Part I The Elements of Bank Financial Supervision, 2 Why Are Banks Supervised?
- Bank resolution and insolvency — Bank supervision — Credit risk — Basel committee on Banking Supervision
2.01 The starting point for a discussion of bank regulation is to ask why banks should be regulated to any greater extent than, say, car manufacturers. It would be reasonable (although cowardly) to avoid this issue altogether, and to begin from the proposition that they just are. However, it is helpful in understanding some of the issues which arise in the context of bank regulation to have a clear idea of what it is that the banking regulatory system is ostensibly trying to achieve.
2.02 The locus classicus of bank supervision is the statement of the Basel Committee as to the Core Principles for Effective Banking Supervision, produced by the Basel Committee in September 1997, reissued in a revised version in October 2006, and further revised in the light of the crisis in 2012. It is helpful to consider these here, since they set out in a short but comprehensive fashion the basic structure of a bank regulatory regime.
Principle 1—Responsibilities, objectives and powers:
2.03 An effective system of banking supervision has clear responsibilities and objectives for each authority involved in the supervision of banks and banking groups. A suitable legal framework for banking supervision is in place to provide each responsible authority with the necessary legal powers to authorise banks, conduct ongoing supervision, (p. 16) address compliance with laws and undertake timely corrective actions to address safety and soundness concerns.
Principle 2—Independence, accountability, resourcing and legal protection for supervisors:
2.04 The supervisor possesses operational independence, transparent processes, sound governance, budgetary processes that do not undermine autonomy and adequate resources, and is accountable for the discharge of its duties and use of its resources. The legal framework for banking supervision includes legal protection for the supervisor.
Principle 3—Cooperation and collaboration:
Principle 4—Permissible activities:
Principle 5—Licensing criteria:
2.07 The licensing authority has the power to set criteria and reject applications for establishments that do not meet the criteria. At a minimum, the licensing process consists of an assessment of the ownership structure and governance (including the fitness and propriety of Board members and senior management) of the bank and its wider group, and its strategic and operating plan, internal controls, risk management and projected financial condition (including capital base). Where the proposed owner or parent organisation is a foreign bank, the prior consent of its home supervisor is obtained.
Principle 6—Transfer of significant ownership:
Principle 7—Major acquisitions:
2.09 The supervisor has the power to approve or reject (or recommend to the responsible authority the approval or rejection of), and impose prudential conditions on, major acquisitions or investments by a bank, against prescribed criteria, including the establishment of cross-border operations, and to determine that corporate affiliations or structures do not expose the bank to undue risks or hinder effective supervision.
Principle 8—Supervisory approach:
2.10 An effective system of banking supervision requires the supervisor to develop and maintain a forward-looking assessment of the risk profile of individual banks and banking groups, proportionate to their systemic importance; identify, assess, and address risks emanating from banks and the banking system as a whole; have a framework in place for early intervention; and have plans in place, in partnership with other relevant authorities, to take action to resolve banks in an orderly manner if they become non-viable.
Principle 9—Supervisory techniques and tools:
Principle 10—Supervisory reporting:
Principle 11—Corrective and sanctioning powers of supervisors:
2.13 The supervisor acts at an early stage to address unsafe and unsound practices or activities that could pose risks to banks or to the banking system. The supervisor has at its disposal an adequate range of supervisory tools to bring about timely corrective actions. This includes the ability to revoke the banking licence or to recommend its revocation.
Principle 12—Consolidated supervision:
Principle 13—Home–host relationships:
2.15 Home and host supervisors of cross-border banking groups share information and co-operate for effective supervision of the group and group entities, and effective handling of crisis situations. Supervisors require the local operations of foreign banks to be conducted to the same standards as those required of domestic banks.
Principle 14—Corporate governance:
2.16 The supervisor determines that banks and banking groups have robust corporate governance policies and processes covering, for example, strategic direction, group and organisational structure, control environment, responsibilities of the banks’ Boards and senior management, and compensation. These policies and processes are commensurate with the risk profile and systemic importance of the bank.
