lessthan 25%.

Chart 5.6 Dow Jones EURO STOXX insuranceindex and its impl ied volat i l ity

(Jan. 2003 - Nov. 2005)

Source: Bloomberg.Note: The Dow Jones EURO STOXX insurance indexcomprises the 19 largest insurance companies in the euroarea. The implied volatility is the average of the volatilityextracted from call and put option prices with a delta equalto 0.5.

implied volatility one-month maturity (%, left-hand scale)Dow Jones EURO STOXX insurance index (index value, right-hand scale)

10

20

30

40

50

60

70

2003 2004 2005100

120

140

160

180

200

220

240

260

119cECB



SYSTEM

For the insurance industry sub-sectors,different patterns can be identified in theperformance of stock prices. The life insurancestock index underperformed the EUROSTOXX index in the first half of the 2005 butoutperformed thereafter (see Chart 5.8). Thechallenging low interest rate environment andto a lesser extent the need to revise reservesupwards to face revisions in longevity riskassessment has led life insurers to retain profitsrather than distributing dividends since 2002.However, the strong financial results of euroarea life insurers released in mid 2005 boostedstock prices in the second half of the year.

Regarding the reinsurance industry, the worst-performing segment of the entire insurancesector until mid-September 2005, one drivingfactor appears to have been reluctance on thepart of companies to return excess capital toshareholders. The reinsurers’ decision to fundreserves rather than target some level of ROEhas therefore not been welcomed by investors.The US investigations into finite riskreinsurance may also have weighed onreinsurance stock prices. Finally, from mid-September 2005, the outperformance of thereinsurance stock index became morepronounced. This may possibly be explained byexpectations of international investors of a

flight from primary insurers towards euro areareinsurers rather than US competitors afterHurricane Katrina. The announcement of ROEtargets by most euro area reinsurers may alsohave played a role.

In the non-life insurance industry, stock pricescontinued to outperform the EURO STOXX in2005, despite the position in the cycle ofpremium prices, which usually indicates thestart of a period of deteriorating underwritingprofits. Three reasons can be offered inexplanation for the strength of stock prices inthis sub-sector. First, share prices are forward-looking and have therefore already anticipatedthe impact of an impending decline in policyprices. Second, firms sent clear signals toinvestors regarding targeted levels for combinedratios and ROE. Third, generous cash dividendsand buybacks in this sector have been viewed asa sign of management confidence in the positiveoutlook regarding underwriting results.

5.3 OVERALL ASSESSMENT

The outlook for the euro area insurance sectorremains favourable, although some risks canstill be identified, mainly relating touncertainties about the likely future path ofeuro area long-term interest rates.

Chart 5.7 Risk-neutral probabi l ity densityfunction on the Dow Jones EURO STOXXinsurance index


6 May 2005 4 November 2005

0.00

0.01

0.02

0.03

0.04

0.07

0.06

0.05

0.00

0.01

0.02

0.03

0.04

0.07

0.06

0.05

150 175 200 225 250 275 300

value of the Dow Jones EURO STOXX insurance index


Chart 5.8 Cumulative change in the euroarea insurance stock price indices relat iveto the Dow Jones EURO STOXX(Jan. 2005 – Nov. 2005, % points, base, Jan. 2005 = 0)


non-life insurerslife insurersreinsurance

-20Jan.

2004 2005Jan.July July

-15

-10

-5

0

5

10

15

-20

-15

-10

-5

0

5

10

15


The prospects for the life insurance industry inthe euro area have improved. Decisions takenby some Member States to offload pensionfunding from the fiscal budget are likely toshift more pension business to the lifeinsurance industry. Hence, growth in the sale oflife policies is expected to pick up gradually inthe future.

The outlook for the non-life industry in the euroarea also appears broadly positive. The risk of asignificant decline in premium prices appearscontained in the period ahead, as investmentincome should remain modest given the currentlow investment return environment. Growth inpremium written should expand further witheconomic recovery.

In the reinsurance sector, underwriting resultsin 2005 are likely to be strong, with premiumprices declining only slowly. In the periodahead, owing to important claims fromHurricane Katrina, reinsurance premium pricesmay halt declining and possibly even increaseslightly, depending on the final amount ofcapital depletion in the sector worldwide. Inthe euro area, the losses incurred by reinsurersfrom the hurricane are expected to only dentcapital positions.

121cECB



SYSTEM6 STRENGTHENING FINANCIAL SYSTEMINFRASTRUCTURES

Key financial infrastructures – includingpayment systems, such as TARGET, andsecurities clearing and settlement systems –remained robust and continued to facilitate asmooth allocation of financial resources. Thissection also reports on the oversight of SWIFTand the ECB’s adoption of an oversightframework for retail payment systems operatingin euro. With regard to securities clearing andsettlement systems, no further consolidatingoperations have been reported over the last sixmonths, as has indeed been the case sincethe launch of the euro. This section alsodiscusses the CPSS-IOSCO Recommendationsfor Central Counterparties, which explicitlyaim at strengthening financial stability.

6.1 PAYMENT SYSTEMS

Any events which could adversely impact onthe smooth functioning of payment systems,particularly large-value interbank fund transfersystems, are undesirable from a financialstability perspective. Examples of such eventsinclude insolvency or liquidity problems of oneor more system participants, or deficiencies ininformation systems. If such events domaterialise, they might, in a best-case scenario,“only” lead to minor disruptions. However, ifthey are severe enough and/or contagious, theycould have an impact on other systemparticipants, other market infrastructures andtheir participants, and financial markets on awide scale, possibly leading to broad andnegative implications for the economy orcurrency area as a whole. As a corollary, it is ofthe utmost importance that payment systemsare designed, operated and overseen in the mostprudent manner. Whereas the design andoperation of individual payment systems lieswith the operators of those systems, centralbanks have the responsibility for overseeingthat systems which can transmit systemicdisruptions comply with the Core Principles forSystemically Important Payment Systems (“theCore Principles”).

Oversight of payment systems, especiallysystemically important payment systems suchas TARGET and EURO1, is one of theEurosystem’s main tasks. It is worthwhileemphasising that it is the Eurosystem’soversight policy to conduct oversight not onlyon the operation of payment systems, but alsoon the design of payment systems with a viewto contributing to the compliance of thesesystems with all relevant Core Principles fromthe outset. This therefore also contributes tolong-term financial stability.

SETTLEMENT OF LARGE-VALUE PAYMENTSIN EUROThe Eurosystem encourages large-valuepayments in euro to be settled safely in systemsthat use central bank money as the settlementasset, thereby contributing to the stabilityof the euro area financial system. Marketparticipants have so far responded positively tothis encouragement. Among the four eurolarge-value payment systems operating in theeuro area, the TARGET system is the one that ismost frequently used to settle, for instance,money market transactions or the cash legof securities transfers.1 TARGET, whichcommenced operations on 4 January 1999, hasbecome the payments backbone of the euroarea. In 2005, the volumes and values ofpayments settled via TARGET grew stillfurther (see Chart 6.1). Over the six monthssince the last Review (between April andSeptember 2005), the system settled an averagedaily value of €1,878 billion.

EURO1 is the second most important eurolarge-value payment system both in terms ofvalue and volume, although it is positioned along way behind TARGET. Between April andSeptember 2005, it settled an average dailyvalue of €165.2 billion.

RECENT DEVELOPMENTS IN TARGETOver the six months since the last Review(from April to September 2005), the fivelargest national real-time gross settlement

1 The other systems are the pan-european EURO1 system, thePNS system in France, and the POPS system in Finland.


systems (RTGS) that are part of the TARGETsystem had a collective share of 82.1% in termsof volume, (those of Germany, Italy, Spain,United Kingdom and The Netherlands) and82.8% in terms of value (the first four countriesand France) of all transactions sent viaTARGET (see Chart 6.2). Given theircollective importance, it is imperative thatthese systems are particularly reliable in orderto ensure that they do not adversely affect thesmooth functioning of TARGET as a whole.

FUTURE DEVELOPMENTS IN TARGETThe TARGET2 system will replace thecurrent TARGET system in November 2007,as discussed in the December 2004 FSR.Despite the complexity of the TARGET2project, the design of the new system hasbeen progressing well. As with any othereuro large-value payment system in the euroarea, TARGET2 will have to comply withthe Eurosystem’s oversight standards. Aspart of the TARGET oversight function,a comprehensive preliminary oversightassessment of TARGET2 in its design phase isplanned in the coming months. The results ofthese oversight activities with regard toTARGET2 will be published in a future editionof the FSR.

In March 2005, the ECB Governing Councilapproved the connection of the new Polish euroRTGS system SORBNET-EURO, operated byNarodowy Bank Polski, to the currentTARGET system.2 A further enlargement of thecurrent TARGET system is expected to takeplace in light of the planned adoption of theeuro by some of the new EU Member States in2007. From a financial stability perspective,any connection of RTGS systems to the currentTARGET system requires the proven systemicstability of the TARGET system to be fullymaintained. Therefore, these new TARGETcomponents – as well as all other systemicallyimportant infrastructures in countriesintending to operate in euro – will be subject tooversight assessments in accordance with theEurosystem’s common oversight policy. TheECB and all relevant NCBs will carry out theseoversight assessments in the coming months,and the results will be covered in a future issueof the FSR.

CONTINUOUS LINKED SETTLEMENT (CLS)The Continuous Linked Settlement (CLS)system aims at substantially reducing foreignexchange (FX) settlement risk by settling bothlegs of FX transactions simultaneously as soon

2 For more details, see ECB (2005), Financial Stability Review,June.

Chart 6.1 Large-value payments processedvia TARGET

(Q1 1999 - Q3 2005)

Source: ECB.

50

60

70

80

90

50

60

70

80

90

1999 2000 2001 2002 2003 2004 2005

value (% of total value of EUR transactions)volume (% of total number of transactions)

Chart 6.2 Large-value payments processedvia TARGET by country

(Apr. 2005 - Sep. 2005, % of the NCBs’/ECB’s share in termsof value and volume)

Source: ECB.

0

10

20

30

40

50

0

10

20

30

40

50

AT BE DE DK ES ECB FI FR GB GR IE IT LU NL PL PT SE

volumevalue

123cECB



SYSTEMas sufficient funds are available. This multi-currency system settles a significant number oftransactions in euro, and is, in terms of value,the second-largest payment system settlingeuro transactions after TARGET (notablyexceeding the value of transactions settledthrough EURO1). The functioning of CLS is ofinterest to the Eurosystem because instabilitiesin the CLS system could have systemicimplications for the euro area.

In 2005, the settlement values of FX transactionsprocessed via the CLS system rose further (seeChart 6.3). In September 2005, CLS settled theequivalent of USD 2.3 trillion, thus eliminatingFX settlement risk equivalent to USD 2.2trillion. The euro values settled via CLSamounted to €378 billion in September 2005,eliminating FX settlement risk of approximately€360 billion.

SWIFTS.W.I.F.T. SCRL, the Society for WorldwideInterbank Financial Telecommunication(henceforth simply SWIFT), is a limited liabilitycooperative company owned by its members. Inaddition to providing secure messaging servicesto more than 7,500 financial institutions, SWIFTis also actively engaged in the messagestandardisation process. SWIFT cooperates withits user community in order to refine existingmessage types and thus to implement messagestandards for new types of transactions or otherfinancial information needs.

The Group of Ten (G10) central banks performoversight on SWIFT in a cooperative way. Thisis due to the importance of SWIFT as a networkservice provider for most of the systemicallyimportant payment and securities clearingsystems, as well as their participants.

In 2004, the G10 central banks agreed onthe need to strengthen practical oversightarrangements (see Box 18). However, the twocore concepts in the initial set-up of theoversight of SWIFT have remained valid,i.e. the concept of cooperative oversight with theNationale Bank van België/Banque Nationalede Belgique (NBB) as lead overseer, and theconcept of moral suasion to induce changes.

Chart 6.3 Volumes and values of foreignexchange trades sett led via CLS inUSD bi l l ion equivalent(Jan. 2003 - Sep. 2005)

Source: ECB.

value in USD billions (left-hand scale)volume in thousands (right-hand scale)

0

200

Jan.2003 2004 2005July July JulyJan. Jan.

400

600

800

1,000

1,200

0

20

40

60

80

100

120

Box 18

THE OVERSIGHT OF SWIFT: OBJECTIVES, SCOPE AND STRUCTURE

The oversight of SWIFT focuses on the security, operational reliability, business continuityand resilience of the SWIFT infrastructure. The oversight activities performed by centralbanks aim at ensuring that SWIFT has in place appropriate governance arrangements,structures, processes, risk management procedures and controls that enable it to manageeffectively the risks it may pose for financial stability and the soundness of financialinfrastructures. This oversight does not grant SWIFT any certification, approval orauthorisation, and SWIFT continues to bear responsibility for the security and reliability of itssystems, products and services.


OVERSIGHT ASSESSMENT OF RETAIL PAYMENTSYSTEMSIn August 2005, the Eurosystem finalised theassessment of 15 euro retail payment systems

that fall within the scope of its oversight policyon retail payment systems (see Box 19).

Box 19

THE OVERSIGHT OF RETAIL PAYMENT SYSTEMS

The Eurosystem’s task of promoting the smooth operation of payment systems consists ofensuring the safety and efficiency of payment systems and the security of payment instruments.In pursuing this objective, to start off with, the Eurosystem concentrated on large-valuepayment systems, as these systems were regarded as the most relevant for financial stability inthe euro area. However, turnover data for euro retail payment systems suggested that some ofthese systems had likewise reached a size, and thus relevance, where disruptions could triggersystemic risks.

Given that SWIFT is incorporated in Belgium, the NBB is the lead overseer. It conducts theoversight of SWIFT in cooperation with other central banks. Currently, these comprise theECB and the following G10 central banks: the Bank of Canada, the Deutsche Bundesbank, theBanque de France, the Banca d’Italia, De Nederlandsche Bank, the Bank of Japan, SverigesRiksbank, the Swiss National Bank, the Bank of England, and the US Federal Reserve System,represented by the Federal Reserve Bank of New York and the Board of Governors of theFederal Reserve System. The NBB and SWIFT have formalised this oversight relationship in aprotocol arrangement. The relationship between the NBB and the other cooperating centralbanks in relation to participation in the oversight of SWIFT has been laid down in bilateralMemoranda of Understanding (MoUs) between the NBB and each of those central banks.

The SWIFT oversight structure comprises of two senior-level groups and a technical oversightgroup:

i) The SWIFT Cooperative Oversight Group – which is composed of all G10 central banks,the ECB and the chairman of the G10 Committee on Payment and Settlement Systems – isthe forum through which central banks conduct cooperative oversight of SWIFT and, inparticular, discuss oversight strategy and policies related to SWIFT.

ii) The Executive Group is composed of only five Oversight Group members: the Bank ofJapan, the Federal Reserve Board, the Bank of England, the ECB and the NBB. On behalfof the Oversight Group, it discusses with SWIFT’s board and management the centralbanks’ oversight policy, any issues of concern, SWIFT’s strategy regarding oversightobjectives, and conclusions.

iii) At the technical level, the Technical Oversight Group carries out the technical preparatorywork for the oversight of SWIFT and reports its findings and recommendations to theOversight Group.

125cECB



SYSTEMTherefore, in June 2003 the Governing Council of the ECB adopted an oversight framework forretail payment systems operating in euro. It is important to note that this oversight frameworkis intended to ensure that retail payment systems cannot become vectors of systemic risks oreconomic malfunctioning in the euro area. The framework contained criteria for classifyingretail payment systems into three different categories: systemically important retail paymentsystems (SIRPS), prominently important retail payment systems (PIRPS), and “other retailpayment systems”. The decisive factor in defining the classification criteria was the degree ofdisruption that a malfunctioning in one of these systems could cause in the financial marketsand/or the economy in general.

The distinguishing feature of a SIRPS is that it can trigger severe disruptions or transmitshocks across the financial system. The main determinants in this respect are the value and thenature of the payments that the system processes. A payment system is likely to be of systemicimportance if at least one of the following is true: (i) it is the only payment system in thecountry, or the principal system in terms of the aggregate value of payments; (ii) it mainlyhandles payments of high individual value; and/or (iii) it is used for the settlement of financialmarket transactions or the settlement of other payment systems. If the disruption of a retailpayment system could threaten the stability of financial markets, the system is considered to beof systemic importance (i.e. a SIRPS). With regard to the criteria for a retail payment systembeing classified as a SIRPS, the Eurosystem took into account three factors: the marketpenetration within the respective retail payment market, the financial risks pertinent to thesystem, and the risk of a domino effect. The following three quantitative indicators are used inthis respect:

– a market share of more than 75% of the respective retail payments market, i.e. the paymentsprocessed via interbank retail payment systems and via other payment arrangements(“market penetration”);

– a processing of payments of more than 10% of the value of the national RTGS system or aprocessing of payments with an average daily value of more than €10 billion (“aggregatefinancial risk”); and

– a concentration ratio (i.e. the market share of the five largest participants) of 80%, or anetting ratio of 10% or less, or a net debit position of participants of at least €1 billion (“therisk of a domino effect”).

Any systems fulfilling all of these criteria were considered to be SIRPS.

If the disruption of a retail system does not have systemic implications, but could nonethelesshave a severe impact, such a system is considered to be of prominent importance for thefunctioning of the retail economy (i.e. a PIRPS). PIRPS are characterised by the fact that theyplay a prominent role in the processing and settlement of retail payments, and that their failurecould have a major economic impact that could undermine the confidence of the public inpayment systems and in the currency in general. In seeking to classify PIRPS, the focus wastherefore on the concentration of the retail payments market and, in particular, the degree ofmarket penetration of the respective system, on the basis of the following quantitativeindicator:


– a market share of more than 25% of payments processed in the respective retail paymentsmarket, i.e. the payments processed via interbank retail payment systems and via otherpayment arrangements.

There are other retail payment systems that do not belong to either of the two previouscategories. These systems have a lesser impact on the financial infrastructure and the realeconomy and therefore do not necessarily have to comply with the Core Principles or the RetailStandards. However, such systems do have to comply with the relevant oversight standards, asand if defined for them. Examples in this respect are the common oversight standards for e-money schemes and the standards defined at the national level by each NCB.

It was decided that euro retail payment systems have to comply with different sets of standardsdepending on their classification. SIRPS have to comply with the whole set of Core Principles,while PIRPS have to observe a sub-set of the Core Principles, namely Core Principles I, II, andVII to X. The oversight standards for other systems were not however further harmonised, andthese other systems continue to be assessed against any applicable standards determined by therelevant overseer.

The Eurosystem’s overseers identified 15 euro retail payment systems which either take theform of an automated clearing house (ACH) or a multilateral interbank agreement, andtherefore fall within the scope of the Eurosystem’s policy on retail payment systems. Six ofthese systems were classified as SIRPS, seven as PIRPS, while two fell into the category of“other systems”. Subsequently these systems were assessed against the applicable CorePrinciples. The results of this assessment have been made available on the ECB website.

The assessment reports prepared by thenational overseers reflected the status of thesystems as at end-June 2004. Two systemicallyimportant retail payment systems (SIRPS),LIPS-Net and PMJ, and one prominentlyimportant retail payment system (PIRPS),STEP2, observed all relevant Core Principles.3

The assessment of the other systems revealedshortcomings with respect to one or more CorePrinciples. In general, the level of observancewas better for the SIRPS than for the PIRPS.

All SIRPS observed Core Principles IV(Prompt final settlement), VI (Settlementassets), VIII (Efficiency), IX (Access criteria)and X (Governance). The shortcomings of theSIRPS were concentrated in the areas of CorePrinciples I (Legal basis), III (Management offinancial risks) and V (Settlement inmultilateral netting systems).

For the PIRPS, the observance of the CorePrinciples and shortcomings was more uneven,with a certain concentration of shortcomings inthe area of legal soundness (Core Principle I).

A number of shortcomings identified duringthe assessment process were immediatelyaddressed by the respective system operators,sometimes in cooperation with the relevantoverseer. Therefore the status after finalisingthe assessments has already improved since thestart of the assessment process. The systemowners are expected to remedy the remainingshortcomings in a timely manner.

3 See BIS (2001), “Core Principles for Systemically ImportantPayment Systems”, January.

127cECB



SYSTEM6.2 SECURITIES CLEARING AND SETTLEMENTSYSTEMS

The European securities and derivativesclearing and settlement infrastructure ischaracterised by ongoing changes that aremostly the result of consolidation activities andthe expansion of activities of centralcounterparty clearing houses (CCPs). CCPs,while traditionally most active in derivativesand repo markets, are now increasingly to befound serving outright securities4 markets aswell. Such developments can have an effect onrisks for financial stability and must thereforebe monitored closely.

EXPANSION OF CCPs’ ACTIVITIESCCPs have been expanding into new businessactivities, with an increasing number of CCPsnow also serving outright securities markets (seeTable 6.1). As reported in the June 2005 FSR, the

latest example is in Austria, where the CCPoperated by the Vienna Stock Exchange wasabolished and its activities were transferred tothe newly created “CCP Austria”, a 50%subsidiary of the Vienna Stock Exchange and theAustrian central securities depository (CSD)OeKB respectively, in January 2005. On thisoccasion, it was decided that the new CCPshould clear not only derivatives, but also allcash market transactions of the Vienna StockExchange.

As CCPs have the potential to reallocate risks infinancial markets in an efficient way, providedthat they manage their own risks adequately, thistrend might contribute to financial stability (seealso Special Feature F on “Central counterpartyclearing houses and financial stability”).

Table 6.1 Euro area CCPs for f inancial instruments

Source: ECB (2005).1) MEFF Renta Fija and MEFF Renta Variable belong to the same holding company.

Country January 1999 December 2005

Belgium BELFOX (derivatives) none

Germany Eurex Clearing (derivatives) Eurex Clearing (derivatives, repos, securities)

Greece ADECH (derivatives) ADECH (derivatives)

Spain MEFF Renta Fija (derivatives on MEFF Renta Fija (repos, gov. bonds,debt instruments) derivatives on debt instruments)MEFF Renta Variable (derivatives on equities) MEFF Renta Variable (derivatives

on equities)1)

France Bourse de Paris (SBF) (equities and options) LCH.Clearnet SA (derivatives, repos,securities, also for markets in BE, NL, PTand for MTS markets)

Matif (derivatives; subsidy of SBF)Clearnet (repos, gov. bonds; subsidy of Matif)

Ireland none none

Italy CC&G (derivatives) CC&G (derivatives, securities, alsofor MTS Italy and EuroMTS)

Luxembourg none none

Netherlands Effectenclearing (securities) noneEOCC (derivatives)

Austria Vienna Stock Exchange (derivative) CCP Austria (derivatives, securities)

Portugal BVLP (derivatives) none

Finland HEX (derivatives) none

4 An outright securities transaction is a transaction wherebysecurities are bought and sold outright in the spot market.


INTERNATIONAL STANDARDS ANDRECOMMENDATIONSClearing and settlement is increasingly aninternational activity, and it has therefore beenrecognised that regulatory and oversightefforts in this field may also requireinternational coordination and cooperation. Ajoint task force of the Committee on Paymentand Settlement Systems (CPSS) of the central

banks of the G10 countries and the TechnicalCommittee of the International Organization ofSecurities Commissions (IOSCO) publishedrecommendations for securities settlementsystems (SSSs) in 2001, and for CCPs in 2004.These CPSS-IOSCO recommendationsexplicitly aim at strengthening financialstability (see Box 20).

1 For additional information, see Special Feature F on “Central counterparty clearing houses and f inancial stability” in this FSR.

Box 20

THE CPSS-IOSCO RECOMMENDATIONS FOR CENTRAL COUNTERPARTIES

In November 2004, the CPSS of the central banks of the G10 countries and the TechnicalCommittee of IOSCO published a report entitled “Recommendations for CentralCounterparties”, which contains 15 recommendations for CCP clearing houses. Most of therecommendations aim at reducing risks that CCPs may be exposed to. As CCPs are systemicallyimportant institutions for the financial markets that they serve, these recommendations alsocontribute to strengthening financial stability. This Box briefly describes some of the aspectsaddressed by the recommendations with respect to the most important risks faced by CCPs thatare particularly relevant from a financial stability perspective.1

Counterparty credit and liquidity risksThe most important types of risk for a CCP are counterparty credit risk and liquidity risk.Counterparty credit risk is the risk that a CCP participant cannot fulfil an obligation to deliverassets to the CCP, typically due to insolvency. Liquidity risk is the risk that the participantcannot fulfil such an obligation in time, but only at a later stage, for example owing tooperational problems.

Recommendation 2 advises CCPs to select their participants carefully. In particular,participants should fulfil adequate capital requirements and should have robust operationalfacilities in place to ensure timely settlement of obligations.

Recommendation 2 aims at reducing the risk that a CCP participant could fail to fulfil anobligation towards the CCP. Other recommendations aim at ensuring that the CCP will notitself fail to fulfil its own obligations even if its participants fail to fulfil their obligationstowards the CCP. Recommendation 10, for example, asks CCPs to use delivery-versus-payment (DVP) facilities to settle transactions with its participants. DVP implies that assetsare transferred from the seller to the buyer if and only if the payment is transferred from thebuyer to the seller. If, for example, the CCP is the seller to a defaulting participant and DVP isnot used, then the CCP is in danger of delivering the assets to the participant without receivingpayment, i.e. the CCP risks losing the full principal value of the assets. Recommendations 3, 4and 5 urge CCPs to have adequate and sufficiently liquid financial resources available, inparticular collateral posted by CCP participants, so that the CCP can be used in case of default

129cECB



SYSTEMby participants. It should be noted, however, that a CCP incurs liquidity risk even when allparticipants pay or deliver in time, as there is not always an obligation for CCP participants topay or deliver at the same moment, for instance because of a lack of DVP during settlement.

Settlement bank risk and custody riskCCPs use various settlement service providers to settle transactions with their participants.They use settlement banks to settle payment obligations. Payments are transferred from aparticipant’s cash account with the settlement bank to the CCP’s cash account or vice versa.Settlement bank risk is the risk that the settlement bank could fail. As cash on the CCP’s cashaccount is a liability of the settlement bank towards the CCP, the CCP may lose up to the fullamount of cash on such an account if the settlement bank fails. In addition, the CCP would needto appoint a new settlement bank, a process that may delay clearing activities. It is thereforeadvisable, as per Recommendation 9, that the CCP carefully selects its money settlementarrangements. In particular, the use of central banks is suggested.

Similarly, CCPs use CSDs and custodians to hold and settle other assets, for example securitiesposted to the CCP as collateral. Custody risk is the risk that assets could be lost while held onaccounts with CSDs or custodians. Recommendation 7, among others, emphasises that theseassets should be protected against the claims of creditors of the CSD or custodian.

Investment riskCCPs typically invest their own financial resources and cash posted by their participants ascollateral in one way or another. Again, Recommendation 7 emphasises the need to invest insound assets so as to limit the losses the CCP could incur from its investment strategies.Moreover, financial resources should be invested in relatively liquid assets to enable the CCPto make swift use of these resources in case of urgent need.

Operational risksCCPs, CCP participants, and the CCPs’ settlement banks, CSDs and custodians use a variety oftechnical systems that must not be prone to operational problems. Recommendation 8, dealingwith operational risks, explains that all systems involved should be reliable and secure andshould have adequate capacity. Furthermore, business continuity arrangements should be inplace to permit a timely resumption of activities.

Legal risksFinally, Recommendation 1 addresses legal risks. The legal framework should be well-founded and transparent. With a view to cross-border clearing activities as well,Recommendation 1 furthermore emphasises the need for the legal framework to be enforceablein all relevant jurisdictions.

In October 2001, the ESCB and the Committeeof European Securities Regulators (CESR) setup a joint working group to design new standardsfor securities clearing and settlement systemswith the aim of deepening and strengtheningthe CPSS-IOSCO recommendations in the

European context. In October 2004, theGoverning Council of the ECB and CESR inprinciple approved a ESCB-CESR workinggroup report with 19 standards.5


131ECB


I V S P E C I A L F E ATUR E SA MEASUREMENT CHALLENGES IN

ASSESSING FINANCIAL STABILITY1

Financial stability assessment as currentlypractised by central banks and internationalorganisations probably compares with the waymonetary policy assessment was practised bycentral banks three or four decades ago –before there was a widely accepted, rigorousframework. The measurement challenges thatlie ahead for financial stability assessmentare formidable. However, it is important toacknowledge that significant progress hasbeen made in recent years. Even though thereis no obvious framework for summarisingdevelopments in financial stability in a singlequantitative measure, a growing number ofcentral banks around the world are makingfinancial stability assessments and publishingfinancial stability reports, many of them basedon a broad and forward-looking conception offinancial stability.

INTRODUCTION

Financial stability is a difficult concept todefine. Although it is often seen only from theperspective of avoiding financial crises, it alsohas a positive dimension. It is a condition wherethe financial system is capable of performingwell all of its normal tasks and where it isexpected to do so for the foreseeable future.From this viewpoint, financial system stabilityrequires the principal components of the system– including financial institutions, markets andinfrastructures – to be jointly capable ofabsorbing adverse disturbances. It also requiresthat the financial system is facilitating a smoothand efficient allocation of financial resourcesfrom savers to investors, that financial risk isbeing assessed and priced reasonably accuratelyand that risks are being efficiently managed.Financial stability also has an importantforward-looking dimension: inefficiencies inthe reallocation of capital or materialshortcomings in the pricing of risk can, by layingthe foundations for future vulnerabilities,compromise future financial system stabilityand, therefore, economic stability.

