fsr_08_special_01_tcm16-23536

Austrian banks� subsidiaries in Central and Eastern Europe deliver a higher contribution to their bankinggroups� overall pre-tax profit than their relative share in assets would suggest. One widespreadexplanation for this is that the margins between credit and deposit rates are higher in the Centraland Eastern European countries than in Austria or the euro area. This paper presents an overviewof the margins in four Central European new EU Member States and analyzes the major differencesin the structure of deposit, lending and overall margins compared with the euro area and Austria.

Introduction1

Austrian banks� subsidiaries located inCentral and Eastern Europe deliver ahigher contribution to their bankinggroups� overall pre-tax profit thantheir relative share in assets wouldsuggest. One widespread explanationfor this is that margins between creditand deposit rates are higher in theCentral and Eastern European coun-tries (CEECs) than in Austria or theeuro area. At the same time, it is oftenclaimed that these margins have al-ready declined significantly.

This paper aims to present anoverview of the margins in four Cen-tral European new EU Member States(the Czech Republic, Hungary, Polandand Slovakia, the so-called NMS-4),based primarily on a first systematicevaluation of the recently launchedMFI (monetary financial institutions)interest rates statistics (MIR) pub-lished by the ECB and the nationalcentral banks of the EU MemberStates. First, we take a close look atthe current level of interest marginsin these countries, comparing themto those in Austria and in the euroarea. Then, we analyze recent margindevelopments in the NMS-4 and sum-

marize the longer-term margin devel-opment in the euro area. Next, twomethods are applied to decomposethe overall margin (spreads) in orderto get a better insight into their driv-ing components. First, adjusted loanand deposit rate spreads in the NMS-4 against the euro area are derivedby subtracting in particular the premiafor sovereign and foreign exchangerisk. Second, the average overall mar-gins are decomposed into the averagelending and deposit margins by usingspecific reference rates for alternativeinvestments/financing, chosen on thebasis of appropriate maturity. This isdone both on an aggregated and disag-gregated (by sector and maturity)level. The resulting lending and de-posit margins are compared with theeuro area average. This method pro-vides a better understanding of therelative importance and developmentof business on the asset and on the li-abilities side. Finally, these analyses fa-cilitate a better assessment of poten-tial future developments in these fourcountries� banking sectors, which willbe highly relevant from the Austrianperspective.

1 The opinions expressed in this paper are those of the authors and do not necessarily represent the views of theOesterreichische Nationalbank. We would like to thank Vanessa Redak for her valuable comments on thispaper.

Zoltan WalkoThomas Reininger

60 Financial Stability Report 8�

Credit and Deposit Interest RateMargins in Four New EU Member States

NMS-4 Interest RateMargins above EuroArea AverageIn March 2004, average overall inter-est rate margins,2 expressed as thedifference between interest rates onloans to and deposits from nonfinan-cial corporations and households,ranged between 3.8 percentage points

and 6.5 percentage points in theNMS-4 (the Czech Republic, Hun-gary, Poland and Slovakia).3 By com-parison, the euro area average was2.4 percentage points (Austria:2.2 percentage points), resulting inmargin spreads of between 1.4 per-centage points and 4.1 percentagepoints.

Recent MarginDevelopments in the NMS-4Due to the short time series of boththe MFI interest rate statistics (MIR)and the Polish interest rate statisticsthat resemble MIR, we cannot draw

far-reaching conclusions about the de-velopment of margins and marginspreads during the past few years.

On the basis of the MIR, the aver-age overall margin (i.e. the weightedaverage of data for households and

2 In this analysis we use the following terminology:The term ��margin�� is used for the difference between interest rates in an individual country. The overallmargin is defined as the difference between interest rates on loans and interest rates on deposits. The lendingmargin is defined as the difference between the interest rate on loans and a (specified) reference interest rate.The deposit margin is defined as the difference between the interest rate on deposits and a (specified) referenceinterest rate.The terms ��deposit rate spread,�� loan rate spread�� and ��margin spread�� are used for the difference between therespective interest rates or margins in individual countries and the respective average interest rates or margins inthe euro area. Unless specified differently elsewhere, the term �average�� refers to the weighted average of interestrates in business with nonfinancial corporations and households, weighted by the volumes of the stock of loansoutstanding and the stock of deposits taken, respectively.

3 Our calculations for all countries except Poland in this analysis are based on the recently launched MFI(monetary financial institutions) interest rates statistics (MIR) published by the ECB and the national centralbanks of the EU Member States. For Poland we relied on the interest rate statistics published by the Polishcentral bank, which strongly resembles the MIR. Nevertheless, due to lacking data on some outstandingvolumes, we could not calculate some (disaggregated and aggregated) lending margins and lending marginspreads for Poland. MIR statistics represent harmonized data with respect to coverage and definition. Never-theless, some differences still exist in the range of publicly available indicators. In order to make cross-countrycomparisons possible, we therefore restricted our sample to deposits with agreed maturity as data for overnightdeposits, deposits redeemable at notice and repurchase agreements were not available for several countries. Forthe same reason, we also excluded overdrafts from the interest rate statistics on loans. Since we are interested inthe impact of interest rate margins on banks� profitability, we focus on interest rates on outstanding volumesrather than on new business volumes.

Table 1

Lending and Deposit Rates and Overall Margins, March 2004

CZ HU PL SK AT EU-12Percentage points

Average lending rate 5.9 14.9 9.5 7.7 4.5 5.1Average deposit rate 1.9 9.9 3.0 3.9 2.3 2.6Average overall margin 4.0 5.0 6.5 3.8 2.2 2.4

Household lending rate 8.2 15.7 12.0 8.6 5.3 5.6Household deposit rate 2.0 9.7 3.0 3.4 2.4 2.7Household overall margin 6.2 6.0 9.0 5.2 2.9 2.9

Nonfinancial corporate lending rate 4.3 14.1 7.1 7.1 4.2 4.5Nonfinancial corporate deposit rate 1.6 10.8 2.9 5.1 2.1 2.5Nonfinancial corporate overall margin 2.7 3.4 4.2 2.0 2.1 2.0

Source: ECB, OeNB, national central banks.

