Bank Corporate Governance, Beyond the Global Banking Crisis · 2011. 5. 17. · Jean Dermine* March...
Transcript of Bank Corporate Governance, Beyond the Global Banking Crisis · 2011. 5. 17. · Jean Dermine* March...
Bank Corporate Governance, Beyond the Global Banking Crisis
_______________
Jean DERMINE 2011/33/FIN
Bank Corporate Governance,
Beyond the Global Banking Crisis
Jean Dermine*
March 2011 The author acknowledges the comments of I. Dermine, E. Diecidue, H. Hamelin, U. Peyer, and Sir David Scholey.
Tentative. Not to be quoted. Comments welcome * Professor of Banking and Finance at INSEAD, Boulevard de Constance 77305 Fontainebleau
cedex, France. Ph: 33 (0)1 60 72 41 33; Email: [email protected] A Working Paper is the author’s intellectual property. It is intended as a means to promote research tointerested readers. Its content should not be copied or hosted on any server without written permissionfrom [email protected] Click here to access the INSEAD Working Paper collection
Bank Corporate Governance,Beyond the Global Banking Crisis
Abstract
Following up on the publication of the Walker Report (2009) in the United Kingdom,international organizations such as the Basel Committee (2010), the OECD (2010), and theEuropean Union (2010) have proposed guidelines to improve bank corporate governance and,more specifically, risk governance. These international reports vary widely on what the primeobjective of bank corporate governance should be, with one group recommending a shareholder-based approach, and the other a stakeholder-based one. Moreover, the focus of these reports isexclusively on risk avoidance, with little guidance as to how an acceptable level of risk shouldbe defined. Drawing on insights from economics and finance, this paper is intended to contributeto the debate on bank corporate governance.
Our four main conclusions are as follows. Firstly, the debate on bank governance should concernnot only the boards but also the governance of banking supervision with clearly identifiedaccountability principles. Secondly, since biases for short-term profit maximization are numerousin banking, boards of banks should focus on long-term value creation. Thirdly, board membersand banking supervisors should pay special attention to cognitive biases in risk identification andmeasurement. Fourthly, a value-based approach to risk taking must take into account theprobability of stress scenarios and the associated costs of financial distress. Mitigation of thesecosts should be addressed explicitly in the design of bank strategy.
These are gross figures as governments can hope to make a return on their1
investment.
Following up on the banking crisis and the Walker Report (2009) published in the United
Kingdom, international organizations such as the Basel Committee (2010), the OECD (2010),
and the European Union (2010) have proposed guidelines to improve banks’ corporate
governance and, more specifically, risk governance - all with a common objective: never again!
These international reports vary widely on what the prime objective of bank corporate
governance should be, with one group recommending a shareholder-based approach and the other
a stakeholder-based one. Moreover, the focus of these reports is exclusively on risk avoidance,
with little guidance as to how boards should define an acceptable level of risk. In this paper, we
contribute to the debate on bank corporate governance, drawing insights from economics and
finance.
The paper is divided into five sections. In Section 1, we remind readers of the significance of the
banking crisis, both in terms of private and public costs. Some ‘high level corporate governance
principles’ are identified in Section 2, and a discussion of the merits of the shareholder- and
stakeholder-based approaches follow. In the next three sections we discuss more applied issues:
problems with the identification of risk, issues in measurement of performance and design of
compensation schemes, and the identification of criteria to guide the board when defining an
acceptable level of risk. This is followed by four conclusions related to the need for governance
of financial supervision, a focus on long-term value creation, attention to cognitive biases in risk
identification and measurement, and a value-based approach to risk taking.
Section 1. The Financial Crisis and the Call for Better Bank Corporate
Governance
The state-led resolution of the 2007-2008 financial crisis has proven to be costly. Upfront
government financing amounts to 5.2% of GDP for advanced G20 economies (IMF, 2009). This
includes capital injection, purchase of assets and lending by Treasury, and central bank support1
with Treasury backing. If one adds to this the guarantees put up by Treasury, the figure rises to
24.9%. In emerging economies, up-front government financing amounts to only 0.4% of GDP.
2
These figures do not include the lost output and social costs for which the global recession is to
blame. In the third quarter of 2010, OECD-average youth unemployment represented 18.5% of
the labor force aged 16-24, an increase of 5.5% over the past three years (OECD, 2010). For the
OECD as a whole, the gross government debt-to-GDP ratio is projected to increase by more than
20 percentage points by 2011 (Furceri and Zdziencka, 2010). In the OECD, countries that host
both large and small banks can be affected (Dermine and Schoenmaker, 2010). As documented
in Table 1, the United Kingdom which hosts large banks relative to GDP (such as Royal Bank
of Scotland), has seen up-front government financing of 20% of GDP. Greece and the USA,
which host smaller banks relative to GDP, also face high bailing costs of respectively 5.4% and
6.7%.
The advent of a common shock (such as a bubble in the real estate market or a currency
devaluation) that simultaneously affects many smaller banks (a case of correlated default risk or
default clustering) explains why countries with smaller banks can face large bailout costs at a
time of crisis. Good corporate governance in banking is essential to both small and large
institutions. With regards to the private costs incurred by bank shareholders during the crisis, the
peak-to-trough evolution of bank share prices between June 2007 and February 2009 is reported
in Table 2. In the disastrous case of Royal Bank of Scotland, following its € 71 billion
consortium bid for ABN AMRO in 2007, its share price felt by 95%. Banks diversified in
emerging markets did better, as illustrated by the share price of HSBC. Although heavily exposed
to the US subprime crisis with its US consumer finance subsidiary Household International, it
felt by 45%, helped in part by diversification in Asian and Latin American markets.
The severity of the global crisis, the huge public costs of bailing out the banks, burgeoning
budget deficits, and the large losses incurred by private shareholders have prompted a chorus of
“Never Again”. Bank corporate governance had failed in several cases, a view reinforced by the
fact that banking crises seem to be a recurrent phenomenon, as reported in Table 3, with 11 major
financial crises observed over the last 30 years. It is the succession of financial crises which has
inspired the call for a review of corporate governance in the banking sector.