There are three important aspects to producinga comprehensive assessment of financialstability. The first entails forming a judgementabout the individual and collective strength androbustness of the constituent parts of thefinancial system – institutions, markets andinfrastructures. The second involvessystematically identifying the plausible and(systemically) important sources of risks andvulnerabilities that could pose challenges tofinancial stability in the future. The third is anappraisal of the potential costs – that is, theability of the financial system to cope – shouldsome combination of these identified risks andvulnerabilities materialise. In practice, thisrequires an ability to measure (and model)strength and robustness, or to calibrate theplausibility and importance of the variousrisks, or to appraise quantitatively the potentialcosts should risks materialise. However, eachof these areas entails formidable measurementand modelling challenges, so much so that inpractice many shortcuts and qualitativejudgements must be made to produce an overallassessment. This Special Feature is thecompanion to a Special Feature in the last issueof this Review, and discusses some of the mainmeasurement challenges involved in practicalfinancial stability assessment.2

The rest of this Special Feature is organised asfollows. Section 2 discusses some of thepractical challenges involved in implementinga framework for financial stability assessment.It outlines criteria for disciplining the processof information gathering, monitoring andassessing, and it highlights the formidablemeasurement challenges faced. Section 3briefly outlines some of the immediate anddifficult challenges that lie ahead in bothassessing and safeguarding financial stability.Finally, Section 4 briefly draws someconclusions.

1 This special feature draws heavily on J. Fell and G. Schinasi(2005), “Assessing Financial Stability: Exploring theBoundaries of Analysis”, National Institute Economic Review,No 192, April, pp. 102-117.

2 See ECB (2005), “Assessing f inancial stability: Conceptualboundaries and challenges”. Financial Stability Review, June,pp. 117-125.


PRACTICAL CHALLENGES IN IMPLEMENTINGA FRAMEWORK FOR FINANCIAL STABILITYASSESSMENT

DISCIPLINING THE PROCESS OF ASSESSMENTOne of the objectives of any financial stabilityassessment should be to determine whether thefinancial system can be judged to be either in azone or corridor of financial stability, asapproaching a boundary of stability/instability,or outside a zone or corridor of stability.Within the third category, the financial systemcould be further judged to be in a position inwhich self-corrective processes andmechanisms are assessed as being likely tomove the system back towards the corridor ofstability or, alternatively, need promptremedial and even emergency measures toreverse the instability.

While categories of possible assessments maybe straightforward to discuss in principle, theyare difficult to identify in practice. How shouldthe boundary of stability be defined andmeasured, for example? When does an isolatedsmall problem threaten to become a systemicone? There would also seem to be a biastowards being prudent and overreaching inidentifying potential sources of risks andvulnerability and therefore towardsoverestimating their likelihood andimportance. Thus, it would be useful toestablish some ground rules or guidelineswhich could discipline the continuous processof information gathering, analysis, andmonitoring; and, most importantly, to identifysources of risks and vulnerabilities. A checklistof disciplining principles for identifying risksand vulnerabilities and for assessing wherealong the stability spectrum the financialsystem might be could include the following:

– Is the process systematic?– Are the risks identified plausible?– Are the risks identified systemically

relevant?– Can linkages and transmission (or

contagion) channels be identified?– Have risks and linkages been cross-checked?

– Has the identification of risks and theassessment been time-consistent?

In practice, the process of assessing financialstability entails a systematic identification andanalysis of the sources of risk and vulnerabilitythat could impinge on stability in thecircumstances in which the assessment is beingmade. For example, consider thecomprehensive list of sources of risk in TableA.1 below.3

An operationally significant distinction ismade between endogenous sources of risk that

Endogenous Exogenous

Institutions-based: Macroeconomic disturbances:Financial risks Economic environment risk

Credit Policy imbalancesMarket Event risk:Liquidity Natural disasterInterest rate Policy eventsCurrency Large business failures

Operational riskInformation technology

weaknessesLegal/integrity riskReputation riskBusiness strategy riskConcentration riskCapital adequacy risk

Market-based:Counterparty riskAsset price misalignmentsRun on markets

CreditLiquidity

Contagion

Infrastructure-based:Clearance, payment and

settlements system riskInfrastructure fragilities

LegalRegulatoryAccountingSupervisory

Collapse of confidenceleading to runs

Domino effects

Table A.1 Sources of r isk to f inancialstabi l i ty

Source: Houben, Kakes and Schinasi (2004).

3 See A. Houben, J. Kakes and G. Schinasi (2004), “Frameworkfor Safeguarding Financial Stability”, IMF Working Paper04/101.

133cECB


IV SPECIALFEATURES

are present within the financial system, andexogenous sources of risk. Defining thefinancial system broadly, endogenous sourcesof risk can arise either in financial institutions,or in financial markets, or in theinfrastructures, or in any combination ofthese.4 For instance, credit, market or liquidityrisks may be present in financial institutionswhich, if they materialise, could hamper theprocess of reallocating financial resourcesbetween savers and investors. Financialmarkets can be a source of endogenous risk notonly because they offer alternative sources offinance to non-financial sectors, but alsobecause they entail systemic linkages betweenfinancial institutions, and more directlybetween savers and investors. Financialinfrastructures are also an importantendogenous source of risk, in part because theyentail linkages between market participants aswell, but also because they provide theinstitutional framework in which financialinstitutions and markets operate. Outside thefinancial system, the macroeconomicenvironment can be an exogenous source ofrisk for financial stability because it directlyinfluences the ability of economic andfinancial actors (households, companies, andeven the government) to honour their financialobligations. Financial stability assessmentsshould entail a systematic and periodic processof monitoring of each of these sources of risk,both individually and collectively, taking intoaccount cross-sector and also cross-borderlinkages.

Calling attention to the main sources of risk andvulnerability to financial stability does notnecessarily aim at identifying the most likelyfuture scenarios. Instead, it entails theidentification of potential sources of risk andnegative events, even if these are remote andunlikely. In order to preserve discipline in anexercise that essentially involves determiningwhat could go wrong, a key consideration is theplausibility of the risks identified.

For example, an analysis of conditions in thehousehold and corporate sectors might reveal

that a sizeable drop in the rate of output growthcould, by significantly lowering income andprofits, cause a notable rise in household andcorporate loan default rates, and therebythreaten the smooth functioning of thefinancial system. However, if the constellationof economic fundamentals underpinning thepace of economic activity suggests that thelikelihood of recession is very low, then suchan assessment would carry limited value.Ideally, if the probability of a disruptive eventoccurring can be estimated reasonably, then theplausibility of a source of risk can be rigorouslydetermined. In current practice, given data,measurement and methodological limitations(which will be discussed later), in most cases aranking of the plausibility of the various risksidentified must be based on qualitativejudgements based on very limited information.

While it is desirable to consider seriously allplausible sources of risk to financial stability, itwould also be desirable to distinguish sourcesthat could prove to be systemically relevant fromsources that are unlikely to prove costly. Forexample, the plausible risk of an asset marketcorrection would be seen as relatively benign if,given the current conjuncture, it was judged toentail only a minor threat to the financialcondition of the household, corporate andfinancial sectors. However, if the risk wasjudged to threaten the solvency of a significantportion of any one of these sectors, it couldprove more costly from a systemic perspective.The challenge is to distinguish between thosethreats to financial stability that, should theycrystallise, carry a high probability of asignificant disruption to real economic activity,and those that are likely to prove self-correctingwithout having a material impact either on thelevel of activity or the process of resourceallocation. As implied by the examples, thesystemic relevance of a particular set of riskscan be determined if a reasonable judgement – ifnot quantitative assessment – can be made about

4 See G. J. Schinasi (2004), “Def ining Financial Stability”, IMFWorking Paper, No 04/187; and G. J. Schinasi (2005),Safeguarding Financial Stability: Theory and Practice, IMF,December.


the likely real economic costs, given thematerialisation of the risks. Ideally, theexpected losses (for example, ones resultingfrom the product of the probability of the eventand the cost, given materialisation) could lead toa ranking of the importance of the variousplausible risks identified. More realistically,formidable practical challenges remain inassessing and estimating the likelihoods of whatare, typically, low probability events and inmeasuring the associated costs. As discussedlater, costs are also difficult to estimate, but atleast there is a history of financial events thatcould, in principle, allow potential costs to becalibrated.

Once plausible and relevant sources of risk andvulnerability to financial stability have beenidentified, it is important to avoid partialequilibrium analysis. For example, incalibrating the financial stability implicationsof the risk of a sharp drop in equity prices, theanalysis would need to go far beyond itspotential impact on financial markets, andadditionally examine the implications forhousehold balance sheets, future corporatefunding, and so on. More generally, aninternally consistent framework for financialstability analysis requires the linkages andchannels of contagion within the financialsector to be identified, as well as those betweenthe financial and non-financial sectors.Because a financial system is comprised ofmany parts (markets, institutions andinfrastructure), the overall degree of financialstability will depend not only on the degree ofstability of each of its constituent parts, butalso on their linkages and channels ofcontagion. This calls for a comprehensiveapproach to collecting and processinginformation on all the important sectors of theeconomy and the financial system.

With regard to cross-checking, since theprocess of identifying sources of risk andvulnerabilities is to some extent contrarian, inthe sense that it identifies what could go wrong,the burden of proof should arguably be higherthan that required for predicting the most likely

outcome. Hence, financial stability analysisshould involve sufficient cross-checking of theassessment by considering a sufficiently widerange of alternative analytical tools, models,and data sources – and importantly, shouldinclude a continuous dialogue with marketparticipants.

Concerning time consistency, furtherdiscipline on the process of identifying risksand vulnerabilities can be achieved if thehorizon over which a given risk is most likelyto crystallise can be assessed. The empiricalliterature has shown that it can be achallenging, if not impossible, task to predictthe timing of crises. This should not stand in theway of judging whether a given plausiblesource of risk has a near, medium or long-termlikelihood of materialising. Doing thissystematically and periodically for the samesets of risks can improve accountability in theprocess of financial stability assessment. Somerisks may ultimately prove to be self-correctingwithout posing any systemic threat, and in suchcases, it is important to understand the reasonswhy. If a “false signal” was sent because of amore orderly than predicted unwinding of animbalance or because of a structural change,such as better risk management thatstrengthened the financial system and therebymitigated the risk, then this information canserve to improve future assessment.

MEASUREMENT AND MODELLING ISSUESFor most macroeconomic or monetary policyobjectives (unemployment, economic growth,external or budgetary equilibrium, pricestability, etc.) there is a widely acceptedmeasurable (set of) indicator(s) that define,and measure deviations from, the objective,even if these indicators are still subject tomethodological and analytical debate and evencontroversy. In the case of macroeconomicsand monetary economics, it took bothdisciplines some several decades of practice,trial and error, measurement and modellingdevelopment, and fundamental research toaccomplish this. Financial stability analysis isstill in its infancy and thus, by contrast, there is

135cECB


IV SPECIALFEATURES

as yet no widely accepted set of measurableindicators of financial stability that can bemonitored and assessed over time. In part, thisreflects the multifaceted nature of financialstability, as it relates to both the stability andresilience of financial institutions, and to thesmooth functioning of financial markets andsettlement systems over time.5 Moreover, thesediverse factors need to be weighed in terms oftheir potential ultimate influence on realeconomic activity. However, this situation alsoreflects the fact that the discipline of assessingfinancial stability is relatively new. Becausemeasurement is not yet highly developed, thecurrent practice of making financial stabilityassessments could be best described more as anart form rather than as a rigorous discipline orscience.

Each of the three main conceptual aspects ofthe notion of financial stability outlined inSchinasi (2004) – resource allocation, riskpricing and management, and absorptivecapacity – poses challenges for measurement.Take the simple example of measures ofsolvency for judging the potential resilienceand absorptive capacity of an individualfinancial institution. Even if balance sheetcapital (that is, the difference between assetsand liabilities) provides a good indication ofnear-term shock absorption capacity, banksolvency may still not adequately capture theforward-looking dimensions of financialstability. If a bank’s high levels of solvencyreflect missed lending opportunities in a highlycompetitive industry, then the foundations maybe laid for future weaknesses in the bankthrough future profit erosion and loss of marketshare. To take a financial market example,while measures of low asset price volatilitycould be indicative of stable conditions in afinancial market, they may alternatively signala failure in the price discovery process. Shouldthis lead to a misallocation of financialresources, it may sow the seeds ofvulnerabilities that could threaten financialstability in the future.

The challenge of measuring financial systemstability extends well beyond the challenge ofmeasuring the degree of stability in eachindividual sub-component of the financialsystem. Financial stability requires theconstituent components of the system – financialinstitutions, markets and infrastructures – to bejointly stable. Weaknesses and vulnerabilities inone component may or may not compromise thestability of the system as a whole, depending onsize and linkages – including the degree andeffectiveness of risk-sharing between differentcomponents. Moreover, as different parts of thesystem perform different tasks, aggregatinginformation across the system represents achallenge. For example, in diversified financialsystems – where both financial institutions andmarkets are important providers of finance –there is no commonly accepted way ofaggregating information on the degree ofstability in both the banking system andfinancial markets in order to form an overallassessment of system stability. If the bankingsystem is functioning well but, at the same time,there are signs of strains in financial markets, theoverall assessment of financial system stabilityis likely to be ex ante ambiguous, particularly ifthe respective shares of the two components asproviders of finance are similar. The morecomplex and sophisticated a financial system is,the more complex the task of measuring overallstability in a precise way is likely to be.

Measurement challenges in identifying therisks and boundaries to financial stabilitycan be illustrated by examining aspects ofthe Minsky (1977) financial instabilityhypothesis.6 In this hypothesis, as an economy

5 Sets of indicators have been developed – and are widely used –for assessing the soundness of banking institutions. See, forexample, IMF (2003), Analytical Tools of the Financial SectorAssessment Program; IMF (2004), Compilation Guide onFinancial Soundness Indicators; and L. Mörttinen, P. Poloni,P. Sandars and J. Vesala (2005), “Analysing Banking SectorConditions – How to Use Macro-prudential Indicators”, ECBOccasional Paper No 26, April.

6 See H. M. Minsky (1977), “The Financial Stability Hypothesis:An Interpretation of Keynes and an Alternative to “Standard”Theory”, Nebraska Journal of Economics and Business, 16 (1),pp. 5-16.


enters into an upswing, risk premia aregradually eroded as managers of firms andbanks discover that the majority ofconservatively financed projects aresucceeding. Gradually, two characteristicsemerge: “Existing debts are easily validatedand units that were heavily in debt prospered: itpays to lever”. As a result, prevailing riskpremia begin to be considered as excessive.Lenders and borrowers begin to take on greaterrisks and, fuelled by credit and optimism aboutfuture profits, this sets off both growth ininvestment and exponential increases in assetprices. At some point, however, excessesoccur, and the conditions that underpinned theboom eventually trigger its collapse.Overinvestment begins to reduce the return oncapital, bankruptcy rates begin to rise, firmsscale back on investment, and consumersreassess their capacity to repay debt. Asoptimism gives way to pessimism, aggregatedemand in the economy falls sharply and assetprices plummet, possibly inducing a financialcrisis.

In practice, the challenges of mapping suchhypotheses into empirical frameworks formeasurement can be significant. Animplication of this hypothesis is that theinferences for risks to financial stability thatcan be drawn from some imbalance indicatorsmay, at certain points in the cycle, be ratherbenign but, with a small change in the samedirection, could suddenly pose a significantthreat following the breaching of a keythreshold. For instance, theory may not offergood answers to questions such as: at what paceof growth does robust and productiveinvestment become overinvestment?Ultimately, the answers to questions such asthese are likely to be settled not theoreticallybut empirically.

Analytical frameworks are required to help inguiding measurement, for example byidentifying and suggesting the sets of variablesand conditions that could underpin threats tofinancial stability. Presently, there is a dearthof general equilibrium models and

comprehensive system-wide approaches foridentifying measures of, and risks to, financialstability.7 Alternatively, some practitionersemploy partial approaches, relying on theanalysis of individual indicators of financialimbalances. Sometimes this entails basingassessments on “rule of thumb” thresholdsderived from longer-term historical averages orfrom cross-country comparisons. Here, too,important measurement (and modelling) issuescan arise. Many, if not most, imbalanceindicators can be interpreted in one of twoways, with each one, which may be cycle-dependent, having different implications forfinancial stability assessments. As discussed,high levels of bank solvency, while possiblyindicating a stable bank, could equally be theharbinger of emerging vulnerabilities. Narrowspreads across a wide range of fixed incomemarkets could indicate perceptions of lowcredit risk in these markets, but also mayreflect a mispricing of risks – as proved to bethe case prior to, and following, the near-collapse of Long-Term Capital Management(LTCM) in 1998. High price-earnings ratios inequity markets might indicate a stock pricebubble but could alternatively represent anaccurate expectation of a future strengtheningof corporate sector profitability. Similarly,while high non-financial sector debt ratiosmight be indicative of heightened credit risksfacing banks, they could also be a reflection ofa welfare-enhancing relaxation of liquidityconstraints, together with a favourableassessment of long-term economic prospectsby private economic agents. These examplesserve to illustrate that in the absence of relyingon a broad range of indicators and anunderstanding of the broader economic andfinancial environment in which indicators arebeing measured, excessive reliance on singleindicator analyses can lead to unsoundfinancial stability assessments.

7 A rare exception can be found in A. Haldane (2004), “DefiningMonetary and Financial Stability”, Bank of England mimeo.Here a general equilibrium model is used to derive a simplef inancial stability “indicator” that is related to monetarystability.

137cECB


IV SPECIALFEATURES

In identifying risks and vulnerabilities, there areways of dealing with the ambiguities that canarise in single indicator analyses. Whileidentifying financial imbalances ex ante can bechallenging, progress can be made by combiningthe information contained in individualindicators such as credit growth and asset prices(see Hargraves and Schinasi (1993), and morerecently and rigorously, Borio and Lowe(2002)).8 Other cross-checking approaches caninvolve looking beneath the surface of aggregatedata by examining micro data. For instance, thequestion whether or not abnormally highaggregate household debt ratios pose acutecredit risks for banks may easily be settled ifmicro data on households reveal that the mostindebted households also have sufficientfinancial buffers to protect them against sharpchanges in interest costs and/or employmentincome.9 Overall, it would appear that the bestassurance of a robust financial stabilityassessment is to base it on eclectic inputs –including a wide range of data sources.

An important component of any financialstability assessment is to assess the ability ofthe financial system to cope with problems,should plausible risks materialise. One of themost common ways to perform suchassessments is stress testing, based on a rangeof techniques – including sensitivity andscenario analyses. These approaches, whichare increasingly used by individual financialinstitutions10, are also being used at anaggregated macro level for assessing systemicstability. The IMF has formalised this throughthe introduction of macroeconomic stresstesting as a key element in its Financial SectorAssessment Program (FSAP).11 Sensitivitytests are ordinarily designed to isolate thelikely impact of selected risk factors such aschanges in interest or exchange rates. Scenarioanalyses tend to be richer, involvingsimultaneous moves in a number of riskfactors. The scenarios can be based onhistorical episodes of financial stress or onhypothetical events that are considered to beplausible, or on sets of events. As suchapproaches often have a high degree of internal

consistency, they can make an importantcontribution to the understanding of thesystemic relevance of financial risks.

While methodological advances have beenmade, as currently practised, macro stress-testing techniques have several limitations. Theimpacts of scenarios can be gauged boththrough bottom-up approaches – aggregatinginformation on how a range of institutionswould weather a plausible but “challenging”scenario – or at an aggregate level, perhapsemploying a macroeconometric model.Combining the two approaches can facilitatecross-checking and more reliable assessment.However, a limitation of both approaches is thatpotential second-round effects of scenarios tendto be ignored because the underlying modelspay insufficient attention to macro-financialinteraction (as discussed in Hoggarth andWhitley (2003)).12 This means that the overallimpacts of adverse disturbances could well beunderestimated. For instance, in the case of adecline in the pace of economic activity that issufficiently large to challenge the robustness ofthe banking system, weakened banks might facean increase in funding costs and/or a withdrawalof deposits that puts further downward pressureon profits. At the same time, faced withdeterioration in the creditworthiness of theircustomers, banks might be inclined to tightenlending terms and conditions. This would mostlikely have second-round effects on aggregate

8 See M. Hargraves and G. Schinasi (1993), “Boom and Bust” inAsset Markets in the 1980s: Causes and Consequences”, inStaff Studies for the World Economic Outlook, IMF, December;and C. Borio and P. Lowe (2002), “Asset Prices, Financialand Monetary Stability: Exploring the Nexus”, BIS WorkingPapers, 114.

9 See, for instance, Sveriges Riksbank (2004), FinancialStability Report, 1; and the Special Feature in this Reviewentitled “Assessing the financial vulnerability of euro areahouseholds using micro-level data”.

10 See Committee on the Global Financial System (CGFS) (2005),“Stress Testing at Major Financial Institutions: Survey Resultsand Practice”, available at http://www.bis.org/.

11 See IMF (2003), “Analytical Tools of the Financial SectorAssessment Program”; and W. Blaschke, M. Jones, G. Majnoniand S. Peria (2001), “Stress Testing of Financial Systems: AReview of the Issues, Methodologies, and FSAP Experiences”,IMF Working Paper, WP/01/88.

12 See G. Hoggarth and J. Whitley (2003), “Assessing the Strengthof UK banks through Macroeconomic Stress Tests”, Bank ofEngland Financial Stability Review, 14.


demand and output, potentially leading tofurther losses in the banking system. Moreover,for a disturbance that was sufficiently large tocause the failure of a large financial institution,this might have a direct impact on the capital, oreven solvency, of other (counterparty) banks.Macro stress testing, as currently practised, isgenerally not capable of assessing theimportance or gauging the magnitude of theseeffects.

Financial stability assessments carry a higherdegree of uncertainty than ordinarily associatedwith forecasts based on macroeconometricmodels. This is because there can be formidablepractical challenges to measuring, modellingand assessing the consequences of rare events.A first practical challenge is that if past criseshad been prevented or tackled by policy actions,assessments of the likely costs of a selectedscenario, based on simulations drawn fromhistorical datasets, would likely prove to bebiased unless sufficient account is taken ofpolicy reaction functions. It is doubtful that pastpolicy responses to episodes of financial stresscould be summarised by a mechanical reactionfunction, particularly if the authorities weremindful of avoiding the moral hazards thattypically follow from predictable behaviour.Moreover, even in cases that did not promptpolicy responses, the frequency of crises inhistorical datasets may be too low to facilitateprecision in estimating the likely policy-neutralconsequences of a stylised scenario.

Second, confidence intervals around theexpected output losses associated with thematerialisation of a specified scenario may benot well-defined statistically, or even notdefined at all. For instance, simulations basedon historical episodes tend to be founded onstatistical relationships that reflect the centraltendency of, rather than the tails of, probabilitydistributions. Moreover, in purely hypotheticalscenarios, it might not be possible to compute aconfidence interval around the simulationbecause the events themselves may be subjectto so-called Knightian uncertainty – orunquantifiable risk.13

Third, most macroeconometric models used forstress testing tend to be built on the basis oflog-linear relationships. For simulations, thismeans that a doubling of the size of a shock willresult in a proportionate change in the effect.However, in reality, it can never be excludedthat in situations of financial stress,unpredictable non-linearities may surface, forinstance due to threshold effects.

Fourth, as witnessed during the near collapse ofLTCM in 1998, unexpected links may surfaceduring crises, such as correlations betweenfinancial markets that do not ordinarily tend tobe correlated. Given such uncertainties, thereal economic costs associated with aparticular scenario could well prove to belarger than those predicted by an empiricalmodel. Such considerations would suggest thatthe output of any stress-testing exercise shouldonly be viewed as indicative of how, or if, thefinancial system would endure such adversedisturbances. To avoid complacency, this callsfor a high degree of caution and judgement informing financial assessments.

Fifth, concerning measurement of the costs offinancial instability, the literature is just in itsinfancy and has tended to focus on theincreasing incidence of bank crises (see Bordoet al. (2001); Garcia-Herrero and Del Rio(2003)) and their considerable costs (seeLindgren, Garcia, and Saal (1996); Hoggarthand Sapporta (2001); and Barrell, Davis andPomerantz (2005)).14 Even defining a systemicfinancial crisis is not straightforward and, oncedefined, there are several elements to take into

13 See F. H. Knight (1921), Risk, Uncertainty, and Profit,Cambridge, Riverside Press.

14 See M. Bordo, B. Eichengreen, D. Klingebiel and M. Martinez-Peria (2001), “Is the Crisis Problem Growing More Severe?”,Economic Policy, 32, pp. 51-82; A. García-Herrero and P. delRio (2003), “Financial Stability and the Design of MonetaryPolicy”, Documento de Trabajo, 0315, Banco de España;C.-J. Lindgren, G. Garcia and M. Saal (1996), “Bank Soundnessand Macroeconomic Policy”, IMF Occasional Papers, 135;G. Hoggarth and V. Sapporta (2001), “Costs of Banking SystemInstability: Some Empirical Evidence”, Bank of EnglandFinancial Stability Review, 10; R. Barrell, E. P. Davis andO. Pomerantz (2005), “Costs of Financial Instability,Household-sector Balance Sheets and Consumption”, NIESR,mimeo.

139cECB


IV SPECIALFEATURES

account in assessing the costs (as Hoggarth andSapporta demonstrate). In measuring the costs,it is particularly important to be mindful offeedback: banking crises can be caused bysluggishness in the pace of economic activity,but they can also be the cause of an economicslowdown or recession. A challenge formeasurement is to disentangle the feedbackeffects and isolate the quantitative impact ofthe crisis on the economy. The costs associatedwith banking crises can include losses faced bystakeholders in the banks which have failed,including shareholders, depositors and othercreditors. Taxpayers may face costs if there is apublic sector resolution of the crisis. If,because of rising risk aversion or the rationingof credit, borrowers lose access to funds or facedifficulties in accessing other sources offinance, economic activity may be adverselyaffected. The incomes of depositors may alsobe adversely affected if banks seek to widenspreads by lowering deposit interest rates inorder to recoup loan losses. Finally, if thefunctioning of the payment system is impairedbecause consumers become reluctant to makedeposits with banks, the overall adverse impacton economic activity may be magnified. Formeasurement, it is not clear-cut whether theoverall costs should be gauged by losses inGDP, fiscal costs or some combination of thetwo. The impact on the broader macroeconomyof some crises may have been avoided becauseof early resolution, resulting in the incurrenceof fiscal costs. For others there may have beenno direct fiscal implication, but instead asignificant impact on economic activity.

Although the wealth effects and costs of thebursting of asset price bubbles can be gauged,less progress has been made in determining thecosts of financial market turbulence anddislocation. Possible channels would includethe direct and indirect effects of loss of accessto funds for borrowers in capital markets and/orthe costs of refunding short-term obligations athigher cost with financial institutions, as wellas the redistributional effects of asset pricechanges which could, in extreme situations,

have a direct impact on the capital, or evensolvency, of banks.

SOME REMAINING CHALLENGES IN DESIGN ANDIMPLEMENTATIONIn order to advance the practice of financialstability assessment from an art towards ascience, progress is necessary on at least threefronts: data, models and the understanding oflinkages. Regarding data, several areas can beidentified which contain shortcomings. Apriority for data gathering must be microbalance sheet data covering financialinstitutions, households and firms. While apicture of the aggregate risks borne within eachof these sectors can be useful for financialstability analysis, far more important is anunderstanding of the way in which the risks aredistributed across sectors, and especiallywhether or not concentrations or pockets ofvulnerabilities can be pinpointed. In matureeconomies, the availability andcomprehensiveness of such data are rathermixed, particularly for the household sector.

It has become fashionable to employ indicatorsbased on the prices of securities for financialstability assessment. In principle, if marketsare efficient, then indicators derived fromsecurities prices – such as credit spreads,distances-to default, volatilities implied byoptions prices, etc. – should contain invaluableinformation for financial stability. This isbecause securities prices should contain thecollective expectations of the multitude ofmarket participants with regard to theunderlying fundamentals governingvaluations. If those market participants alsohave an eye on the possible impacts of the samerisks and vulnerabilities as the publicauthorities then, in principle, market indicatorscould reveal information on the ability of thefinancial system to weather plausible adversedisturbances. For instance, via risk-neutraldensities, options prices can even facilitatethe extraction of market-based probabilitiesof the occurrence of pre-specified assetprice movements over pre-specified horizons.


15 See R. Sahajwala and P. van den Berg (2000), “SupervisoryRisk Assessment and Early Warning Systems”, BaselCommittee on Banking Supervision Working Paper, 4, for anoverview of early warning systems used by some G10authorities, and M. Persson and M. Blåvarg (2003), “The Use ofMarket Indicators in Financial Stability Analysis”, EconomicReview, Sveriges Riksbank, pp. 5-28, on the use of f inancialmarket indicators.

16 See J. Campbell and R. Shiller (2001), “Valuation Ratios andthe Long run Stock Market Outlook: An Update”, NBERWorking Paper No 8221, April.

17 See Blaschke et al. (2001), who review issues of measurementand methodology in stress testing, as well as the IMF’sexperience with FSAP.

However, there can be risks of circularity inthis analysis: a comprehensive financialstability assessment should attempt to gaugewhether there are plausible risks of marketdislocations resulting from mispricing,whereas inferences on market expectations arebuilt on the assumption that prices are always“correct”. More and better data on quantityindicators – such as indicators of liquidity,leverage, market positioning, etc. – would helpin shedding light not only on the indicatorproperties of securities prices for financialstability assessment, but also on thevulnerabilities prevailing within financialmarkets.