Financial Stability Report 8 61�

Credit and Deposit Interest Rate

Margins in Four New EU Member States

nonfinancial corporations) in theCzech Republic increased from 3.1 per-centage points at the start of the timeseries in January 2001 to 4.0 percent-age points in March 2004. This wascaused by a sharper decline in depositrates than in loan rates, especially inbusiness with households, for whichthe overall margin rose by 1.4 per-centage points. The overall corporatemargin rose by 0.6 percentage point.Since January 2003, the average over-all margin has declined by 0.1 per-centage point, as the average loan ratehas fallen slightly more strongly (—0.4percentage point) than the average de-posit rate (—0.3 percentage point).The magnitude of the decline wasbroadly equal for nonfinancial corpo-rations and households.

By contrast, in Poland, the averageoverall margin tumbled from 7.9 per-centage points at the start of the timeseries in March 2002 and a peak ofabout 8.5 percentage points in No-vember 2002 to 6.7 percentage pointsin April 2004. Since January 2003 theaverage overall margin has declined by0.3 percentage point, as interest rateson loans fell more strongly (—1.5 per-centage points) than rates on deposits(—1.2 percentage points). This wasattributable to a decline in the house-hold overall margin (—1 percentagepoint), while the overall margin inthe corporate sector remained stable.

For Hungary, the MIR statisticshave been available only since the startof 2003. From January 2003 to April2004, the average overall margin de-creased by 0.9 percentage point to5.0 percentage points, as the increasein deposit rates since the second halfof 2003 outpaced the increase in lend-ing rates. Thereby, the householdoverall margin declined significantly

more strongly (—3.2 percentagepoints) to 6.1 percentage points. Thiswas primarily caused by the decline ininterest rates on loans to this sector,which contrasted with the increase inother lending rates. The corporateoverall margin narrowed by 1.3 per-centage points.

Comparing the development dur-ing the 15-month period from January2003 on in those three countries ofthe NMS-4 for which time series areavailable (i.e. the Czech Republic,Hungary and Poland), the followingmain features emerge.4 First, theoverall margin decreased, albeit onlyslightly in the Czech Republic andPoland. The strongest decrease tookplace in Hungary, despite an increasein the general interest rate level.Second, the decrease in the overallmargin resulted from a stronger down-ward or weaker upward movement ofloan rates rather than from the respec-tive movement of deposit rates.

What Does the PastDevelopment in theEU-15 Tell Us?Due to the lack of historical MIRstatistics for the EU-15, it is difficultto draw firm conclusions about apotential systemic convergence ofmargins within the EU-15 or the euroarea. A rough estimate can be madeon the basis of the retail interest ratestatistics, which was replaced by theMIR statistics. Nevertheless, the factthat the retail interest rate statisticshad not been harmonized hinderscross-country comparisons. More-over, due to differences in coverageand definitions between the MIR andthe retail interest rate statistics, adirect comparison of the two is notpossible, either.

4 For Slovakia, the MIR statistics cover the period since the beginning of 2004 only.




Retail interest rate data suggeststhat interest rate margins decreasedin several EU countries in the courseof the 1990s, when margins in periph-eral countries like Greece or Portugalgradually approached the EU average.A study prepared by the ECB in 2000comes to the conclusion that lendingmargins5 fell significantly in the EU-15 between 1997 and 2000, whichrepresented the continuation of a lon-ger trend observed throughout the1990s. However, the widening of de-posit margins often compensated forthis effect. The article identifies thefollowing factors which contributedto the decline in banks� overall mar-gins from 1997 to 2000:— the tightening of pricing condi-

tions owing to increased competi-tion and deregulation;

— the impact of falling interest ratelevels (deposit rates cannot fallbelow zero);

— the flattening of the yield curve(bank loans tend to have longermaturities than deposits);

— the shift towards charging explicitfees for services connected withbank accounts (instead of incor-porating them into interest rates);

— the securitization of loans (lowercapital costs);

— the reduction of operating costs;— fiscal or regulatory changes allow-

ing banks to engage in more com-petitive pricing.The above factors may have been

or may in the future become relevantfor the banking sector in the NMS-4.In addition, in these countries thegradual improvement in asset quality

may have already contributed to lowerlending margins and will most likelycontinue to do so in the future.

Decomposition in AdjustedLoan and Deposit RateSpreadsIn the following, we decompose theaverage overall margins and marginspreads observed in the NMS-4 intoseveral components. In a first step,we assess to what extent loan anddeposit rates in the NMS-4 differ fromthose in the euro area and into whichcomponents these spreads can bedecomposed.

We found that the average loanrates in the NMS-4 were between5.9% (Czech Republic) and 14.9%(Hungary) in March 2004. Hence,they stood above the euro area averageby between 0.8 percentage point(Czech Republic) and 9.9 percentagepoints (Hungary) in that month. Thedifference between the highest andthe lowest value was thus 9 percentagepoints. These spreads against the euroarea are supposed to compensate theinvestor for differences in (1) sover-eign risk; (2) foreign exchange risk;(3) bank-related risk; (4) client-related risk and (5) other risks, forexample liquidity risk.

On the basis of the differencebetween five-year local currency-denominated government bond yieldsin the NMS-46 and the benchmarkeuro area government bond yield,we approximated the risk premium forthe combined sovereign and the for-eign exchange risk (�country-specificrisks��). Obviously, bond yield spreads

5 The lending margin refers to the margin between the average lending rate and a reference rate, while thedeposit margin is the margin between the average deposit rate and a reference rate.

6 The five-year maturity was chosen to roughly match the maturity of the average lending rates. Due to dataunavailability in Slovakia, we used the three-year yield spread since August 2003 and the two-year yieldspread for the period before.




are also affected by liquidity risk.Nevertheless, disentangling this com-ponent would go beyond the scopeof this analysis, and in the followingwe leave this risk factor out of consid-eration.