Principle 15—Risk management process:
2.17 The supervisor determines that banks have a comprehensive risk management process (including effective Board and senior management oversight) to identify, measure, evaluate, monitor, report and control or mitigate all material risks on a timely basis and to assess the adequacy of their capital and liquidity in relation to their risk profile and market and macroeconomic conditions. This extends to development and review of contingency arrangements (including robust and credible recovery plans where warranted) that take into account the specific circumstances of the bank. The risk management process is commensurate with the risk profile and systemic importance of the bank.
Principle 16—Capital adequacy:
(p. 18) 2.18 The supervisor sets prudent and appropriate capital adequacy requirements for banks that reflect the risks undertaken by, and presented by, a bank in the context of the markets and macroeconomic conditions in which it operates. The supervisor defines the components of capital, bearing in mind their ability to absorb losses. At least for internationally active banks, capital requirements are not less than the applicable Basel standards.
Principle 17—Credit risk:
2.19 The supervisor determines that banks have an adequate credit risk management process that takes into account their risk appetite, risk profile, and market and macroeconomic conditions. This includes prudent policies and processes to identify, measure, evaluate, monitor, report, and control or mitigate credit risk (including counterparty credit risk) on a timely basis. The full credit lifecycle is covered including credit underwriting, credit evaluation, and the ongoing management of the bank’s loan and investment portfolios.
Principle 18—Problem assets, provisions and reserves:
Principle 19—Concentration risk and large exposure limits:
2.21 The supervisor determines that banks have adequate policies and processes to identify, measure, evaluate, monitor, report and control or mitigate concentrations of risk on a timely basis. Supervisors set prudential limits to restrict bank exposures to single counterparties or groups of connected counterparties.
Principle 20—Transactions with related parties:
2.22 In order to prevent abuses arising in transactions with related parties and to address the risk of conflict of interest, the supervisor requires banks to enter into any transactions with related parties on an arm’s length basis; to monitor these transactions; to take appropriate steps to control or mitigate the risks; and to write off exposures to related parties in accordance with standard policies and processes.
Principle 21—Country and transfer risks:
Principle 22—Market risks:
2.24 The supervisor determines that banks have an adequate market risk management process that takes into account their risk appetite, risk profile, and market and macroeconomic conditions and the risk of a significant deterioration in market liquidity. This includes prudent policies and processes to identify, measure, evaluate, monitor, report and control or mitigate market risks on a timely basis.
Principle 23—Interest rate risk in the banking book:
(p. 19) 2.25 The supervisor determines that banks have adequate systems to identify, measure, evaluate, monitor, report and control or mitigate interest rate risk in the banking book on a timely basis. These systems take into account the bank’s risk appetite, risk profile and market and macroeconomic conditions.
Principle 24—Liquidity risk:
2.26 The supervisor sets prudent and appropriate liquidity requirements (which can include either quantitative or qualitative requirements or both) for banks that reflect the liquidity needs of the bank. The supervisor determines that banks have a strategy that enables prudent management of liquidity risk and compliance with liquidity requirements. The strategy takes into account the bank’s risk profile as well as market and macroeconomic conditions and includes prudent policies and processes, consistent with the bank’s risk appetite, to identify, measure, evaluate, monitor, report and control or mitigate liquidity risk over an appropriate set of time horizons. At least for internationally active banks, liquidity requirements are not lower than the applicable Basel standards.
Principle 25—Operational risk:
2.27 The supervisor determines that banks have an adequate operational risk management framework that takes into account their risk appetite, risk profile and market and macroeconomic conditions. This includes prudent policies and processes to identify, assess, evaluate, monitor, report and control or mitigate operational risk on a timely basis.
Principle 26—Internal control and audit:
2.28 The supervisor determines that banks have adequate internal control frameworks to establish and maintain a properly controlled operating environment for the conduct of their business taking into account their risk profile. These include clear arrangements for delegating authority and responsibility; separation of the functions that involve committing the bank, paying away its funds, and accounting for its assets and liabilities; reconciliation of these processes; safeguarding the bank’s assets; and appropriate independent internal audit and compliance functions to test adherence to these controls as well as applicable laws and regulations.
Principle 27—Financial reporting and external audit:
2.29 The supervisor determines that banks and banking groups maintain adequate and reliable records, prepare financial statements in accordance with accounting policies and practices that are widely accepted internationally and annually publish information that fairly reflects their financial condition and performance and bears an independent external auditor’s opinion. The supervisor also determines that banks and parent companies of banking groups have adequate governance and oversight of the external audit function.