Two areas where more and better analyticalresearch on financial stability modellingappears necessary include models foridentifying risks and vulnerabilities, andmodels for assessing the consequences ofadverse disturbances.15 Concerning theidentification of risks, the literature suggeststhat it is doubtful that models will ever becapable of predicting crises, particularly whenit comes to their precise timing. Nevertheless,this should not stand in the way of developingmodels for assessing vulnerabilities. Evensimple single indicator approaches can beuseful for gauging risks to financial stability(see Campbell and Shiller (2001))16, andongoing work holds out some hope for thedevelopment of more comprehensiveframeworks that could pinpoint the sets ofvariables (see IMF (2004)), as well as theconditions that increase the likelihood offinancial stress (see, for example, Borio andLowe (2002)). As for the prediction of crises, itcannot be excluded that, by borrowing fromadvances made in other disciplines in themodelling of discontinuous processes (such asthe prediction of earthquakes), insights may begained that can benefit financial stabilityassessment.

Ideally, to ensure an accurate assessment of thelikely impact of adverse disturbances, it wouldbe necessary to have dynamic generalequilibrium modelling frameworks capable of

measuring (possibly non-linear) interactionwithin and between financial and non-financialsectors of the economy, including at the globallevel. Although current practices are far fromachieving this, the implementation ofmacroeconomic stress-testing frameworks,such as those increasingly applied in thecontext of the IMF’s FSAPs, have undoubtedlyadvanced the development of internallyconsistent frameworks for assessing theresilience of financial systems to adversedisturbances.17 Sources of risk andvulnerability can be quantitatively mapped intotheir impact on banks’ balance sheets, bothindividually and on a system-wide basis.However, reflecting the limitations ofunderlying models, current practices tend toignore the second-round effects of financialcrises. They also tend to focus exclusively onthe functioning of the banking system, whereasa broader definition of the financial systemrequires an understanding of the likely impactson other financial institutions and on thefunctioning of financial markets andinfrastructures. This calls for further work to beconducted not only on the modelling of real-financial interaction, the complexity of whichexhibits a tendency to increase over time, butalso on interactions within the financial systemitself.

Finally, a good understanding of linkages iscrucial for financial stability analysis. Toensure that important linkages are not missed ina financial stability assessment, both thefinancial system and the sources of potentialrisk and vulnerability should be defined in

141cECB


IV SPECIALFEATURES

sufficiently broad terms. For instance,although alertness grew among marketparticipants and public authorities in the late1990s and in 2000 of the vulnerability of the USstock market to an abrupt correction, generalawareness of the possible impact on theEuropean insurance industry – one of the placeswhere the subsequent market tumble hurt themost – was rather limited. This was mainlybecause little macro-prudential surveillance ofthe industry was being undertaken at the time.Micro balance sheet data, especially onexposures, can be helpful when seeking toidentify the relevance of linkages both betweenreal and financial sectors and within thefinancial system itself. As financialinstitutions strengthen their disclosurepolicies, data availability in this area has thepotential to improve over time. In order to makeinferences on the linkages and channels ofcontagion, cross-correlation analysis ofsecurities prices can also be helpful, althoughsight should not be lost of the fact that duringcrisis periods, correlations may differmarkedly to those prevailing when markets areoperating smoothly.

CONCLUDING REMARKS

Current practice in financial stabilityassessment can probably be compared to theway monetary policy assessment was practisedby central banks three or four decades ago –before there was a widely accepted, rigorousframework. The measurement challenges thatlie ahead for financial stability assessment areformidable, in part because financial stabilityassessments must not only take stock ofdisturbances as they emerge, but also need toidentify and examine the vulnerabilities thatcould lead to such disturbances occurring in thefuture. A forward-looking approach is requiredto identify the potential build-up of financialimbalances and to account for the transmissionlags in policy instruments. The real difficulty isthat financial crises are inherently difficult – ifnot impossible – to predict, in part because ofcontagion effects and likely non-linearities inboth the build-up of imbalances and their

transmission to the real economy. In addition,financial stability risks often reflect the far-reaching consequences of unlikely events. Thisimplies that the focus of attention is not themean, median or mode of possible outcomes,but the entire distribution of outcomes, inparticular the left tail.

While macro stress-testing techniques areimproving knowledge with regard todetermining the systemic relevance ofplausible risks to financial stability, thesetechniques have important limitations –including, most importantly, shortcomings inthe modelling of real-financial interactions andfeedback as well as the uncertainty thatsurrounds estimates of potential costs. Untilthese limitations have been sufficientlyaddressed, the best and most pragmaticassurance of robust financial stabilityassessment is to use an eclectic approach thatdraws upon inputs from a wide range of datasources, indicators and models.

While many conceptual and methodologicalchallenges lie ahead, it is important toacknowledge that significant progress has beenmade in recent years. Even though there is noobvious framework for summarisingdevelopments in financial stability in a singlequantitative measure, a growing number ofcentral banks around the world are makingfinancial stability assessments and publishingfinancial stability reports, many of them basedon a broad and forward-looking conception offinancial stability.


B FINANCIAL MARKET CONTAGION

Recent research has greatly improved theunderstanding of financial contagion. Thereare two main channels though which contagionmay emerge among financial markets: physicalexposure and asymmetric information.Contagion can be empirically identifiedthrough the propagation of extreme negativereturns, the increase in interdependencecompared to normal times, and the distinctionfrom common shocks. The evidence oninternational financial market contagionsuggests that it is a relevant phenomenon thathas indeed occurred in various crises, but insevere form, it is rather rare. In most instancesthe breadth of contagion seems to be limited tospecific countries or geographical regions. Inaddition, it is less frequent across differentasset classes than within the same asset class.Finally, simple measures for market co-movements, such as standard correlationcoefficients, do not usually perform well asindicators of contagion.

INTRODUCTION

Recent research has greatly improved theunderstanding of financial contagion, stressingthe propagation of extreme negative outcomes,the increase in interdependence compared tonormal times, and the distinction of contagionfrom common shocks. This Special Featureexamines the most widely used approaches inthe research literature on how to assessfinancial market contagion phenomena. Thesecond section describes market contagionfrom a theoretical perspective and illustratesits policy relevance. The third section reviewsthe main approaches on how to identifyfinancial market contagion. The fourth sectionprovides some selected evidence about theprevalence and breadth of market contagionphenomena, covering various areas of theworld. Finally, the last section offers sometentative conclusions.

CONCEPT AND POLICY RELEVANCE OFFINANCIAL MARKET CONTAGION

When a crisis in the stock market of onecountry causes a crisis in the stock market ofanother country this can be thought of asfinancial market contagion. There are two mainchannels through which contagion may emergein financial systems: physical exposures andasymmetric information. As an example of theexposure channel, the following scenario canbe considered. Assume that a crash in onefinancial market reduces the wealth of traderswho are also active in other markets. They maythen want to rebalance their portfolios and sellassets in other markets, triggering a crash theretoo, even if the two markets are unrelated interms of their fundamentals (Kyle and Xiong(2001)).1

Asymmetric information across economicagents active in financial systems may alsoresult in contagion. King and Wadhwani(1990)2 argue that traders in internationalfinancial markets face “signal extractionproblems”. Traders from one country may haveonly imperfect information about the situationin other countries. Hence, they have to extractfurther information from observable stockprice movements, reflecting other traders’behaviour. However, sometimes they willconfuse price movements in relation toidiosyncratic problems in a foreign countrywith price movements that also revealinformation about their home country. In thisway, asymmetric information can causeexcessive price spillovers across borders,including crashes. Moreover, Kodres andPritsker (2002)3 show that the transmission ofidiosyncratic shocks across markets throughportfolio rebalancing tends to be reinforcedthrough asymmetric information.

1 A. S. Kyle and W. Xiong (2001), “Contagion as a WealthEffect”, Journal of Finance 56 (4), pp. 1401-40.

2 M. King and S. Wadhwani (1990), “Transmission of Volatilitybetween Stock Markets”, Review of Financial Studies 3 (1),pp. 5-35.

3 L. E. Kodres and M. Pritsker (2002), “A Rational ExpectationsModel of Financial Contagion”, Journal of Finance 57 (2),pp. 769-99.

143cECB


IV SPECIALFEATURES

Contagion is a policy-relevant issue for tworeasons. First, some contagion phenomenahave the character of externalities, resulting inan inefficient allocation of risk in the economy.Agents do not take the effect of their actions onother agents into account and, hence, the levelof risk is too high. Ex ante policies, such asregulating markets, could be used to re-establish efficiency. Moreover, if they are notsuccessful, then ex post intervention could,where necessary, be attempted in order to“neutralise” the trigger of contagion or tocushion the effects on other markets. Second,if contagion is very widespread, then suchpropagation could in theory contribute to ageneral destabilisation of the financial systemand adversely affect growth. In such a worst-case scenario, macroeconomic stabilisationpolicies could help to fight the consequences ofwidespread contagion for the economy as awhole.

HOW CAN CASES OF MARKET CONTAGION BEIDENTIFIED?

The literature has now developed a number ofempirical approaches on how to identifycontagion in financial markets. As differentmethods lead to different results, most of thedebate in the literature and amongpolicymakers is about which approach capturesthe notion of contagion best.

Five main criteria have been proposed so far toidentify contagion: (i) a decline in an assetprice leads to declines in other asset prices; (ii)the relationships between asset price declinesare different from those observed in “normal”times (regular interdependence); (iii) therelationships are in excess of what can beexplained by economic fundamentals; (iv) theyare negative extremes, such as market crashes,so that they correspond to crisis situations; and(v) the relationships are the result ofpropagations over time rather than beingcaused by the simultaneous effects of commonshocks.

Most empirical approaches proposed in theliterature on how to measure market contagioncapture the first criterion, but this is whereagreement usually ends. Authors differ in theirview as to which of the other criteria areessential for identifying cases of contagion.

INCREASED CORRELATION DURING CRISISPERIODSOne influential approach advocating thesecond criterion has been proposed by Forbesand Rigobon (2002).4 The authors argue thatcontagion means that correlations betweendifferent equity markets increase significantlyduring well-known crisis episodes. One reasonmay be the information channel describedabove, which can enhance price spillovers intimes of stress. If correlations do not increase,then any propagation of volatility during thesecrises is nothing more than the expression ofthe regular interdependence between markets,rather than a sign of contagion. The authorsfind no significant increases in equity marketcorrelations during some important crises, suchas the US stock market crash of 1987, theMexican crisis of 1994 or the Asian crisis of1997.

CO-MOVEMENTS IN EXCESS OF ECONOMICFUNDAMENTALSThe idea behind the third criterion in the abovelist (“excess co-movements”) is that iffinancial market prices co-move by more thanwhat would be justified by the fundamentalvariables driving those prices (say, due toasymmetric information), then this would beevidence of contagion. Examples are given invarious studies, such as Shiller (1989)5,Pindyck and Rotemberg (1993)6, and Bekaert,Harvey and Ng (2005).7 Shiller (1989) finds

4 K. Forbes and R. Rigobon (2002), “No Contagion, OnlyInterdependence: Measuring Stock Market Comovements”,Journal of Finance 57 (5), pp. 2223-62.

5 R. J. Shiller (1989), “Comovements in Stock Prices andComovements in Dividends”, Journal of Finance 44 (3),pp. 719-29.

6 R. S. Pindyck and J. J. Rotemberg (1993), “The Comovementsof Stock Prices”, Quarterly Journal of Economics 108 (4),pp. 1073-1104.

7 G. Bekaert, C. Harvey and A. Ng (2005), “Market Integrationand Contagion”, Journal of Business 78 (1), pp. 39-69.


that between 1917 and 1987 US and UK stockmarket indices co-moved by more than whatwould be justified by the relationship betweendividends paid in the US and the UK. Pindyckand Rotemberg (1993) divide 42 US companiesinto six groups, so that in each group thecompanies included produce different goodsand exhibit low earnings correlation with eachother. Then, for each group they runregressions of stock returns on current andlagged macroeconomic fundamentals forquarterly data ranging from 1969 to 1987, andtest whether the residuals of these regressionsare correlated across (within-group) firms. Itturns out that in all cases residuals are highlycorrelated for all groups of companies.Bekaert, Harvey and Ng (2005) estimate a two-factor asset pricing model for stock returns of22 countries, in which risk factors can varyacross specific time periods. Contagion isdefined as an increase in the correlationbetween the model residuals that cannot beexplained by shifts in the common risk factors.In other words, this methodology combines theexcess co-movements approach with theincrease in correlation approach. The authorsfind evidence of such contagion effects amongAsian countries during the Asian crisis, but notduring the Mexican crisis.

CONDITIONAL SPILLOVER PROBABILITIESIn line with the fourth (and first) criterion, afurther group of papers estimates theconditional probabilities of large returns insome markets as a function of large returns inother markets. Three main techniques can bedistinguished in this regard: standard limiteddependent variable estimations, quantileestimations of conditional spillovers, andapplications of extreme value theory.

Limited dependent variable estimationsEichengreen, Rose and Wyplosz (1996)8 wereperhaps the first to estimate the probability thatfinancial crises could spread across countries,using a probit model. 20 industrialisedcountries were covered in their study over atime span between 1959 and 1993. The authorsexamine whether the occurrence of a balance of

payments crisis in one country increases theprobability of a balance of payments crisis inother countries, conditional on political andmacroeconomic country fundamentals. Theresults reject the null hypothesis of nocontagion. Inspired by the epidemiologyliterature, Bae, Karolyi and Stulz (2003)9 applythe multinomial logit model to explainconcurrent large negative and positive returnsamong 17 emerging market countries, the USand Europe between 1992 and 2000 at a dailyfrequency. In other words, they estimate theprobability that a certain number of marketsdecline by more than a certain return thresholdas a function of a number of other marketsdeclining by that much. By controlling for afew fundamentals (interest rates and exchangerates), they can also incorporate some aspectsof the excess co-movements approach(criterion (iii)). They find some evidence ofcontagion between Latin America and Asia, butnone between Asia and the US during the Asiancrisis. Europe seems to be quite sheltered fromshocks occurring in Asia, Latin America andthe US. In this literature large market returnsare usually defined as the 95 percentile, so thatfor weekly data, a large return occurs every 20weeks.

Quantile regressions and co-movement boxCappiello, Gérard and Manganelli (2005)10

estimate conditional spillover probabilitiesbetween two financial markets using quantileregressions. The estimation of conditionalprobabilities in this approach follows a three-step procedure. First, adopting the conditionalquantile regression technique of Engle andManganelli (2004)11, individual time varyingquantiles for returns on each financial market

8 B. Eichengreen, A. Rose and C. Wyplosz (1996), “ContagiousCurrency Crises: First Tests”, Scandinavian Journal ofEconomics 98 (4), pp. 463-84.

9 K. Bae, A. Karolyi and R. Stulz (2003), “A New Approach toMeasuring Financial Contagion”, Review of Financial Studies16 (3), pp. 717-63.

10 L. Cappiello, B. Gérard and S. Manganelli (2005), “MeasuringComovements by Regression Quantiles”, ECB Working PaperNo 501.

11 R. F. Engle and S. Manganelli (2004), “CAViaR: ConditionalAutoregressive Value at Risk by Regression Quantile”, Journalof Business and Economic Statistics 22 (4), pp. 367-81.

145cECB


IV SPECIALFEATURES

are evaluated.12 Second, for each return and foreach quantile, indicator variables that are equalto one if the observed return is lower than theconditional quantile (and zero otherwise) areconstructed. Finally, an ordinary least squaresregression on these indicator functions iscarried out. The regression coefficientsprovide a direct estimate of the conditionalprobabilities of co-movements. This approachcan be used to estimate spillover probabilitiesfor any quantile of the empirical returndistribution, i.e. any size of return, as long as itis not too close to the sample boundaries. Agraphical representation of the spilloverprobabilities for different return sizes (“co-movement box”) allows an assessment onwhether co-movements have increasedsignificantly or not during times of specificcrises. By implementing a statistical test ofsignificant increases in spillovers, the authorsalso integrate the increases in correlationapproach (ii) into their analysis. They applytheir technique to daily data from EMEs inLatin America between 1988 and 2004. Theevidence of contagion during crisis periodsturns out to be mixed.

Applications of the extreme value theoryThe extreme value theory (EVT) literatureargues that in order to identify contagion, onehas to look at much more extreme marketmovements than the 95 or 99 percentile in orderto avoid mixing crisis linkages with non-crisislinkages. For example, the great stock marketcrashes of October 1929 or October 1987 aremuch less frequent, although these are the mostinteresting crises from a financial stabilityperspective. EVT allows conditional spilloverprobabilities to be estimated for these crises,the most dramatic market movements inhistory.

Longin and Solnik (2001)13 were among thefirst to apply bivariate EVT to estimate extremeequity market spillovers. They assume thatequity returns follow a logistic distribution,similar to Bae et al. (2003). This means that theextreme dependence between equity returns isdescribed by the logistic tail copula.14 Under

this assumption and for monthly equity marketreturns of G5 countries between 1958 and1996, they find that the conditional correlationof extreme negative returns (crashes) is higherthan for extreme positive returns (booms).Hartmann, Straetmans and de Vries (2004)15

estimate extreme conditional spilloverprobabilities within and between stock andgovernment bond markets of the G5 countriesfor weekly returns between 1987 and 1999.Looking at crisis linkages across asset classesis important when assessing how widespreadcontagion can be (“systemic risk”). Moreover,they estimate the spillover probabilities semi-parametrically, so that these probabilities (andthe underlying tail copulae) are not fixed tofollow a specific probability law. The resultssuggest that extreme linkages between stockmarkets are higher than extreme linkagesbetween bond markets. Contagion acrossdifferent asset classes is even weaker.Actually, there is evidence of “flight toquality”, which is described by stock marketcrashes being accompanied by boominggovernment bond markets.16

SELECTED EVIDENCE ON INTERNATIONALFINANCIAL MARKET CONTAGION

This section presents some selected evidenceon the prevalence and breadth of contagionphenomena in international financial markets.It covers the three approaches usingconditional spillover probabilities describedabove as applied to different regions in theworld. It starts with the evidence provided by

12 Different quantiles of the return distribution refer to differentsizes of returns.

13 F. Longin and B. Solnik (2001), “Extreme Correlation ofInternational Equity Markets”, Journal of Finance 56 (2),pp. 649-76.

14 For a given bivariate or multivariate distribution, the copula is afunction that describes the dependence between the respectivetwo or more marginal distributions.

15 P. Hartmann, S. Straetmans and C. G. de Vries (2004), “AssetMarket Linkages in Crisis Periods”, Review of Economics andStatistics 86 (1), pp. 313-26.

16 There are also a few papers referring to our last identificationcriterion, the propagation of contagion over time (criterion(v)). They have been surveyed in O. De Bandt and P. Hartmann(2000), “Systemic Risk: A Survey”, ECB Working Paper No. 35,sub-section 4.2.1.1.2, and are not further reviewed here.


EVT on the existence of extreme linkagesbetween the stock and bond markets of G5countries. It then shows to what extentEuropean stock markets are exposed tospillovers from the US, Asia and LatinAmerica, using the multinomial logit model.Last, it reports evidence on contagionphenomena among Latin American EMEsusing the quantile regression approach.

CROSS-ASSET CONTAGION AND FLIGHT TOQUALITY AMONG G5 COUNTRIESTable B.1 reports the results of the EVT cross-asset analysis conducted by Hartmann et al.(2004). The upper panel (panel a.) shows threemeasures of domestic spillovers between stockand government bond markets in France,Germany, the UK, the US and Japan. The“correlation” column shows the estimatedcorrelation coefficient for the respective tworeturn series. The “contagion” column showsthe estimated crisis spillover probability,which is defined as the probability that for agiven country both the stock and thegovernment bond market will crash, assumingthat one of the two has already crashed.17 Thelast column shows the estimated probabilitythat the government bond market will boom,given the stock market crashes (“flight toquality”).

A first observation is that regular correlation isnot a reliable indicator of crisis spillovers. Forexample, the contagion risk between theJapanese stock and bond markets (9%) isalmost twice as high as between the US stockand bond markets (5%). However, the US stockand bond markets are much more highlycorrelated (24%) than those of Japan (5%).Second, contagion risk across both assetclasses is not very high (ranging between 3%and 12%). Third, the “flight to quality”phenomenon is roughly as frequent within thefive countries as contagion. The latter tworesults illustrate some limits to the propagationof market crises within the major industrialcountries.

17 A crash refers to a 20% weekly stock market decline and an 8%weekly bond market decline. This corresponds to the size of the1987 stock market crash and a corresponding bond marketcrash, so as to make the historical frequency of bothapproximately equal.

The lower panel of Table B.1 (panel b.) refersto cross-asset spillovers across borders. It alsodistinguishes the “directions” of spillovers.The country pairs in the left-hand column statefirst the country with a stock market crash andsecond the country with a bond market crash (orboom). For example, the probability in lineFR-US and column “contagion” describes theprobability of a stock market crash in France

Country/ Correlation Contagion Flightcountry pair to quality

a) Domestic linkagesDE 19.0 2.7 3.4FR 24.8 11.5 5.5UK 21.7 5.9 7.3US 23.5 5.2 4.6JP 5.1 9.2 5.0

b) Cross-border linkagesDE-FR 18.7 9.3 5.7FR-DE 17.2 3.9 3.9DE-UK 7.9 7.8 5.9UK-DE 8.3 5.3 5.2DE-US 1.5 3.5 7.9US-DE 12.2 6.0 5.7DE-JP -5.6 9.6 6.8JP-DE 0.0 1.4 3.1FR-UK 16.5 5.2 8.0UK-FR 10.2 6.8 5.1FR-US 10.1 8.0 7.7US-FR 9.7 2.8 3.0FR-JP -0.7 4.1 8.3JP-FR 2.1 3.8 3.6UK-US -5.5 2.5 8.3US-UK 14.1 3.8 5.2UK-JP -1.5 1.6 8.0JP-UK 4.2 4.9 3.2US-JP 6.8 6.9 8.0JP-US -1.1 5.0 3.3

Table B.1 Domestic and internationalextreme stock-bond market l inkages amongG5 countries

Source: Hartmann, Straetmans and de Vries (2004), “AssetMarket Linkages in Crisis Periods”, Review of Economics andStatistics, 86 (1), Table 3, p. 322. ©2004 by the President andFellows of Harvard College and the Massachusetts Institute ofTechnology. Reproduced with kind permission by MIT Press.Note: All f igures in the table are percentages.

147cECB


IV SPECIALFEATURES

given that there is a bond market crash in theUnited States (8%). The line underneath(US-FR) shows the reverse probability, whichis substantially lower (3%). The results showthat the extent of cross-border contagion riskacross assets is quite similar to that of domesticrisk, i.e. not particularly high. Moreover, thereare not any specific geographic patterns. Thismay be interpreted as suggesting that withhighly integrated international capital markets,distance does not shelter countries from crisisspillovers. Finally, there are some indicationsthat the US government bond market hasplayed the role of a safe haven. The flight toquality from other countries to the US bondmarket in the right-hand side column isestimated to be higher than from the US stockmarket to other bond markets, except for Japan.

STOCK MARKET CONTAGION FROM OVERSEASTO EUROPECharts B.1.a-B.1.c select those results fromBae et al. (2003) that provide information aboutthe extent to which European stock markets areexposed to contagion risk from the US, Asiaand Latin America.18 For this application, theyrepresent the probability that a large negativereturn could occur throughout Europe givenlarge negative returns occurring in one, two,three or four Asian or Latin American countries(except for the US). The red areas show theprobabilities of contagion to Europe, whereasthe violet areas show the probabilities of theabsence of contagion. As the number of largestock market downturns overseas increases, thered area becomes larger too, as the likelihoodof adverse effects on European stock marketsalso rises. All in all, the relatively small area inred suggests that Europe is rather insulatedagainst the occurrence of large equity marketdownturns in other regions. Nevertheless, thethree charts also imply that Europe’s exposureto Latin American shocks is still a little bithigher than its exposure to Asian or US shocks.

18 Asia is covered by ten countries (China, Korea, the Philippines,Taiwan, India, Indonesia, Malaysia, Pakistan, Sri Lanka andThailand) and Latin America by seven countries (Argentina,Brazil, Chile, Colombia, Mexico, Peru and Venezuela). As forEurope, the authors use the Datastream International Europeindex, which includes the following countries: Austria,Belgium, Denmark, Finland, France, Germany, Greece, Ireland,Italy, the Netherlands, Norway, Portugal, Spain, Sweden,Switzerland, and the UK.

Chart B.1 Condit ional probabi l ity responsesof European stock markets to large returnson overseas markets

Source: Working paper version of K. H. Bae, G. A. Karolyiand R. M. Stulz (2003), “A New Approach to MeasuringFinancial Contagion”, Review of Financial Studies, 16 (3),pp. 717-36.

0102030405060708090

100

0102030405060708090100

0 1 2 3 4number of co-exceedances of overseas markets

implied conditional probability of co-exceedance event (Y = 0, 1) (%)

P (y = 1)P (y = 0)

a) Spillovers from Asia to Europe

0102030405060708090

100

0102030405060708090100

0 1 2 3 4number of co-exceedances of overseas markets


b) Spillovers from Latin America to Europe

0102030405060708090

100

0102030405060708090100

0 1number of co-exceedances of overseas markets


c) Spillovers from the US to Europe


STOCK MARKET CONTAGION AMONG EMERGINGMARKET COUNTRIESCappiello et al. (2005) represent conditionalspillover probabilities for returns estimatedwith quantile regressions in the so-called co-movement box. This is a square with unit side,where conditional probabilities are plottedagainst the thresholds. When the plot of theconditional probability lies above the 45° line,which represents the case of independencebetween two markets, then this is interpreted asevidence of positive co-movements. In general,the higher the conditional probability, thehigher the co-dependence between two marketreturns. The authors use this methodology toinvestigate the joint impact of the “Tequila”crisis of 1994, the “Asian flu” of 1997 and the“Russian virus” of 1998 on the main LatinAmerican equity markets (Argentina, Brazil,Chile and Mexico).

Charts B.2.a-B.2.d represent the estimatedconditional probability of co-movement for aselected number of country pairs. Two solidlines are plotted together with the case ofindependence. The thin line indicates theconditional probability of co-movements overtranquil times. The thick line, by contrast,shows the conditional probability of co-movements during the three crisis periods.Confidence bands are plotted as dotted lines.For financial stability purposes, the emphasisis on the far left-hand side of the box, i.e. largenegative returns. When the thin line there liesbelow the thick one and outside the confidencebands, this indicates statistically significantcontagion. The results show strong evidence ofcontagion between Argentina and Brazil. Largenegative spillovers also increase for the otherthree cases in the figures, but these changes arenot statistically significant. Overall, it can beconcluded that some EMEs are subject to stockmarket contagion, and others not.

CONCLUDING REMARKS

This Special Feature illustrates that theliterature has now developed a number ofmethods to identify and measure financial

Chart B.2 The co-movement box applied toLatin America – estimated conditionalprobabilities in crisis versus tranquil periods

Source: L. Cappiello, G. Gerard and S. Manganelli (2005),“Measuring Comovements by Regression Quantiles”, ECBWorking Paper, No 501, Figure 4, pp. 25, 26.

crisistranquil

p–(θ)

0.00.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

0.4

0.5

0.5

0.6

0.6

0.7

0.7

0.8

0.8

0.9

0.9

1.0

1.00.00.10.20.30.40.50.60.70.80.91.0

a) Argentina – Brazil

θ

p–(θ)

0.00.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

0.4

0.5

0.5

0.6

0.6

0.7

0.7

0.8

0.8

0.9

0.9

1.0

1.00.00.10.20.30.40.50.60.70.80.91.0

b) Argentina – Chile

θ

p–(θ)

0.00.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

0.4

0.5

0.5

0.6

0.6

0.7

0.7

0.8

0.8

0.9

0.9

1.0

1.00.00.10.20.30.40.50.60.70.80.91.0

c) Mexico – Chile

θ

p–(θ)

θ

0.00.0

0.1

0.1

0.2

0.2

0.3

0.3

0.4

0.4

0.5

0.5

0.6

0.6

0.7

0.7

0.8

0.8

0.9

0.9

1.0

1.00.00.10.20.30.40.50.60.70.80.91.0

d) Mexico – Brazil

149cECB


IV SPECIALFEATURES

market contagion phenomena. While allrelevant features of contagion seem to becaptured, there is still disagreement aboutwhich approach is the best to use. Therefore,it is probably better to employ severalapproaches rather than just a single one. Futureresearch could also help to combine differentapproaches still further.

Keeping the above caveats in mind, thefollowing tentative conclusions may be drawnfrom the evidence provided. As central banksare interested in the prevalence and breadthof contagion from a financial stabilityperspective, the emphasis should be on extrememarket situations. While smaller correlationchanges or excess co-movements may beinefficient, they will usually not be veryimportant in terms of financial instability.Overall, international financial marketcontagion seems to be a relevant but relativelyinfrequent phenomenon. It does not occurwith vehemence in each market crisis, butoccasionally contagion phenomena are present.In most instances the breadth of contagionseems to be limited to specific countries orgeographical regions. Moreover, the extent ofcontagion is easily overestimated if only stockmarkets are considered, which tend to be themost highly interlinked asset class. Manyother asset classes, conversely, tend to be lessinterlinked. In addition, crisis propagationacross different asset classes is much weakerthan within the same asset class. The flightto quality is an economically relevantphenomenon that tends to limit the breadth ofcontagion. Finally, correlations are not a goodindicator of contagion.