Five-year yield spreads ranged be-tween 0.4 percentage point (CzechRepublic) and 6.0 percentage points(Hungary) in March 2004. Taking intoaccount that the premium for puresovereign risk (estimated by thespreads on sovereign eurobondsdenominated in euro, Euro-EMBIGlobal) ranged between 15 and 45 ba-sis points, the five-year yield spreadsseem to mostly offset foreign ex-change risk. Stripping this compensa-tion for sovereign and foreign ex-change risk off the average loan rateleft us with adjusted lending rates be-tween 5.3% and 5.9% in the CzechRepublic, Poland and Slovakia. InHungary this adjusted rate amountedto 8.9%.

We took the difference betweenthe five-year swap rates and the five-year government bond yields7 as aproxy for additional commercial bank-specific risks compared with the sover-eign risk in individual countries. Nev-ertheless, this comparison must betreated with some caution. For exam-ple in Hungary, the swap rate has beenmodestly below the bond yield sinceearly 2003. Partly, this probably mir-

rors a difference in credit risk be-tween Hungarian banks and the Hun-garian sovereign, as a lot of Hungarianbanks are subsidiaries of AAA-ratedforeign banks. Partly, it may also be at-tributable to maturity mismatch, todifferent liquidity and price transpar-ency as well as to calculation differen-ces (day count) between the swap rate(spread) and the government bondyield (spread). In general, the differ-ences between swap rates and govern-ment bond yields were very low in theNMS-4, ranging between —0.1 per-centage point and +0.2 percentagepoint. Comparing these compensa-tions for additional bank risks in theNMS-4 with that in the euro area(+0.2 percentage point) leaves uswith the compensation for relativebank risk in the NMS-4 (comparedwith the euro area). Given that inHungary, Poland and Slovakia thecompensation for bank risk (com-pared with the sovereign risk) is zeroor even negative, the relative compen-sation (compared with the euro area)is negative. Hence, stripping this fac-tor off the adjusted (i.e. for sovereignand foreign exchange risk) lendingrates leads to different results. In theCzech Republic, this adjustment leadsto a roughly unchanged lending rate,while in the other three countries, thiscorrection raises the adjusted lendingrates modestly.

7 Due to data unavailability in Slovakia, we used the three-year yields and swap rates since August 2003 andthe two-year yields and swap rates for the period before.




Through these adjustments wearrived at lending rates in the NMS-4 which are adjusted for sovereignrisk, foreign exchange risk and (withsome reservation) relative bank riskin the NMS-4. The adjusted lendingrates ranged from about 5.5% in theCzech Republic, to the modestlyhigher level of 6.1% in Poland andto the significantly higher level of9.1% in Hungary in March 2004. Bycomparison, the average loan rate inthe euro area was 5.1%.

The resulting adjusted lending ratespreads should compensate for client-related risks, other country-specificfactors (for example differences infinancial supervision rules, tax treat-ment, etc.) or differences in marketstructure (for example the level ofcompetition). Obviously, in Hungary,Poland and Slovakia the adjusted lend-ing rate spreads are far less ��impres-sive�� than the unadjusted spreads.Thus, the differences between averagelending rates in the NMS-4 and theeuro area are to a large extent due

to the compensation for sovereignand in particular for foreign exchangerisk. In the Czech Republic, the dif-ference between unadjusted andadjusted spreads is smaller, but it stillamounted to slightly less than half ofthe unadjusted spread.

The adjusted lending rate spreadsamounted to between 0.5 percentagepoint and 1.0 percentage point inthe NMS-4 in March 2004, with theexception of Hungary. The Hungarianspread had been relatively stable sincethe beginning of 2003, hoveringmostly between 3.5 and 4 percentagepoints.8 At the same time, the qualityof the banking portfolio in Hungary isrelatively good compared with theother NMS-4, the market structureis less concentrated than in the CzechRepublic and Slovakia (and roughlythe same as in Poland) and there areno major legislative or supervisoryfactors which put Hungarian banks ata disadvantage compared with theother three countries. Therefore, thehigh adjusted lending rate spread

Table 2

Loan Rates Adjusted for Country-Specific and Bank Risk

CZ HU PL SK AT EU-12

Percentage points

Average loan rate 5.9 14.9 9.5 7.7 4.5 5.1Average loan rate spread against EU-12 0.8 9.9 4.4 2.7 —0.5 . .

Average loan rate less sovereign and foreignexchange risk (five-year bond yield spread) 5.5 8.9 5.9 5.3 . . . .

Average loan rate less sovereign and foreignexchange risk (five-year bond yield spread)against EU-12 0.5 3.8 0.8 0.3 . . . .

Average loan rate less sovereign, foreignexchange and additional bank risk 5.5 9.1 6.1 5.6 . . . .

Average loan rate less sovereign, foreignexchange and additional bank risk against EU-12 0.5 4.1 1.0 0.5 . . . .


8 One factor that might partly explain the relatively high adjusted lending rate spread in Hungary in March2004 is based on the fact that a (possibly sizeable) part of the outstanding volume of loans, which is taken asthe basis for calculating the (weighted) average loan rate, consists of short-term loans and of long-term loanswith variable interest rates. This fact combined with the marked inversity of the yield curve in Hungary (asopposed to flat or upward sloping yield curves in the other countries in March 2004) may have contributed to arelatively higher adjusted lending rate (i.e. after adjustment by the five-year bond yield spread). However, thisexplanation does not help to explain the relatively higher adjusted lending rate in the first half of 2003.




may be related to the high absolutelevel of interest rates (which lowersthe margin in relative terms) or signalexcessive return potential. It may alsoresult from the fact that governmentbond yields may not fully reflect do-mestic inflation expectations as theyare influenced by the widespread par-ticipation of foreign investors in theHungarian bond market. For these in-vestors exchange rate expectationsplay a more important role than infla-tion expectations. By contrast, do-mestic inflation expectations can beassumed to dominate in banks� lend-ing business. As a result, to the extentthat the inflation risk premium ex-ceeds the exchange rate risk pre-mium, lending rates can be higherthan government bond yields, pushingup adjusted lending rates.