Section 2. Bank Corporate Governance, the’ High Level Principles’
Corporate governance has been defined in the following ways. According to the OECD
Principles of Corporate Governance (OECD, 2004 ; Mulbert, 2010), “Corporate governance
involves a set of relationships between a company’s management, its board, its shareholders and
3
other stakeholders. Corporate governance also provides the structure through which the
objectives of the company are set, and the means of attaining those objectives and monitoring
performance are determined. Good corporate governance should provide incentives for the board
and management to pursue objectives that are in the interests of the company and its
shareholders, and should facilitate effective monitoring.” Professors Schleifer and Vishny (1997)
offer a more succinct definition: “Corporate governance deals with ways in which suppliers of
finance to corporations assure themselves of getting a return on their investment.”
In a market-based economy, several types of corporate governance mechanisms may exist and
co-exist. For example, the board of a bank can consist exclusively of representatives elected by
shareholders. In countries such as Germany, it will also include representatives of employees.
In a cooperative structure, such as a credit cooperative, it may include representatives of clients:
borrowers and depositors. In several countries, members of the board of savings banks are locally
elected politicians. What is striking to observe from a reading of the international proposals for
better bank corporate governance is the division between two camps, representing a shareholder-
based view of governance and a stakeholder-based approach. According to the first view,
governance should serve the shareholders, the owners of the corporation. According to the
second view, corporate governance should serve both shareholders and stakeholders, including
depositors, employees, clients, taxpayers and society. Both views are discussed next.
The Walker Report (2009, p.23), which builds on the UK Companies Act, is resolutely based in
the shareholder-based camp: “The role of corporate governance is to protect and advance the
interest of shareholders through setting the strategic direction of a company and appointing and
monitoring capable management to achieve this.” But it does not imply that other stakeholders
are to be ignored. Annex 3 to the Walker Report (pp 135-136) includes a list of sound principles
of management:
“To promote the success of a company, directors must have regard, amongst other matters, to the
following six factors:
- The likely consequence of any decision in the long term- The interests of the company’s employees- The need to foster business relationships with suppliers, customers and others (fiduciaryresponsibility)- The impact of the company on the community and environment- The desirability to maintain a reputation (long-term franchise)- The need to act fairly between members of the company”
However, according to the Walker Report, a single goal should be pursued by the directors of a
4
A discussion of public failures in banking and the need for banking regulation and2
supervision, national and international, is available in Baltensperger and Dermine (1987) andDermine (2003). The traditional argument for bank regulation is the need to limit moral
corporation: to promote the welfare of the shareholders, the owners of the corporation.
In sharp contrast, the proposals for improved bank corporate governance drawn up by the Basel
Committee (2010) and the European Union (2010) favor the stakeholder view. For instance, the
Basel Committee’s Principles for Enhancing Corporate Governance (2010, pages 5 and 10)
state “ …how the board and senior management:
- Set the bank’s strategy and objectives- Determine the bank’s risk tolerance/appetite-Operate the bank’s business on a day-to-day basis-Protect the interest of depositors, meet shareholder obligations, and take into account theinterests of other recognized stakeholders.- Align corporate activities and behavior with the expectation that the bank will operate in a safeand sound manner, with integrity and in compliance with applicable laws and regulations.”
« In discharging these responsibilities, the board should take into account the legitimate interestsof shareholders, depositors, and other relevant stakeholders.”
While the Walker Report refers to a single objective for bank corporate governance, the Basel
Committee and the European Union refer to multiple objectives, that of serving the welfare of
shareholders, depositors, employees, clients, suppliers and society.
As mentioned above, different forms of corporate governance can prevail in a market-based
economy. The question is which form of corporate governance leads to the most efficient
outcome for society? The proposals of both the Basel Committee and the European Union have
been greatly influenced by the global financial crisis and the cost to society. The prevailing view
is that, as some banks have taken excessive risk leading to large losses, costly bailouts by
taxpayers, large budget deficits and increase in unemployment, pressure should be put on the
boards to take into account not only banks’ shareholders, but also depositors and society as a
whole.
While the public reaction and calls for changes to corporate governance are legitimate, it should
not prevent a dispassionate and lucid analysis of the merits of various forms of governance
mechanisms, not only at the level of banks, but also at the level of bank regulators and
supervisors. Indeed, the need for efficient banking regulation and supervision was identified
many years ago, and one mission of bank regulators and supervisors is the development of a2
5
hazard arising out of deposit insurance and the inability of depositors to assess bank riskiness.An additional argument brought out by the crisis is the need to internalize the systemic costsresulting from financial distress. Regulatory tools include control of management and rules ofconduct, such as level of capital, a list of permissible activities, and group structure (branchesvs. subsidiaries).
Exceptions include the head of banking supervision in Ireland, or the president of the3
central bank in Iceland.
sound and safe banking system. Our discussion starts with the governance of bank regulators and
supervisors, followed by a discussion of corporate governance by the boards of banks.
2.1 Governance of Bank Supervision
With regard to the solvency and stability of banking systems, what is needed, in our opinion, is
independence and accountability of banking supervisors and appropriate legal mechanisms to
privatize bank losses. Regulators with rules and legislation codify what is permissible. The
supervisors, based on custom, practice and judgement of the rules and regulations, control what
and how it should be done. The need for independence and accountability of banking supervisors
was addressed in Dermine and Schoenmaker (2010). Unlike CEOs of banks, very few heads of
national supervisory authorities have been asked to step down over the past three years, which3
inevitably calls into question the accountability and independence of banking supervisors. In
several countries in Central and Eastern Europe (ECB, 2006, 2007), for example, banks were
allowed to lend massively in foreign currency (mostly Swiss francs, euro, and yen) on the
individual mortgage market. This created a large source of systemic risk, as the devaluation of
the local currency would raise the default rate across the entire banking system.
Why was this source of systemic risk allowed to develop? It is not difficult to imagine that, for
many ministers, a strong real estate market was helping to buoy up the economy, employment,
real estate developers, and the public budget with increased tax receipts. It would require a very
brave bank supervisor to put a break on foreign currency lending, slowing down the economy and
hurting real-estate developers. Other examples include the Financial Sector Assessment Programs
(FSAP) undertaken jointly by the IMF and the World Bank or the EU stress tests. The FSAP had
been undertaken since 1999, yet they did not prevent the crisis. It is striking to observe that, in
the wake of the December 2010 Irish banking crisis, European bank supervisors have recognized
that the European stress tests conducted in July 2010 were too lenient. Given that it is not a lack
of regulations but poor enforcement by banking supervisors that contributed to the crisis (Levine,
6
It is symptomatic to observe that the Financial Services Authority in the UK did not4
publish the details of its investigation on RBS and wrote “We did not identify any instancesof fraud or dishonest activity in RBS senior individuals or a failure of governance on the partof the board.” Under public pressure, it announces a forthcoming new report that wouldinclude a discussion of any failings in the FSA’s approach (WSJ, 16/12/2010).