Principle 28—Disclosure and transparency:
2.30 The supervisor determines that banks and banking groups regularly publish information on a consolidated and, (p. 20) where appropriate, solo basis that is easily accessible and fairly reflects their financial condition, performance, risk exposures, risk management strategies and corporate governance policies and processes.
Principle 29—Abuse of financial services:
2.31 The supervisor determines that banks have adequate policies and processes, including strict customer due diligence rules to promote high ethical and professional standards in the financial sector and prevent the bank from being used, intentionally or unintentionally, for criminal activities.
2.32 The 2012 revision of these principles focused on four major areas—first, a significantly greater focus on corporate governance within banks; second, an obligation on supervisors to ensure that banks are appropriately prepared for resolution; third, an obligation for supervisors to assess bank risks in the context of the macroeconomic environment; and, fourth, the idea that supervisors should have higher expectations of banks which are globally systemically significant than for other banks.
2.33 These principles set out the general objectives which a bank supervisor should seek to achieve as regards its supervision of the management of the undertaking of each of its banks. However, although these objectives are all correct and desirable, they are all indications of general direction rather than statements of specific limits. Possibly more importantly, all of these principles would also be the objectives of the board and senior management of any well-run bank. Once the regulator has satisfied itself that the bank has systems in place to achieve these objectives, and that those systems are reasonably effective and in line with best industry practice, it is tempting for the bank regulator to assume that everything possible has been done.
2.34 This is not, however, the practice of bank regulators. The reason for this is that, for the reasons set out in paras 2.38 to 2.44, bank regulators seek to reduce the overall risk exposure of the banking system to a level below that which the directors of the bank would otherwise maintain. This is done by requiring the maintenance of a specified amount of regulatory capital.
2.35 It should be noted that the requirement by a regulator that a bank maintain a specific amount of capital relative to the risk which it runs is exactly the same mechanism as that which the bank management employs within the institution when imposing economic capital limits on individual businesses (see paras 1.40 to 1.41). In both cases the specification of a limited amount of capital is imposed as a proxy for the imposition of a risk limit—however, rather than the regulator specifying in detail the amount of each type of risk which the bank may run, the (p. 21) regulator in practice gives the bank a ‘risk budget’ by specifying the amount of capital which can be recognized and the mechanism by which capital may be allocated to risk. A bank which perceives itself as having access to particularly advantageous business is free to raise more capital, but until it has done so the regulator will not allow it to engage in that business unless it simultaneously decreases its risk exposure by reducing or exiting other businesses.
2.36 This approach makes capital requirements the primary tool of bank regulation. The idea that capital regulation is the proper study of bank regulators is deeply embedded within regulators, and in the same way that Sir Ernest Rutherford believed that ‘science is either physics or stamp-collecting’, many senior bank regulators quietly believe that bank regulation can be divided into quantitative supervision and box-ticking. It is fair to say that the usefulness of regulatory capital requirements is frequently rejected amongst academic commentators,1 who object that there is no academically validated proof to show that capital regulation has in fact reduced risk. Put simply, the argument is that the level of capital maintained by a particular bank is not necessarily a good guide to how likely it is to survive a financial crisis and, in particular, it is clearly not the case that banks with capital levels substantially above the regulatory minima will necessarily survive financial turbulence. Banks with low capital but good risk controls are more likely to survive a period of financial turbulence than banks with high capital levels but poor risk controls.
2.37 It should be noted, however, that it does not follow from the fact that capital requirements are justified that capital regulation in the form in which it has existed has always been useful. Most importantly, a regulatory capital regime which sets a requirement lower than that which the bank would otherwise maintain is a bureaucratic burden which is of no value. Second, an insufficiently risk-sensitive capital requirement is likely to create perverse incentives which outweigh any benefit which it may create (arguably the position of Basle I from about 2001 onwards). Thus we cannot simply say that capital regulation is a good thing per se—it is necessary to explain what the objectives of such regulation should be.