Cross-country comparisons, shouldalso take into account that our analysisfocuses on average lending rates. Thismeans that differences in the structureof the loan portfolio of individualcountries with respect to the maturitystructure along with differences in the

term structure, and to the weight ofbusiness with households and nonfi-nancial corporations have an impacton average loan rates and hence oncross-country differences in loan rates.9

Turning to the deposit side, averagedeposit rates in the NMS-4 ranged be-tween 1.9% and 9.9% in March2004. Hence, the average deposit ratein the Czech Republic was 0.8 per-centage point below the euro areaaverage, while in the other threecountries the spread amounted to0.3 percentage point and 7.3 percent-age points, respectively.

The adjusted deposit rates, i.e. therates after similar adjustments forsovereign risk, foreign exchange riskand relative bank risk premium asoutlined above for the lending rates,were between —0.5% in Poland and4.1% in Hungary in March 2004.

The resulting adjusted deposit ratespreads were negative in the CzechRepublic, Poland and Slovakia, as theadjusted deposit rates were lower thanthe euro area average and positive inHungary at 1.5 percentage points.

9 For example, the bias introduced by a maturity mismatch between five-year government bond yields acrosscountries (as a measure for country-specific risks) can be aggravated if the yield curve is upward sloping inone country and downward sloping in the other. Similarly, the lengthening of the average maturity of theloan portfolio in one country compared with another country will have a different impact on the average loanrate and the average loan rate spread, depending on differences in the term structure. Nevertheless, we believethat even the analysis of the aggregated data gives a good impression of the true situation. The analysis of thedisaggregated data (by individual loan categories and by sectors) would go beyond the scope of this stocktakingexercise. Moreover, the availability of data necessary for the calculation of disaggregated adjusted lending rates(in order to match the maturity of the loan portfolio with the maturity of the proxy for the country-specific andbank risks) would cause difficulties.

Table 3

Deposit Rates Adjusted for Country-Specific and Bank Risk


Percentage points

Average deposit rate 1.9 9.9 3.0 3.9 2.3 2.6Average deposit rate spread against EU-12 —0.8 7.3 0.3 1.3 —0.3 . .

Average deposit rate less sovereign, foreignexchange and additional bank risk 1.5 4.1 —0.5 1.8 . . . .

Average deposit rate less sovereign, foreignexchange and additional bank risk against EU-12 —1.1 1.5 —3.1 —0.9 . . . .





The relatively low or — in Poland —even negative adjusted deposit ratesmay be an indication of less fiercecompetition for deposits in theNMS-4 than in the euro area. Anotherexplanation for relatively low adjusteddeposit rates could be households�preference for bank deposits, as op-posed to ��more sophisticated�� formsof savings. Also, the structural liquid-ity surplus in the NMS-4 banking sys-tems, supported by a greater weightof equity — and in some countries ofnet foreign liabilities — in banks� over-all liabilities, may be a reason forNMS-4 banks� relatively relaxed atti-tude towards domestic deposits.Moreover, cross-country differencesin the maturity structure of depositstogether with differences in the termstructures as well as differences inthe weight of business with householdsand nonfinancial companies may causedifferences in average adjusted depositrates. Also, it should be noted that theadjustment of deposit rates for coun-try-specific (i.e. sovereign and foreignexchange) risks introduces some biasas well. The reason for this is thatthe average maturity of deposits isshorter than the five years which areused for the approximation of thecountry-specific risk.

The average overall margins in theNMS-4 ranged between 3.8 percent-age points (Slovakia) and 6.5 percent-age points (Poland) in March 2004,resulting in a span of 2.7 percentagepoints between the maximum andthe minimum.

These margins were above theeuro area average of 2.4 percentagepoints, delivering average overall mar-gin spreads between 1.4 percentagepoints (Slovakia) and 4.1 percentagepoints (Poland). The average overallmargin spread can also be seen asthe difference between the loan ratespread and the deposit rate spread.As average (adjusted) loan rate spreadswere clearly higher than the mostlynegative average (adjusted) depositrate spreads across the NMS-4, weobtained positive average overallmargin spreads of between 1.4 per-centage points (Slovakia) and 4.1 per-centage points (Poland). With the ex-ception of Hungary, the adjusted de-posit rate spreads contributed a largerpart to the average overall marginspreads than the adjusted loan ratespreads, as the former were to a largerextent negative than the latter werepositive.

Table 4

Summary of Adjusted Average Loan and Deposit Rates and Spreads

CZ HU PL SK

Percentage points

Average loan rate less sovereign, foreign exchange and additional bank risk 5.5 9.1 6.1 5.6EU-12 average loan rate 5.1 5.1 5.1 5.1

Average deposit rate less sovereign, foreign exchange and additional bank risk 1.5 4.1 —0.5 1.8EU-12 average deposit rate 2.6 2.6 2.6 2.6

Average overall margin 4.0 5.0 6.5 3.8EU-12 average overall margin 2.4 2.4 2.4 2.4

Average loan rate less sovereign, foreign exchangeand additional bank risk against EU-12 0.5 4.1 1.0 0.5

Average deposit rate less sovereign, foreign exchangeand additional bank risk against EU-12 —1.1 1.5 —3.1 —0.9

Average overall margin against EU-12 1.6 2.6 4.1 1.4

Source: ECB, OeNB, national central banks, Bloomberg.




Regarding the segmentation bysectors, the overall margin spreadswere especially high in business withhouseholds, suggesting a potentiallymore competitive business environ-ment with nonfinancial corporations.

Decomposition intoLending and DepositMargin SpreadsAn alternative way of looking at theoverall margin (spread) is to decom-pose the average overall margin intothe average lending margin and theaverage deposit margin. Subsequently,these margins can be compared withthe euro area average. The referencerates used for the calculation of lend-ing and deposit margins represent thereturn on alternative investment/financing. Therefore, we have chosenthe reference rates in a way that theirmaturity roughly matches the matur-ity of the average lending rate andthe average deposit rate.