The use of a panel of experts may partly solve the issue of confidentiality.5
Confidentiality, a principle enshrined in legislation, implies that prudential supervisors arenot allowed to publish confidential information about individual banks. So, supervisorscannot publish successful actions: cases in which failure was averted because of promptintervention by the supervisor (Dermine and Schoenmaker, 2010).
2010), a reinforcement of banking supervisors’ accountability is a must.4
The need to evaluate bank supervisors calls for the development of measures of performance. In
the case of central banking, these are simpler to identify: inflation or inflation expectations. In
the case of banking supervision, they are more difficult as no single index of financial stability
is available. A comparison can be drawn with the evaluation of the risk department of a bank for
which no single index of bank riskiness exists. To evaluate performance, quantitative data (such
as spread on banks’ CDS contract or subordinated debt, and the probability of bank default
provided by external firms) will need to be combined with ‘soft’ data (such as evaluation of
regulation and supervision by a panel of experts). A clear linkage between results and
promotion/remuneration appear necessary to build the correct incentive structure.5
The second tool available to reinforce financial stability is to develop a legal mechanism that
forces debt holders to bear bank losses. It is only when debt holders are at risk that they spend
resources to analyze the risks taken by bank. Two such mechanisms have been proposed:
financing by bail in securities that can absorb losses on a going concern basis (such as equity and
contingent convertible bonds), or a swift bankruptcy resolution mechanism that can deal with
bank default and impose a ‘haircut’ on debt holders. The latter would be facilitated by ‘living
wills’, plans that facilitate a rapid resolution in a situation of financial distress (Avgouleas,
Goodhart and Schoenmaker, 2010).
In other words, appropriate mechanisms can be put in place to reduce the likelihood of a banking
crisis and the potential consequences for the public finances. If the governance of bank
regulation and supervision allows for independence and accountability, the debate on bank
corporate governance can focus on the efficiency of financial institutions and economic
development.
7
2.2 Bank Corporate Governance
Three arguments can be put forward for the shareholder-based view of corporate governance,
centering on organizational efficiency, entrepreneurial innovation and economic welfare.
The first refers to the efficiency of an organization. The pursuit of a single objective - the welfare
of shareholders- provides a clearer goal than the pursuit of multiple objectives. To use a biblical
reference, “ No one can serve two masters”, (Matthew 6: 24). Or, as the old adage goes, “By
pursuing multiple objectives, you will achieve none.” As it is difficult to develop an indicator of
aggregated stakeholders’ welfare, there is a danger of a lack of accountability of management
who might prefer to maximize their own welfare, for example with empire building (Tirole,
2006).
Michael Jensen (2001, p.38) develops the notion of Enlightened Value Maximization:
“It is a basic principle of enlightened value maximization that we cannot maximize the long-
term market value of an organization if we ignore or mistreat any important constituency. We
cannot create value without good relations with customers, employees, financial backers,
suppliers, regulators and communities.”
Shareholder value maximization is the scorecard objective, but to achieve this, common sense
tells us that proper care must be taken of stakeholders; in banking it is known as a fiduciary duty
vis-à-vis depositors and clients. Moreover, the useful maturity transformation performed by
banks with the financing of illiquid assets with short-term deposits (Diamond-Dybvig, 1983)
creates the risk of a run on the bank by depositors. To reduce the likelihood of a costly run, a
bank has an incentive to protect short-term depositors with a cushion of long-term securities
such as equity or subordinated debt that can absorb losses as a going concern. The protection of
short-term depositors is thus compatible with shareholder value maximization
The second argument refers to economic development and risk taking. Progress in our free-
market societies is due in part to private firms taking risk, in accordance with the Shumpeterian
view of the role of innovations by private firms. But risk taking implies necessarily the
occurrence of failure. If one of the objectives of corporate governance is to serve debtholders and
banking supervisors, then one should reduce risk to a minimum, avoiding actions that can lead
8
In theory, an increase of riskiness of asset can be accompanied by an increase in the6
deposit rate or the insurance premium to compensate depositors or deposit insurers, but thismight be difficult to do if terms have been fixed ex ante.
An example of security design is the proposal for COERC, a Call Option Enhanced7
Reverse Convertible (Pennacchi et al. 2010), a form of contingent convertible bonds that willact as efficient bail in securities in case of bank distress.
to failure. Ceteris paribus, a reduction of risk will increase the value of debt and reduce the
liability borne by the deposit insurance system.6
While risk taking may lead to economic progress, it may not always be compatible with serving
debtholders and deposit insurers. Numerous innovations have occurred in the banking domain
such as microfinance, lending to subprime borrowers, junk bonds issued by non-investment
grade firms, financial derivatives, guaranteed-funds, credit derivatives, and alternative investment
vehicles. In some cases, these have led to large losses and failures, but few would deny the
usefulness of derivatives in hedging risks or in security design.7
The third strand of argument, from welfare economics, is that under certain conditions (such as
competitive markets), the pursuit of shareholder value maximization leads to economic welfare
with the equalization of marginal revenue to marginal cost. Well-identified frictions such as
imperfect competition, imperfect information or distribution of income call for public
interventions. Similarly in banking, the fragile nature of the industry caused by the useful
transformation of short-term deposits into illiquid assets requires public intervention (deposit
insurance, central banks acting as lender-of-last-resort), but these frictions do not necessarily call
for a change in corporate governance.
We favor a dual governance system based on clear objectives and accountability: the governance
of banking supervision should insist upon a clear objective (stability of the banking system) and
accountability of supervisors, and the governance of banks should concern itself with the
maximization of the welfare of shareholders. Devolving the responsibilities of regulators to the
board of banks, as the proposals of the Basel Committee and European Union imply, could hurt
economic efficiency, innovation and development, and further extend the lack of accountability
of banking supervisors and ministers of finance. Understandably, the global crisis has generated
a vast amount of emotion. What is needed is an efficient regulatory/bankruptcy system to
promote stability and an efficient corporate governance system to promote economic
development and risk taking.