While very widespread severe financial marketcontagion is extremely rare, this does notmean that policymakers should disregard italtogether. Policies to maintain internationalfinancial stability are there to keep thelikelihood of such extreme events – potentiallyrelated to general losses of confidence in thesystem – as low as possible. Policymakers mustbe prepared to face the consequences whensuch events do nevertheless occur and

risk affecting the functioning of the economyas a whole. A first step is that individualcountries should “keep their own house inorder” by establishing a stable macroeconomicenvironment and a resilient domestic financialsystem. In a second step – in the absence of aglobal central bank or supervisory authority –international financial surveillance and thesetting of standards by the Financial StabilityForum and the IMF are important.


C ASSESSING THE FINANCIALVULNERABILITY OF MORTGAGE-INDEBTED EURO AREA HOUSEHOLDSUSING MICRO-LEVEL DATA

From a financial stability viewpoint, thecondition of household sector balance sheetscan have an important bearing on the creditrisks that banks face. As in other matureeconomies, increasing household sectorindebtedness in the euro area over recent yearshas raised some concerns about sustainabilityand, as a corollary, creditworthiness. Drawingupon survey information contained in theEuropean Community Household Panel (ECHP)database, this Special Feature highlights someof the characteristics of indebted households inthe euro area, and analyses the degree ofvulnerability of mortgage-indebted households.The picture that emerges from an analysis ofmicro data covering euro area households overthe period 1994-2001 suggests an overallimprovement in resilience. In particular,mortgage debt appeared to be held mostly byhigh-income households, which tend to havegood prospects for servicing debt. Nevertheless,considering the substantial increase in houseprices and the significant accumulation ofmortgage debt in some Member States after theperiod covered by the data examined in thisSpecial Feature, continued monitoring ofhousehold sector indebtedness is called for.

INTRODUCTION

Against a background of improvements in theability of larger groups of households to accesscredit, and as prospects for debt servicingimproved in a low interest rate environment,the aggregate indebtedness of euro areahouseholds began to swell in the late 1990s.1

Between 1997 and 2004 the stock of euro areahousehold debt grew at an annual average rateof about 7%, and the debt-to-GDP ratio ofthe euro area household sector rose from 45%to 55%. With this, concerns about thesustainability of household sector debt and thepossible risks posed to the stability of the euroarea banking system began to surface.

Significant changes in financial aggregates,such as domestic credit growth, have oftenserved as early-warning indicators of financialcrises in mature economies.2 In the euro area,while aggregate household sector debt rose at afaster pace than disposable income over thepast decade, the increase in the debt-to-assetratio was more muted.3 Indeed, the ability ofhouseholds to repay their debts out of liquidfinancial assets remained comfortable, despitesome deterioration. At the same time, asinterest rates fell and remained low, the totaldebt servicing burden – expressed as a ratio ofdisposable income – also remained broadlystable. Hence, assessments based on macro-level indicators of the risks and vulnerabilitiesposed by growing euro area household sectordebt for financial stability have tended to befairly sanguine (see sub-section 2.3).

Any analysis of developments in the aggregateindicators of household balance sheetconditions can mask important disparities infinancial conditions across different segmentsof the household sector. For instance, not allhouseholds are indebted and, in order to drawaccurate conclusions about the sustainabilityof household sector debt, which has to berepaid out of the income of those holding thedebt, it is important to consider the debt ratio ofthe indebted population only. Moreover, a risein the aggregate debt-to-income ratio couldeither be due to a rise in the actual debt ratio ofindebted households, or it could merely reflectthat the proportion of indebted households hasincreased. These two developments can havevery different implications for financialstability, and can only be disentangled byexamining micro data.

1 Household borrowing has increased considerably in a numberof developed countries over the past two decades. See G.Debelle (2004), “Household Debt and the Macro-economy”,BIS Quarterly Review, March.

2 See ECB (2005), “Indicators of f inancial distress in matureeconomies”, Financial Stability Review, June, pp. 126-131.

3 This is a trend shared by many countries, both within andoutside the euro area. For a comprehensive survey, see IMF(2005), “Household Balance Sheets”, Chapter III, GlobalFinancial Stability Report, April.

151cECB


IV SPECIALFEATURES

More generally, the sensitivity of aggregatehousehold balance sheets to changes indisposable income, interest rates or houseprices depends crucially on the relativeimportance and characteristics of differenttypes of indebted households in the totalpopulation. Since these characteristics do nottend to be uniformly distributed across, forinstance, the income spectrum of the householdsector, an analysis based on macro indicatorswill not be capable of detecting growingpockets of fragility within the sector as theyemerge. Hence, when seeking to form acomprehensive view of the risks andvulnerabilities posed by the household sectorfor financial stability, analysis of micro datacan serve to complement macro-basedassessments.

For this purpose, this Special Feature usessurvey information from the EuropeanCommunity Household Panel (ECHP)database. The ECHP is a survey based on astandardised questionnaire that involvesannual interviewing of a representative panelof households and individuals in the 12 euroarea Member States.4 By late 2005, the resultsfrom surveys spanning eight years, runningfrom 1994 to 2001, had become available. Awide range of topics are covered by thequestionnaire, such as income statements,health status and educational background,housing conditions, demographics andemployment characteristics.5 In addition, thesurvey includes an indication of households’own perception of their financial situation, forboth owners and tenants. A set of variablesdescribing households’ financial situation canalso be used to assess financial distress, whichcould be related to an excessive debt burden,and might provide a measure of defaultrisk.6

AN OVERVIEW OF HOUSEHOLD INDEBTEDNESS

Data from the ECHP survey make it feasibleto focus the assessment of the financialconditions on indebted households only.Notably, despite a significant increase in the

amount of household debt outstanding between1997 and 2001, the proportion of indebtedhouseholds in the euro area increased onlyslightly from 33.6% to 36.4% of the householdpopulation over that period (see Table C.1).This contrasts with the situation in the US,where as much as 75.1% of households held atleast one type of debt in 2001, up from 71.3% in1998.7 The distribution of debt by type –mortgage or non-housing-related debt –remained fairly stable in the euro area over theperiod covered by the survey. Approximately21% of euro area households had a mortgageloan in 2001, and another 22% held some typeof non-housing-related debt. The share ofhouseholds carrying both types of debt washowever rather small (6.9% in 2001).8

In order to analyse in more depth theindebtedness situation of households and their

4 The survey also includes Denmark, Sweden and the UK.5 Despite the rich information content of the database, the short

time-span of the coverage and the timeliness of the findingspresent obvious shortcomings. Furthermore, the ECHP surveydoes not contain stock variables on households’ balance sheetitems (outstanding debt, holding of assets, etc).

6 For a complete description of the ECHP database, see Eurostat(2003), “ECHP UDB: Description of Variables – DataDictionary, Codebook and Differences between Countries andWaves”, December. The ECHP has now been discontinued (thelast wave was carried out in 2001), and will be replaced by theEU Statistics on Income and Living Conditions (EU-SILC).Unlike the ECHP, the EU-SILC will be harmonised ex post.

7 See A. Aizcorbe, A. Kernickell and K. Moore (2003), “RecentChanges in US Family Finances: Results from the 1998 and 2001Survey of Consumer Finances”, Federal Reserve Bulletin,January.

8 As a comparison, in 2001 49.3% of US families held home-secured debt (or other residential property debt), while 45.2%had instalment loans.

Table C.1 Distr ibution of household debt inthe euro area

(% of total)

Sources: ECHP database and ECB calculations.Note: Proportions calculated based on population weights.

1995 1997 1999 2001

mortgage debt 19.1 20.1 20.6 21.2non-housing debt 14.4 19.3 20.9 22.1both debt categories 4.3 5.8 6.5 6.9indebted 29.2 33.6 35.1 36.4no debt 70.8 66.4 64.9 63.6


ability to service their debt, the householdscovered by the survey were divided into fiveequally sized categories according to theirmonthly net income.9 As might be expected, theproportion of indebted households in the euroarea has tended to be larger for higher incomesegments. In 2001, the proportion of indebtedhouseholds ranged from 15% in the lowestincome segment to 53% in the highest. Therewere also differences in the distribution ofindebtedness across income levels for thedifferent categories of debt. The share ofmortgage-indebted households rose sharplywith income, from 6% in the lowest incomecategory to 37% in the highest category,suggesting that the bulk of mortgage debt wasconcentrated in the higher income segments.By contrast, non-housing debt tended to bemore evenly distributed, at least in the higherincome categories. Most of the householdswith the lowest incomes appeared to have nodebt at all, and in those cases where they did,they tended to hold non-housing debt.Households with both types of loans belongedgenerally to the higher income segments.10

Considering patterns over time, the share ofmortgage-indebted households increasedsomewhat between 1997 and 2001, especiallyfor the mid-income categories, while itdecreased slightly for the highest incomecategory. The proportion of householdsholding non-housing debt, however, increasedsomewhat more over the same period (seeTable C.1). In the four highest incomecategories, the proportion of householdscarrying non-housing debt increased by 3to 4%.11

Within the euro area, the proportion of indebtedhouseholds, as well as the stock of debt, varies toa wide degree across countries. At one extreme,only about 10% of the total number ofhouseholds in Italy and Greece carried mortgagedebt in 2001; at the other, almost every secondhousehold in the Netherlands had a mortgageloan (see Chart C.1). Analysis at the countrylevel shows that between 1994 and 2001, nosignificant changes occurred in euro area

countries in the debt composition across incomecategories, including those which experiencedcomparatively high growth in debt over theperiod.

FINANCIAL CHARACTERISTICS OF MORTGAGE-INDEBTED HOUSEHOLDS

Lending for house purchase is of particularrelevance to the banking sector, as it representsthe bulk of bank lending to households in theeuro area and it has expanded the most rapidlyin recent years. Mortgage lending accountedfor 69% of the outstanding amount of loans tohouseholds in the second quarter of 2005, upfrom 59% in 1998. The other loan categories –consumer credit and other lending – are mostlyunsecured, which might imply an additionalelement of risk for the banks. However, thesecategories only represented 13% and 18%

Chart C.1 Proportion of indebtedhouseholds across euro area countries

(2001, % of total)

Sources: ECHP database and ECB calculations.

mortgage debtother debt

0NL IE LU BE FI FR ES AT euro

areaDE PT GRIT

10

20

30

40

50

0

10

20

30

40

50

9 The grouping of households, and the calculation of the statisticspresented, was conducted by applying population weights to thecountry samples. The aggregate results for the euro area wereproduced using country weights based on population size.

10 These qualitative f indings for the euro area broadly mirrorthose found in the 1998-2001 US Survey of ConsumerFinances. The distribution of home-secured loans in the US isalso skewed towards the higher income segments, while thepercentage of families with instalment loans appears to be moreevenly distributed.

11 The proportion of households with non-housing debt in thelowest income category – representing a group likely to be moresensitive to variations in unemployment and interest rates –only increased by 1% over the same period.

153cECB


IV SPECIALFEATURES

respectively of the total volume of creditoutstanding in the second quarter of 2005,decreasing from 16% and 25% respectively in1998. Moreover, the average amount of anindividual loan for house purchase issubstantially higher than in the other loancategories. Therefore, the rest of this SpecialFeature will focus on the financial situation ofhouseholds carrying mortgage debt, thesustainability of which bears the strongestimplications for financial stability.

In terms of financial resources, mortgage-indebted households tended to have a higheraverage level of income than the totalpopulation. These households also tended toreport a better ability to save, regardless of theirincome level. On the other hand, mortgageborrowers had a lower average capital incomethan the household sector in total, and were lesslikely to possess a holiday home. The householdsin the highest income segment, however,represent an exception, reporting significantlyhigher capital income on average, and having anabove-average income from rental activities.

Although the ECHP survey does not provideinformation on amounts, inferences can bemade about where debt holdings areconcentrated. First, considering both theproportion of mortgage-indebted householdsand the level of repayment burdens faced, thetwo highest income categories appeared to bethe holders of the bulk of mortgage debt in theeuro area. Second, the relationship betweenthe share of mortgage-indebted households inthe highest income category and the overallshare of mortgage-indebtedness is indeedstrongly proportional across individual euroarea countries. Chart C.2 illustrates that thelarger the proportion of mortgage-indebtedhouseholds in a given country, the larger theshare of mortgage-indebted households inthe highest income category. For instance, inthe case of the country having the highestproportion of mortgage-indebted (close to 50%of the total population) in 2001, the associatedproportion of mortgage-indebted householdsin income category 5 was as high as 80%.

All else being equal, the higher the volatility ofhousehold income, the higher is the credit risklikely to be for banks who have extended loansto these households. In this respect, it wasnotable that the households in the highestincome category also displayed the lowestincome volatility throughout the sample. Atthe euro area level, the average volatility ofincome for the total population was 26%.The corresponding figure for the mortgage-indebted households was 22%, and as low as16% for the mortgage-indebted households inthe highest income category.12 This patternproved to hold true for all individual countrysamples, even if the level of volatility differedto some extent.

Because of the considerable dispersion in thelevels of indebtedness across countries withinthe euro area, the level of mortgage paymentburdens differs as well. In 2001, for instance,the average level of mortgage payments toincome ranged between slightly more than 14%in Greece to 27% in the Netherlands.13

Chart C.2 Concentration of mortgage- indebtedhouseholds in the highest income categoryacross euro area countries(2001, %)


12 The volatility estimates (representing the average one standarddeviation in the annual percentage change in household income)were compiled using the panel dimension of the data, trackingindividual households between 1996 and 2001.

13 The level of mortgage debt outstanding and the mortgageservicing burden depend, to a large extent, on differences in thenational mortgage markets and the domestic tax treatment ofmortgage debt.

0

10

20

30

40

50

60

70

80

90

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50% of mortgage-indebted in the country population

% of mortgage-indebted in income category 5

euro area


THE SUSTAINABILITY OF MORTGAGE DEBT

From a financial stability viewpoint, whatultimately matters is the risk that banks face frommortgage lending. In this context, banks face therisk of being confronted with a higher rate ofdefault on mortgage credit than they have setaside in provisions. The key determinant of theprobability of such an event is the risk facingindividual mortgage-indebted households of beingunable to meet their debt servicing obligations.

There are many factors that can determine thefinancial risks affecting the debt servicingcapacity of households. The sources of risk canbe systematic or idiosyncratic. Among thesystematic sources of risk, there are adversedisturbances at the regional, national or euroarea-wide levels (such as changes inunemployment, interest rates or house prices).Idiosyncratic sources of risk (such as illness anddivorce) might affect the financial situation ofindividual households. In the remainder of thisSpecial Feature, the focus will be on shocks

triggered by macroeconomic events that arelikely to impact banks’ balance sheets the most.Financial risks also differ regarding their effecton financial resources (income flows, financialassets or home equity) or financial commitments(interest payments and family-related expenses).

Financial risk only matters to households if they,rather than for instance banks14, are ultimatelyexposed to its consequences. Vulnerability is anex ante measure of this sensitivity, which couldbe defined as the degree to which householdswould be able to cope with the adverse effects ofa shock, should it crystallise.

All else being equal, the total debt-at-risk oflenders’ loan portfolios will have increased ifbanks have extended credit to vulnerableborrowers, or if an increasing proportion ofindebted households become vulnerable. Thedata allow us to construct several indicators

14 For instance, when mortgages are extended at f ixed rates ofinterest, it will be the bank that extended the loan, rather thanthe borrower, which ultimately faces interest rate risk.

Table C.2 Household r isk exposure and vulnerabi l ity indicators

Source: ECB.

trigger event financial variable vulnerability threshold of vulnerability vulnerabilityof financial impacted indicators financial indicators indicatorsrisk (quantitative) distress (subjective) = (qualitative) =

perception of inability to makefinancial distress payment

rise in interest monthly interest monthly debt above 30% housing costs inability to payrates rate payment servicing burden- are a burden mortgage

on outstanding to-income ratio paymentsmortgage

inability to pay

monthly debt above 30%

mortgage

service burden-

payments

rise inmonthly income

to-income ratioproblems making

inability to

unemploymentends meet

face basic

rate financial margin below zero expenses

inability topay bills

decrease in housing equity mortgage debt belowhouse prices (market value to housing equity loan-to-value(in combination of the house) (or net housing ratiowith one of the equity)two other risks)

155cECB


IV SPECIALFEATURES

that measure the financial strength ofhouseholds. A set of indicators representingmeasures of vulnerability, which correspondsto each of the different types of risk discussed,is presented in Table C.2. Each indicator is alsoassociated with a certain threshold that maysignal financial distress.

In terms of its ability to capture the differentdimensions of risk faced by households, thetotal debt servicing burden-to-income ratiomay have the highest information content forgauging debt-related vulnerability. This isbecause it can be impacted both by shocks tointerest rate payments and to income.15 Someevidence suggests that households facing debtservicing burdens in excess of 30% of theirincome might be classified as riskyborrowers.16 This can define a threshold for thetotal debt servicing burden-to-income ratio inindicating risks of financial distress related tothis indicator.

The vulnerability of the household sector tofinancial strains can also increase in theabsence (or because of weaknesses) of risk-reducing elements which could prevent orminimise the loss associated with theoccurrence of a particular event. A usefulindicator in this respect is the financial marginof mortgage-indebted households, which isdefined as the difference between current totalmonthly income and the reported minimummonthly income necessary to make ends meet.A shock to income caused by deterioratingeconomic conditions (loss of or decrease inincome owing to unemployment) couldsqueeze this margin to a level that would pushhouseholds into financial distress, andeventually result in them being unable to meetrepayment obligations. A negative financialmargin is assumed to indicate financialdistress.

For each indicator, the relevance of the above-mentioned associated quantitative threshold offinancial distress is tested. For this purpose,two variables were used as proxies for financialdistress: one indicating the perception of a high

financial burden related to housing costs, andanother reporting the household’s inabilityto repay mortgage instalments over the last12 months. One would expect the proportionof mortgage-indebted households reportingfinancial distress, as captured by the proxyvariables, to increase significantly at, oraround, the theoretical thresholds.

However, there appears to be no straightforwardway to quantify the probability of beingfinancially distressed on the basis ofrespondents’ answers. The data do not supportthe existence of any uniquely defined thresholdfor financial vulnerability. In terms of themortgage debt servicing burden, the share ofhouseholds reporting that mortgage paymentsare a heavy burden does not significantlyincrease at the predefined theoretical thresholdof 30% (see Chart C.3). As much as 60% of thosewho declared their mortgage payment to be

Chart C.3 Perception of housing costs beinga heavy burden

(2001)


mortgage payment perceived as somewhat a heavy burdenmortgage payment perceived as a heavy burden

0.00

0.1

10

0.2

20

0.3

30

0.4

40

0.5

50

0.6

60

0.7

70

0.8

80

0.9

90

1.0

1000.00.10.20.30.40.50.60.70.80.91.0

mortgage payment burden (% of monthly net income)

cumulative share

15 The impact of a change in interest rates depends among otherfactors on the interest rate variation regime of the mortgagecontract, which varies widely across countries (see ECB(2004), Financial Stability Review, December, Box 6).

16 Some research based on individual US household balancesheets has associated the ratio of annual payments of principaland interest on all outstanding debt obligations (consumer andmortgage debt) to annual disposable income. A ratio of higherthan 30% was found to be a statistically significant predictorof future household insolvency. See S. A. DeVaney andR. H. Lytton (1995), “Household Insolvency: A Review ofHousehold Debt Repayment, Delinquency, and Bankruptcy”,Financial Services Review, 4 (2), pp. 137-56.


somewhat a heavy burden in 2001 had amortgage debt servicing burden ratio less thanthe euro area average of 23%. The correspondingshare for those perceiving mortgage payments asbeing a heavy burden was lower, but still ratherhigh, at about 40%. Likewise, it does not seempossible to establish a clear relationshipbetween a default on payments and the fact ofhaving a negative financial margin.

A detailed picture of changes in thevulnerability indicators of mortgage-indebtedhouseholds across income categories suggeststhat, notwithstanding the increase in householdindebtedness, financial resilience improvedsomewhat in the euro area as a whole between1994 and 2001. Both the payment ratio and thefinancial margin exhibited a stable pattern overthe sample period in all five income categories,improving only slightly. On the other hand, theproportion of mortgage-indebted householdsreporting mortgage payment difficulties fellconsiderably in the lower and middle incomecategories. In the third income category, forexample, this share fell from 4.8% to 3.3%between 1994 and 2001.

Chart C.4 Proportion of vulnerablehouseholds among euro area mortgageborrowers across income categories(% of total)

Sources: ECHP database and ECB calculations.Note: Vulnerable households are def ined as having amortgage payment burden in excess of 30%, a negativef inancial margin, and a reported inability to pay theirmortgage. Germany and Luxembourg have been excludedowing to data limitations.

1995199719992001

0

1

2

3

4

5

6

7

8

0

1

2

3

4

5

6

7

8

category 1 category 2 category 3 category 4 category 5

Chart C.5 Vulnerabi l ity of mortgage-indebted households versus the totaldebt-to-GDP ratio in the euro area(1994 - 2004, % of total)

Sources: ECHP database and ECB calculations.Note: The proportion of euro area mortgage-indebtedhouseholds that reported diff iculties in meeting mortgagepayments has been calculated excluding Germany andLuxembourg, owing to data limitations.

proportion of households unable to meet mortgage payments (right-hand scale)total debt-to-GDP ratio of the household sector (left-hand scale)

40

45

50

55

60

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 20042

3

4

5

6

Of particular interest from a financial riskperspective is the category of mortgage-indebted households that are impaired byall three above-mentioned vulnerabilitycharacteristics, i.e. those with a mortgagepayment burden in excess of 30%, togetherwith a negative financial margin and a reportedinability to pay their mortgage. Chart C.4shows that the proportion of those mostvulnerable mortgage borrowers tended to belower in the higher income categories, anddeclined throughout the sample period. Thisproportion was the lowest in the two highestincome categories, which accounted for almost65% of all mortgage-indebted households in2001 (compared to 6% for the lowest incomecategory). Again, households in these twocategories are likely to carry the bulk ofoutstanding mortgage debt.

Several other variables in the dataset alsosupport the view that financial resilienceimproved between 1994 and 2001. Forinstance, the ability to pay for utilities andother loans increased in all income categories.Likewise, households in all income segmentsreported much better prospects of making endsmeet.

157cECB


IV SPECIALFEATURES

17 See O. May and M. Tudela (2005), “When Is MortgageIndebtedness a Financial Burden to British Households? ADynamic Probit Approach”, Bank of England Working PaperNo 277, October.

18 See De Nederlandsche Bank (2004), “Financial Behaviour ofDutch Households”, Quarterly Bulletin, September.

19 See Sveriges Riksbank (2005), “Swedish Households’ Debt-Servicing Ability 1997-2003”, Financial Stability Report, 1,May.

20 See Banca d’Italia (2004), Annual Report 2003, May.21 This applies to interest rates at both the short and the long end

of the yield curve.22 See ECB (2005), Financial Stability Review, June, Box 6.

Overall, the analysis suggests that borrowers’vulnerability decreased in all countries wheremortgage debt grew rapidly between 1997 and2001. It should however be emphasised thatthis analysis does not cover the period from2001 onwards. This has to be borne in mindwhen considering the results, since this periodhas seen the most dramatic growth in mortgagedebt (see Chart C.5).

The lack of survey data for the period after2001 does not allow any conclusions to bedrawn about how patterns of householdvulnerability have changed since then.Nevertheless, it is possible to highlight a fewfacts that characterise latest developments.

First, the results from the ECHP data for theperiod 1994-2001 appear to be broadly in linewith a number of country-level studiesencompassing more recent data. These studiesdid not find any major recent deterioration inthe financial situation of households, andconfirmed – at least in the countries surveyed –that most of the debt still appears to be carriedby the highest income households. Analysis ofdata from the British Household Panel Survey(BHPS) leads to the conclusion that theprobability of mortgage payment problemsamong UK households, and the amount of debtat risk, decreased between 1994 and 2002.17 Inthe case of the Netherlands, analysis based onmicro-level data is available up to 2004.18 Moreupdated survey information is also availablefrom Sweden, showing that the resilience of theSwedish household sector remained unchangedbetween 2001 and 2003.19 Furthermore,analysis based on the Banca d’Italia survey ofhousehold income and wealth (up to 2002)concludes that the largest part of Italianhousehold borrowing is accounted for bywealthier households.20

Second, the data show that the most significantchanges in the vulnerability indicatorsdiscussed here appear to have followeddevelopments in household income that weredriven by macroeconomic developments.Following the slowdown in economic activity

between 2000 and 2002, euro area growthpicked up in the second half of 2003. Thisdevelopment, together with the fact thatinterest rates continued to decline over theperiod since 2001,21 suggests that the paymentburden of euro area households has beencontained since then.

Third, the most recent accumulation of euroarea mortgage debt seems to have taken placein the countries where the financial resilienceof the household sector improved between1994 and 2001.

CONCLUDING REMARKS

The level of household indebtedness in the euroarea does not necessarily pose a material risk tofinancial stability in itself. Households outsidethe euro area have been able to carry muchheavier debt burdens than those in the euroarea.22 The picture emerging from the analysisbased on micro data suggests that the rise ofhousehold mortgage indebtedness between1994 and 2001 did not create any major pocketsof vulnerability within the euro area householdsector over the period. The households mostlikely to carry most of the debt were those withthe highest incomes. In addition, this categoryof households had sufficient financial marginsto cope with an unexpected decrease in income,and held significant wealth buffers. Asexpected, this category also showed the bestability to cope with the adverse effects ofa financial shock, as measured by thevulnerability indicators.

Based on macro indicators, e.g. the estimateddebt service burden of the household sector


(see Chart S37), there is little evidence tosupport concerns that euro area householdsector resilience has deteriorated since 2001given the decline in interest rates. A rise ininterest rates could, however, alter the assessment.Unfortunately, it is not straightforward toextrapolate the results from the micro databeyond the sample period. The distinctivefeatures of the national markets for housingfinance in the euro area tend to makecomparisons of vulnerability across countriessomewhat difficult, especially as increases inhouse prices and the stock of mortgage debthave accelerated in many Member States since2001. To the extent that households have beenmyopic concerning their expectations forfuture developments in interest rates andincome growth, an unanticipated increase inrepayment burdens could strain householdbalance sheet’s, ultimately posing credit risksfor banks. Against this background, continuedmonitoring of household sector indebtedness iscalled for.

159cECB


IV SPECIALFEATURES

D WHAT DETERMINES EURO AREA BANKPROFITABILITY?

Banks are key components of the euro areafinancial system. Understanding the interplaybetween banks and their operatingenvironment assists in identifying sources ofrisk and vulnerability within the system. ThisSpecial Feature attempts to examine theempirical importance of bank-specific, marketstructure and macro-financial factors on euroarea banks’ financial performance over thelast decade or so.

INTRODUCTION

Healthy and sustainable banking sectorprofitability is vital for maintaining thestability of the financial system. Even ifsolvency is robust, weak profitability can, byweakening the capacity of the system to absorbadverse disturbances, sow the seeds of futurevulnerabilities. This Special Feature empiricallyexamines factors that may drive profitability,measured by return on equity (ROE), among apanel of large banks in the euro area, based onindividual banks’ annual accounting data overthe period 1993-2004. It builds on previouswork in this area by trying to incorporate bank-specific, market structure and macroeconomicfactors simultaneously in an empirical modeland over a longer time period than previousstudies.

The main findings are that bank profits tend tobe persistent over time, though the inclusion ofdifferent explanatory variables weakens thestatistical significance of this finding. Growthin total assets is positively related toprofitability. Banks’ equity capital appears tobe positively related to profitability, althoughthe evidence for this is somewhat mixed,depending on the control variables included.Finally, the macroeconomic environment, ascaptured by real GDP growth, positivelyinfluences bank profitability, a findingreported by banks themselves.1 Overall, theresults point to a need to improveunderstanding of the interplay between the

macroeconomic environment and the bankingsector.

The reminder of this Special Feature isorganised as follows: first a brief review ofthe relevant literature is provided; then, itprovides an overview of the data and empiricalmethodology; and finally, it summarises theresults and conclusions.

FACTORS INFLUENCING BANK PROFITABILITY

Banks’ earnings, or profitability, are one ofthe main indicators used to make assessmentsof the health of individual banks and, atthe aggregate level, the banking system as awhole.2 The question as to what determinesbank profitability can, of course, beapproached from several different angles. Forsimplicity, these factors are discussed underthree main headings: bank-specific factors;market structure factors; and macro-financialfactors identified in the previous work in thearea.3

BANK-SPECIFIC FACTORSBanks may differ in terms of their competitivestrategy, efficiency, asset and liabilitydiversification, and the way they managecapital and credit risk. The strategy andinternal operations of an international bankwhose balance sheet is measured in billions of

1 See Box 12 “Survey on major EU banks’ perception of risks inthe year ahead” in this Review.

2 A complementary approach, not employed in this SpecialFeature, utilises measures of prof itability derived from equitymarkets, and relates these to bank-specific and business cyclevariables. See L. Baele, R. Vander Vennet and O. De Jonghe(2004), “Bank Risks and the Business Cycle”, University ofGhent, Department of Economics Working Paper No 264. Otherapproaches use banks’ equity and debt prices as inputs in orderto calculate forward-looking market indicators and measuresof contagion risk from one bank to another; see for exampleR. Gropp, J. Vesala and G. Vulpes (2005), “Equity and BondMarket Signals as Leading Indicators of Bank Fragility”,Journal of Money, Credit and Banking, forthcoming.