Average lending margins, derived asthe difference between the average

loan rate and the country-specificreference rate, varied more widely,between 2.3 percentage points and5.9 percentage points in March2004. The average lending marginwas highest in Hungary and lowest inthe Czech Republic. By comparison,in the euro area, it was 1.8 percentagepoints. Due to their construction, theaverage lending margins �neutralize��country-specific risk factors (like sov-ereign and foreign exchange risk) andrelative bank risk factors.

The lending margin spreads takeinto account the spread between thereference rates both in the NMS-4and the euro area. Thus, the fact thatthe lending margins in the NMS-4were higher than in the EU-12 andlending margin spreads were hencepositive in the NMS-4 may be ex-plained by some of those factors whichwere identified for the decline ofbanks� margins in the EU-15 citedabove (for example, different intensityof competition, higher interest ratelevel, less widespread securitization).

In addition, differences in thecountry-specific mismatches betweenthe maturity structure of the average

loan portfolio and the reference ratesas well as the already noted differencesin the sectoral structure of the loan

Table 5

Overall Margins and Margin Spreads by Economic Sectors


Percentage points

Average overall margin 4.0 5.0 6.5 3.8 2.2 2.4Average overall margin spread 1.6 2.6 4.1 1.4 —0.2 . .

Household overall margin spread 3.3 3.1 6.2 2.4 0.0 . .Nonfinancial corporate overall margin spread 0.7 1.4 2.2 0.0 0.1 . .


Table 6

Average Lending Margins and Margin Spreads


Percentage points

Average lending margin 2.3 5.9 2.8 3.0 1.3 1.8Average lending margin spread 0.5 4.1 1.0 1.1 —0.5 . .





portfolio and in the term structure,which might be reflected to variousdegrees in lending rates and in the ref-erence rates, may provide some fur-ther explanation for the differencesin the lending margin spreads amongthese four countries.

Therefore, we also looked intolending margins and lending marginspreads at a more disaggregated level.A comparison of the Czech, Hungar-ian and Slovak markets (due to lack

of data, disaggregation was not possi-ble for Poland) showed that with afew exceptions lending margins inthe NMS-3 were higher than in theeuro area. Thus, the aggregated lend-ing margin spread that was calculatedon the basis of disaggregated data con-firmed the above analysis based on theaverage lending margin spread, withHungary showing the highest lendingmargin spreads.10

10 It should be noted that the relatively high average lending margin spread in Hungary in March 2004 might bepartly explained by the marked inversity of the reference yield curve (as opposed to flat or upward sloping yieldcurves in the other countries) combined with the fact that a (possibly sizeable) part of the outstanding volumeof longer-term loans carries variable interest rates. Similarly, the same factors may contribute to explain theincrease of the margins of the loans to the non-financial corporations with maturity.

Table 7

Disaggregated Lending Margins and Margin Spreads


Percentage points

Household loan for house purchase < 1y margin 3.1 4.2 . . 1.6 3.1 2.8Household loan for house purchase < 1y margin spread 0.3 1.3 . . —1.3 0.2 0.0Household loan for house purchase 1—5y margin 4.2 5.3 . . 3.2 1.4 2.1Household loan for house purchase 1—5y margin spread 2.2 3.2 . . 1.1 —0.7 0.0Household loan for house purchase > 5y margin 2.2 5.1 . . 2.1 1.0 1.4Household loan for house purchase > 5y margin spread 0.8 3.8 . . 0.7 —0.4 0.0

Household loan for consumption and other purpose< 1y margin 10.1 6.2 . . 5.8 6.1 6.1

Household loan for consumption and other purpose< 1y margin spread 4.0 0.1 . . -0.3 0.1 0.0

Household loan for consumption and other purpose1—5y margin 10.8 14.1 . . 9.6 3.1 4.5

Household loan for consumption and other purpose1—5y margin spread 6.3 9.6 . . 5.1 —1.4 0.0

Household loan for consumption and other purpose> 5y margin 5.9 6.0 . . 4.7 1.5 2.2

Household loan for consumption and other purpose> 5y margin spread 3.6 3.8 . . 2.4 —0.7 0.0

Nonfinancial corporate loan < 1y margin 1.6 2.4 . . 1.2 2.0 2.6Nonfinancial corporate loan < 1y margin spread —0.9 —0.2 . . —1.3 —0.5 0.0Nonfinancial corporate loan 1—5y margin 1.3 4.2 . . 2.6 1.8 1.3Nonfinancial corporate loan 1—5y margin spread 0.1 3.0 . . 1.3 0.6 0.0Nonfinancial corporate loan > 5y margin 1.2 5.3 . . 2.6 0.5 0.9Nonfinancial corporate loan > 5y margin spread 0.2 4.4 . . 1.7 —0.4 0.0





In all three countries, the lowestlending margin spreads were observedfor �loans to nonfinancial corporationswith a maturity of less than one year,��which have a considerable weight inthe banks� loan portfolios (between17% and 28%). Indeed, the lendingmargin spread on these loans wasnegative in all three countries.11

Again in all three countries, thehighest lending margin spread wasobserved for �loans to households forconsumption and other purposes witha maturity between one and fiveyears,�� which also have a considerableweight in the banks� loan portfolios(between 6.5% and 9%).

With respect to segmentation bymaturity, margin spreads in lendingto nonfinancial corporations increasedwith maturity, while margin spreadsin household lending were the highestin the medium (between one and five

years) maturity range, with the excep-tion of Hungarian household loansfor house purchases, for which thelongest maturity segment showed thehighest spreads.

More generally, lending marginswere higher in business with house-holds than with nonfinancial corpora-tions in particular in the Czech Re-public and Hungary and to a lesser de-gree in Slovakia. As the difference be-tween household lending margins andcorporate lending margins was by farmore accentuated in the NMS-3 thanin the euro area average, householdlending margin spreads also exceededcorporate lending margin spreads sig-nificantly. Indeed, in the Czech Re-public and in Slovakia, the corporatelending margin was already more orless at the euro area level, implying acorporate lending margin spread closeto zero.