9
Under Basel 2 capital regulations, an optimistic assessment of the probability of8
default of a borrower or of a loss-given-default will reduce the reported weighted assets. Financial Times, 11/12/ 2009. 9
Having developed argument in favor of shareholders-based governance, it is important to address
concern about whether the pursuit of value maximization may lead to short-term objectives. In
theory, share prices should reflect the present value of future expected cash flows. So, the
monitoring of share prices would be one way to assess the expected long-term outcome of
strategic decisions. However, if because of imperfect information or inefficiency, the share price
is driven by short-term results, there could be an incentive to increase short-term reported profit
at the expense of the long-term well-being of the firm. In banking, there are numerous ways to
improve short-term reported profits. Firstly, an increase in leverage (debt/equity) can increase
Return on Equity and earnings-per-share. This higher leverage could be hidden by manipulating
the measure of risk-weighed assets commonly used to assess capital adequacy. Secondly,8
delaying the recognition of provisions on expected credit losses will improve reported profit.
Thirdly, higher credit risks can lead to large interest rate margin in the short term, with credit
losses realized subsequently. Fourth, a plain maturity mismatch (cash and carry trade) at a time
of a rising yield curve can report positive profits in the short term, followed by losses later on
when interest rates increase.
Although the extent to which financial markets can see through remains to be proven, it appears
that opacity is particularly prevalent in banking due to the complex nature of many transactions.
In Graph 1 we report the evolution of the share price of the Irish Anglo-Irish Bank, that
precipitated the economic crisis in Ireland. The bank pursued reckless growth in mortgage
lending, soon to be followed by competitors in a classical herding saga. As is apparent from the
graph, the stock market rewarded the short-term profit of Anglo-Irish Bank until the crisis broke
out in the Summer 2007. Martin Taylor, former CEO of the British bank Barclays expressed a
harsh judgment on his peers, compensation practice and the pursuit of short-term profits:
“Innumerate bankers were ripe for a reckoning. Any industry that pays out in cash (to employees)
accounting profits that are largely imaginary will go bust quickly”.9
In a world in which financial markets reward short-term reported profits, it is the responsibility
of the bank’s board to take care of long-term value creation, even it that means hurting reported
revenue and the share price in the short term. Executives should drive the business within
regulations in accordance with the strategy and manner (ethics/culture) set and supervised by the
board.
10
As the Roman poet Juvenal wrote in Satires : “Quis custodiet ipsos custodies” - Who will guard
the guards? Transposed to a corporate governance context, how does one ensure that board
members fulfill their mission? Some have observed that non-executive or independent members
have been negligent in their duties, frequently due to their lack of understanding of the businesses
for which they were responsible. The Walker Report (2009) addresses the issue with a discussion
of the stewardship of institutional investors. Since due to free-riding, it is unlikely that individual
investors will spend resources exercising control, one has to rely on long-term institutional
investors who should make public their intent to exercise their governance rights.
Having examined the ‘ high level principles of governance’, we now turn to three key issues in
bank governance: the identification of risk, financial compensation, and the balance between risk-
avoidance and risk-taking.
Section 3. Issues with the Identification of Risks
The case for improved bank corporate governance rests in part on the perception that risks taken
by banks were not identified properly by their boards. In this section, we look at the measurement
of risks by bankers and banking supervisors as it refers to four issues: identification of relevant
stress tests, a clear distinction between situations of risk and those of uncertainty, psychological
cognitive biases in risk measurement, and lessons from a 20-year experience with bankers using
a bank simulation.
Identification of relevant stress tests
A necessary condition for bank soundness is that it can survive and keep the going-concern value
of the franchise at a time of significant economic shock. Bank capital must able to absorb losses
at a time of a stress scenario (also referred to as tail risk). During the 2007-2009 banking crisis,
the United States attempted to reassure financial markets on the soundness of its banking system
by submitting them to stress testing in May 2009. In July 2010, the European Union disclosed
information on stress tests conducted on 91 banks (CEBS, 2010). All but seven banks passed the
test of a minimum 6% Tier 1 ratio, including the two large Irish banks Allied Irish Bank and
Bank of Ireland.
In the case of the July 2010 European stress test, an assumption was made of a decline in house
prices in Spain of 15%, at a time when The Economist indicators had reported a potential
overvaluation of 35%, based on a standard rent-to-value ratio. Moreover, it is troubling that the
EU stress tests evaluated the magnitude of potential losses at 1.1 % of risk-weighted assets of
11
In March 2011, the European Banking Authority (EBA) announced the publication10
of more demanding stress tests in June 2011 (FT, 3 March 2011).
banks (aggregate Tier 1 falling from 10.3% to 9.2%), while the Basel 2 capital regulation, which
measures risk with 99.9% confidence, evaluated them at 8 % of risk-weighted assets. Six months
later in December 2010, the Irish government called for EU/IMF support to help bail out its
banks. Proposals for producing super stress tests followed.10
In a similar manner, securitized vehicles such as Collaterized Debt Obligations (CDOs) were
created on a very large scale. In a securitized vehicle, assets such as mortgages are funded with
several types of ‘waterfall’ securities. Equity and mezzanine (subordinated) tranches absorb the
first loan losses, protecting the senior tranches. The art in securitization design is to select the
size of equity and mezzanine tranches needed to protect senior debt holders. Intuitively, the
higher the level of loan losses in cases of stress (a case of correlated defaults), the larger the
equity and mezzanine tranches should be. Securitization experts and rating agencies rely heavily
on mathematics of credit risk. In Figure 2, we present the loan loss probability distribution for
a case of constant probability of loan default of 2%, but with default correlation varying from
0.02 to 0.1. One can see that, as correlation increases from 0.02 to 0.1, the potential for large
losses increases very rapidly from 5% to 9%. Despite major empirical uncertainty with estimates
of credit default correlation, there was an explosion of securitization deals. In 2005, world-
leading academics had already expressed concerns with the lack of empirical knowledge on
default correlation (Das et al., 2005).
Both examples - the design of stress tests and the growth of the securitization market- underline
the need for independence in the design of stress tests, possibly with input from external
independent experts.
Risk and Uncertainty
Subprime lending, similar to the case of junk bonds with the financing of non-investment grade
corporations in the 1980s, was a great innovation. It allowed more risky borrowers to access
credit and home ownership. In a similar manner, microfinance allows borrowers to finance small
businesses. However, while these are useful innovations, one must recognize that it is difficult
to assess potential losses on these new ventures because, being new, few historical data are
available, unlike lending to wealthier prime customers for whom a large set of historical data are
available to evaluate the volatility of credit losses over a business cycle.