3 More expansive reviews can be found in A. Berger (1995), “TheRelationship between Capital and Earnings in Banking”,Journal of Money, Credit and Banking, 27 (2), May; J. Goddard,P. Molyneux and J. Wilson (2004), “Dynamics of Growth andProfitability in Banking”, Journal of Money, Credit andBanking, 36 (6); and C. Northcott (2004), “Competition inBanking: A Review of the Literature”, Bank of Canada WorkingPaper No 24.


euro is unlikely to be similar to a community-based savings bank with a balance sheetmeasured in millions. However, size does notnecessarily say anything about the banks’relative profitability. Rather, profits are morelikely driven by the competitive strategychosen by the respective banks. Size, inbalance sheet terms, may be a poor proxy forstrategy, which more often tends to bedetermined by the bank’s corporate ownershipmodel.4

This may be an important consideration for theeuro area, given that the euro area bankingsector is composed of a fairly diverse group ofinstitutions, both in terms of size andownership structure. Indeed, banks in the euroarea range from large bank holding companiesand commercial banks to small savings,cooperative and mortgage banks. In addition,there is a large number of specialisedgovernment-owned banks. This complicatesthe analysis of profitability in the euro areabanking sector when using bank-specificcharacteristics such as size and ownership asexplanatory factors.

Just as productivity is an importantdeterminant of macroeconomic performance,efficiency at the firm level is an obviousdriver of bank profitability. When measuringefficiency in banking, one typically tries togauge how a particular set of prices andquantities of inputs and outputs vary, inaccordance with the banks’ chosen strategy,and how this impacts on bank profitability.

Findings from the literature suggest that amongcertain bank categories, such as commercialbanks, large banks tend to be more efficientthan smaller ones. This result however may nothold for banks with other types of ownershipstructures, such as savings banks. Owing to thediffering sample periods, variables andestimation techniques adopted in the variousstudies, it is difficult to draw any generalconclusions concerning the efficiency of theEuropean banking sector as a whole.5

A recent additional line of research isconcerned with the effects of diversification onbank profitability. The intuition here is thatmore diverse sources of income may contributeto smoother/higher profitability. One of theways that diversification has been measured inthe empirical literature is to use off-balancesheet items as a proxy for non-interest income.Some evidence has been found that bankprofitability is positively related to the extentof off-balance sheet business. Suchconclusions need to be qualified, however,since the benefits of diversification may beoutweighed by the exposure to non-interestincome activities. These may be more volatileand less profitable than income generatedthrough lending.6

Adequate management of bank capital can alsobe important in determining bank profitabilitybecause it potentially has a bearing on theavailability of funding for future lendingdecisions. The empirical literature focusing onissues of a regulatory nature, such as capitaladequacy, has found, based on US data, thatcapitalisation and profitability are positivelyrelated. It is also a key determinant of bankcredit ratings, thereby directly affecting thecosts of funding faced by banks. Highercapitalisation contributes to higher earnings,mainly through a reduction in interest rates

4 See R. DeYoung and T. Rice (2004), “How Do Banks MakeMoney? A Variety of Business Strategies”, Federal ReserveBank of Chicago Economic Perspectives, Q4. One notableexception in the European context is Y. Atlumbas and D.Marques Ibanez (2004), “Mergers and Acquisitions and BankPerformance in Europe: The Role of Strategic Similarities”,ECB Working Paper No 398.

5 See P. Schure, R. Wagenvoort and D. O’Brien (2004), “TheEff iciency and Conduct of European Banks: Developmentsafter 1992”, Review of Financial Economics, 13. For acomprehensive overview of the relationship betweencompetition, eff iciency and prof itability, see J. Bikker and J.Bos (2004), “Trends in Competition and Profitability in theBanking Industry: A Basic Framework”, DNB Working PaperNo 18.

6 On the topic of off-balance sheet items and profitability, seeJ. Goddard, P. Molyneux and J. Wilson (2004), “TheProfitability of European Banks: A Cross-sectional andDynamic Panel Analysis”, Manchester School, 77 (3). For thedownside of diversif ication, see K. Sitroh and A. Rumble(2005), “The Dark Side of Diversif ication: The Case of USFinancial Holding Companies”, Journal of Banking andFinance, forthcoming.

161cECB


IV SPECIALFEATURES

charged on deposits not covered by depositinsurance, such as interbank deposits. Theoptimal management of bank capital managesto balance this constraint against that offoregoing profitable and riskier lendingactivities. Studies based on EU banks also findtentative evidence of a positive relationshipbetween capitalisation and profitability,though the significance of this relationshipvaries across the countries in the sample.7

For banks, unexpected losses are deductedfrom capital and expected losses from creditrisk are managed through the use of loan lossprovisions. Most institutions set aside apredetermined amount to cover expectedlosses, and other amounts to cover lossesrelated to specific loans. Increased provisionsreduce profitability by increasing expenses onbanks’ profit and loss accounts. Someempirical evidence suggests that banks mayunder-provision during business cycle upturns,and delay provisioning until the downturn hasset in.8

THE ROLE OF MARKET STRUCTUREThe particular structure of the market in whichbanks operate may also influence bankprofitability in two main ways. The first isthat more market power, as proxied byconcentration measures, tends to be associatedin most industries with high levels ofprofitability, as firms collude to extractrents. The second explanation stresses theimportance of potential competition, whichdepends on the barriers to entry to variousbanking markets. This could imply that marketpower as proxied by concentration may notmatter as much as the threat of entry by newcompetitors.9

Work based on euro area micro data hasfound evidence for both of these hypotheses,albeit for particular banking products.10 Inthe present context where different types ofbanks and countries are being considered,market structure measures may be importantin explaining some of the cross-sectionalvariation in profitability across countries, as

well as the finding in the empirical literaturethat profits tend to be highly persistent.

MACRO-FINANCIAL FACTORSThe macroeconomic environment may alsoimpact on bank profitability through its effectson net income, on credit risk through therepayment abilities of borrowers, and on thevalue of collateral, all of which may vary withthe economic cycle.

For example, deteriorating macroeconomicfundamentals, possibly combined withdeclining asset prices, could cause loan lossesfor banks. These losses may induce banksto reduce lending, which in turn furtherexacerbates asset price declines, possiblyresulting in financial instability.11 Theincentives banks face over the business cyclemay also change. Banks could also be temptedto assume greater risks if their franchise orcharter value is threatened by loan lossesinitially caused by a macroeconomic downturn.

This suggests that macroeconomic variablesmay be important in this context. Indeed,empirical work on the causes of financialdistress has focused on identifying commonpatterns in macroeconomic variables before theonset of banking crisis episodes. In someinstances, this may take the form of a decline inthe GDP growth rate below its trend value,particularly if the macroeconomic downturn

7 For the US, see Berger (1995), op. cit., who also finds that thisrelationship breaks down for the period 1990-1992, possiblybecause banks overshot their optimal capital ratios. For Europe,see Goddard et al., (2004), op. cit.

8 See L. Laeven and G. Majnoni (2003), “Loan LossProvisioning: Too Much, Too Late?” Journal of FinancialIntermediation, Vol. 12.

9 In the euro area, individual bank behaviour and marketstructure may also have been affected by important regulatorychanges, such as the introduction of the First Banking Directivein 1993 and additional directives since then such as the LargeExposures Directive (92/121/EEC), the Capital AdequacyDirective (C152/6/EEC), and the Investment Services (93/22/EEC) Directive, which came into force in 1994, 1996 and 1996respectively.

10 See R. Gropp and S. Corvoisier (2002), “Bank Concentrationand Retail Interest Rates”, Journal of Banking and Finance,Vol. 26 (11).

11 See G. von Peter (2004), “Asset Prices and Banking Distress: AMacroeconomic Approach”, BIS Working Paper No 167.


was preceded by strong credit growthcombined with rapid growth in propertyprices.12

DATA AND METHODOLOGY

Euro area banks of differing ownership typesoperate and compete with each other in variousmarket segments across the euro area.Therefore, this analysis relies on theclassification provided in a private sectordataset which, in turn, is based on informationprovided by banks in their annual reports.13

This section provides an overview of thevarious measures of financial performance ofeuro area banks, before moving on to describethe empirical methodology and results.

MEASURES OF FINANCIAL PERFORMANCEVariations in accounting measures of financialperformance by ownership type may emergebetween the various types of banks for severalreasons. For example, commercial banksgenerate a higher return, given their focus onfor-profit activities, but possibly at the cost ofgreater variability in returns. Mutually ownedinstitutions or government-owned banks mayhave additional objectives to that ofmaximising profit, such as economic or socialdevelopment goals for specific geographicregions.14 One widely used measure ofperformance is ROE, which is defined in thisstudy as the post-tax net income a bank hasmade during a given year, divided by theaverage shareholder equity during that year.The advantage of using this measure is that itcaptures income that the bank generates fromtraditional intermediation activities and fromoff-balance sheet activities, such as tradingactivity, and the provision of risk managementsolutions to clients.15

Chart D.1 plots the average ROE for varioustypes of euro area financial institutions in thesample against its standard deviation. Bankholding companies and commercial banksshow nearly identical ROE over the sampleperiod, but with greater variability than otherownership types, indicating that higher return

Chart D.1 Euro area banks’ mean andstandard deviation ROE

(%)

Sources: Bureau van Dijk (Bankscope) and ECBcalculations.Note: The sample period covers the period 1993-2004. Themeans and standard deviations are calculated for each groupof banks across euro area Member States.

mortgage

cooperative bank

commercial bank bank holding company

saving

governmental

0

2

4

6

8

10

12

0

2

4

6

8

10

12

0 1 2 3 4standard deviation of ROE

mean ROE

12 See ECB (2005), “Indicators of f inancial distress in matureeconomies”, Financial Stability Review, June, pp. 126-131; andC. Borio and P. Lowe (2002), “Asset Prices, Financial andMonetary Stability: Exploring the Nexus”, BIS Working PaperNo 114.

13 The individual bank accounting information used in this featureas well as the type of bank is drawn from Bankscope, a privatesector database produced by Bureau van Dijk. Data for large euroarea banks were selected and subsidiaries operating in the euroarea were excluded, as were subsidiaries of foreign banksoperating in the euro area. Observations lying in the 1st and 99thpercentiles were discarded, as were institutions with implausiblevalues such as a loans-to-total assets ratio of greater than 100%.Data are deflated using the GDP deflator for each country.

14 It is worth noting that for the US, there is relatively littleagreement in the literature on whether ownership type mattersfor profitability. See J. Hughes, W. Lang, L. Mester, C. Moon andM. Pagano (2003), “Do Bankers Sacrifice Value to BuildEmpires? Managerial Incentives, Industry Consolidation, andFinancial Performance”, Journal of Banking and Finance,Volume 27 (3).

15 As a robustness check, profit before tax and country dummieswere used to ensure that differences in corporate tax were notdriving results. Both produced very similar results to thosepresented in this Special Feature. In addition, as the estimation iscarried out in differences, it should not be affected to any largeextent by different corporate taxes.

strategies are also associated with greater riskin returns. Moreover, the level and variabilityof profits experienced by these type of banksmay be influenced by the experience ofsubsidiaries within the group. By contrast,cooperative banks and government-ownedlending institutions show a much lowervariability in return but a mean return only halfof that recorded by bank holding companiesand commercial banks.

163cECB


IV SPECIALFEATURES

Three main reasons can be identified for thesedifferences. First, some banks may benefitfrom a diversification effect if their sources ofincome are not concentrated on one particularmarket. Second, some institutions may bemore efficient in terms of producing a givenamount of output at minimum cost ormaximising profit. Third, differences in thelevel and management of capital may lead todiffering financial performances.16

The degree to which a bank is diversified mayaffect its ability to generate revenues throughthe business cycle. More diversified banks maybe able to maintain consistent profitability overthe business cycle because they are not relianton any one particular market. However, asnoted in some of the banking literature,diversity may also increase exposure to morevolatile revenue sources without achieving anysignificant increase in profitability. Twomeasures of diversity are considered: onebased on income, the other on assets.

Income diversity attempts to gauge a bank’sreliance on income from traditionalintermediation versus more fee-based

Chart D.2 Euro area banks’ mean incomedivers ity by s ize group

Sources: Bureau van Dijk (Bankscope) and ECBcalculations.Note: The sample period covers the period 1993-2004.Income diversity is calculated as 1-absolute value[(net interest income minus non-interest income)/(totaloperating income)]. A score of 0 indicates no diversity.

0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

governmentalsavingmortgagebank holdingcompany

cooperativebank

commercialbank

≤1bn>1bn≤5bn>5bn≤15bn>15bn≤30bn>30bn

Chart D.3 Euro area banks’ mean assetdivers ity by s ize group

Sources: Bureau van Dijk (Bankscope) and ECBcalculations.Note: The sample period covers the period 1993-2004.Income diversity is calculated as 1-absolute value [(loansminus other earning assets)/(total earning assets)]. A scoreof 0 indicates no diversity.

0.00

0.25

0.50

0.75

1.00

0.00

0.25

0.50

0.75

1.00

governmentalsavingmortgagebank holdingcompany

cooperativebank

commercialbank


activities. Chart D.2 shows one diversitymeasure by size and type for euro area banks.17

According to this measure, commercial banksand bank holding companies are the mostdiversified, while government-owned andmortgage banks are the least diversified. It isnotable that, on the basis of this measure,cooperative banks seem to be nearly asdiversified as commercial banks, though this

16 It is possible that accounting ratios may also differ acrosscountries due to differences in national taxation policy, possibleearnings smoothing and variations in national accountingpractices that are particularly related to the treatment ofgoodwill. (See Special Feature E on the effects of IFRS in thisReview for further detail.) These differences are sometimescited as a reason to use only market-based indicators. Theseindeed have some advantages. They are forward-looking,available at a higher frequency, and reflect relevant informationon individual institutions. Market indicators based on equityprices, on the other hand, suffer from the drawback that equityinvestors may be willing to assume more risk and share in thebenef its of a bank’s management taking more risk thandepositors. A more practical problem is that market indicatorsare typically only available for large listed banks. Theavailability of market indicators is a particular problem for euroarea banks, as of the 300 or so banks used in this study, only 72were in 2004 quoted institutions, limiting the coverage of thesample across countries and overtime. Gropp, Vesala, andVulpes (2005), op. cit., use a sample of 84 EU15 banks.

17 These diversity measures are based on R. Levine and L. Laeven(2005), “Is There a Diversif ication Discount in FinancialConglomerates?”, University of Minnesota, mimeo.


conclusion depends both on the size and type ofinstitution.

Asset diversity looks at the specialisation ofthe institution in terms of the intermediationactivities it undertakes, and is based on balancesheet variables. For example, a high value ofasset diversity indicates a better balancebetween loans and other assets. Chart D.3suggests that asset diversity roughly increaseswith size. Commercial banks, bank holdingcompanies and cooperative banks seem allquite diversified according to this measure.

The productive efficiency of banks may alsoinfluence profitability. One proxy commonlyused to measure efficiency is the ratio ofoperating costs to income. As Chart D.4 shows,in some cases smaller institutions appear to beless efficient. However, no clear pattern can beidentified in terms of mean levels of efficiencyover the sample period.

Finally, capital management could affectprofitability, and the literature suggests apositive relationship, given that retained profits,after subtracting operating costs and provisions,can be added to banks’ reserves to boost capitalif they are not paid out in dividends to

shareholders or used to cover unexpected losses.Chart D.5 plots the mean ROE and mean equity-to-total asset ratio for euro area banks. The chartsuggests a slight positive relationship betweencapitalisation and profitability, although thisappears to vary across bank type.

ECONOMETRIC ANALYSIS

The estimation method adopted in this study isthat of a dynamic panel data model. Panel datamodels combine a cross-section component(many banks observed at one point in time)with a time dimension (the same banksobserved over different years). The cross-section nature of the panel controls for bank-specific factors and how these vary acrossbanks. The addition of a time dimensionallows other external factors – such as marketstructure and macroeconomic developments –potentially to impact on bank profitability.

A dynamic panel model builds on this byincluding a lag of the dependant variable as anadditional right-hand-side variable. This hasthe advantage of allowing short-run dynamicsto be explored. The main hypothesis to betested is that that ROE is related to bank-specific characteristics such as lagged ROE

Chart D.4 Euro area banks’ mean costincome ratio by s ize group

Sources: Bureau van Dijk (Bankscope) and ECB calculations.Note: The sample period covers the period 1993-2004.


bank holdingcompany

commercialbank

cooperativebank

mortgagesavinggovernmental

90

80

70

60

50

40

30

20

10

0

90

80

70

60

50

40

30

20

10

0

Chart D.5 Euro area banks’ mean return onequity and capital rat ios

Sources: Bureau van Dijk (Bankscope) and ECB calculations.Note: The sample period covers 1993-2004.

mortgage

cooperative bank

commercial bank bank holding company

saving

governmental

0

2

4

6

8

10

12

0

2

4

6

8

10

12

0 1 2 3 4 5 6 7mean equity (% of total assets)

mean ROE

165cECB


IV SPECIALFEATURES

(+), size (+), capital (+), off-balance sheetitems (+), provisions (-), and diversitymeasures (+ or -).

All these variables are treated as endogenous inthe estimations, taking into account thepotential relationship between the independentvariables and the error term. Bankspecialisation, market structure characteristics(concentration, Herfindahl index, (both +), andmacroeconomic variables (real GDP growth,real property prices (both +)) are treated asexogenous in the estimations.

The estimation period covers the period 1993-2004, using an unbalanced panel of databased on 329 banks with a minimum of fiveyears of consecutive data. Given the sign andsignificance of lagged profits, the results fromthe baseline model suggest first of all that thechange in profits is persistent (see Table D.1).18

This is a common finding based on the resultsof previous studies. On the basis of thesample considered, the change in profitabilityis also influenced by capital. This may be dueto retained profits added back to capital;alternatively, well-capitalised banks may beable to pursue a wider range of business,including off-balance sheet business, owing totheir higher creditworthiness.

Loan loss provisions have the expected(negative) sign, but the coefficient is

18 After taking lags and differencing, this left approximately1,400 bank-year observations available for estimation,depending on the independent variables used. The one stepstandard errors are used for inference. The estimations werecarried out using the Arellano and Bond GMM estimator. Theone step standard errors are used for inference. The diagnostictests referred to in Table 1 are the Sargan test for the validity ofover-identifying restrictions, and tests for f irst and secondorder autocorrelation. For more details, see M. Arellano and S.Bond (1991), “Some Tests of Specif ication for Panel Data:Monte Carlo Evidence and an Application to EmploymentEquations”, Review of Economic Studies, 58.

19 As an additional robustness check on the estimations, thesystem GMM estimator of Blundell and Bond (1998) was used;the results were similar to those presented here. For moredetails, see R. Blundell and S. Bond (1998), “Initial Conditionsand Moment Restrictions in Dynamic Panel Models”, Journalof Econometrics, 87; and R. Blundell and S. Bond (1999),“GMM Estimation with Persistent Panel Data: An Applicationto Production Functions”, IFS Working paper 99/4.

Table D.1 Empir ical results

Source: ECB calculations.Note: The diagnostic tests’ “yes” refers to non-rejection of both the Sargan test for the validity of over-identifying restrictions andtests for second order autocorrelation. Signif icance denotes results that are significant at a 10% level or lower.

variable sign significance sign significance sign significance

lagged profit + yes + yes + nocapital + yes + yes + yesoff-balance sheet + no + no + nosize + yes + yes + yesreal GDP growth + yes + yes + yesloan loss provisions - no - nospecialisation dummies + noconcentration + no

diagnostic tests yes yes yes

insignificant. The positive and significantrole for real GDP growth tends to confirm theview that macroeconomic developments areimportant for bank profitability. For example,the cyclical effect of real GDP growth could beoverwhelming the provisioning cycle. Finally,the change in the size variable also has apositive effect, suggesting that profitabilityis positively related to an increase in theinflation-adjusted size of a bank’s balancesheet.19

Given that the estimation method takes intoaccount bank-specific differences, it is notsurprising that the variables measuringbanks’ specialisation are insignificant. Finally,concentration measures – such as the ratio of


variable sign significance sign significance sign significance

lagged profit + no + no + yescapital + no + yes + yesoff-balance sheet . . . . . .size + yes + yes + yesreal GDP growth + yes + yes . .income diversity - no - - . .asset diversity - noreal property prices - no

diagnostic tests yes yes yes

Table D.2 Empir ical results

Source: ECB calculations.Note: The diagnostic tests’ “yes” refers to non-rejection of both the Sargan test for the validity of over-identifying restrictions, andtests for second order autocorrelation. Signif icance denotes results that are signif icant at a 10% level or lower.

the five largest banks’ assets to the assets ofeach country’s banking system and theHerfindahl index – were positively signed butinsignificant. Given that market structurechanges only slowly over time, the lack ofvariation in these variables within countriesand over time is probably the reason for itsstatistical insignificance. In this case, areduced sample owing to the unavailability ofmarket structure indicators before 1997 mayalso be a contributing factor. While thesemeasures are standard indicators of marketstructure, it cannot be ruled out that alternativemeasures may reveal a different relationship.Investigation of this topic is beyond the scopeof this Special Feature. The inclusion of thevariable leads to the lagged profit variablebecoming marginally insignificant, thuspointing towards some relationship betweenmarket structure and profitability.

Alternative specifications were tried asrobustness checks. Two main types of checkswere carried out: ones based on bank-specificfactors, and ones based on macroeconomicfactors.

The inclusion of asset and income diversitymeasures weakens the significance of thelagged profit variable, perhaps indicating thatthe previous findings of profit persistence maybe due in part to an omitted variables

problem.20 Although the data were screenedcarefully before estimation, idiosyncraticevents relating to certain banks could havedriven the results. To check this, dummyvariables based on data from Gropp et al.(2004) were therefore used to control for this.The results were unchanged. To control for thepossibility that mergers or takeovers could beresponsible for the role that growth in sizeappears to play, dummy variables wereconstructed for banks that were involved inM&As. Their inclusion did not however affectthe results, probably because only a smallnumber of observations in the sample wereaffected.

One country in the sample experienced abanking crisis at the beginning of the sample,and an interaction dummy for the country andreal GDP growth was used in the estimation.While the dummy was significant, the overalleffect on the results was similar to thosereported in Column 3 of Table D.2 below.Finally, an experiment was carried out toreplace real GDP growth with real residentialproperty prices; however, the variable provedto be insignificant.

20 Additional instrument lags for the independent variables werealso used in this instance. The lagged prof it variable continuedto remain insignif icant.

167cECB


IV SPECIALFEATURES

CONCLUDING REMARKS

This Special Feature set out to review variousfactors identified in the literature that mayaffect bank profitability. Based on micro data,stylised facts concerning euro area bankprofitability were presented. Finally, aneconometric analysis based on a dynamicpanel data approach was carried out to identifyfactors that could influence bank profitabilityin the euro area.

Both macroeconomic and bank-specificfactors appear to have a role to play, with realGDP growth and bank size being the mostimportant determinants. A positive but weakerrelationship was found between bank equitycapital and profits. It is important to note thatthe estimation method takes into accountthe potentially endogenous nature of therelationship between lagged profits andcapital. On the other hand, the regressionmodel is a reduced form model and not derivedfrom a structural economic model. This meansthat it is difficult to identify the exact nature ofthe links between size, capital and profitabilitybased on the current approach.

Overall, for the purposes of financial stabilitymonitoring, the results point towards a need toanalyse and understand better the interplaybetween bank-specific factors and themacroeconomic environment before any firmpolicy conclusions can be drawn. Additionalwork in two particular areas could provevaluable in this regard. First, the relationshipbetween size and profitability could beanalysed further in order to determinewhether this operates via economies of scaleand scope. Second, further analysis of theempirical effects of income and assetdiversification on bank profitability couldimprove understanding of the overall effect ofgrowth in non-interest income on bankprofitability and stability.


E MAIN EFFECTS FROM THE NEWACCOUNTING FRAMEWORK ON BANKS

The EU Regulation requiring all listedcompanies, including banks, to prepareconsolidated financial statements inaccordance with International FinancialReporting Standards (IFRS) has been apositive development that will increase thetransparency and comparability of financialstatements in the EU. However, the first-timeapplication of these new rules will have asignificant impact on financial statementswhich should be taken into account whenanalysing the accounting figures. The aim ofthis Special Feature is to provide a briefoverview of the main ways in which IFRS willaffect banks’ primary financial statements.

INTRODUCTION

Regulation 1606/2002/EC1 concerning theapplication of IFRS2 was adopted by theEuropean Parliament and the EU Council on19 July 2002. According to this Regulation,for each financial year starting on or after1 January 2005, all companies listed in the EU,including banks, are required to comply withthe accounting and financial reportingstandards issued by the InternationalAccounting Standards Board (IASB) withregard to their consolidated accounts. In orderto become effective in the EU, each individualIASB standard is required to be endorsed by theEU Commission. At this juncture, however, itshould be noted that the standard concerningthe recognition and measurement of financialinstruments (IAS 39), which is extremelyimportant to the banking industry, has onlybeen partially endorsed and, in that context,certain hedging provisions have been carvedout. It should in addition be noted that theRegulation also contains the option forMember States to extend the application ofIFRS to financial statements of individualfirms and to unlisted companies.

The purpose of this Special Feature is tohighlight the likely main effects from the

introduction of the new accounting rules forbanks. Changes in the new EU accountingframework may potentially affect theindicators used in the Financial StabilityReview in two main ways. First, one-off effectscould arise in the financial statements owing tothe transition from national accountingprinciples to the new framework. Such one-offeffects would affect the components of themacro-prudential indicators. This sort ofchange in the indicators is independent ofunderlying changes in the stability of thefinancial system. Second, following theintroduction of the new framework, balancesheet items are likely to display different timeseries behaviour compared to that under thecurrent national rules. Such differences couldfor instance materialise in higher or lowervolatility or in altered sensitivity of accountingfigures to market factors such as changes ininterest rates or prices for shares.Consequently, changes in the prudentialindicators need to be interpreted with careduring the transition phase.

The aggregate impact of the changesintroduced by IFRS on the banking sector as awhole is, however, impossible to assess exante. Indeed, the overall impact very muchdepends on the composition and structure ofeach bank’s balance sheet. The accountingrules currently in place also play an importantrole, as some national standards are quitesimilar to IFRS, while others significantlydiffer. Finally, the overall impact also dependson the accounting practices of individual firmsand their use of different options incorporatedinto the accounting rules.

Notwithstanding the above-mentioneddifficulties in assessing the impact of theintroduction of the new accounting framework,this Special Feature tries to identify the mainchanges in the accounting rules which are

1 Generally known as the “IAS Regulation”.2 International accounting standards issued by the London-based

International Accounting Standards Board (IASB).

169cECB


IV SPECIALFEATURES

relevant for banks and may have a significantimpact on their balance sheets and incomestatements.

It is important to note that the effects of the newaccounting framework will not materialise toan equal extent in all the various financialindicators or reporting. Indeed, prudentialreporting based on individual accounts wouldnot necessarily3 be affected by the new rules,contrary to consolidated prudential ratiosbased on consolidated accounting figures.Furthermore, some effects in the calculation ofregulatory capital will be mitigated throughinternationally agreed “prudential filters” (seesection below concerning the impact onregulatory capital). Hence, indicatorsconcerning regulatory capital (e.g. solvencyratios) which are associated with prudentialfilters may be affected by the new accountingframework in a different way than indicatorsthat rely on the accounting definition of equitycapital (e.g. ROE).

This Special Feature focuses upon thefollowing areas: (i) reclassification ofinstruments as debt or equity; (ii) accountingfor business combinations; (iii) valuation offinancial instruments, which also includes therecognition of derivatives, available-for-salesecurities, hedging provisions, the fair valueoption, share-based payments and allowancesfor credit losses and own credit risk;(iv) measurement of post-employmentbenefits; (v) de-recognition of special purposeentities; (vi) dividend adjustment; and(vii) software and other intangibles.

Finally, the article ends with a brief discussionon the prudential filters used by bankingsupervisors with the aim of safeguarding thedefinition and quality of regulatory capital.

RECLASSIFICATION OF DEBT AND EQUITYINSTRUMENTS

The definition of debt and equity classificationprinciples applicable to capital instrumentsdiffer under IFRS compared to most national

accounting rules. According to IFRS, issuedinstruments are classified as liabilities whenthe issuer has a present obligation to delivercash or another financial asset to the holder ofthe instrument.

Hence, the introduction of IFRS has entailedthe reclassification of certain debt and equityinstruments. For example, certain capitalinstruments such as preference shares that werepreviously treated and recognised as equitywill now need to be reclassified as liabilities.IFRS distinguish equity from non-equitypreference shares on the basis of whether thedividends paid out on the share are mandatoryor discretionary. In the former case, preferenceshares are required to be reclassified asliabilities, which will result in a negativeadjustment in equity. It should be noted thatnon-equity minority interest, which isrecognised under equity on the balance sheet,may also need to be reclassified into liabilities.This may have a potentially negative impact onnet income, as the reclassification results willaffect the interest expense.

Conversely, certain instruments (e.g. reservecapital instruments) previously recognised asliabilities will be reclassified as minorityinterest, which is presented withinshareholders’ equity on the balance sheet. Thisreclassification will result in a decrease ininterest expense.

The overall impact of the reclassification ofdebt and equity instruments on equity and onnet income greatly depends on the specificcomposition of individual banks’ balancesheets.

It should be noted that for prudential purposes,the current definition of own funds will bemaintained, and hence the potentially differentaccounting classifications of debt and equityinstruments will not affect regulatory capital.

3 In cases where f irms are required or provided with an option touse IFRS for their individual accounts, prudential reportingbased on individual accounts could be similarly affected byIFRS.