Average deposit margins, derived asthe difference between the average de-posit rate and the country-specific ref-erence rate, varied between —1.0 per-centage point (Slovakia) and —3.2 per-centage points (Poland) in the NMS-4.These negative deposit margins (i.e.

the average deposit rate is below thereference rate) contrasted with a zeromargin in the euro area. This may bean indication of less fierce competi-tion for deposits in the NMS-4 thanin the euro area.

11 In addition, in Slovakia the margin spreads for �loans to households for house purchase with a maturity of lessthan one year�� and �loans to households for consumption and other purposes with a maturity of less than oneyear�� were negative. However, these two types of loans have only a very small share in the overall volume ofloans to households and nonfinancial corporations.

Table 8

Aggregated Lending Margins and Margin Spreads


Percentage points

Household lending margin (calculated from disaggregated data) 4.6 6.8 . . 3.9 1.9 2.1Nonfinancial corporate lending margin

(calculated from disaggregated data) 1.4 3.9 . . 1.9 1.1 1.5Average lending margin (calculated from disaggregated data) 2.7 5.4 . . 2.8 1.4 1.8

Household lending margin spread(calculated from disaggregated data) 2.5 4.7 . . 1.8 —0.2 . .

Nonfinancial corporate lending margin spread(calculated from disaggregated data) —0.1 2.4 . . 0.4 —0.4 . .

Average lending margin spread (calculated from disaggregated data) 0.9 3.6 . . 1.0 —0.5 . .





However, as noted above in con-nection with the average lending mar-gin, some residual specification bias inthe construction of the reference ratemay have remained in place, despitethe country-specific selection of thereference rates.

Data disaggregation showed thatdeposit margins in the NMS-4 tendedto be larger (i.e. more negative: de-posit rates below the reference rate)than in the euro area. Thus, the aggre-gated deposit margin spread that wascalculated on the basis of disaggre-gated data confirmed the above analy-sis based on the average depositmargin spread. In particular, Polandwas confirmed as the country withthe largest negative deposit marginand deposit margin spread.

Looking at table 10, the moststriking observation is the positivedeposit margins, in particular forlong-term corporate deposits (in theCzech Republic, Slovakia, Austria

and the euro area) but also to a sig-nificant extent for long-term house-hold deposits in Slovakia. On the basisof opportunity cost considerations,positive deposit margins can be re-garded as a rather exceptional phe-nomenon. Apart from potential dataissues involved, one possible explana-tion for the positive corporate depositmargins may relate to the fact that formost banks customer relationships arepredominantly lending relationships.Thus, banks may be willing to incurlosses (calculated on the basis of theopportunity cost concept) in thelong-term deposit business with theircorporate clients, as this may accountfor only a minor part of their overallcustomer relationships and help fosterprofitable relations in lending. For theeuro area, one possible explanationfor positive deposit margins of long-term corporate deposits may also bea possibly still underdeveloped cross-country interbank lending business.

Table 9

Average Deposit Margin and Margin Spread


Percentage points

Average deposit margin —1.2 —1.7 —3.2 —1.0 —0.3 0.0Average deposit margin spread —1.1 —1.6 —3.2 —1.0 —0.3 . .


Table 10

Disaggregated Deposit Margins and Margin Spreads

CZ HU PL SK AT EU—12

Percentage points

Household deposit < 2y margin —1.3 —1.5 —2.9 —2.1 —0.4 —0.2Household deposit < 2y margin spread —1.1 —1.4 —2.8 —1.9 —0.2 0.0Household deposit > 2y margin —0.8 —3.4 —3.1 0.7 0.1 0.1Household deposit > 2y margin spread —0.9 —3.4 —3.2 0.6 0.0 0.0

Nonfinancial corporate deposit < 2y margin —0.7 —0.7 —3.0 —0.3 —0.2 0.0Nonfinancial corporate deposit < 2y margin spread —0.7 —0.6 —3.0 —0.3 —0.2 0.0Nonfinancial corporate deposit > 2y margin 0.5 —3.6 —2.5 0.5 0.5 0.9Nonfinancial corporate deposit > 2y margin spread —0.4 —4.5 —3.4 —0.5 —0.4 0.0





More generally, deposit marginswere more negative in business withhouseholds than with nonfinancialcorporations in particular in Hungaryand Slovakia and to a lesser degree inthe Czech Republic. This was true inparticular for deposits with a maturityof less than two years, which accountfor the bulk of total deposits of bothsectors in the NMS-4. As there washardly any difference between house-hold and corporate deposit marginsin the euro area, the differencebetween the sectoral deposit marginsin these three countries is reflectedalso in the deposit margin spreads.By contrast, in Poland, the two busi-ness lines offered roughly the samedeposit margins and deposit marginspreads.

Turning again to the overall margin,the average overall margin can becalculated as the difference betweenthe lending margin and the depositmargin. As the average lending mar-

gins were clearly positive throughoutthe NMS-4 and the euro area, whilethe deposit margins were negative inthe NMS-4 and about zero in the euroarea, we obtained positive averageoverall margins between 3.5 percent-age points (Czech Republic) and 7.6percentage points (Hungary).12 Withthe exception of Poland, average lend-ing margins delivered a larger contri-bution to the average overall marginsthan average deposit margins, as inthe NMS-4 (except Poland) the for-mer were to a larger extent positivethan the latter were negative, and inthe euro area average deposit marginswere about zero. In Poland, the con-tribution of both components to theaverage overall margin was aboutequal.

However, the situation is differentfor overall margin spreads (calculated asthe difference between average lend-ing margin spreads and averagedeposit margin spreads) due to the

Table 11

Aggregated Deposit Margins and Margin Spreads


Percentage points

Household deposit margin (calculated from disaggregated data) —1.0 —1.7 —2.9 —1.4 —0.2 0.0Nonfinancial corporate deposit margin

(calculated from disaggregated data) —0.7 —0.8 —3.0 —0.3 —0.1 0.2Average deposit margin (calculated from disaggregated data) —0.9 —1.5 —3.0 —1.0 —0.2 0.0

Household deposit margin spread(calculated from disaggregated data) —1.0 —1.7 —2.9 —1.3 —0.2 . .

Nonfinancial corporate deposit margin spread(calculated from disaggregated data) —0.9 —0.9 —3.1 —0.4 —0.3 . .