12
In such as situation, an explicit distinction between risk and uncertainty would seem useful
(Knight, 1921 and Keynes, 1921). In a situation of risk, the probability distribution of losses can
be identified with relevant data. In the case of uncertainty, the distribution of losses cannot be
measured, as the situation being new, no relevant data are available. This is not to say that
bankers should necessarily avoid situations of uncertainty; as entrepreneurs they should look at
new business opportunities. But it would seem that cases of uncertainty should receive very
special attention by the board of banks and banking supervisors, and that, at a minimum, the scale
of exposure should be limited until more information on risk becomes available. The fatal error
of quite a few banks was not, in our opinion, their decision to enter the subprime market: rather
it was rather their inability to limit the growth of this highly uncertain activity.
Psychological Biases in Risk Measurement
Experts in decision science (Hammond et al., 1998) have identified several cognitive biases with
the measurement of risks. A few of these include:
Anchoring: The brain appears to have a bias in ‘fixing’ an estimate close to a reported value.
The EU stress test offers an example. The design of stress testing was based on two steps: it first
reported the evolution of EU economies and banks based on a current forecast designed by the
European Commission, and then estimated a stress deviation from that initial forecast. The
initial forecast acted as an anchor which might have prevented the design of tougher stress tests.
Framing formulation: This refers to the way a risk measure is expressed. For example, a 99%
maximum loss of X, or a 1% chance that a loss will be higher than X. The first expression, the
99% maximum loss, can be interpreted by the bank as a maximum loss, while the second
expression calls to the attention that losses could be larger.
Availability: This looks at risk in relation to memory. Due to limited memory we may believe
that the recent past describe potential volatility, ignoring historical cases of higher volatility.
Memory is one source of availability. In general, events are judged more likely to the extent that
they are vivid or easily recalled.
As evidence of these cognitive biases mounts, it seems important for boards of banks and
banking supervisors to take these explicitly into account in measuring risk. This is one more
reason why the help of independent experts could help.
13
Lessons from 20-year experience with a banking simulation
Over the past 20 years, the author has observed senior bankers, banks’ board members and
central bankers acting as members of an Asset & Liability Committee (ALCO) in a banking
simulation setting, ALCO Challenge. Although a banking simulation is an imperfect
representation of reality, several lessons can be drawn from these experiences (Dermine, 2008):
1. Success can bring over-confidence and complacency. If a group succeeds with interest rate
forecasts over a few quarters, confidence in their ability to forecast increases, complacency sets
in, and much large positions are taken, leading eventually to large losses.
2. Controlling the consequences of extreme events. Related to the ‘anchoring’ bias discussed
earlier, bankers have a hard time to imagine an economic world different from the current one.
3. Over-confidence in numbers and graphs. Numbers, formulas and mathematics give a false
sense of comfort and precision in measuring risk, when the relevant questions should concern
the key assumptions driving the numbers. Board member should spend more time validating the
assumptions driving reports, rather than analyzing the results.
4. Relative positioning. In a competitive setting, the focus of bankers is more on relative
performance (how do I rank vis-à-vis competitors) than on absolute performance. This behavior,
characteristic of Western societies (Layard, 2005), can lead to excessive risk taking to outperform
the competitors.
5. Group decision making as opposed to individual decision making. Group discussion allows
a better, more complete identification of risks, leading to reduced risk taking.
6. Separate risk management function. Forcing a group to create a specific risk management
committee, separate from the trading unit, leads to more attention paid on risk at the expense of
profit opportunities, and reduced risk-taking.
To summarize this discussion of risk measurement, board members and bank supervisors should
be very careful with cognitive biases in risk assessment and wary of the supposed comfort given
by reports, graphs and risk figures. They should be more cautious about the underlying
assumptions driving these reports.
Section 4. Measurement of Performance and Compensation
All the above-mentioned international reports on bank corporate governance devote considerable
attention to the design of performance evaluation and compensation schemes. A view common
14
“Multiple surveys find that over 80 percent of market participants believe that11
compensation practices played a role in promoting the accumulation of risks that led to thecurrent crisis” (FSF, 2009). In their review of the empirical literature, Ferrarini andUngureanu (2010) report that it is far from proven that pay structures generally contributed toexcessive risk taking before the recent crisis.
Note that benchmarking can only be made if the level of riskiness of the loan can be12
identified ex ante.
among the public, the press and many politicians is that compensation schemes have been, in
part, responsible for excessive risk taking and the global banking crisis. 11
We begin by reviewing the reasons for the difficulty in measuring performance in banking and
then discuss the compensation mechanisms proposed. For example, in granting a loan, the end
result can be positive if the economy performs well, but negative in case of a recession. Similarly,
when an asset manager invests in shares, the results can be positive if the market goes up, and
negative if the overall market goes down. The possible outcomes resulting from taking a risky
decision are shown in Figure 3, which represents the probability distribution of economic profit,
that is, the profit of a transaction net of a cost of allocated equity. If one was able to observe ex
ante, at the time the decision is taken, the expected economic profit, this measure could be used
to evaluate performance and design a bonus scheme. However, observation of the complete
probability distribution ex ante will often be difficult. What is observed, ex post, is the outcome,
the result of the decision. If economy is doing well (or if the stock market goes up), there will be
a profit; but if the economy goes into a recession (stock market goes down), there will be a loss.
As a consequence, the observation of a realized economic profit does not allow us to identify
whether the result is due to good luck or to good management. Similarly, the observation of a
loss does not allow to observe if it is due to bad luck or bad management. Not only is it difficult
to evaluate superior performance, but a second related issue is that bonuses calculated on realized
ex post performance can create an incentive to increase the riskiness of an activity. If the outcome
is positive, a high bonus is paid, while if the outcome is negative, the loss is borne by the bank
or by tax payers in case of bailout. This is the well-known parallel between bonus design and call
options, known to increase in value with increased volatility.