BUSINESS COMBINATIONS

When combining businesses through theacquisition of another business, the acquirertypically pays a price that differs from the netbook value of the assets and liabilities of thisbusiness. This difference, which is typically apositive amount, is referred to as goodwill.Prior to the introduction of IFRS, most nationalstandards required such goodwill to beamortised according to a predeterminedschedule or for it to be fully written offimmediately after the acquisition. Hence, thetreatment of goodwill differs across entitieswithin a jurisdiction.

Two main observations may be made fromcomparison of the new framework with currentnational rules. First, the measurement ofgoodwill deserves attention. Under IFRS 3, thecurrently applicable standard for businesscombinations under the new framework,goodwill is measured by allocating the cost ofthe acquisition to the fair values of identifiableassets and liabilities of the acquired business.The excess of costs then constitutes thegoodwill. This can differ from current nationalrules, where merger accounting4 may implyzero goodwill, where more flexibility may beavailable in allocating the cost of theacquisition to balance sheet assets, and wheregoodwill might not reflect fair values.

Second, under IFRS, goodwill is recognised asan asset that must not be regularly amortised.Rather, it needs to be tested for impairment, i.e.it is regularly tested to establish whether thepresent value of the business units still justifiesthe reported goodwill. If not, an impairmentloss is recognised that cannot be recoveredlater. By contrast, negative goodwill at the timeof the acquisition is immediately recognised inprofit and loss.

The ongoing effect of IFRS 3 is that income-based financial indicators are not necessarilyweakened following an acquisition or a merger.Furthermore, the ongoing amortisation of pastacquisitions will cease, improving earnings in

principle. In the long run, it is however unclearwhether the impairment tests will effectivelylead to a quicker writing off of the goodwillthan regular amortisations, although in a lesssmooth fashion. If the economics of a mergerhave been overestimated and the acquisitionwas overpriced, this may become apparentbefore the end of the regular amortisationschedule, thus triggering a full write-off of thegoodwill. It is moreover safe to assume thatchanges in goodwill and related effects onequity and income will in the future occur in amore discrete, volatile fashion and will, in linewith the expected returns from the acquired ormerged unit, behave cyclically. It is likely thatduring a recession, goodwill positions willevaporate faster from banks’ balance sheetsthan under the current long-term amortisationschedules.

An interesting feature of possible one-offeffects is that in principle, IFRS 3 accountingcan be, but does not have to be, applied to pastacquisitions, which means that the originalgoodwill could be reactivated, undoing pastamortisation through profit and loss withcorresponding effects on income and, mostnotably, on book equity.

Goodwill accounting clearly influencesindicators that are based on equity and assets,and changes in the amortisation of goodwillinfluence earnings. However, regulatory ownfunds will not be affected because thedefinition of own funds excludes intangibleassets. Furthermore, goodwill is not risk-weighted and thus does not affect thenumerator of the solvency ratio. Consequently,the solvency ratio remains unaffected bychanges in goodwill accounting. For instance, aone-off increase in reported goodwill at thefirst-time application of IFRS would increaseshareholders’ equity; however, for the ownfunds, the effect is eliminated because of anincrease in the deductions from the own fundsof the same amount.

4 This is also referred to as pooling of interest accounting.

171cECB


IV SPECIALFEATURES

FINANCIAL INSTRUMENTS

DERIVATIVESIFRS require all derivatives to be recognised onthe balance sheet and measured at fair value.Gains and losses from changes in the fair valuewill flow through the income statement, withthe exception of derivatives qualified forhedging (see section on hedging provisions).Prior to the adoption of IFRS, derivatives heldfor trading were already valued at fair value inmany European countries and recognised onthe balance sheet. However, this is a newfeature for derivatives recognised in thebanking book, as these were formerly onlyregistered off-balance sheet at cost.

Additionally, derivatives that are embedded inhybrid financial instruments, but which haveeconomic characteristics and risks that are notclosely related to the economic characteristicsof the underlying financial instrument, arerequired to be separated from the hybridinstrument, to be valued at fair value andrecognised on the balance sheet on a stand-alone basis.

The recognition of derivatives at fair value willresult in an increase in the overall size of thebalance sheet. Furthermore, changes in the fairvalue will cause additional volatility in theincome statement and, therefore, also in equity.However, it should be stressed that therecognition of derivatives on the balance sheetat fair value, as opposed to the current situationof generally simply being registered off-balance sheet, can be considered a significantstep forward for users of financial statements,as it increases understanding of the underlyingrisks incurred by banks which may be dealingwith and exposed to derivatives transactions ona large scale.

AVAILABLE-FOR-SALE SECURITIESUnder IFRS, available-for-sale (AFS)securities are required to be recorded at fairvalue. Under certain national accounting rules,AFS securities could include long-terminvestments which were carried at cost, less

provisions for impairment. For these assets, theintroduction of IFRS seems, in most cases andat this juncture, to result in an increase in valuefrom the recognition of these assets at fairvalue.5 The increase from cost to fair value willbe recognised in a specific reserve ofshareholders’ equity. Hence, this increased useof fair value for AFS securities may potentiallyresult in an increase in the overall asset size ofthe balance sheet and in an increase in thevolatility of equity.

The accounting treatment of AFS securities hasindeed prompted banking supervisors todevelop a prudential filter to neutralise theeffect on regulatory capital (see section on theimpact on regulatory capital).

HEDGING PROVISIONSHedge accounting rules allow the hedging itemto follow the accounting treatment of thehedged item, which is generally known asaccruals accounting. Under this treatment, thegain or loss on the hedging instrument isrecognised in the income statement when theoffsetting gain or loss on the hedged instrumentis recognised. Hence, given the possibility todefer or anticipate income recognition, strictrequirements need to be complied with inorder to qualify for hedge accounting so as toprevent discretionary income manipulation bymanagement. To qualify for hedge accounting,IFRS require, inter alia, specific identificationand documentation of the hedging and hedgedinstruments, identification of the risk beinghedged, and effectiveness testing of the hedgeitself. IFRS also allow macro-hedging (on aportfolio basis) in a fair value hedge for interestrate risk.

IFRS distinguish between two main types ofhedges: cash-flow hedges and fair valuehedges. Cash-flow hedges aim to cover the riskof variability of future cash flows (e.g. variablerate financial instruments), and the valuation of

5 Fair values can be higher or lower than amortised cost,depending on the changes in interest rates and the amount thatthe asset was recognised as on the balance sheet.


the hedging derivative is recognised at fairvalue in shareholders’ equity. As the gains andlosses from the changes in the fair value of thehedged instrument are recognised in theincome statement, the fair value of the hedginginstrument recognised in equity is adjusted andthe corresponding gains and losses are“recycled” through the income statement. Itshould be noted that the accounting treatmentof cash-flow hedges was also subject to aprudential filter to safeguard the quality ofregulatory capital (see section on the impact onregulatory capital).

Fair value hedges are designed to coverchanges in the price of a financial instrument.This can be accomplished by hedging thetransaction (micro hedging) or on a portfoliobasis (macro hedging). Under micro fair valuehedging, changes in the fair value of thederivative and changes in the fair value of thehedged item are recognised in the incomestatement symmetrically. For macro hedging,the change in the fair value of the hedged itemis recognised in the balance sheet on a separateline item.

Hedge accounting rules were in place prior tothe introduction of IFRS. However, the IFRScriteria seem to be tighter than existingnational hedging rules. Therefore, someexisting hedging relationships may fail tocomply with the IFRS hedging criteria, andthus will no longer qualify for hedgeaccounting, which may subsequently result inartificial volatility in net income.

FAIR VALUE OPTIONThe new accounting rules introduce thepossibility to designate irrevocably atinception a financial asset or financial liabilityas at fair value through profit and loss – the so-called fair value option. However, this optionmay only be applied if: (i) it eliminates orsignificantly reduces a measurement orrecognition inconsistency (sometimes referredto as an accounting mismatch), or (ii) when agroup of financial assets, financial liabilities or

both is managed and its performance isevaluated on a fair value basis in accordancewith a documented risk management strategy.6

Although the introduction of this option mayincrease the use of fair value, which couldpotentially entail additional volatility in netincome from the changes in the fair value, italso allows the elimination of an accountingmismatch of an economically hedged position(thus reducing “artificial” volatility).

SHARE-BASED PAYMENTSUnder IFRS, banks are required to recognise intheir income statements the fair value of shareoptions and other share-based paymentsawarded by banks to their employees andexecutives. Under current rules there is no suchrequirement, and such share-based paymentsare kept off-balance sheet. This expense willhave a one-off negative impact on net income.

ALLOWANCES FOR CREDIT LOSSESLoans are traditionally recognised on thebalance sheet at cost. Banks have somediscretionary leeway in classifying certainloans as doubtful or non-performing, and incalculating the related provision for loanlosses, which could be either general orspecific. Specific provisions cover losses onindividual or on a portfolio of loans which havebeen specifically identified as impaired or non-performing. The nature of general provisionsvaries significantly across Member States,from statistically supported allowances forlosses inherent in the loan portfolio to countryrisk reserves or reserves for general bankingrisks (which are not associated with animpairment).

Under IFRS, banks will be required to assess ateach balance sheet date whether there isobjective evidence that the loan or group ofloans is impaired. An impairment loss should

6 See IAS 39 “Financial Instruments: Recognition andMeasurement”, Amendment to IAS 39 for Fair Value Option,June 2005.

173cECB


IV SPECIALFEATURES

be recorded when it is probable that the bankwill not receive the payment of interest andprincipal according to the original contractualterms. The amount of the loss is the differencebetween the carrying amount and the netpresent value of expected future cash flowsdiscounted at the loan’s original interest rate.This loss, recognised as an allowance for loanlosses, will flow through profit and loss. Inaddition, for collateralised loans, banks willneed to recognise collateral at fair value.

The recording of impairment losses willincrease the potential pro-cyclical effects onbanks’ profit and loss. Furthermore, these newrules may result in a reduction in the overalllevel of allowances for credit losses, as bankswill only be allowed to create reserves7 tocover losses which have been incurred. Thisneed not, however, lead to insufficientprovisions. The new rules require a two-stepassessment of incurred credit losses, wherebyloans are assessed individually andcollectively. An individual loan which hasbeen determined as not impaired is included ina group of loans with similar credit riskcharacteristics for collective impairmentassessment. The process considers all creditexposures, not only those of low credit quality.Where observable data are limited, the newrules require the use of experienced judgementin the estimation process. A potential reductionin the allowance expense for credit losses willhave a positive impact on net income.

POST-EMPLOYMENT BENEFITS

Pensions are the most significant position inthe category of post-employment benefitsunder the IFRS. In this context, a pension planasset or liability is recognised on the balancesheet only if the employer bears the investmentand actuarial risks of the pension plan. If this isthe case (e.g. in the case of defined benefitschemes), either a net asset or a net liability isreported, based in principle on the differencebetween the fair value of pension plan assetsand the actuarial, discounted present value offuture pensions.

The one-off effects of the first-time applicationof IFRS can basically materialise in threedifferent respects. First, pension plan assetsand liabilities may currently not be recognisedon the balance sheet at all, even though theplan is of a defined benefit nature. In theseinstances, the one-off expense and theadditional liability appearing on the balancesheet will obviously be significant andwill strongly influence both income-basedindicators and capital ratios. However, evenif currently applicable rules require therecognition of pension liabilities, it is likelythat the amount of the liabilities will increaseunder IFRS, given that the actuarial factors tobe accounted for are rather extensive comparedto the often less binding or limited guidancegiven by national rules. Second, themeasurement of the pension plan assets may incertain cases produce precisely the oppositeeffect. Where pension plan assets were notdesignated as such in the past under nationalaccounting rules and did not offset pensionliabilities, the effect may be that the netliability under IFRS is actually lower than theliability currently reported, implying a positiveeffect on income and equity in the transition toIFRS. Third, a similar situation may also arisewhen there are considerable hidden reserves inpension plan assets that would be disclosed andnetted against pension plan liabilities uponfirst-time application.

With regard to ongoing effects, it can beassumed that the measurement of pension planassets at fair value and the relatively stringentactuarial methods in connection with therequirement to update regularly thecalculations will tend to lead to increasedvolatility of profit and loss compared tomeasurement under most of the currentnational rules. However, it should be noted thatif actuarial gains or losses exceed certainthresholds, IFRS allow them to be spread atmaximum over the average remaining serviceperiod of the employees, which would

7 The term “reserve” is meant as a provision or allowance in thecontext of loan impairment.


contribute to smoothing their impact on netpension assets or liability and on profit andloss. There is also the possibility that firmsmight try to avoid these effects on their balancesheets in the medium term by increasinglyopting for defined contribution pensionschemes.

SPECIAL PURPOSE ENTITIES

The main issue in this context is whether theassets and liabilities of a special purpose entity(SPE) should be included in the individualaccounts of a bank, and whether the SPE needsto be consolidated. SPEs are of particularrelevance in the banking sector because theyare used as a conduit for securitisations ofbanks’ assets such as credit portfolios (in suchcases, the bank sells the assets to an SPE thathas issued securities, and pays those assetswith the proceeds from this issuance). Whileideally such a transaction, also referred to as a“true sale”, would insulate the bank from therisks and returns of those assets and would thusjustify their de-recognition from the bank’sbalance sheet, there are various issues thatimply continued risks for the selling bank, suchas retained tranches of the securitisation orimplicit support for the SPE. Other potentialuses of SPEs for banks include the selling ofnon-core activities such as real estate holdings.

IFRS contain specific provisions onsecuritisation. The possibility that securitisedassets may be de-recognised from the bank’sbalance sheet requires a case-by-case analysis.As a first step, it needs to be analysed whetherthe bank bears the risks and returns of theassets. The level of control the bank has overthe SPE also has to be assessed. Consolidationas opposed to inclusion in individual accountsis, by contrast, required if the bank controls theSPE, i.e. when it has for instance the ability toappoint its management or issue orders to theSPE.

In some cases IFRS appear somewhat moreconcrete and binding than some currentnational accounting principles (for instance the

German and French). Others (for instance theDutch and British) are more similar to IFRS.Consequently, it is likely that in the first cases(but rather not in the latter) that some SPEs willneed to be included in the individual and/orconsolidated accounts for the first time. Thisassessment is obviously a static one in thesense that IFRS also allow SPEs in principle tobe structured in a way that they do not have tobe included. Consequently, some reportingentities may choose to restructure transactionsrather than to recognise them on their balancesheets.

In those cases where existing, off-balance sheetSPEs need to be included on the balance sheet,there will be a one-off increase in assets andliabilities and a consequent change in return ontotal assets and in the debt-to-equity ratio.Solvency ratios, by contrast, would remainunaffected as the capital treatment of thesecuritised assets and the retained tranchesdepends on the respective prudential rules,irrespective of the accounting treatment.

DIVIDEND ADJUSTMENT

Under national accounting rules, dividends arerecognised as soon as they are declared.However, under IFRS dividends are onlyrecognised later when approved and not wheninitially declared. This results in a positiveadjustment in equity for year-end accounts.However, this adjustment is temporary, giventhat it will be corrected for in interim accounts,when the declared dividends are effectivelyapproved for distribution. This adjustment isparticularly large for some countries.

SOFTWARE DEVELOPMENT COSTS AND OTHERINTANGIBLES

According to most current national accountingrules, banks have the option either to expenseor to capitalise certain software developmentcosts. Under IFRS, these internally developedsoftware and other intangible assets can becapitalised and amortised, but only if certainconditions are met. Therefore, for banks which

175cECB


IV SPECIALFEATURES

8 www.c-ebs.org and www.bis.org/bcbs/index.htm.

had previously chosen to expense theirsoftware development costs, the transition toIFRS and the retroactive application of this rulewould imply an increase in the asset size of thebalance sheet from the capitalisation of thesecosts and an increase in equity from the relatedpositive adjustment.

However, the annual amortisation of thesecosts, which are normally amortised over ashort period of time, will subsequently, at year-end, have a negative effect on the incomestatement.

PRUDENTIAL FILTERS

The impact of the application of the newaccounting standards may in certain cases besignificant on equity and on the incomestatement. Given that these accounting figuresare normally used as the basis for prudentialreporting, banking supervisors deemed itnecessary to develop some prudential filters.

Prudential filters are designed to maintain thecurrent definition and quality of regulatorycapital. It should be noted that with theobjective of maintaining a level playing-fieldacross the EU and G10 countries, the prudentialfilters proposed by the Committee of EuropeanBanking Supervisors are consistent withthose of the Basel Committee on BankingSupervision.

A brief description of some of these prudentialfilters is provided below; a detailed descriptionof all the prudential filters developed by theabove-mentioned committees can be found ontheir respective websites.8

OWN CREDIT RISKBanking supervisors advise the exclusion fromregulatory capital of any cumulative unrealisedgains and losses arising from changes in aninstitution’s own credit standing – so-calledown credit risk – as a result of the potentialfuture application to liabilities of the fair valueoption. When issued liabilities are recognisedat fair value in a bank’s balance sheet, a

deterioration in the bank’s credit quality leadsto an increase in the discount, which results in areduction in the value of the liabilities and inturn to the recognition of an accounting gain.Conversely, an improvement in the bank’screditworthiness leads to an increase in the fairvalue of the liabilities (discounted at a lowerrate), which results in the recognition of anaccounting loss. Banking supervisors advisethat these gains and losses should be extricatedfrom regulatory capital.

CASH-FLOW HEDGESIt is recommended that fair value reservesrelated to cash-flow hedges of financialinstruments measured at amortised cost shouldnot be included in regulatory capital, given thatthis fair value reserve will be subsequentlyadjusted and the related gains and lossesrecognised through profit and loss.

AVAILABLE-FOR-SALE PORTFOLIOThe AFS portfolio comprises equities, loansand receivables and other financialinstruments. For equities, unrealised lossesshould be deducted from regulatory capital(more specifically from tier one), whileunrealised gains should only partially beincluded (in tier two). For loans andreceivables, unrealised gains and losses – apartfrom those related to impairment – are notrecognised in regulatory capital. Other AFSassets are either treated as equities or as loansand receivables.

CONCLUDING REMARKS

To sum up, the changeover from nationalaccounting rules to IFRS may raise issues ofinterpretation or comparison in the near future.This is particularly true at the EU-wide macrolevel, given that the concrete nature and size ofthe effects from the transition to IFRS willdepend on both the pre-existing national rulesin each Member State and the current practicesand specific features of individual firms thathave been applying the national rules. From a


financial stability perspective, however, suchissues that may arise during the initial phase ofthe changeover to IFRS are only temporary innature and by no means outweigh the long-termbenefits of an accounting regime which is bothmore harmonised and better reflects theunderlying risks that an individual firm isexposed to. The application of the newaccounting rules across individual institutionsin different Member States clearly benefitscross-country comparisons and aggregation,which in turn results in cross-country macro-prudential indicators that are more meaningfulin the longer term.

177cECB


IV SPECIALFEATURES

F CENTRAL COUNTERPARTY CLEARINGHOUSES AND FINANCIAL STABILITY

Central counterparty clearing houses (CCPs)play an important role in efficientlyreallocating counterparty credit risks andliquidity risks in financial markets. However,as systemically important players, they mustmanage their risks in an adequate way in orderto avoid creating new risks for financialstability.

INTRODUCTION

In financial markets, the clearing oftransactions involves the calculation, usuallyon a net basis, of the obligations of marketparticipants that result from their tradingactivities. Clearing takes place after thematching of buy and sell orders and prior to thelegal fulfilment of the respective obligation. Inmany markets, clearing is performed by a CCP,in which case the CCP interposes itselfbetween the original buyer and seller, acting asthe buyer to each seller and the seller to eachbuyer. In recent years, CCPs have been playingan increasingly important role in the clearing oftransactions in financial markets. In particular,against a background of rising trading volumes,derivatives and repo markets have becomeheavily reliant on CCPs for the clearing oftransactions. In addition, CCPs have beenincreasingly serving outright securitiesmarkets, including OTC markets. In manymajor markets, traders are obliged to use a CCPto clear all of their trades, either as direct orindirect participants of the CCP.

CCPs can play an important role in thefunctioning of financial markets, as they havethe potential to reduce the counterparty creditrisks that financial market participants facewhen they enter into transactions. In addition,they can contribute to improving efficiency infinancial markets by providing multilateralnetting of trades and by facilitating anonymoustrading. However, because a CCP alsoconcentrates risks, significant disruptions inthe financial markets that they serve could arise

if the risk management procedures they have inplace prove inadequate. Thus, a CCP’s riskmanagement procedures play a crucial role insafeguarding financial stability.

This Special Feature discusses the ways inwhich the core functions of CCPs cancontribute to financial stability. It alsodescribes the risks that CCPs are exposed to,and what CCPs can, or should, do to managesuch risks appropriately.

COUNTERPARTY CREDIT RISK AND LIQUIDITYRISK

Transactions in the financial markets involve atrading phase and one or more settlementphases. The trading phase is the moment whentwo parties conclude an agreement. In anoutright securities transaction, for example,the parties agree to exchange securities forfunds typically within one or two days. In thecase of derivatives transactions, for example afutures contract, the parties will agree toexchange the underlying security for funds at alater (expiry) date. And in the case of arepurchase agreement, the parties agree toexchange the underlying security for fundswithin one or two days and to redeliver theunderlying security at a later date.

The settlement phases of a transaction are whenobligations from the trading phase are fulfilled,i.e. when assets are exchanged for funds and– in the case of a repurchase agreement –redelivered when due. Outright transactionsare characterised by a single settlement phase,while for example repurchase transactionshave two settlement phases, first as assets aredelivered, and then as they are redelivered lateron.

There is a time-lag between the trading and thesettlement phases in particular for derivativesand repurchase transactions, and even in thecase of outright transactions. This time-lagappears to be the main reason why the twoparties in a transaction are exposed tocounterparty credit risk and to liquidity risk.


Counterparty credit risk is the risk that oneparty in a transaction is unable to fulfil itsobligations, typically as a consequence ofinsolvency between the trading and thesettlement phase. Liquidity risk is the risk thatthe trading party cannot fulfil its obligationswhen due, but only with a delay, for examplebecause of operational problems.

Nowadays, the settlement of cash markettransactions typically takes place in DVPmode; i.e. when assets are to be exchanged forfunds, the assets are delivered if and only if thefunds are delivered. Accordingly, the risk thatthe non-defaulting trading party delivers to adefaulting party while the defaulting party doesnot deliver to the non-defaulting party (so-called principal risk) should be negligible, sothat the non-defaulting party should not losethe full principal value of the assets or fundsdelivered.

However, counterparty credit and liquidityrisks can still imply significant losses for thenon-defaulting party. For example, if the non-defaulting party urgently needs the assets thatthe defaulting party failed to deliver, it has toreplace the failed trade by a new one. The priceof the new trade can however be less favourablethan that of the failed trade.

Counterparty credit risk and liquidity risk canpose risks for financial stability, especiallythrough a domino effect. For instance, supposethat two parties, A and B, conclude a trade andthat A fails to deliver. B, however, in theexpectation of receiving assets from A, may inthe meantime have assumed in another trade theobligation to deliver the assets to a third partyC. The failure of A may then also entail afailure of B to deliver to C, and so on.

REALLOCATION OF RISKS BY CCPs

In order to limit the potential impact ofcounterparty credit and liquidity risks, CCPshave been established in many financialmarkets. A CCP is a special purpose entity thatinterposes itself between the buyer and the

seller in a securities transaction, acting as theseller to the buyer and as the buyer to the seller.In the simple case of an outright securitiestransaction, the seller must deliver thesecurities when settlement is due to the CCPrather than to the buyer. Similarly, the sellerreceives the funds from the CCP, the buyerdelivers the funds to the CCP, and receives thesecurities from the CCP. In doing so, the CCPassumes the counterparty credit and liquidityrisk from the trading parties. If, for example,the buyer fails to pay, then the CCP must stillsettle the transaction with the seller, while thetransaction between the buyer and the CCP iscancelled or settlement is postponed. Hence,the seller will not be affected by a default of thebuyer. The CCP thus acts as guarantor for thefulfilment of obligations from trades.

Historically, most CCPs tended only to befound in derivatives and repo markets, as thetime-lag between the trading and the settlementphase is longer in these markets than in outrightsecurities markets. This longer time-lagimplies that the risk of one party becominginsolvent before settlement (the counterpartycredit risk) is also greater in derivatives andrepo markets than in outright markets.1

However, many CCPs have recently startedserving outright securities markets as well.

It should be noted that CCPs do not eliminatecounterparty credit risk and liquidity risk;instead, they reallocate it. The risk that, forexample, the buyer will not be able to fulfil itsobligations will now be borne by the CCPrather than by the seller. The seller is only leftwith the risk that the CCP cannot fulfilits obligations towards the seller. However,CCPs specialise in managing exposure tocounterparty credit and liquidity risks. Ifadequate procedures are in place, then they arein a better position than the trading partnersbehind the transactions to cope with such risks.

1 On the other hand, liquidity risk may under certaincircumstances decline as the time-lag between the trading andsettlement phase increases. A greater time-lag gives a tradingparty which is short in an asset that it has to deliver more time toclose its position.

179cECB


IV SPECIALFEATURES

CCPs are therefore expected to reallocate theserisks in an efficient way, thereby contributingto financial stability.

RISKS AND RISK MITIGATION IN CCP CLEARING

To ensure that CCPs do indeed contribute tofinancial stability, it must be ensured that theycannot default on their own obligations. CCPsshould and indeed do use various measures tothis end, some of which are discussed below.

FINANCIAL RESOURCESSuppose that two parties, A and B, conclude anoutright trade according to which at settlementday, A has to deliver assets to B and B has tomake a payment to A. If there is a CCPinterposed between A and B, and A nowdefaults, then the CCP is released from theobligation to make a payment to A, although itwill not receive assets from A either. Despitethis, the CCP is obliged to deliver the assets toB (and B has to make a payment to the CCP). Tofulfil its obligation towards B, the CCP mightnow have to buy the assets in the market from athird party. However, the price of the assetsmay in the meantime have increased so that theCCP will incur a loss. To avoid the risk that

such losses could result in insolvency and, as aconsequence, that CCPs could default on theirown obligations, CCPs typically use a varietyof financial resources for protection.

As a first line of protection, CCP participantsare normally subject to margin requirements,i.e. they must post collateral in the form of cashor other assets. Several types of margins can bedistinguished, depending on how margins aredetermined. Initial margins, for example, aremargins that are to be posted to the CCP when aparticipant opens a position, for example whenit buys a futures contract. The amount to beposted typically depends on the volatility of therespective futures price. If the participantdefaults, then the CCP uses the margins postedby the defaulting participant as compensationfor its losses from such a default. Variationmargins are margins that are to be posted whenthe price of an earlier opened position varies.Participants whose positions have lost valuewill post collateral to the CCP; the CCP thenpasses the collateral on to participants whosepositions have gained in value. A stylisedexample of the variation margining process fora futures contract is provided in Box F.1 below.

Box F.1

STYLISED EXAMPLE OF A TYPICAL LIFECYCLE OF A FUTURES CONTRACT WITH AND WITHOUTVARIATION MARGINING

Consider a derivatives exchange that offers the trading of a futures contract. The first tradingday is Day 1, the last trading day and delivery day is Day 3. The underlying security is agovernment bond. A CCP clears all trades on the exchange.

Three parties (B1, B

2 and S) trade the futures contract. On Day 1, B

1 buys 10, B

2 buys 20 and

S sells 30 contracts. For simplicity, it is assumed that throughout the day the price of thecontract remains f

1. Thus, the CCP buys 30 contracts from S and sells 10 to B

1 and 20 to B

2 at

price f1 respectively. At the end of Day 1, B

1 has a long position of 10 contracts, B

2 has a long

position of 20 contracts and S has a short position of 30 contracts.

On Day 2, B1 sells 10 contracts and S buys 10 contracts while B

2 does not trade. Again, it is

assumed that the price of the contract remains the same throughout Day 2, now at f2. Thus, the

CCP buys 10 contracts from B1 at f

2 and sells 10 contracts to S at f

2. At the end of Day 2, B

1 has


Table F.1.1 Net asset value f lows withoutvariat ion margining, no default

Table F.1.2 Net asset value f lows withoutvariat ion margining, S defaults after Day 2

Table F.1.3 Net asset value f lows withvariat ion margining, no default

Table F.1.4 Net asset value f lows withvariat ion margining, S defaults after Day 2

Day 1 Day 2 Day 3 Sum

B1

-10 f1

10 f2

0 -10 (f1-f

2)

B2

-20 f1

20 f3

-20 (f1-f

3)

S 30 f1

-10 f2

-20 f3

30 f1-10 f

2

-20 f3

CCP 0 0 0 0


B1

-10 f1

10 f2

0 -10 (f1-f

2)

B2

-20 f1

0 20 f3

-20 (f1-f

3)

S 30 f1

-10 f2

0 30 f1-10 f

2

CCP 0 0 -20 f3

-20 f3


B1

0 -10 (f1-

-f

2) 0 -10 (f

1-f

2)

B2

0 -20 (f1-

-f

2) -20 (f

2-

-f

3) -20 (f

1-f

3)

S 0 30 (f1-

-f

2) -20 (f

2-

-f

3) 30 f

1-10 f

2

-20 f3

CCP 0 0 0 0


B1

0 -10 (f1-

-f

2) 0 -10 (f

1-f

2)

B2

0 -20 (f1-

-f

2) -20 (f

2-

-f

3) -20 (f

1-f

3)

S 0 30 (f1-

-f

2) 0 30 (f

1-f

2)

CCP 0 0 20 (f2-

-f

3) 20 (f

2-

-f

3)

accordingly closed its position, B2’s position remains unchanged, and S has reduced its short

position by 10 to 20 contracts.