2.38 We begin with the drivers of bank capital. In reality, a bank’s decision as to how much capital it needs will be driven by the question of how to maximize its returns by minimizing its costs of borrowing. Put simply, if a bank has too little capital, lenders will demand an increased interest rate for lending to it. However, it is also possible for a (p. 22) bank to have too much capital—beyond a certain level, the decrease in borrowing costs will no longer be sufficient to compensate for the dilution of return on equity.2 Consequently, for any given bank, ordinary business analysis will produce an optimum level of capital which maximizes return on equity, and that is the capital level to which the bank will naturally gravitate. The question for regulators is as to why they should seek to require a bank to maintain more capital than this.
2.39 The primary reason for this is known as the ‘asymmetry’ problem. This is the fact that bank profits accrue to private shareholders but bank losses in practice fall either on their creditors or, if banks are rescued, on the public purse. There are a number of explanations for this, but the simplest (and the most convincing) is that although managers and directors of banks are highly incentivized to avoid their bank failing, they are not particularly incentivized to differentiate between small and large failures—bank managers will be unemployed, and their reputation destroyed, whether their bank fails owing £1 or £100bn. This results in the odd situation that the individuals within a bank are incentivized not to fail, but, if they do fail, to fail big.
2.40 In this context it may be noted that recent attempts to create mechanisms by which banks can in theory fail without having to be rescued by the taxpayer are not the end of this discussion.3 Because of the integral role which banks play in the economy, a state may well find itself confronted with the problem that the cost to its economy of restructuring the institution will be significantly greater than the size of the capital injection needed to avoid such restructuring. In such cases, even where a perfectly workable and effective mechanism exists for allowing a bank to fail, optimal public policy may still dictate a public rescue.
2.41 This phenomenon is not, however, unique to banks, and is to some extent present in all corporations. However, unlike most other forms of commercial firms, banks perform a public as well as a private function—the transmission of money and credit around the economy is a function which is similar in some ways to that which is performed by sewerage or electricity companies. If an electricity transmission company were to fail, government would take measures to ensure that it continued operating (analogous in our case to implementing a special resolution regime), and this is comparable to the steps which are taken in practice on a bank failure to ensure the continued operation of the payment and deposit-taking functions of the bank. In reality, government has little choice but to act to keep these operating come what may, since the maintenance of these services is part of the irreducible minimum of services which electors regard governments as created and elected to ensure. However, if an electricity company were to fail, there is no (p. 23) obvious reason why government should support its ordinary commercial creditors. Why should banks be any different? The answer is that creditors of banks are, in this regard, generally other banks. To allow one bank to fail in this way would create a knock-on impact for other banks, which would be likely to result in further failures and the necessity for further intervention. Further, since banks are the mechanism by which credit is supplied to the economy, the collapse of even one substantial bank would imperil short-term economic prospects, and the collapse of multiple banks would be a fortiori in this regard. Thus governments act to save troubled banks (and therefore support their existing creditors) not because they choose to but because there is no acceptable policy of choosing not to.
2.42 The issue, therefore, is as to how to reduce to an acceptable minimum the chance that government may be obliged to commit public resources to support failing banks, and the amount of public resources that may have to be committed in such a case. The primary technique must necessarily be to restrict the risks taken by a bank relative to its resources, and this is the task which capital regulation performs (another important part of the process of minimizing impact on the public purse is the establishment of ‘resolution regimes’—specialized insolvency regimes which permit government to intervene early and effectively in the affairs of a failing institution in order to minimize the harm resulting from its demise. These are considered at paras 2.59 to 2.62).
2.43 The aim of bank capital regulation is therefore not to increase bank capital per se, but to limit risk exposures relative to the bank’s capital. Since risk and reward are two sides of the same coin, if a bank is to be limited in the level of risk which it takes on for a given amount of capital, it is also necessarily limited in the return which it can make on that capital. Because banks ultimately compete for capital with other corporations, the perception is therefore that regulators are limited in their powers to restrict bank capital, since if a bank’s permitted return on capital is restricted below that of other types of business, no capital will be invested in the banking sector and the sector itself will disappear. However, this is not quite correct. Banks can be required to maintain any level of capital that the regulator chooses and may remain profitable (even super-profitable) provided that they have a sufficient degree of monopoly over at least one of the services which they provide. Banks do in practice have a statutory monopoly in almost all jurisdictions on the acceptance of deposits, but in general do not have such a monopoly on lending.4 Regulators have the option of imposing almost any level of capital requirement which they choose on banks, provided they are prepared to co-operate in ensuring that banks are able to charge their customers sufficient to enable them to raise that capital. Thus the idea of meaningful capital requirements is inextricably connected with the idea of a protected statutory monopoly.