Average deposit margin spread(calculated from disaggregated data) —1.0 —1.5 —3.0 —1.0 —0.2 . .


12 It should be noted that the thus calculated average overall margins differ from the average overall marginsdescribed earlier as the difference between (adjusted) loan and deposit rates shown in tables 1 and 4. Thisis due to the difference in the reference rates which are used for the calculation of the lending and the depositmargins. More precisely, the difference between the average overall margin as the difference between averageloan and deposit rates and the average overall margin as the difference between average lending and depositmargins is equal to the difference between the longer-term reference rate for the average loan rate and theshorter-term reference rate for the average deposit rate. Thus, in case of an upward sloped term structure thisdifference is positive (Czech Republic, Poland), while in case of an inverse interest rate curve it is negative(Hungary).




fact that in the euro area lendingmargins were clearly positive, whiledeposit margins were zero.13 In theCzech Republic and Poland, the aver-age deposit margin spreads accountedfor a larger portion of the averageoverall margin spreads than the aver-age lending margin spreads, as the for-

mer were to a larger extent negativethan the latter were positive. In Slova-kia, the contribution of the two com-ponents was roughly equal, while inHungary, the highly positive averageoverall margin spread stemmed pri-marily from the positive average lend-ing margin spread.

Main Findings of theDecompositionsThe following picture emerges formargins and margin spreads in theNMS-4 in March 2004:

(1) Country-specific risk factors(sovereign risk, foreign exchange risk)account for a large portion of theoverall average lending rate spreadbetween the NMS-4 and the euroarea. If the compensation for theserisk components is left aside, thespread is far less impressive, exceptfor in Hungary.

(2) The average deposit rate in theCzech Republic was below the euro

area average, and in Poland, it wasclose to the euro area average level.The extraction of the country-specificrisk premium from the average de-posit rate further enhanced the nega-tive spread against the euro area inthe Czech Republic and led to a neg-ative spread in Poland and Slovakiaas well. The relatively low or, inPoland, even negative adjusted depositrates may be an indication of lessfierce competition for deposits in theNMS-4 than in the euro area.

(3) In the NMS-4, the averageoverall margins were higher than inthe euro area. With the exception of

13 Similar to what was stated in the previous footnote for the average overall margins, it should be noted that alsothe thus calculated average overall margin spread differs from the average overall margin spread describedearlier and shown in table 4. The difference between the previous average overall margin spread and the averageoverall margin spread as the difference between lending and deposit margin spreads is equal to the differencebetween the country-specific difference between the loan reference rate and the deposit reference rate in theNMS-4 and the euro area-specific difference between the loan reference rate and the deposit reference ratein the euro area.

Table 12

Summary of Lending and Deposit Margins and Margin Spreads


Percentage points

Average loan rate 5.9 14.9 9.5 7.7 4.5 5.1Reference rate for average loan rate 3.6 9.0 6.7 4.8 3.2 3.2Average lending margin 2.3 5.9 2.8 3.0 1.3 1.8

Average deposit rate 1.9 9.9 3.0 3.9 2.3 2.6Reference rate for average deposit rate 3.1 11.6 6.2 4.9 2.7 2.7Average deposit margin —1.2 —1.7 —3.2 —1.0 —0.3 0.0

Average overall margin(difference of lending margin and deposit margin) 3.5 7.6 6.1 4.0 1.6 1.9

Average lending margin spread 0.5 4.1 1.0 1.1 —0.5 . .Average deposit margin spread —1.1 —1.6 —3.2 —1.0 —0.3 . .Average overall margin spread

(difference of lending margin and deposit margin spread) 1.6 5.7 4.2 2.1 —0.2 . .





Hungary, the adjusted deposit ratespreads contributed a larger part tothese average overall margin spreadsthan the adjusted loan rate spreads.

(4) This overall margin spread wasparticularly large in the business withhouseholds, suggesting a potentiallymore competitive business environ-ment with nonfinancial corporations.

(5) Similarly, business with house-holds in particular in the Czech Re-public and Hungary and to a lesser de-gree in Slovakia (no data for Poland)offered higher lending margins andlending margin spreads (comparedwith the euro area) than the corporatebusiness. Indeed, in the Czech Repub-lic and in Slovakia, the corporate lend-ing margin was already more or less ateuro area level, implying a corporatelending margin spread close to zero.The spread in lending margins wasthe highest for �loans to householdsfor consumption and other purposeswith a maturity between one and fiveyears.�� On the other hand, the spreadin lending margins for �loans to non-financial corporations with a maturityof less than one year,�� which have aconsiderable weight in the loan port-folio (between 17% and 28%), waseven negative.

(6) While the average depositmargins in the NMS-4 were negative(i.e. the deposit rates were lower thanthe reference rates), the average de-posit margin in the euro area wasabout zero. This may be an indicationof less fierce competition for depositsin the NMS-4 than in the euro area.

(7) In Hungary and Slovakia inparticular and to a lesser degree inthe Czech Republic, household de-posit margins and margin spreadsagainst the euro area were more neg-ative than the corporate deposit Mar-gins and margin spreads. The twobusiness lines offered roughly the

same deposit margins and marginspreads in Poland.

(8) With the exception of Poland,the average lending margins delivereda larger contribution to the averageoverall margins than the average de-posit margins. In Poland, the contri-bution of both components to theaverage overall margin was aboutequal.

(9) By contrast, in the Czech Re-public and Poland, the average depositmargin spreads accounted for a largerportion of the average overall marginspreads than the average lendingmargin spreads, as the former wereto a larger extent negative than thelatter were positive. In Slovakia, thecontribution of the two componentswas roughly equal, while in Hungary,the highly positive average overallmargin spread stemmed primarilyfrom the positive average lendingmargin spread.

(10) On the basis of disaggregateddata, the dominance of the lendingmargin spread in Hungary as well asa slight dominance of the deposit mar-gin spread in the Czech Republiccould be confirmed. (Due to the lackof disaggregated data, this comparisoncould not be made for Poland.)