The contract is not traded on Day 3. The price of the government bond is f3. The futures contract

stipulates that any trader with a long position of x at the end of Day 3 will receive x bonds fromthe CCP, while any trader with a short position of y must deliver y bonds to the CCP.

Table F.1.1 shows the asset value flows in a case where variation margining is not applied andwhere there are no defaults. On Day 1, B

1 pays 10f

1 and B

2 pays 20f

1 to the CCP, while the CCP

pays 30f1 to S. The flows for Day 2 are interpreted in a similar way. On Day 3, S must deliver 20

bonds to the CCP. As the price of the bond is f3, this implies an asset value flow of 20f

3 from S

to the CCP. Finally, the CCP must deliver 20 bonds to B2.

Now consider again a case where the CCP does not apply variation margining, but S defaultsafter Day 2 and thus cannot fulfil any obligations on Day 3, thus obliging the CCP to step in.The resulting asset value flows are presented in Table F.1.2. The CCP now makes a loss of 20f

3.

As compensation, it can now claim the initial margins posted by S when S opened its positionon Day 1. However, the calculation of the initial margins was based on the price f

1 of Day 1. If

f3 is significantly higher than f

1, then the initial margins may not be sufficient to cover the

CCP’s losses.

Tables F.1.3 and F.1.4 show the effects of introducing variation margining. On Day 1, no assetflows occur (except for initial margins which are not considered in the tables). On Day 2, B

1

pays 10(f1-f

2) to the CCP. If S defaults after Day 2 so that it cannot fulfil its obligations, then the

CCP realises a gain of 20(f2-f

3) which is negative (a loss) if f

2 is smaller than f

3. A comparison

of the situations described in Table A.2 and in Table A.4 shows that variation marginingreduces the CCP’s potential losses.

181cECB


IV SPECIALFEATURES

Margins are collateral posted by a CCPparticipant and are used by the CCP in casethis participant defaults. As a second layerof protection, many CCPs use clearing funds.A clearing fund is a pool of collateral towhich every participant contributes. Thus, itconstitutes a type of mutual insurance. If theCCP is forced to have recourse to the clearingfund, then all participants will share in thelosses incurred by the CCP.

Some CCPs buy insurance against losses fromdefaulting participants or have contingentclaims on a participant’s resources or on theresources of a participant’s parent company.Finally, if all other layers are exhausted, theCCP’s own capital must counterbalance allremaining losses. The CCP should thereforehave sufficient own capital to cope withextreme losses.

Credit lines and liquidity of financialresourcesWhen a CCP participant defaults and the CCPis forced to step in instead, the obligations thatarise from the participant’s default must befulfilled in a timely manner. Ideally, the assetsthat the CCP must deliver to the non-defaultingparticipants are already part of the CCP’sfinancial resources. If this is not the case, thenit is important that the CCP can easily buy orborrow the assets in the market. An adequatepart of the CCP’s financial resources shouldtherefore be sufficiently liquid to be used tobuy any required assets or to be used ascollateral to borrow them. Sufficient creditlines should allow the CCP to borrow what itneeds.

The way in which CCPs hold financialresources not only determines whether they canfulfil all obligations that arise owing todefaulting participants in a timely manner, butalso determines the extent to which they canincur losses from investments. Risky assetsexpose CCPs to additional risks. It may, forexample, be appropriate that CCPs hold cashpositions mainly in central bank money, i.e. onaccounts with a central bank.

Participation requirements and limitsTo a certain extent, it might be advisable torestrict participation in a CCP by imposingparticipation requirements. Institutions thatare characterised by a relatively highprobability of default, for example becausethey are undercapitalised, may be excludedfrom participation in a CCP. At the same time,position limits may be in place, i.e. limits onthe amount that the CCP is ready to guarantee.

Setting the optimal level of participationrequirements and limits is a difficult task. Ifthey are too demanding, then too few tradeswill be cleared through the CCP, and marketparticipants will then be exposed tocounterparty credit and liquidity risks. MostCCPs, however, allow their participants notonly to clear their own obligations through theCCP, but also those of market participantswhich do not participate directly in the CCP.2

Operational proceduresFinally, it is important to note that CCPs relyon technologically sophisticated proceduresfor transferring assets from, or to, participantsand for calculating collateral requirements.This not only involves procedures operated bythe CCP, but also those of cash and securitiessettlement systems. All of these proceduresmust be operationally reliable. This isespecially important given that manytransactions are cleared and settled “straightthrough”, i.e. automatically in a central routineprocedure. If such a procedure fails, major –though hopefully only temporary – disruptionsof financial markets could ensue. Businesscontingency facilities should support theoperational reliability of the CCP.

CONCLUDING REMARKS

As CCPs are now starting to serve an increasingnumber of markets – including outrightsecurities markets – their systemic importance

2 CCP participants that are able to clear only their ownobligations are often called “individual” or direct clearingmembers, whereas those that are also able to clear obligationsof their clients are typically called general clearing members.


3 See Section 6.

has grown in recent years. Additionally,consolidation has significantly reduced thenumber of CCPs in Europe3, leading to aconcentration of more risk in each of theremaining CCPs. Insolvency or operationalproblems of a CCP could therefore lead tosevere disruptions in the financial markets.

CCPs apply sophisticated risk managementmeasures and are highly regulated by publicauthorities. However, in an ever-changingenvironment, new risks may occur that must bedetected in time and adequately monitored.

ECBFinancial Stability Review

December 2005

S TAT I S T I C A LA N N E X

S 1

Chart S1: US non-farm, non-financial corporate business liabilities S3Chart S2: US non-farm, non-financial corporate business net equity issuance S3Chart S3: US household debt-to-disposable income ratio S3Chart S4: US household debt burden S3Chart S5: Share of adjustable rate mortgages in the US S4Chart S6: US general government debt-to-GDP ratio S4Chart S7: Japanese banks’ non-performing loans S4Chart S8: International positions of all BIS reporting banks vis-à-vis emerging markets S4Chart S9: Consolidated claims on non-banks in offshore financial centres S5Chart S10: Hedge fund inflows S5Chart S11: Nominal broad USD effective exchange rate index S6Chart S12: One-month implied volatility for USD/EUR, JPY/EUR and JPY/USD S6Chart S13: US risk aversion index S6Chart S14: Stock prices in the US S6Chart S15: Price-earnings (P/E) ratio for the US stock market S7Chart S16: VIX implied volatility for the S&P 500 index S7Chart S17: Option-implied probability distribution function for the S&P 500 index S7Chart S18: US mutual fund flows S7Chart S19: Debit balances in New York Stock Exchange margin accounts S8Chart S20: Open interest in options contracts on the S&P 500 index S8Chart S21: Gross equity issuance in the US S8Chart S22: Spreads on US high-yield corporate bonds S8Chart S23: Sovereign bond spreads in major emerging regions S9Chart S24: Equity market indices in major emerging regions S9Chart S25: Precious metal prices S10Chart S26: Share of non-commercial futures positions in overall crude oil futures

positions S10Chart S27: Net lending/borrowing of non-financial corporations S11Chart S28: Total debt of non-financial corporations in the euro area S11Chart S29: Total debt-to-financial asset ratio of non-financial corporations

in the euro area S11Chart S30: Annual growth of loans to non-financial corporations in the euro area

for selected maturities S11Chart S31: Annual growth of debt securities issued by non-financial corporations

in the euro area S12Chart S32: Euro area non-financial corporations’ expected default frequency (EDF)

distributions S12Chart S33: Expected default frequency (EDF) distributions for large and small

euro area non-financial corporations S12Chart S34: Household debt-to-GDP ratio in the euro area S12Chart S35: Ratio of household debt to financial assets and liquid financial assets

in the euro area S13Chart S36: Annual growth of loans to households in the euro area S13Chart S37: Total debt servicing burden of the euro area household sector S13Chart S38: Euro area spreads between interbank deposit and repo interest rates S14Chart S39: Bid-ask spreads for EONIA swap rates S14Chart S40: Option-implied skewness coefficient for ten-year bond yields in Germany S14Chart S41: Stock prices in the euro area S14Chart S42: Price-earnings (P/E) ratio for the euro area stock market S15

S TAT I S T I C A L ANNEX

ECB cFinancial Stability ReviewDecember 2005S 2

Chart S43: Implied volatility for the Dow Jones EURO STOXX 50 index S15Chart S44: Option-implied probability distribution function for the Dow Jones

EURO STOXX 50 index S15Chart S45: Open interest in options contracts on the Dow Jones

EURO STOXX 50 index S15Chart S46: Gross equity issuance and pipeline deals in the euro area S16Chart S47: Corporate bond spreads in the euro area S16Chart S48: Spreads on euro area high-yield corporate bonds S16Chart S49: Number of euro area banking sector mergers and acquisitions (M&As) S17Chart S50: Value of euro area banking sector mergers and acquisitions (M&As) S17Chart S51: Number of mergers and acquisitions (M&As) between banks and

insurance companies in the euro area S17Chart S52: Value of mergers and acquisitions (M&As) between banks and

insurance companies in the euro area S17Chart S53: Cross-border activity of euro area MFIs S18Chart S54: Annual growth in euro area MFI loans extended by sector S18Chart S55: Annual growth in euro area MFIs’ securities and shares issuance S18Chart S56: Euro area MFIs’ foreign currency-denominated assets, selected

balance sheet items S19Chart S57: Lending margins of euro area MFIs S19Chart S58: Deposit margin of euro area MFIs S19Chart S59: International exposure of euro area banks to Latin American countries S19Chart S60: International exposure of euro area banks to Asian countries S20Chart S61: Expected default frequencies (EDF) for large euro area banks S20Chart S62: Distance-to-default for large euro area banks S20Chart S63: European financial institutions’ credit default swaps on senior and

subordinated debt S20Chart S64: Large euro area banks’ earnings per share (EPS) S21Chart S65: Price-earnings (P/E) ratios for large euro area banks S21Chart S66: Implied volatility for Dow Jones EURO STOXX total market

and bank indices S21Chart S67: Euro area corporate bond and bank loan spreads S21Chart S68: Subordinated bond spreads and expected default frequencies (EDF)

for the euro area insurance industry S22Chart S69: Expected default frequencies (EDF) for the euro area insurance industry S22

Table S1: Selected financial vulnerability indicators for some of the mainemerging market economies S5

Table S2: Total international bond issuance (private and public) in selectedemerging markets S9

Table S3: Euro area banking sector structure S18Table S4: Euro area consolidated foreign claims of reporting banks on individual

countries S23Table S5: Euro area banks’ profitability and efficiency S24Table S6: Euro area banks’ balance sheet and off-balance sheet items S25Table S7: Euro area banks’ non-performing loans and provisioning S26Table S8: Euro area banks’ regulatory capital ratios and risk-adjusted items S27Table S9: Financial conditions of a set of large euro area banks S28Table S10: Euro area banks’ exposures at risk to seven aggregate sectors S28


December 2005


S 3

1 EX T E RNA L ENV I RONMENT

Chart S3 US household debt-to-d isposableincome rat io



0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

140

1980 1984 1988 1992 1996 2000 2004

totalhome mortgagesconsumer credit

Chart S4 US household debt burden



10

12

14

16

18

20

10

12

14

16

18

20

1980 1984 1988 1992 1996 2000 2004

financial obligations ratiodebt service ratio

Chart S1 US non- farm, non- f inanc ia lcorporate bus iness l iab i l i t ies

(Q1 1960 - Q2 2005, %)

Sources: US Federal Reserve Board and Bureau of EconomicAnalysis.

20

40

60

80

100

120

140

160

20

40

60

80

100

120

140

160

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

liabilities to financial assetsliabilities to GDPcredit market liabilities to GDP

Chart S2 US non- farm, non- f inanc ia lcorporate bus iness net equity i ssuance

(Q1 1990 - Q2 2005, USD billions, seasonally adjustedquarterly annualised data)


-500

-400

-300

-200

-100

0

100

200

-500

-400

-300

-200

-100

0

100

200

1990 1992 1994 1996 1998 2000 2002 2004


Chart S5 Share of ad justable ratemortgages in the US

(Jan. 1998 - Oct. 2005, % of total new mortgages)

Source: Mortgage Bankers Association.

0

10

20

30

40

50

60

0

10

20

30

40

50

60

1998 1999 2000 2001 2002 2003 2004 2005

number of loansdollar volume

Chart S6 US genera l governmentdebt-to-GDP rat io

(Q1 1980 - Q2 2005, %)

Source: US Federal Reserve Board.Note: This refers to the consolidated federal, state and localgovernment debt.

40

50

60

70

80

40

50

60

70

80

1980 1984 1988 1992 1996 2000 2004

Chart S8 Internat iona l pos it ions of a l lB IS report ing banks v is -à-v is emerg ingmarkets(Q1 1999 - Q1 2005, USD billions)

Source: Bank for International Settlements (BIS).

650

700

750

800

850

900

950

1,000

1,050

1999 2000 2001 2002 2003 2004 2005120

140

160

180

200

220

240

260

280

loans and deposits (left-hand scale)holdings of securities (right-hand scale)

Chart S7 Japanese banks ’ non-per formingloans

(Mar. 1998 - Mar. 2005, % of total loans)

Source: Japan Financial Services Agency.

012345678910

0123456789

10

Mar.1998

Sep. Mar.1999

Sep. Mar.2000

Sep. Mar.2001

Sep. Mar.2002

Sep. Mar.2003

Sep. Mar.2004

Sep. Mar.2005


December 2005


S 5

Chart S10 Hedge fund in f lows

(Q1 1994 - Q2 2005, USD billions)

Source: TASS Research.Note: Excluding funds of hedge funds.

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005-10

0

10

20

30

40

50

-10

0

10

20

30

40

50

otherevent-drivenmarket-neutraldirectional

Chart S9 Consol idated c la ims on non-banks in of f shore f inanc ia l centres

(Q1 1996 - Q1 2005, USD billions)

Source: BIS.

0

200

400

600

800

1,000

0

200

400

600

800

1,000

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

all reporting bankseuro area banks

Current account balance External debt Short-term external debt Foreign reserves(% of GDP) (% of GDP) (% of reserves) (in months of imports)

2002 2005(e) 2002 2005(e) 2002 2005(e) 2002 2005(e)

Latin AmericaArgentina 8.5 1.0 157 73 170 57 5.3 6.8Brazil -1.7 1.7 50 27 65 31 5.4 5.7Chile -0.9 0.6 60 38 36 42 7.4 4.2Colombia -1.7 -1.5 46 34 17 19 6.9 6.2Mexico -2.1 -1.3 25 23 71 50 3.0 3.2Venezuela 8.2 15.3 41 30 78 29 4.6 6.3AsiaChina 2.7 6.7 13 13 16 12 9.9 13.4India 1.2 -1.3 23 19 18 14 9.8 8.8Indonesia 3.9 1.6 66 45 54 51 6.0 4.7Malaysia 7.6 10.0 52 58 25 26 4.1 7.0South Korea 1.0 2.3 26 24 41 31 7.6 7.7Thailand 5.5 -2.5 47 30 31 28 5.8 4.2Emerging EuropeRussia 8.0 5.7 42 31 56 40 6.9 11.4Turkey -0.8 -5.1 69 48 78 141 5.2 3.8

Tab le S1 Se l e c ted f i nanc i a l vu lnerab i l i t y i nd i ca tor s fo r some o f the ma in emerg ingmarket e conomies

Source: Institute of International Finance.Note: Data for 2005 are estimates.


2 I N T E RNAT I ONA L F I N ANC I A L MARKE T S

Chart S11 Nominal broad USD ef fect iveexchange rate index

(Jan. 2002 - Oct. 2005, index: Jan. 2002 = 100)


80

85

90

95

100

105

80

85

90

95

100

105

2002 2003 2004 2005

Chart S12 One-month impl ied volat i l i tyfor USD/EUR, JPY/EUR and JPY/USD

(Jan. 2002 - Nov. 2005, %)

Source: Reuters.

6

7

8

9

10

11

12

13

14

6

7

8

9

10

11

12

13

14

2002 2003 2004 2005

USD/EURJPY/EURJPY/USD

Chart S13 US r i sk avers ion index

(Jan. 1990 - Oct. 2005)

Source: Goldman Sachs.Note: The risk aversion index ranges between 0 and 10, andmeasures investors’ willingness to invest in risky assets asopposed to risk-free securities.

0123456789

10

012345678910

1990 1992 1994 1996 1998 2000 2002 2004

risk-loving

risk-neutral

risk-averse

Chart S14 Stock pr ices in the US

(Jan. 2002 - Nov. 2005, S&P 500, index: Jan. 2003 = 100)

Source: Reuters.

80

90

100

110

120

130

140

150

80

90

100

110

120

130

140

150

2002 2003 2004 2005


December 2005


S 7

Chart S15 Pr ice-earn ings (P /E) rat io forthe US stock market

(Jan. 1983 - Nov. 2005, %, ten-year trailing earnings)

Sources: Thomson Financial Datastream and ECB calculations.Note: The P/E ratio is based on prevailing stock prices relative toan average of the previous ten years of earnings.

10

20

30

40

50

10

20

30

40

50

1983 1986 1989 1992 1995 1998 2001 2004

Chart S16 VIX impl ied vo lat i l i ty for theS&P 500 index

(Jan. 2002 - Nov. 2005, %)

Source: Thomson Financial Datastream.Note: Data calculated by the Chicago Board Options Exchange(CBOE).

0

5

10

15

20

25

30

35

40

45

0

5

10

15

20

25

30

35

40

45

2002 2003 2004 2005

Chart S17 Opt ion- impl ied probabi l i tyd istr ibut ion funct ion for the S&P 500index


0.000.020.040.060.080.100.120.140.160.180.200.22

0.000.020.040.060.080.100.120.140.160.180.200.22

-25 -15 -5 15 255

October 2005May 2005

% change


Chart S18 US mutual fund f lows

(Mar. 1998 - Sep. 2005, USD billions, three-month movingaverage)

Source: Investment Company Institute.

-30

-20

-10

0

10

20

30

40

50

-30

-20

-10

0

10

20

30

40

50

1998 1999 2000 2001 2002 2003 2004 2005

stock fundsgovernment bond funds


Chart S19 Debit ba lances in New YorkStock Exchange marg in accounts

(Jan. 1992 - Sep. 2005, USD billions)

Source: New York Stock Exchange (NYSE).Note: Borrowing to buy stocks “on margin” allows investors touse loans to pay for up to 50% of a stock’s price.

0

50

100

150

200

250

300

0

50

100

150

200

250

300

1992 1994 1996 1998 2000 2002 2004

Chart S20 Open interest in opt ionscontracts on the S&P 500 index

(Jan. 1999 - Oct. 2005, millions of contracts)

Source: Chicago Board Options Exchange (CBOE).

0

1

2

3

4

5

6

7

0

1

2

3

4

5

6

7

1999 2000 2001 2002 2003 2004 2005

Chart S21 Gross equity i ssuancein the US

(Jan. 2000 - Sep. 2005, USD billions, 12-month moving sums)


50

100

150

200

250

50

0 0

100

150

200

250

2000 2001 2002 2003 2004 2005

Secondary Public Offerings – pipelinesSecondary Public Offerings – realisedInitial Public Offerings – pipelinesInitial Public Offerings – realised

Chart S22 Spreads on US h igh-y ie ldcorporate bonds


Source: JP Morgan Chase & Co.Note: Spread between the yield to maturity of the US domestichigh-yield index (BB+ rating or below, average maturity of 7.7years) and US ten-year government bond yield.

200

400

600

800

1,000

1,200

200

400

600

800

1,000

1,200

1999 2000 2001 2002 2003 2004 2005


December 2005


S 9

Chart S23 Sovere ign bond spreads in majoremerg ing reg ions


Source: JP Morgan Chase & Co.Note: The series shown is the Emerging Market Bond Index Plus(EMBI+) “performing” index.

0

500

1,000

1,500

2,000

2,500

3,000

3,500

0

500

1,000

1,500

2,000

2,500

3,000

3,500

1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

EMBI+EMBI+ AsiaEMBI+ EuropeEMBI+ Latin America

Chart S24 Equity market indices in majoremerg ing reg ions

(Jan. 2002 - Nov. 2005, index: Jan. 2002 = 100)

Source: Bloomberg.

50

100

150

200

250

300

50

100

150

200

250

300

2002 2003 2004 2005

MSCI emerging marketsMSCI AsiaMSCI eastern EuropeMSCI Latin America

2001 2002 2003 2004 2004 2004 2005 2005 2005 2005H1 H2 total Q1 Q2 Q3 Jan.-Oct.

Total major EMEs 66,078 65,220 98,982 56,178 54,365 110,543 27,657 24,995 32,190 94,341

Latin America 29,154 18,963 32,635 19,825 16,877 36,702 9,087 6,444 12,573 29,245

of which:Argentina 3,328 - - 915 - 915 150 - - 150Brazil 7,417 5,736 11,803 4,621 4,726 9,346 3,402 2,490 9,262 15,829Chile 2,150 1,399 1,000 750 557 1,307 - - - -Colombia 4,004 1,000 1,265 500 1,044 1,544 447 - 1,000 1,447Mexico 7,552 6,098 11,226 9,223 6,278 15,501 3,363 1,475 800 6,103Venezuela 1,729 1,049 4,478 2,380 2,000 4,380 1,325 1,604 150 3,079

Non-Japan Asia 31,677 35,629 49,942 26,534 27,074 53,608 10,951 10,861 14,533 42,769

of which: China 2,552 860 2,979 352 5,837 6,188 500 195 1,500 3,272

Hong Kong 9,267 1,989 12,631 3,362 2,907 6,268 1,678 2,280 650 5,033India 99 153 450 1,863 2,554 4,417 1,018 500 1,238 3,270South Korea 6,545 11,843 11,193 8,507 7,496 16,003 3,725 2,717 3,023 10,589Malaysia 1,766 5,965 1,442 1,325 2,115 3,440 1,053 1,095 1,100 4,248Singapore 7,400 812 3,885 3,267 3,794 7,061 425 1,025 3,134 4,841Thailand - 48 300 1,000 400 1,400 150 650 650 1,550

Emerging Europe 5,247 10,629 16,406 9,818 10,414 20,232 7,620 7,690 5,083 22,327

of which:Russia 1,353 3,713 8,285 4,060 6,430 10,490 3,466 3,967 4,288 12,681Turkey 2,159 3,560 5,454 3,843 2,634 6,477 3,794 2,875 795 7,464Ukraine - 399 1,250 808 1,350 2,158 100 234 - 1,308Bulgaria 223 1,248 62 10 - 10 260 - - 260Romania 794 1,062 814 - - - - 614 - 614Croatia 718 647 541 1,098 - 1,098 - - - -

Tab le S2 Tota l i n te rnat iona l bond i s suance (pr i va te and pub l i c ) i n se l e c ted emerg ingmarke t s(USD millions)

Source: Dealogic (Bondware).Note: Regions are defined as follows: Latin America: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador,Guatemala, Honduras, Mexico, Panama, Paraguay, Peru, Uruguay, Venezuela. Non-Japan Asia: Brunei, Burma, China, SpecialAdministrative Region of Hong Kong, Indonesia, Laos, Macau, Malaysia, Nauru, North Korea, the Philippines, Samoa, Singapore,South Korea, Taiwan, Thailand, Vietnam. Emerging Europe: Bulgaria, Croatia, Romania, Russia, Turkey and Ukraine.


Chart S25 Prec ious meta l pr ices

(Jan. 1999 - Nov. 2005, index: Jan. 1999 = 100, prices in USD)

Source: Bloomberg.

6080

100120140160180200220240260

6080100120140160180200220240260

1999 2000 2001 2002 2003 2004 2005

goldsilverplatinum

Chart S26 Share of non-commerc ia lfutures pos it ions in overa l l crude o i lfutures pos it ions(Jan. 2001 - Oct. 2005, %)

Source: Bloomberg.

0

5

10

15

20

25

0

5

10

15

20

25

2001 2002 2003 2004 2005


December 2005


S 11

3 EURO A R E A ENV I RONMENT

Chart S27 Net lending/borrowing ofnon- f inanc ia l corporat ions

(1995 - 2004, financing gap, % of GDP)

Source: ECB.Note: Data for 2003 and 2004 are estimates using flow-of-fundsprojections.

-4

-3

-2

-1

0

1

2

3

-4

-3

-2

-1

0

1

2

3

1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

Chart S28 Tota l debt of non- f inanc ia lcorporat ions in the euro area

(Q1 1998 - Q3 2005, %)

Source: ECB.

50

52

54

56

58

60

62

64

1998 1999 2000 2001 2002 2003 2004 200550

70

90

110

130

150

170

190

210

230

debt-to-GDP ratio (left-hand scale)debt-to-equity ratio (right-hand scale)

Chart S29 Tota l debt-to- f inanc ia l assetrat io of non- f inanc ia l corporat ions inthe euro area(Q1 1998 - Q1 2005, %)

Source: ECB.

60

70

80

90

60

70

80

90

1998 1999 2000 2001 2002 2003 2004 2005

Chart S30 Annual growth of loans tonon- f inanc ia l corporat ions in the euroarea for se lected matur it ies(Q1 1999 - Q3 2005, % per annum)

Source: ECB.Note: Data are based on financial transactions of MFIs’ loans.

-5

0

5

10

15

20

-5

0

5

10

15

20

1999 2000 2001 2002 2003 2004 2005

up to 1 yearover 1 and up to 5 yearsover 5 years


Chart S31 Annual growth of debtsecur it ies i ssued by non- f inanc ia lcorporat ions in the euro area(Mar. 1991 - Aug. 2005, % per annum, three-month movingaverage)

Source: ECB.

-20

-10

0

10

20

30

40

50

-20

-10

0

10

20

30

40

50

1991 1993 1995 1997 1999 2001 2003 2005

totalshortlong

Chart S32 Euro area non- f inanc ia lcorporat ions ’ expected defaultf requency (EDF) d istr ibut ions

Sources: Moody’s KMV and ECB calculations.Note: The EDF provides an estimate of the probability of defaultover the following year.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

December 2004March 2005June 2005September 2005

% probability

expected default frequency (% probability)

Chart S33 Expected defau lt f requency(EDF) d istr ibut ions for large and smal leuro area non- f inanc ia l corporat ions

Sources: Moody’s KMV and ECB calculations.Note: The EDF provides an estimate of the probability of defaultover the following year. Size is determined by the quartiles ofthe value of liabilities: small if in the lower and large if in theupper quartile of the distribution.

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0

small, March 2005small, September 2005large, March 2005large, September 2005

% probability

expected default frequency (% probability)

Chart S34 Household debt-to-GDP rat ioin the euro area

(Q1 1998 - Q3 2005, %)

Sources: ECB and Eurostat.Note: Data for Q2 and Q3 2005 are estimated on the basis ofmonetary data.

42

44

46

48

50

52

54

56

58

42

44

46

48

50

52

54

56

58

1998 1999 2000 2001 2002 2003 2004 2005


December 2005


S 13

Chart S37 Tota l debt serv ic ing burden ofthe euro area household sector

(1991 - 2004, % of disposable income)

Source: ECB calculations.Note: Data for 2004 are estimates.

0

2

4

6

8

10

12

14

0

2

4

6

8

10

12

14

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004

reimbursement flowsinterest payments

Chart S35 Rat io of household debt tof inanc ia l assets and l iqu id f inanc ia l assetsin the euro area(1995 - 2003, %)

Source: ECB.

60

65

70

75

80

85

90

1995 1996 1997 1998 1999 2000 2001 2002 200324

25

26

27

28

29

30

household debt-to-financial assets ratio (right-hand scale)household debt-to-liquid financial assets ratio (left-hand scale)

Chart S36 Annual growth of loans tohouseholds in the euro area

(Q1 1999 - Q3 2005, % per annum)

Source: ECB.Note: Data are based on financial transactions of MFIs’ loans.

0

5

10

15

0

5

10

15

1999 2000 2001 2002 2003 2004 2005

consumer credithouse purchaseother lending


4 EURO A R E A F I N ANC I A L MARKE T S

Chart S38 Euro area spreads betweeninterbank depos it and repo interest rates

(Jan. 2000 - Oct. 2005, basis points, 20-day moving average)

Source: ECB.

-202468

1012141618

-2024681012141618

2000 2001 2002 2003 2004 2005

one weekone monthone year

Chart S39 Bid-ask spreads for EONIAswap rates

(Jan. 2003 - Oct. 2005, basis points, 20-day moving average,transaction weighted)

Source: ECB.

0.0

0.5

1.0

1.5

2.0

2.5

0.0

0.5

1.0

1.5

2.0

2.5

2003 2004 2005

one monththree monthone year

Chart S41 Stock pr ices in the euro area

(Jan. 2002 - Nov. 2005, Dow Jones EURO STOXX,index: Jan. 2003 = 100)

Source: Reuters.

70

80

90

100

110

120

130

140

150

160

70

80

90

100

110

120

130

140

150

160

2002 2003 2004 2005

Chart S40 Opt ion- impl ied skewnesscoef f i c ient for ten-year bond y ie lds inGermany(Jan. 1999 - Oct. 2005, average monthly skewness)

Sources: Eurex and ECB calculations.

-0.2

0.0

0.2

0.4

0.6

-0.2

0.0

0.2

0.4

0.6

1999 2000 2001 2002 2003 2004 2005


December 2005


S 15

Chart S42 Pr ice-earn ings (P /E) rat io forthe euro area stock market

(Jan. 1983 - Nov. 2005, %, ten-year trailing earnings)

Source: Thomson Financial Datastream.Note: The P/E ratio is based on prevailing stock prices relative toan average of the previous ten years of earnings.