(p. 24) 2.44 This monopoly must, however, be a regulated monopoly. It could be argued that we could dispense with the detail of financial regulation and simply permit the financial services markets to be dominated by a small, closed cartel. The cartel would act in its own best interests by restricting its rate of increase of risk in good times in order to avoid suffering losses in bad times. There are two objections to this, one theoretical and one practical. The theoretical objection is that conventional economic theory tells us that such a cartel would act in its own interests, and would require regulatory intervention to require it to act in the best interests of its customers. The practical objection is that in reality the position in the banking market is exactly the opposite of this—the market is highly fragmented and highly competitive, and produces low costs to users. In such a market, even assuming that bank managers were individually convinced of the merits of holding higher capital levels, they could not do so voluntarily—if one bank were to do so it would be competed out of business, and if banks were to agree collectively to do so this would almost certainly constitute illegal anticompetitive behaviour. As a result the cumulative private interests of market participants fall short of an outcome which is optimal for the market as a whole. Thus in order to promote the social benefit which arises from an efficiently functioning market, it is necessary to impose regulation to require the market as a whole to preserve itself.
2.45 If this is the justification for bank capital requirements, the next issue for regulators is as to the level of those requirements. The issue may be regarded as an insurance problem—would the public prefer to pay for future bank crises up front in the form of higher bank charges or after the event in the form of bank rescues leading eventually to higher taxes? The problem for the regulator is nothing more than to set the insurance premium level correctly. The remainder of this book is devoted to a detailed consideration of the technical and scientific rules which are used to carry out this calculation. However, it must be accepted that there is no technical or scientific technique which can answer the basic question as to how high these levels should be. The initial Basel 8 per cent requirement was famously set because ‘7 per cent sounded too low and 9 per cent too high’, but it must be accepted that the determination of the ‘right’ regulatory capital level comes down, in the end, to guesswork.
2.46 If prudential supervision is a good thing, then the immediate question to ask is as to why it seems to have failed so spectacularly in the banking crisis which began (p. 25) in 2008. In order to answer this question we need to do two things. One is to understand what the cause of the 2008 crisis actually was, and the second is to understand how bank regulatory capital requirements interacted with that cause.
2.47 Bank failures can broadly be divided into general and idiosyncratic failures. An idiosyncratic failure is a failure which results from factors which are unique to a single bank (or a group of banks). Idiosyncratic failure can have systemic consequences if the bank concerned is large enough. However, in most cases of idiosyncratic failure the failure is of risk control within the individual institution. A general failure, by contrast, occurs where there is a failure of part of the intellectual underpinning of the market. An example of this is the Dutch tulip mania of the seventeenth century. Whilst the market proceeded on the common belief that tulip bulbs had a particular value, stability was assured. However when this belief evaporated, the consequent collapse in values affected a large number of industry participants simultaneously. Similar events can be discerned in some other historic systemic crises—the 1929 Wall St Crash is an instance. Sometimes such events are entirely external to the financial markets, and the impact on banks is simply an instance of the impact of such developments on businesses generally. However, from time to time it is the internal mechanism of banks themselves which fail. The typical instance of one of these latter episodes is a general conviction amongst bankers that a particular asset class is highly valuable—railway stocks in the nineteenth century, sovereign debt in the 1980s, commercial real estate in the 1970s, and dot-com shares during the 1990s are all examples. In these cases an analytical error becomes a collective error and after a while begins to derive its validity from the fact that it is collective—‘it must be so because all the other banks think it so’. The asset class which occupied the position of most overvalued asset in 2007–8 was the residential mortgage—in particular the US sub-prime residential mortgage—and it was the collapse in the valuations of this asset class which led to some of the most significant write-downs in bank balance sheets. However, the reason that the crisis appeared at the time to be different from those which had preceded it was that the problem was not simply a collective misapprehension of the value of a particular asset class. What appeared in the early stages to have been uncovered was a flaw in the entire market’s fundamental approach to risk and risk quantification, a flaw which could potentially have invalidated almost every risk control paradigm used in every financial institution. For the second half of 2008, the primary concern of the market was whether any bank could ever be trusted again.