In the asset management industry, the superior realized performance of an asset manager is
evaluated against a passive benchmark (a portfolio with comparable risk). In the same way, we
have argued that a similar technique could be used in lending, in comparing realized performance
to a benchmark corporate bond with similar risk (Dermine, 2009). An alternative to12
benchmarking in the asset management industry is to report the average performance of a fund
over several years, in the hope that the lucky and unlucky draws of nature will cancel out,
15
revealing the expected economic profit and true performance. The proposals for compensation
scheme designs in banking (for example, Financial Stability Board (2009), Basel Committee
(2010), Bebchuk (2010), European Union (2010)) are all inspired by a desire to reduce incentives
for risk taking and option-like compensation structures. It is hoped that linking compensation
to performance evaluated over several years will succeed in rewarding superior performance, not
just luck. Leaving aside measures taken to increase disclosure and transparency of senior
executive pay, the key measures on compensation proposed in international reports include:
- Reducing the part of the bonus paid in cash immediately and increasing the part deferred over
a minimum 3-year period.
- Including a clawback (malus) clause that allows the deferred bonus pool to be reduced if a loss
materializes later.
- Paying a significant part of the bonus in restricted shares that cannot be sold over a certain
retention period.
Let’s analyze the efficiency of this proposal. The deferral period with a clawback clause goes
in the right direction of evaluating average performance and avoiding the option-like payoffs
discussed earlier. Indeed, it will create a symmetry between gains and losses as the malus clause
will reduce the deferred bonus pool. However, the three-year period appears quite short in the
credit area, given that positive economic cycle lasts often much longer than 3 years. The
investment in bank shares is more problematic if, due to opacity or market inefficiency discussed
above, the share price is driven by short-term reported profits. This could in fact encourage
management to focus on short-term results at the expense of long-term value creation. To reduce
this risk a fairly long retention period should be enforced to ensure that share prices reflect long-
term value (Bebchuck,2010).
Beltratti and Stultz (2009) observe that banks which were operating under ‘good governance
principles’ with significant restricted equity stakes owned by management did not perform better
than other institutions. This has three interpretations: the first is that these banks were not aware
of the risk. The second is that the incentive scheme was not appropriate because the retention
period was too short. As shareholders, they had an incentive to take risk. The third is that motives
other than financial compensations were at work, such as a herding mentality in competing for
market share.
An alternative to payment of bonus in restricted shares is payment of bonus in restricted bonds.
One example of this is that of Barclays’s 2011 proposal to pay bonus in contingent convertible
16
Financial Times 25/01/1112
Note that the proposal to pay bonus in bonds defeats the objective pursued with the13
payment of bonuses in shares: the alignment of management’s interest with that ofshareholders.
FT 7 March 2005.14
bonds -bonds that convert automatically into shares if a certain level of distress (such as a
minimum capital ratio) is reached. Like regular bond holders who receive a fixed interest12
payment and only fall in value when the bank suffers, holders of contingent convertible bonds
would have an incentive to reduce the overall riskiness of the bank, aligning their interest with
those of depositors and debtholders. 13
In the debate on compensation, it is implicitly assumed that behavior and risk-taking is driven
mostly by compensation incentives. However, it is the author’s belief that, while some
individuals are driven mostly by financial incentives, sometimes other factors may be at work,
such as competition among large egos, the race to be in first position, or a tendency for
conformity and herding behavior. Ireland provides a good example when two large historically
prudent banks, Bank of Ireland and Allied Irish Bank, abandoned strict credit standards policy
to engage in market share fight with a reckless risk-taker Anglo-Irish Bank. In this compensation
discussion, one likes to quote the legendary investor Warren Buffet: ”It is only when the tide
goes away that one can see who has been swimming naked”.14
Difficulty in linking compensation to superior performance, the race for first position, and the
possible herding behavior of bank senior executives reinforce the call for a focus of the board
on long-term value creation, even if it can have a negative effect on share price performance in
the short term. A remarkable case is that of the major Canadian banks. Although not far from
Wall Street, they managed to avoid the worst of the mortgage crisis (Freeland, 2010).
Section 5 Risk Avoidance or Risk-Taking
It will always be possible to engineer a stress test, a tail risk that brings a bank down. So is there
any guidance as to how much risk is acceptable? Indeed, international reports on bank corporate
governance seem concerned with risk avoidance, providing no guidance on risk taking.
Conceptually, a long-term value-based approach should provide guidance on how much risk is
acceptable.
17
Long-term value creation implies that in evaluating a risky decision, such as an investment in
mortgage loans, an acquisition, or investment in shares, one should assess the discounted value
of expected short- and long-term cash flows. One would evaluate the probability of positive
outcomes, such as repayment of a mortgage loans or the liquidity available to finance an
acquisition, with the probability of less positive outcomes, such as default payment or
disappearance of liquidity. In the event of negative outcomes, one should include the additional
expected costs resulting from financial distress.
The concept of cost of financial distress needs to be clarified. In a recession it is normal to report
low profit and to observe a fall in the share price. Costs of distress are additional costs that arise
because the company is in a difficult situation. In the banking industry, the costs of financial
distress can be quite significant. First, due to the nature of its activities - borrowing short to lend
long- the short-term depositors may panic and run away. Second, an undervalued stock will raise
the cost of external finance (earnings dilution if the bank issues stock at an undervalued price)
and prevent investment such as an acquisition. A third type of cost to arise from financial distress
might be the need to sell assets at fire-sale price to meet a minimum regulatory capital ratio.
Finally, and specific to the European Union, government bail-out subsidies lead to remedies
(various divestments) imposed by the European Union’s competition authorities.
To identify the expected cash flows resulting from a risky decision it is useful to separate what
are reversible decisions from irreversible ones. A reversible decision entails a risky decision that
can be unwound fairly rapidly. One example would be a position in a stock or in currency traded
on liquid markets. In this case the probability of a stress scenario or tail risk is likely to be small
over a short-term holding horizon. However, in the case of irreversible decisions, such as holding
of long positions in illiquid credit instruments or physical investment in a country, the probability
of a large shock becomes much higher over a long-term holding horizon. Indeed, Table 3 shows
that large shocks have occured fairly regularly over the past 30 years.
In the context of irreversible investments, it seems helpful for a bank to reduce the probability
of distress and the resulting costs of distress listed above. A first option could be to avoid risk,
but that implies missing value-creating activities. A second possibility is to fund the bank with
a significant amount of securities that can absorb losses on a going-concern basis (the bail-in
securities). Equity or contingent convertible bonds would be needed. This approach is feasible,
but likely to be expensive. A third option is to be diversified across products and geographies.