5

10

15

20

25

30

35

5

10

15

20

25

30

35

1983 1986 1989 1992 1995 1998 2001 2004

Chart S43 Impl ied volat i l i ty for the DowJones EURO STOXX 50 index

(Jan. 2002 - Nov. 2005, %)

Source: Bloomberg.

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

2002 2003 2004 2005

Chart S44 Opt ion- impl ied probabi l i tyd istr ibut ion funct ion for the Dow JonesEURO STOXX 50 index


0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

-25 -15 -5 5 15 25

October 2005May 2005


% change

Chart S45 Open interest in opt ionscontracts on the Dow Jones EURO STOXX50 index(Jan. 1999 - Oct. 2005, millions of contracts)

Source: Eurex.

02468

10121416182022

0246810121416182022

1999 2000 2001 2002 2003 2004 2005


Chart S46 Gross equity i ssuance andpipe l ine dea ls in the euro area

(Jan. 2000 - Aug. 2005, EUR billions, 12-month moving sums)


Secondary Public Offerings – pipelinesSecondary Public Offerings – realisedInitial Public Offerings – pipelinesInitial Public Offerings – realised

20

0

40

60

80

100

120

140

160

180

20

0

40

60

80

100

120

140

160

180

2000 2001 2002 2003 2004 2005

Chart S47 Corporate bond spreads in theeuro area


Source: Thomson Financial Datastream.Note: Spread between the seven to ten-year yield to maturity andthe euro area seven to ten-year government bond yield.

0

50

100

150

200

250

300

350

0

50

100

150

200

250

300

350

1999 2000 2001 2002 2003 2004 2005

BBBAAAAAA

Chart S48 Spreads on euro areahigh-y ie ld corporate bonds


Source: JP Morgan Chase & Co.Note: Spread between the yield to maturity of the euro area high-yield index (BB+ rating or below, average maturity of 5.5 years)and the euro area f ive-year government bond yield.

200

400

600

800

1,000

1,200

1,400

1,600

1,800

200

400

600

800

1,000

1,200

1,400

1,600

1,800

1999 2000 2001 2002 2003 2004 2005


December 2005


S 17

5 EURO A R E A F I N ANC I A L I N S T I T U T I ON S

Chart S49 Number of euro area bankingsector mergers and acquis i t ions (M&As)

(1985 - 2005, number of deals)

Sources: Thomson Financial SDC and ECB calculations.Note: M&As include both controlling and minority stakes anddeals with and without reported value. “Cross-border” refers tointer-euro area M&As; “inward” denotes M&As by non-euroarea banks in the euro area; and “outward” stands for M&Aactivity of euro area banks outside the euro area. (*) Data untilOctober 2005, annualised.

01985 1987 1989 1991 1993 1995 1997 1999 2001 2005*2003

25

50

75

100

125

150

175

200

0

25

50

75

100

125

150

175

200

outwardinwardcross-borderdomestic

Chart S50 Value of euro area bankingsector mergers and acquis i t ions (M&As)

(1985 - 2005, value of deals, EUR billions)

Sources: Thomson Financial SDC and ECB calculations.Note: M&As include both controlling and minority stakes.“Cross-border” refers to inter-euro area M&As; “inward”denotes M&As by non-euro area banks in the euro area; and“outward” stands for M&A activity of euro area banks outsidethe euro area. (*) Data until October 2005, annualised.


0

20

40

60

80

100

0

20

40

60

80

100

1985 1987 1989 1991 1993 1995 1997 1999 2001 2005*2003

Chart S51 Number of mergers andacquis i t ions (M&As) between banks andinsurance companies in the euro area(1985 - 2005, number of deals)

Sources: Thomson Financial SDC and ECB calculations.Note: The number of deals includes both deals with and withoutreported value, and records both minority and controllingstakes. (*) Data until October 2005, annualised.


0

10

20

30

40

50

0

10

20

30

40

50

1985 1987 1989 1991 1993 1995 1997 1999 2001 2005*2003

Chart S52 Va lue of mergers andacquis i t ions (M&As) between banks andinsurance companies in the euro area(1985 - 2005, value of deals, EUR billions)

Sources: Thomson Financial SDC and ECB calculations.Note: Deals include both controlling and minority stakes.(*) Data until October 2005, annualised.


1985 1987 1989 1991 1993 1995 1997 1999 2001 2005*2003

0

10

20

30

40

50

60

0

10

20

30

40

50

60


Chart S53 Cross-border act iv i ty o f euroarea MFIs

(Q1 1999 - Q2 2005, % of total domestic outstanding amounts)

Source: ECB.Note: Cross-border activity refers to cross-euro area activity(i.e. it excludes international activities in the non-euro area andthird countries).

cross-border non-bank securitiescross-border interbank loanscross-border loans to non-banks

0

20

40

60

80

100

0

20

40

60

80

100

1999 2000 2001 2002 2003 2004 2005

Chart S54 Annual growth in euro area MFIloans extended by sector

(Q1 1999 - Q3 2005, % per annum)

Source: ECB.Note: Data are based on f inancial transactions of MFIs’ loans.

householdsnon-financial corporationsMFIs

1999 2000 2001 2002 2003 2004 20050

2

4

6

8

10

12

0

2

4

6

8

10

12

Chart S55 Annual growth in euro areaMFIs ’ secur it ies and shares i ssuance

(Jan. 2003 - Aug. 2005, % per annum)

Source: ECB.

securities other than shares (all currencies)securities other than shares (EUR)quoted shares

2003 2004 20050123456789

10

012345678910

Changefrom 2003

Number of credit institutionsStand-alone credit institutions 4,102 -219Banking groups 458 -35Credit institutions 4,551 -259

Domestic credit institutions 3,681 -234Foreign-controlled subsidiariesand branches 870 -25

Total assets (EUR billions)Domestic credit institutions 18,963 5.9

of which (%):Large 69.4 2.5Medium-sized 26.5 -2.0Small 4.2 -0.6

Foreign-controlled subsidiariesand branches 2,936 8.7

Table S3 Euro area banking sectorstructure

(2004)

Source: Banking Supervision Committee.Note: Changes from 2003: for the number of institutions, theyare in absolute numbers; for total assets, in percentages; forthe size distribution breakdown of total assets, in percentagepoints.


December 2005


S 19

Chart S56 Euro area MFIs ’ fore igncurrency-denominated assets , se lectedbalance sheet i tems(Q1 1998 - Q2 2005)

Source: ECB.

USD securities other than shares (% of total foreigncurrency-denominated securities other than shares)USD loans (% of total foreign currency-denominated loans)

40

50

60

70

80

40

50

60

70

80

1998 1999 2000 2001 2002 2003 2004 2005

Chart S57 Lending marg ins of euro areaMFIs

(Jan. 2003 - Aug. 2005, % points)

Source: ECB.Note: The weighted lending margins are the difference betweenthe interest rate on new lending and the interest rate swap rate,where both have corresponding maturities.

household lendinglending to non-financial corporations

0.0

0.5

1.0

1.5

2.0

2.5

0.0

0.5

1.0

1.5

2.0

2.5

2003 2004 2005

Chart S58 Depos it marg in of euro areaMFIs

(Jan. 2003 - Aug. 2005, % points)

Source: ECB.Note: The weighted deposit margins are the difference betweenthe interest rate swap rate and the deposit rate, where both havecorresponding maturities.

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

2003 2004 2005

Chart S59 Internat ional exposure of euroarea banks to Lat in Amer ican countr ies

(USD billions)

Source: BIS.

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

140

Q1 2004Q3 2004Q1 2005

Mexico Brazil Chile Argentina


Chart S60 Internat ional exposure of euroarea banks to As ian countr ies

(USD billions)

Source: BIS.

Q1 2004Q3 2004Q1 2005

0

5

10

15

20

25

30

35

0

5

10

15

20

25

30

35

SouthKorea

India China Taiwan Indonesia Malaysia Philippines Thailand

Chart S61 Expected default f requenc ies(EDF) for large euro area banks

(Jan. 1999 - Sep. 2005, % probability)

Sources: Moody’s KMV and ECB calculations.


0.0

0.2

0.4

0.6

0.8

0.0

0.2

0.4

0.6

0.8

1999 2000 2001 2002 2003 2004 2005

Chart S62 Distance-to-default for largeeuro area banks

(Jan. 1999 - Sep. 2005)

Sources: Moody’s KMV and ECB calculations.Note: An increase in the distance-to-default reflects animproving assessment.

median10th percentile

4

5

6

7

8

9

4

5

6

7

8

9

1999 2000 2001 2002 2003 2004 2005

Chart S63 European f inanc ia l inst i tut ions ’credit defau lt swaps on senior andsubordinated debt(May 2002 - Nov. 2005, basis points, five-year maturity)

Source: JP Morgan Chase & Co.Note: “European financial institutions” corresponds to thedefinition of JP Morgan Chase & Co.

subordinatedsenior

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

140

2002 2003 2004 2005


December 2005


S 21

Chart S67 Euro area corporate bond andbank loan spreads


Sources: ECB and Thomson Financial Datastream.Note: Spread between the rate on loans to non-f inancialcorporations with one up to five years’ initial rate f ixationbelow (small) and above (large) 1 EUR million, and the three-year government bond yield.

BBB-rated corporate bond spread spread on large loansspread on small loans

02003 2004 2005

50

100

150

200

250

0

50

100

150

200

250

Chart S66 Impl ied vo lat i l i ty for Dow JonesEURO STOXX tota l market and bankindices(Jan. 2002 - Nov. 2005, %)

Source: Bloomberg.

Dow Jones EURO STOXXDow Jones EURO STOXX bank index

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

2002 2003 2004 2005

Chart S64 Large euro area banks ’ earn ingsper share (EPS)

(Q1 1999 - Q3 2005, %)


simple averageweighted average10th percentile

0

1

2

3

4

0

1

2

3

4

1999 2000 2001 2002 2003 2004 2005

Chart S65 Pr ice-earn ings (P /E) rat ios forlarge euro area banks

(Jan. 1999 - Oct. 2005, %)


simple averageweighted average10th percentile

0

4

8

12

16

20

0

4

8

12

16

20

1999 2000 2001 2002 2003 2004 2005


Chart S69 Expected default f requenc ies(EDF) for the euro area insurance industry

(Jan. 1992 - Sep. 2005, % probability)

Source: Moody’s KMV.


0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Chart S68 Subordinated bond spreads andexpected default f requenc ies (EDF) for theeuro area insurance industry(Jan. 2003 - Nov. 2005)

Sources: Moody’s KMV and JP Morgan Chase & Co.

EDF median (% probability, left-hand scale)subordinated debt spreads (basis points, right-hand scale)

0.1

0.2

0.3

0.4

0.5

0.6

2001 2002 2003 2004 200550

100

150

200

250

c ECBFinanc ia l S tabi l i ty Rev iew

December 2005 S 23

Q1 2003 Q2 2003 Q3 2003 Q4 2003 Q1 2004 Q2 2004 Q3 2004 Q4 2004 Q1 2005

Total all countries 3847.3 4164.7 4146.5 4345.3 4824.6 4815.5 4900.4 5588.5 5774.3

Total non-developedcountries (incl.offshore centres) 962.5 993.7 1045.3 1129.7 1185.6 1208.7 1238.5 1408.4 1414.1

Hong Kong 26.3 30.3 30.0 31.9 35.3 36.3 36.7 41.1 35.9Singapore 31.4 31.0 31.6 29.1 34.8 34.1 34.2 36.2 35.8

Total offshore centres 269.4 272.8 290.6 302.4 331.6 343.6 364.7 416.9 423.5

China 18.5 19.0 20.2 19.0 20.4 22.5 20.6 23.8 25.3India 14.7 15.9 17.6 18.4 21.4 21.1 21.6 24.2 25.8Indonesia 14.7 15.8 15.0 15.2 15.2 14.4 15.5 15.8 15.4Malaysia 7.4 8.2 8.7 8.7 8.4 7.9 8.1 9.9 10.1Philippines 6.6 6.9 7.5 7.5 8.8 8.7 9.0 8.4 9.2South Korea 23.6 27.0 30.0 29.9 32.9 31.4 29.2 33.3 34.6Taiwan China 11.7 13.6 17.2 17.9 22.1 23.7 20.5 23.6 20.9Thailand 9.6 9.5 10.4 9.9 10.1 9.3 6.3 6.3 6.7

Total Asia andPacific EMEs 121.6 130.9 142.9 145.1 160.3 162.0 151.5 168.6 172.1

Cyprus 21.4 24.2 25.9 31.5 30.4 33.7 33.7 37.8 17.6Czech Republic 24.3 25.7 26.3 39.0 40.4 41.1 39.1 46.0 45.5Hungary 27.1 29.3 31.9 36.0 37.2 39.5 41.4 49.8 50.4Poland 56.4 57.3 59.0 64.1 62.9 65.2 69.4 87.2 88.5Russia 24.3 25.8 28.0 33.3 37.1 34.2 34.2 40.7 40.0Turkey 20.6 20.5 20.8 22.5 22.7 23.3 23.7 26.1 26.9

Total European EMEsand new EUMember States 244.2 256.0 270.6 322.8 330.1 342.0 354.4 419.5 408.2

Argentina 23.5 23.1 22.9 21.6 20.3 19.8 19.8 19.8 18.1Brazil 51.2 54.4 57.1 59.4 59.1 58.4 62.7 67.2 73.8Chile 29.3 29.2 29.9 32.6 31.9 31.0 32.5 35.0 35.1Colombia 6.8 6.7 6.7 6.4 6.8 6.7 6.9 8.1 7.4Ecuador 0.6 0.6 0.7 0.7 0.8 0.8 0.9 0.9 0.9Mexico 98.2 100.7 100.7 103.9 106.6 107.2 105.5 120.0 121.8Peru 8.7 9.8 9.2 9.5 9.3 9.5 9.6 10.0 9.9Uruguay 2.2 2.0 2.1 2.0 1.9 2.0 2.0 2.0 2.3Venezuela 10.5 10.8 11.7 13.1 12.1 12.5 12.8 14.7 14.3

Total Latin America 239.9 245.9 249.8 258.4 258.0 256.5 261.3 288.0 294.2

Iran 6.4 7.4 7.8 8.7 9.5 9.5 10.1 11.7 12.2Morocco 10.4 9.2 9.7 11.3 10.5 11.0 11.4 12.6 12.6South Africa 9.2 10.6 10.7 11.3 11.3 11.2 11.6 13.5 12.5

Total Middle Eastand Africa 87.4 88.1 91.4 101.0 105.6 104.5 106.6 115.4 116.0

Table S4 Euro area consol idated foreign claims of reporting banks on individual countries

(USD billions)

Source: BIS.

ECB cFinanc ia l S tabi l i ty Rev iewDecember 2005S 24

All Change Large Change Medium Change Small Change Foreign Changedomestic from domestic from domestic from domestic from banks from

banks 2003 banks 2003 banks 2003 banks 2003 2003

Income(% of total assets)Net interest income 1.21 -0.05 0.95 -0.06 1.70 0.02 2.42 0.00 0.88 -0.15

Interest receivable 3.71 -0.24 3.47 -0.21 4.22 -0.22 4.41 -0.24 3.95 -0.62Interest payable 2.50 -0.18 2.52 -0.15 2.53 -0.25 1.99 -0.23 3.06 -0.46

Net non-interest income 1.08 -0.03 1.16 -0.06 0.86 0.03 1.26 -0.01 0.97 -0.02Fees and commissions (net) 0.61 0.02 0.61 0.01 0.57 0.03 0.88 0.08 0.58 0.00Trading and forex results 0.19 -0.01 0.25 -0.02 0.05 0.00 0.04 -0.01 0.16 -0.01Other operating income (net) 0.29 -0.03 0.30 -0.05 0.23 0.01 0.33 -0.07 0.23 -0.01

Total income 2.29 -0.08 2.11 -0.12 2.55 0.05 3.68 -0.01 1.85 -0.17

Expenditure structure(% of total assets)Staff costs 0.85 -0.03 0.80 -0.05 0.88 0.01 1.47 0.02 0.59 -0.09Administrative costs 0.50 -0.03 0.48 -0.03 0.48 -0.01 0.90 -0.01 0.44 -0.05Other 0.11 -0.01 0.10 -0.01 0.13 0.00 0.21 -0.02 0.10 0.00Total expenses 1.46 -0.07 1.38 -0.09 1.49 -0.01 2.58 -0.01 1.13 -0.14

Profitability(% of total assets)Operating profits 0.83 -0.01 0.73 -0.03 1.06 0.06 1.09 -0.01 0.72 -0.03Specific provisions 0.26 -0.11 0.18 -0.13 0.45 -0.03 0.37 -0.10 0.15 -0.05Funds for general banking risks 0.00 0.00 0.00 -0.01 0.02 0.01 0.02 0.00 -0.01 -0.03Extraordinary items (net) 0.00 0.00 -0.02 0.01 0.07 0.01 0.03 -0.05 0.04 -0.01Tax charges 0.17 0.01 0.15 0.02 0.21 0.01 0.24 -0.02 0.13 0.01Profits (before tax andextraord. items) 0.59 0.12 0.57 0.13 0.62 0.09 0.71 0.10 0.58 0.04Profits (after tax andextraord. items) (ROA) 0.42 0.11 0.39 0.12 0.49 0.09 0.49 0.06 0.49 0.02

Return on equityProfits (after tax andextraord. items) (% Tier 1) (ROE) 10.54 2.78 11.68 3.79 9.50 1.45 6.52 0.90 10.46 0.14

Income structure(% of total income)Net interest income 52.77 -0.47 45.07 -0.13 66.50 -0.54 65.76 0.15 47.92 -3.22Net non-interest income 47.23 0.47 54.93 0.13 33.50 0.54 34.24 -0.15 52.08 3.22

Fees and commissions (net) 26.50 1.71 28.73 2.07 22.21 0.65 24.03 2.14 31.32 2.83Trading and forex results 8.25 -0.22 11.83 -0.41 2.12 -0.20 1.13 -0.36 8.40 -0.05Other operating income (net) 12.49 -1.02 14.37 -1.53 9.17 0.09 9.09 -1.93 12.35 0.45

Expenditure structure(% of total costs)Staff costs 57.98 0.69 57.76 0.55 58.77 0.97 56.96 0.86 52.58 -1.48Administrative costs 34.26 -0.20 34.99 0.00 32.32 -0.73 34.83 -0.27 38.97 0.64Other 7.77 -0.48 7.25 -0.55 8.91 -0.24 8.21 -0.58 8.45 0.84

EfficiencyCost-to-income ratio(% of total income) 63.67 -0.88 65.34 -0.70 58.52 -1.52 70.31 0.13 61.08 -1.76Asset share of banks with acost-to-income ratio of over 80% 5.18 -3.29 5.91 -3.51 2.26 -3.25 11.51 -1.26 9.04 2.27

Table S5 Euro area banks’ prof itabi l ity and eff ic iency

(2004)

Source: Banking Supervision Committee.Note: Changes from 2003 are in percentage points.


December 2005 S 25

Table S6 Euro area banks’ balance sheet and off-balance sheet items

(2004)

Source: Banking Supervision Committee.Note: Changes from 2003 are in percentage points. Concerning the item “debt securities”, some countries only provided information onthe total amount and not on the split between the two sub-items, i.e. “issued by public bodies” and “issued by other borrowers”.Consequently, the sum of the two sub-items can be smaller than the total amount.



Assets(% of total assets)Cash and balances 1.31 -0.08 1.24 0.00 1.41 -0.23 1.90 -0.11 0.88 -0.12Short-term government debt 1.78 0.53 1.48 0.44 2.17 0.83 3.13 0.17 1.39 0.48Loans to credit institutions 16.52 -0.69 17.70 -0.74 14.03 -1.05 12.64 -0.13 28.17 -1.95Debt securities 20.50 -0.06 23.56 0.16 13.58 -1.47 13.54 -0.01 21.31 2.68

Debt securities(public bodies) 7.37 2.85 9.43 3.82 4.57 1.36 2.09 1.49 5.79 0.41Debt securities(other borrowers) 7.62 -8.57 9.94 -7.95 4.30 -9.02 3.34 -11.57 12.29 1.74

Loans to customers 48.39 -0.01 43.92 -0.16 58.68 1.56 57.25 0.35 37.90 -1.69Shares and participatinginterest 3.41 -0.02 3.04 0.04 4.09 -0.11 5.38 0.46 2.78 -0.11Tangible and intangibleassets 1.34 -0.03 1.24 -0.06 1.51 0.07 1.84 -0.02 0.88 -0.06Other assets 7.11 0.59 8.30 0.47 4.70 0.61 2.50 0.00 7.02 1.26

LiquidityLiquid asset ratio 1(cash and short-termgovernment debt) 2.60 0.28 2.24 0.29 3.20 0.41 4.85 0.06 1.87 0.21Liquid asset ratio 2(ratio 1 + loans to cred. inst.) 19.12 -0.41 19.94 -0.45 17.22 -0.63 17.48 -0.06 30.05 -1.74Liquid asset ratio 3(ratio 2 + debt sec. by publicbodies) 22.13 -0.53 23.24 -0.67 19.84 -0.64 18.20 0.17 34.08 -1.40

Liabilities(% of total assets)Amounts owed to creditinstitutions 22.18 -0.79 24.50 -0.72 17.53 -1.52 13.09 -1.57 38.52 0.13Amounts owed to customers 40.59 -0.06 36.92 0.03 46.22 0.66 65.75 1.71 30.15 -1.42Debt certificates 21.36 0.42 21.99 0.57 22.04 -0.27 6.49 0.55 15.54 0.92Accruals and other liabilities 7.91 0.31 9.39 0.31 4.67 -0.03 3.87 -0.37 8.53 1.26Funds for general banking risks 0.17 -0.01 0.16 -0.03 0.21 0.04 0.19 0.02 0.19 -0.03Provisions for liabilitiesand charges 1.34 0.02 1.19 -0.09 1.77 0.33 1.08 0.01 0.70 -0.09Subordinated liabilities 1.91 0.06 1.95 -0.07 2.01 0.35 0.63 -0.11 1.51 -0.19Equity 3.68 0.00 3.07 -0.07 4.68 0.32 7.44 0.15 4.30 -0.49Other liabilities 0.44 0.03 0.44 0.00 0.49 0.09 0.11 0.02 0.18 -0.01Profit or loss for thefinancial year 0.41 0.03 0.31 0.06 0.51 0.08 1.56 -0.41 0.46 -0.04

Selected off-balance sheetitems (% of total assets)Credit lines 12.67 0.83 14.72 0.91 8.66 0.34 4.06 -1.22 12.14 -0.38Guarantees and othercommitments 5.91 -0.11 5.26 -0.38 7.73 0.62 5.02 0.32 7.14 0.28Derivatives 2.47 -6.43 3.33 -12.33 1.75 -0.15 0.32 0.09 8.45 -5.48




Asset quality(% of loans andadvances)Non-performing anddoubtful assets (gross) 3.05 -0.36 2.56 -0.29 3.55 -0.47 7.00 0.12 1.24 -0.19

Asset quality(% of own funds)Non-performing anddoubtful assets (gross) 49.23 -6.49 46.86 -5.00 49.65 -10.56 65.41 -1.22 16.37 1.09Non-performing anddoubtful assets (net) 13.34 -4.60 8.65 -3.03 17.54 -8.19 31.22 -2.53 -0.37 1.06

Provisioning (stock)(% of loans and advances)Total provisions 2.19 -0.09 2.09 -0.12 2.20 -0.03 3.64 0.07 1.25 -0.17

Provisioning (stock)(% of non-performingand doubtful assets)Total provisions 72.90 5.09 81.55 4.05 64.68 7.01 52.28 0.23 103.49 4.47

Table S7 Euro area banks’ non-performing loans and provisioning

(2004)

Source: Banking Supervision Committee.Note: Changes from 2003 are in percentage points. Definitions of non-performing and doubtful assets differ between countries.Consequently, these data should be interpreted with caution.


December 2005 S 27



Overall solvency ratio 11.49 -0.12 10.92 -0.30 12.11 0.17 14.33 0.27 14.77 -0.70Tier 1 ratio 8.41 0.05 7.92 -0.01 8.73 0.16 12.19 0.15 12.47 -0.53

Risk-adjusted items(% of total risk-adjusted assets)Risk-weighted assets 81.70 -0.28 78.15 -0.25 86.66 -0.48 92.56 0.07 79.11 -0.82Risk-weighted off-balancesheet items 12.23 0.41 13.86 0.50 10.14 0.35 6.07 -0.03 12.40 -1.27Risk-adjusted trading book 6.07 -0.14 7.99 -0.25 3.20 0.13 1.37 -0.03 8.48 2.09

Table S8 Euro area banks’ regulatory capital ratios and risk-adjusted items

(2004)

Source: Banking Supervision Committee.Note: Changes from 2003 are in percentage points.

All Changebanks from

2003

Overall solvency ratio 11.76 -0.15Tier 1 ratio 8.75 0.02

Distribution of overallsolvency ratioOverall solvency ratio < 7% 0.03 0.01Overall solvency ratio 7%-8% 0.03 0.02Overall solvency ratio 8%-9% 2.49 -1.35Overall solvency ratio 9%-10% 14.80 7.10Overall solvency ratio 10%-11% 21.74 -0.90Overall solvency ratio 11%-13% 43.31 -2.24Overall solvency ratio > 13% 17.60 -2.65

Overall solvency ratio below 9%Number of banks 67 -27Asset share (% of total banking sector assets) 0.76 0.30

Risk-adjusted items(% of total risk-adjusted assets)Risk-weighted assets 81.49 -0.33Risk-weighted off-balance-sheet items 12.24 0.28Risk-adjusted trading book 6.27 0.05

Composition of trading bookown funds requirement (% oftotal trading book own fundsrequirement under CAD)Own funds requirement for tradeddebt instruments 49.07 0.49Own funds requirement for equities 9.96 0.27Own funds requirement forforeign exchange risk 7.78 -0.02Own funds requirement for othertrading book items 33.19 -0.74


Table S9 Financial condit ions of a set of large euro area banks

(2004 - H1 2005)

Sources: Banks’ annual accounts and interim results and ECB calculations.

IFRS reporting banks non-IFRS reporting banks

min. 1st avg. 3rd max. min. 1st avg. 3rd max.quartile quartile quartile quartile

Return on equity 2004 7.40 12.45 13.58 19.35 31.80 -23.50 -2.15 4.50 8.25 10.80H1 2005 9.60 16.54 20.81 25.28 35.60 14.20 14.50 15.30 15.10 15.40

Net interest income 2004 0.70 1.19 1.55 1.93 2.75 0.32 0.51 0.61 0.86 1.13(% of total assets) H1 2005 0.41 0.85 0.93 1.40 1.98 0.28 0.40 0.55 0.58 0.63

Provisions 2004 0.05 0.14 0.16 0.36 0.76 0.04 0.09 0.09 0.15 0.24(% of total assets) H1 2005 0.01 0.02 0.08 0.17 0.38 0.04 0.06 0.08 0.14 0.20

Tier 1 ratio 2004 6.60 7.50 7.53 8.46 11.40 6.50 7.55 8.06 8.33 8.70H1 2005 6.40 7.46 8.24 9.55 11.10 6.70 7.25 8.28 8.45 9.10

Overall solvency 2004 10.10 10.81 11.36 12.40 13.01 10.50 10.73 11.55 12.08 12.30ratio H1 2005 10.00 10.59 11.92 12.54 18.40 10.80 11.25 12.57 12.65 13.60

Cost-to-income 2004 45.10 59.43 63.52 70.53 81.10 59.20 65.60 75.05 75.70 79.40ratio H1 2005 43.40 53.65 61.44 66.15 82.20 68.20 69.40 71.93 71.80 73.00

Sources: Banking Supervision Committee and Moody’s.Note: The sectors are as follows: basic materials and construction (BaC), capital goods (Cap), consumer cyclicals (CCy), non-cyclicals(CNC), energy and utilities (EnU), financial (Fin), and technology and telecommunications (TMT).1) Changes are based on seven euro area countries to ensure comparability with 2004 data. The countries included in the 2005 data areAT, BE, DE, ES, FI, IE, PT, IT and FR.

BaC EnU Cap CCy CNC Fin TMT

9 euro area countriesTotal exposure, March 2005 EUR billions 608.9 223.9 223.9 1,682.5 805.3 6,360.5 218.2Sectoral EDF, June 2004 % probability 0.63 0.97 0.90 0.58 0.23 0.13 2.42Sectoral EDF, March 2005 % probability 0.25 0.08 0.5 0.46 0.23 0.07 1.13Exposure at risk, March 2005 EUR billions 1.5 0.2 1.1 7.7 1.9 4.5 2.5Change in exposure at risk,March 2005 - June 2004 1) % -66.9 -87.6 -50.3 -14.5 32.1 46.3 -14.5Change in EDF,March 2005 - June 2004 % -60.3 -91.8 -44.4 -20.0 2.2 -46.2 -53.2

Table S10 Euro area banks’ exposures at r isk to seven aggregate sectors

FINANCIAL STABILITY REVIEW DECEMBER 2005€¦ · INDEBTED EURO AREA HOUSEHOLDS USING MICRO-LEVEL...

Documents

Transcript of FINANCIAL STABILITY REVIEW DECEMBER 2005€¦ · INDEBTED EURO AREA HOUSEHOLDS USING MICRO-LEVEL...