2.48 The technique which had apparently failed so spectacularly was statistical risk modelling. The use of risk modelling to assess and control bank exposures has been a fast-developing field for the last twenty years, and the significant advances made in financial mathematics enabled very sophisticated statistical techniques to be applied to portfolios of financial assets. Eventually these techniques graduated from being used by banks to assess their risk exposure to being used by structured (p. 26) product engineers to create securities with defined risk characteristics. In effect, these structures reversed the ordinary process of risk analysis—instead of starting with a portfolio of risks and assessing its riskiness, they started with a target level of riskiness and structured the portfolio to deliver it. The result was the creation of a deep and liquid market in securities whose pricing was in effect determined by the output of risk models. These securities were, reasonably enough, treated as interchangeable with ordinary securities—thus a AAA bond issued by a corporation was regarded as broadly similar to a AAA bond issued by a structured vehicle.5
2.49 The difficulty with this was—as we now know—that the output of the risk models was wrong, and significantly understated the riskiness of these structured securities. This discovery had a devastating impact not only on the market for securities based on risk structures, but also on the balance sheets of those who owned large proportions of these securities—a group which included the banks. Lead had been sold as gold, and once this was realized the balance sheets of financial institutions suffered badly from the write-downs. Even worse, the institutions which believed that they had found the way to turn lead into gold had made the cardinal error of keeping that apparent gold on their own balance sheets and borrowing against it—as was frequently pointed out, the problem with the ‘originate to distribute’ model was that it had become an ‘originate to not distribute’ model. Banks had in effect become so enamoured of the high-return low-risk securities which they believed that they were creating that they could see no reason not to keep those returns for themselves.
2.50 It is worth examining at this stage why the output of the models turned out to be so wrong. Statistical risk modelling was in many ways a victim of its own success, in that the more work was done in the field, the more robust the models appeared, until eventually it became impossible to doubt that which had been so clearly demonstrated by so many able mathematicians. This created a self-reinforcing cycle of confidence, such that—as explained in the Senior Supervisors Group report6—in some financial institutions the output of the statistical risk model became the sole guide for the making of trading and investment decisions throughout the bank. The problem, of course, was that the validity of a statistical model is an entirely separate issue from the validity of the output of the model. A model must be mathematically valid to be of any use at all—no useful output will be produced by a flawed model. However, the output of a model is only as good as the inputs to that model, (p. 27) and in the context of financial mathematics the data set of historical information about markets and prices is surprisingly weak. The reasons for this weakness are multifarious, but one of the most important is that if one goes back more than twenty years in most institutions the bulk of the available data will be on paper in filing cabinets (if it has been retained at all). Even where such data can be recovered and turned into usable electronic form, it is unlikely to pass any very stringent test of formal validity—for example, the exact definition of ‘default’ used may not be clear from the record, the extent of the exposure at default may not be clear, and the amount ultimately recovered may not be capable of being established. In many cases the inclusion of such data in the default database would have created serious validation problems for institutions seeking to demonstrate to their regulators that their systems were internally consistent and data was subject to rigorous review before being used in the models.
2.51 Aggregate market index numbers are available for an extended historical period spanning several severe economic downturns. However, specific pricing information about securities in electronic form—and even more importantly default and loss data about loans and recoveries in electronic form—is rare, and what is in existence is relatively recent. At this point it is important to remember that the name for the period of ten years to 2007 amongst economists is ‘the “NICE” decade’ (noninflationary, constant expansion), and that this period is generally reckoned to have been one of the most benign economic environments of the preceding century. In other words, the data which the models had to work on was not only incomplete, but was in fact drawn almost exclusively from a period in which default and crisis were notably absent.
2.52 The point here is, of course, that statisticians are only as good as their data. In the absence of hard historical data it is impossible to just make up some more. Manful efforts were made to take the data that was available and ‘flex’ it to reflect harder economic times. However, although it is possible in this way to project trends, it is very difficult to construct a statistical mechanism which reflects the fact that the data on which it operates is wholly unrepresentative, not least because beyond a certain point the output would lose the primary merit that a statistical analysis has—i.e. that it is based on hard data.