This ensures that a shock in one market can be absorbed by the diversified group. Examples of
diversified financial group that have, so far, done relatively well in the crisis are HSBC and
18
Note that this conclusion runs counter to finance wisdom according to which there is15
no benefits in financial diversification as shareholders can diversify themselves. Theargument is that financial diversification is one way to avoid the expected cost of financialdistress.
Santander. As mentioned earlier, despite severe losses in its US consumer finance subsidiary
Household Intl, HSBC’s share price fell by 45% in the crisis. In the case of Santander exposed
to the major recession in Spain and the United Kingdom, diversification in Latin America helped
to maintain a good performance. It is worth noting that these two groups have made acquisitions
recently.
It is interesting to observe the pendulum of opinion among bank analysts. When the economy
is doing well and memories of crisis fade, praise is heaped upon focused companies that operate
in one market. When the economy falls into a recession, the praise is reserved for diversified
groups that can sustain heavy losses. As large tail risk seems to occur at frequent interval, it is
recommended to run a diversified financial group. In a recent proposal on a rating methodology15
for banks, Standard & Poor (2010) refer explicitly to diversification, with an analysis of both the
benefits and the capacity of management to manage complexity.
Conclusions
The global economic crisis with its impact on unemployment and public debt-to-GDP ratios has
naturally generated anger and emotion. In its wake, several proposals on bank corporate
governance have been put forward by international organizations, such as the OECD, the Basel
Committee of Banking Supervision and the European Union.
Our discussion of bank corporate governance leads to four main conclusions. Firstly, the debate
on bank governance should not only focus on their boards but also on banking supervisors with
clearly identified accountability principles. Adequate governance of banking supervision should
have as objective a sound banking system, while adequate governance by board should focus on
the welfare of shareholders, the owners of corporation. As discussed earlier a single objective
will help to create a more efficient organization, which, to succeed, needs to be concerned about
its fiduciary duty to its customers and its relationship with employees and suppliers. Secondly,
since biases for short-term profit maximization are numerous in banking, boards are advised to
focus on long-term value creation. Thirdly, board members and banking supervisors should pay
special attention to several cognitive biases recognized in risk identification and measurement.
Fourthly, while the governance papers focus very much on risk avoidance, little guidance is
19
offered to evaluate an adequate level of risk. A value-based approach to risk taking must take into
account the probability of stress scenarios and the associated costs of financial distress, likely to
be quite significant in banking. In this respect, a useful distinction should be made between
reversible and irreversible investments, the latter being a more likely source of financial distress
costs. We have argued that diversification of risks is one way to reduce the probability of distress
and its associated costs. This will require the building up of an organization that can manage
complexity.
20
Countries with
large banks
wit
Equity/GDP > 4%
Bailout cost
(% of GDP)
Countries with
small banks
with
Equity/GDP < 4%
Upfront
Government
Financing
(% of GDP)
Austria 5.3% France 1.5%
Belgium 4.7% Germany 3.7%
Denmark 5.93% Italy 1.3%
Ireland 5.3% USA 6.3%
Spain 4.6% Greece 5.4%
Netherlands 6.2%
Sweden 5.8%
Switzerland 1.1%
United Kingdom 19.8%
Table 1: Bank Size and Public Bailout Cost (Dermine and Schoenmaker, 2010).
Source: IMF (2009), IMF (World Economic Outlook), Thomson One Bankers Analytics.
Public bailout costs refers to Upfront Government Financing. It includes capital injection, purchase of assets and
lending by treasury and central bank support provided with treasury backing. For Denmark, the source is Wall Street
Journal (2/7/09)
21
Bank Share Price Trough
as a % of Share Price Peak
15/06/2007- 15/02/2009
Royal Bank of Scotland 5%
Citigroup 5%
Commerzbank 8%
Lloyds TSB 10%
UBS 17%
Deutsche Bank 20%
BNP Paribas 30%
Credit Suisse 34%
BBVA 35%
Santander 40%
Standard Chartered 50%
HSBC 55%
Table 2: Stock Performance of Private Banks 2007-2009
Source: Thomson One Bankers Analytics (author’s calculation)
22
Date Event
1982 Global Emerging Market LATAM Crisis
1982 US S&L Crisis
1987 Stock Market Crash
1991 Global Real Estate Crises
1992 Japan Bank Losses, Scandinavia
1997 Asia Financial Crisis
1998 Russia
2000 End of Tech & Telecom Bubble
2001 Turkey
2002 Argentina Default
2007 Subprime Crisis
Table 3: A list of Significant Events that led to Severe Bank Losses
24
Figure 2: Loan Loss Distribution (probability of default = 2% ; correlation = 0.02)
Figure 2: Loan Loss Distribution (probability of default = 2% ; correlation = 0.1)
Source: Dermine (2009)
25
Economic profit
Figure 3: Probability Distribution of Economic Profit
Probability
Expected Economic Profit
26
Selected Literature on Bank Corporate Governance
Arnaboldi Francesca and Barbara Casu (2011): “Corporate Governance in European Banking”,mimeo, Cass Business School, City University, 1-26.
Avgouleas E., C. Goodhart and D. Schoenmaker (2010):” Living Wills as a Catalyst for Action”,mimeo, pp1-23.
Baltensperger Ernst and Jean Dermine (1987): "Banking Deregulation in Europe", EconomicPolicy, 4.
Basel Committee on Banking Supervision (2010): ”Compensation Principles and StandardsAssessment Methodology”, January, 1-31
Basel Committee on Banking Supervision (2010): “Principles for Enhancing CorporateGovernance”, Final, October, 1-34
Basel Committee on Banking Supervision (2010): “Range of Methodologies for Risk andPerformance Alignment Remuneration”, 1-53.
Bebchuk Lucian and Holger Spamann (2009): “Regulating Banker’s Pay, Harvard Law School,1-47
Bebchuk Lucian and Jesse Fried (2010): “How to Tie Equity Compensation to Long-termResults”, Journal of Applied Corporate Finance, Vol. 22 (1), 99-106, Winter
Bebchuk Lucian A. (2010): “How to Fix Bankers’ Pay”, Daedalus, 139 (4), forthcoming.
Beltratti A. and R. Stultz (2009): “Why did Some Banks Perform Better during the Credit Crisis?A Cross-country Study of the Impact of Governance and Regulation”, ECGI Working Paper 254.
Committee of European Banking Supervisors (2010): “High Level Principles for RiskManagement”, February, 1-6
Committee of European Banking Supervisors (2010): “Results of the 2010 EU-wide StressTesting Exercise”, 1-55.