Overall, the analysis of the availa-ble data suggests that the marginswere wider in the NMS-4 than inthe euro area in particular in businesswith households on both the lendingand the deposit side. This may be at-tributable to higher costs related tobusiness with households in theNMS-4 than in the euro area, forexample due to a more widespreaduse of labor-intensive banking ser-vices. On the other hand, lower com-petition in business with householdsin the NMS-4 compared with the euroarea may offer an alternative explana-tion.




Other Factors AffectingBanks� Net Interest IncomeApart from the interest rate marginsin business with nonfinancial corpora-tions and households, the net interestmargin (defined as net interest incomeas a percentage of average total assetsor average total equity in a certainperiod) in the NMS-4 may also beaffected by other factors. Thereby,specific factors in the NMS-4 on theasset side of commercial banks tendto reduce net interest income com-pared with the euro area average,while specific factors on the liabilityside tend to have a positive impact.

On the asset side, one specific fea-ture of the NMS-4 banking sectors isthat claims on the general governmentaccount for a higher percentage oftotal assets than on average in the euroarea. While this share was around10% in the euro area at the end of2003, it ranged between 15% (Hun-gary) and 26% (Slovakia) in theNMS-4. In Poland and the Czech Re-public, the share stood at around20%. As lending to the general gov-ernment tends to be less risky thanlending to other sectors, and henceusually carries lower interest rates,this factor can be expected to lowernet interest income in the NMS-4compared with the euro area.

Another characteristic of theNMS-4 is the structural liquidity sur-plus in the money market. Therefore,central banks are generally liquidityabsorbers. Since the interest rateon the liquidity-absorbing facilitiesmostly represents the benchmark orthe floor for interbank interest rates,investment in these central bank instru-ments usually carries lower interestrates (albeit at practically no counter-

party risk) than business with otherclients.

On the liability side, the differencein the share of deposits of domestic resi-dents other than the central governmentand monetary financial institutions be-tween the NMS-4 and the euro areais striking. While this share amountedto around 32% at the end of 2003 inthe euro area, it ranged between 51%(Hungary) and 66% (Slovakia) in theNMS-4. As we pointed out above,the adjusted deposit rate spreads anddeposit margin spreads in the NMS-4can be identified as a major sourceof the overall margin spreads. Thehigh share of domestic deposits intotal liabilities multiplies this effect.

Similarly, the share of capital andreserves in total liabilities is signifi-cantly higher in the NMS-4 than inthe euro area. While in the euro areathis share amounted to 5.8% at theend of 2003, it stood at 8.9% in Hun-gary, 11% in the Czech Republic,13.5% in Slovakia and around 16%in Poland. This enables commercialbanks in the NMS-4 to finance inter-est-bearing assets by equity financing— a relatively cheap form of financing— which adds to their overall profita-bility.

Another question is whether andto what extent banks in the NMS-4utilize the positive interest rate differen-tial between local and foreign cur-rency — in particular euro — interestrates. Borrowing in foreign currencyand lending in local currency at higherinterest rates would enhance profits,but also increase banks� exposure toforeign currency risk. Unfortunately,data on banks� overall net FX position14

are not readily available in all coun-tries. Czech banks had a small overall

14 The overall net FX position comprises both the on-balance and the off-balance net FX position.




net short foreign exchange position15

(0.6% of total balance sheet liabilities)at the end of 2003. By contrast, dataof the Hungarian central bank showedthat Hungarian banks in the pasttended to close their long balancesheet position by off-balance sheetitems. Indeed, Hungarian banks main-tained a small long (around 0.5% oftotal balance sheet liabilities) overallnet foreign exchange position overthe last year. Similarly, Polish bankstended to close their long balancesheet FX positions by short off-bal-ance sheet items. Therefore, theyhad a minimum overall net long for-eign exchange position of less than0.1% of total balance sheet liabilitiesat the end of 2003. Hence, the generalconclusion that banks in these coun-tries take advantage of the interestrate differential by incurring higherexposure to foreign exchange riskcannot be drawn.

Concluding RemarksIn this analysis we found that country-specific risk factors, in particularrelated to foreign exchange risk,account for a significant part of thelending and deposit rate spreads be-tween four new EU Member States(NMS-4) and the euro area. More-over, we found that interest ratemargins in the NMS-4 were indeedhigher than in the euro area, contribu-ting significantly to the favorable

relative net income development ofbanks. Margin spreads compared tothe euro area were particularly highin business with households on boththe lending and the deposit side.

We also found that in the NMS-4(with the exception of Hungary), de-posit margin spreads compared withthe euro area delivered a larger contri-bution to the overall margin spreadsthan lending margin spreads. Thismay be taken as an indication thatcompetition for (household) depositsis less fierce in the NMS-4 than inthe euro area and also less fierce thanin the lending business (to nonfinan-cial corporations). Nevertheless, ow-ing to a potential specification bias,conclusions from deposit marginsshould be drawn with some caution.

However, as the deepening of fi-nancial intermediation will fuel banks�appetite for financial resources, whilealternative forms of savings are likelyto gain significance, a more competi-tive environment can be expected inthe deposit business. Similarly, it canbe anticipated that EU accession,which has pushed the door wide openfor cross-border banking services, willincrease competition during the nextfew years. Thus, it can be expectedthat interest rate margins in theNMS-4 will converge towards euroarea levels over the medium to longterm, in line with the developmentsthat were observed for the EU-15.

ReferencesCeska na«rodnı« banka. MIR Statistics. Prague.

European Central Bank. 2000. EU Bank�s Margins and Credit Standards. Frankfurt/Main.

European Central Bank. MIR Statistics. Frankfurt/Main.

Magyar Nemzeti Bank. MIR Statistics. Budapest.

Narodna« banka« Slovenska. MIR Statistics. Bratislava.

Narodowy Bank Polski. Weighted Average Interest Rates on Deposits and Credits. Warsaw.

15 A short position is defined as excess liabilities over assets, a long position is defined as the opposite.




fsr_08_special_01_tcm16-23536

Documents

Transcript of fsr_08_special_01_tcm16-23536