2.53 The point which many argue ought to have alerted the banks to the fact that they were basing decisions and products on unacceptably rosy scenarios was the level of returns which they were making on this business. A system which produced apparently AAA bonds with yields far above the equivalent yields paid on ordinary AAA bonds should have sent a clear warning signal that something was wrong. This became particularly clear when repackaging was taken into consideration—if your model is underpricing risk, if you then put the output of that model back into it you will end up with a greater underpricing, and so on. This phenomenon, and the proliferation of ‘squared’ and ‘cubed’ products which it generated, led to (p. 28) repackaging of structured paper being almost a licence to print money. These developments did raise alarm. However, by that stage no one was prepared to take the view that the output of the statistical risk models on which these securities were based was anything less than perfect—not least because, as just seen, the models themselves could be exhaustively validated mathematically.
2.54 The defaults which occurred within structured vehicles in 2008 constituted a very public demonstration that something was wrong with the architectural theory on which these constructs had been based. The crisis of confidence which this created had the effect of closing many of the financial markets almost completely, on the—quite reasonable—basis that investors knew that something was wrong but did not know what, and as a result avoided anything which might be tainted with the structured finance tar brush. The resulting market panic resulted in massive falls in the market price of apparently safe assets, and it was the consequences of these falls—and their impacts on bank balance sheets—which provided the links in the chain to the meltdown of late 2008.
2.55 This brings us neatly to the interplay between the market crisis and regulation. The market meltdown was in fact nothing more than a particular instance of an asset price bubble, with the distinguishing feature (there is always a distinguishing feature in every asset bubble) being the very high levels of confidence which banks had in their models translating into very high levels of confidence in their outputs, and as a result into very high levels of confidence in the valuation of securities based on those outputs. The revaluation of these assets, when it came, was therefore particularly severe and particularly unexpected.
2.56 Now any sufficiently severe asset revaluation will break a bank. Banks are typically at least 25 times geared (ie their total liabilities equal 96 per cent of their equity capital), and consequently even a relatively small move in aggregate values is sufficient to wipe out that capital. However, what the regulatory system does is to assess (roughly) the riskiness of assets, and to arrange that relatively little capital need be held against high-quality assets (such as government and AAA bonds), whilst larger amounts are held against more volatile assets. The problem in 2007–8 was that the regulatory system had accepted the banks’ models own assessment of the value of structured bonds, and had accepted that, being low risk, these should require relatively small amounts of capital. Thus, when these bonds started to suffer substantial liquidity and valuation impacts, the amount of capital held against those positions was rapidly consumed, and the shortfalls radiated out across the banking system as a whole.
2.57 It may be argued that this proves that the Basel system—which is itself based on statistical modelling—must also be fundamentally flawed. However, this is to (p. 29) confuse the method and the application. The fact that statistical models, applied to the wrong data, give the wrong results does not demonstrate that the technique of statistical modelling is itself invalid. It remains the case that there is a substantial difference in risk profile between a US government bond and an equity interest in a start-up, and a system which disregards this difference is unlikely to be an improvement on one which does. It remains the case that the regulatory system is likely to be at its most effective where it is based on an assessment of actual risk exposure. In retrospect, it is not surprising that regulators should have based their work on approaches which were accepted across the industry and across academe—and, indeed, it is positively desirable in general terms that they should do so. In many respects all that this proves is that where everyone in a particular market suffers from a common error, the regulator in that market is almost certain to suffer in the same way from the same error. The error has now been corrected (at least within the regulatory system), but there will of course be others. Nonetheless, if capital requirements are necessary for society (and it does appear that they are), then even a failure of the system to assess risk correctly can justify the adoption of a system which rejects risk analysis completely. If we are to have capital regulation then we must have risk-based capital regulation, and if we are to have risk-based capital regulation then we must accept that the analysis of risk within the regulatory system is likely to follow the same lines as the analysis of risk in the market which the regulator regulates. If the whole industry is wrong then the regulator will be wrong too, but even so it is better for the regulator to be occasionally wrong in line with the industry than invariably wrong by refusing to adopt sensible risk-based criteria to regulation.