Das Sanjiv .R., Darrell Duffie, Nikunj Kapadia and Leandro Saita (2005): “Common Failings:How Corporate Defaults are Correlated “ published in Journal of Finance, 2007, 62(1), 93-117.
Dermine Jean (2003): ”European Banking, Past, Present, and Future”, in The Transformation ofthe European Financial System (Second ECB Central Banking Conference), eds V. Gaspar, P.Hartmann, and O Sleijpen, ECB, Frankfurt.
27
Dermine Jean (2008): “Gerenciamento de Ativos et Passivos” (Asset and Liability Management,Lessons from 18 years of Experience with the ALCO Challenge Banking Simulation”), KPMGBusiness Magazine 11, March.
Dermine Jean (2009): Bank Valuation and Value-based Management, McGraw-Hill, New York.
Dermine Jean and Dirk Schoenmaker (2010): ”In Banking, is Small Beautiful?” (Journal of)Financial Markets, Institutions and Instruments, Vol. 19 (1), 1-19.
Diamond Douglas and Phil Dybvig (1983): “Bank Runs, Deposit Insurance and Liquidity”,Journal of Political Economy 91, 401-419.
European Central Bank (2006):”EU Banking Sector Stability”, November
European Central Bank (2007a):”EU Banking Sector Stability”, November
European Union (2010): “Green Paper on Corporate Governance in Financial Institutions andRemuneration Policies”, 1-19.
European Union (2010):Commission Staff Working Document on Corporate Governance inFinancial Institutions: Lessons to be drawn from the Current Financial Crisis, Best Practices”, 1-45.
Fahlenbrach Rüdiger and Rene Stultz (2010): “Bank CEO Incentives and the Credit Crisis”,Journal of Financial Economics,1-32.
Faulkender Michael, Dalida Kadyrhanova, N. Prabhjala, and Lemma Senbet (2010): “ExecutiveCompensation: An Overview of Research on Corporate Practices and Proposed Reforms”, Journalof Applied Corporate Finance, Vol. 22 (1), 107-118, Winter
Ferrarini Guido and Maria Cristina Ungureanu (2010): “Economics, Politics, and the InternationalPrinciples for Sound Compensation Practices. An Analysis of Executive Pay at European Banks”,ECGI working papers 169, 1-61.
Ferreira Daniel, Tom Kirchmaier and Daniel Metzger(2010): “Boards of Banks”, FinancialMarkets Group, London School of Economics, 1-46.
Financial Reporting Council (2010): The UK Corporate Governance Code, 1-40
Financial Stability Forum (2009): “FSF Principles for Sound Compensation Practices”, Basel, 1-16.
Freeland Chrystia (2010): “What Canada Can Teach the World”, FT.Com/Magazine, January30/31.
Furceri Davide and Aleksandra Zdzienicka (2010): “The Consequences of Banking Crises forPublic Debt”, OECD Working Papers 801, 1-26.
28
Hammond John, Ralph Keeney and Howard Raiffa (1998): “The Hidden Traps in DecisionMaking”, Harvard Business Review, September-October, 3-13
Jensen Michael C. (2001): “Value Maximization, Stakeholder Theory, and the CorporateObjective Function”, Journal of Applied Corporate Finance, 14, 8-21. Reprinted in Journal ofApplied Corporate Finance, 2010, Vol. 22 (1), Winter, 32-42.
Keynes John Maynard (1921): A Treatise on Probability.
Knight Frank (1921): Risk, Uncertainty and Profit, Boston, MA: Hart, Schnaffner & Marx,Houghton Miffin Cy.
Krozner Randall (2004): “ Economics of Corporate Governance Reform”,Journal of AppliedCorporate Finance, Vol. 16 (2-3), 42-50, Spring/Summer
Ladipo David and Stilpon Nestor (2009): “Bank Boards and the Financial Crisis. A CorporateGovernance Study of the 25 Largest European Banks”, Nestor Advisors Ltd, May, 1-123.
Layard Richard: ”Happiness: Lessons from a New Science”, Penguin Books.
Levine Ross (2010: “The Governance of Financial Regulation: Reform lessons from theRecent Crisis”, BIS Working papers 329, 1-24
Loderer Claudio, Lukas Roth, and Urs Waelchli (2010): “Shareholder Value: Principles,Declarations and Actions”, 1-53.
Mülbert Peter O. (2010): “Corporate Governance of Banks after the Financial Crisis-Theory,Evidence and Reforms”, European Corporate Governance Institute (ECGI), 1-40.
Nestor Advisors (2010): “Principles of Enhancing Corporate Governance - Response to theConsultative Document”, London, 1-7.
Nestor Stilpon (2010): “Avoiding Pitfalls in the New Bank Governance Framework”, TheBanker, August, p. 8.
OECD (2004): OECD Principles of Corporate Governance, Paris, 1-67.
OECD (2010): “Corporate Governance and the Financial Crisis”, OECD Steering Group onCorporate Governance”, Paris, 1-34
OECD (2010): “Off to a Good Start? Jobs for Youth, Paris, 1-5.
Pennacchi George, Theo Vermaelen and Christian C.P. Wolff (2010): “Contingent Capital: theCase for COERCs”, mimeo, 1-38.
Przemyslaw Koblut, Josep Tapies and Rafael Fraguas (2008): “Review of Selected GenerallyAccepted Corporate Governance Codes”, IESE business School Occasional paper 151, 1-11
29
Senior Supervisors Group (2009): “Risk Management Lessons from the Global Banking Crisisof 2008", www.newyorkfed.org/newsevents/news/banking/2009/SSG_report.pdf, 1-32.
Shivdasani Anil and Marc Zenner: ” Best Practices in Corporate Governance: What TwoDecades of Research Reveals”, Journal of Applied Corporate Finance, Vol. 16 (2-3), 29-41,Spring/Summer.
Shleifer and Vishny (1997): “A Survey of Corporate Governance”, Journal of Finance, Vol.LII (2), June, 737-783
Standard & Poor (2010): ”Banks: Rating Methodology”, 6 January, 1-73.
Taylor Martin (2009): “Innumerate bankers were ripe for a reckoning”, FT, 11 December
Tirole Jean (2006): The Theory of Corporate Finance, Princeton University Press.
Walker David (2009): “A Review of Corporate Governance in UK Banks and Other FinancialIndustry Entities”, 26 November, 1-175