2.58 However, what does follow from this is that it is arguable that the regulator’s focus on the use and output of risk-based models, although it followed the industry’s own best practice, may have been to some extent flawed in the same way. Quantitative risk regulation may be essential, but it is not sufficient, and regulators must follow bank management into a position where a broader overview may be taken of the risk and reward profile of the bank as a whole. This may in some cases be presented as a retreat from an over-reliance on quantitative modelling, and there may be something in this position. However, it should not be presented (or considered) as an abandonment of risk-based regulation, any more than any bank management team should be permitted to cease to attempt to quantify risk for internal purposes.
2.59 Given that bank failures are inevitable in any system, it is also important for the public sector to take steps to mitigate the impact on the public purse of bank (p. 30) failure. When a bank is in danger, or fails, in principle it should be subject to the ordinary insolvency regime of the jurisdiction in which it is established. However, it may be necessary for the public sector to intervene. This is because—as already noted—bank failure has social consequences above and beyond the immediate impact on creditors of the bank, and the application of ordinary insolvency procedures may lead to sub-optimal outcomes for society as a whole. Such intervention is often necessary and sometimes socially desirable even though technically unnecessary. Consequently governments should ensure that they have the necessary legislative, administrative, and legal powers in place to enable themselves to conduct such interventions in a swift and effective fashion, and in particular should ensure that they have in place all of the relevant powers and authorities identified in the IMF/World Bank Overview of the Legal, Institutional, and Regulatory Framework for Bank Insolvency.7
2.61 Intervention may take the form either of rescue or, if rescue is impossible, of some form of resolution or procedure.8 However, these are not mutually exclusive—a rescue may well involve the liquidation of some part of the rescued group, and a good bank/bad bank rescue will inevitably result in the eventual liquidation of the bad bank. There are therefore in broad terms three possible policy responses to bank failure, shown in Table 2.1.(p. 31)
2.62 There are a variety of reasons why government should seek to rescue an individual bank. However, the most commonly encountered in recent practice has been the public function of the banking system as a conduit for payments, and the ambiguity which that creates as regards deposits. The difficulty is that it is not possible to draw a bright line between deposits maintained for the purpose of payment services and deposits maintained as savings. Thus governments which seek to protect the payment services elements of an institution will necessarily be required to protect the deposits maintained with that institution. It should also be noted that the determination of which institutions should be supported because they are significant providers of payment services is an entirely different determination from that as to whether an institution is systemically significant. This can be seen in the UK in the case of institutions such as Northern Rock, where the basis of the government decision to support depositors was based on the extreme dislocation that the suspension of bank accounts would create, and despite the fact that the institution concerned was not systemically significant. What this demonstrates is that governments may frequently find themselves in a position where they feel forced to intervene in the failure of an institution which is neither systemically significant nor ‘too big to fail’. Thus the current paradigm—that banks should ordinarily go into ‘normal’ insolvency, and that special regimes are the exception—may in practice be the mirror image of the true position.(p. 32)
1 See, for example, Scott, Capital Adequacy Beyond Basel (OUP, 2005), Barth, Caprio, and Levine, Rethinking Banking Regulation (Cambridge University Press, 2006), and Tarullo, Banking on Basel (Petersen Institute, 2008).
2 The capital asset pricing model states that in pure theory a bank should be indifferent to this, since (again in pure theory) the two should balance each other out. This is not, however, the experience of most banks in the real credit markets.
5 In fact market prices tended to show a significant spread in the yield of such bonds over the ‘true’ AAA rate, suggesting that the market maintained a degree of scepticism as to the value of such bonds. However this level of scepticism proved insufficient.
6 ‘Observations on Risk Management Practices during the Recent Market Turbulence’, 6 March 2008, followed by ‘Risk Management Lessons from the Global Banking Crisis of 2008’, 21 October 2009, available on the Financial Stability Board’s website at <http://www.fsb.org/2009/10/r_0910a>.
8 There are a wide range of available procedures in the major jurisdictions, ranging from administration regimes whose object is to rescue the underlying business to liquidation regimes whose object is to realize assets and distribute them amongst creditors. For the purposes of this book we refer to all of these collectively as ‘insolvency proceedings’. These proceedings almost invariably involve the appointment of one or more persons to conduct the proceedings under the supervision of the court. We refer to such persons as ‘office holders’.