Speeches and Presentations

Check against delivery.

1. Introduction1)

Many economic policy debates during the past two decades have been guided by a principle that could be characterized as “regulatory skepticism.” This skepticism has its roots in historic experience with the limits of economic policy and the daily experience with tight constraints policymakers have to face in a globalized world. This general skepticism has affected some areas in economics more than others. Financial regulation is an issue that has been relatively unaffected. It is a widely held conventional wisdom that the financial system has to be firmly embedded in quite an extensive regulatory framework, and debates seem to revolve around details of this framework rather than focus on the principle itself. 

This general attitude to financial regulation is of course rooted in historic experience of episodes of financial crises and the difficulties to deal with them as well as in insights from a huge body of research in economics and finance. We know that the costs of financial instability can be huge and affect a large number of people. 

Nevertheless it might be worthwhile to take the occasion of this lecture to have a look at financial regulation in the spirit of regulatory skepticism. Regulators always have to be careful not to represent special interests and to devise policies that promote theses special interests in the name of the common goal “financial stability.” 

For these reasons, clear ideas about what exactly financial regulation aims to achieve and the analysis of the exact mechanisms by which certain regulatory policies are supposed to work are essential. In particular, this requires pinning down as precisely as possible the nature of market failures which regulatory intervention tries to correct. Furthermore, it is important to consider not only the direct but also the indirect consequences of financial regulation. 

Given the fundamental role of the financial system in directing funds from households to productive investments of firms and in allocating and sharing risks, the indirect consequences of regulation are most likely to occur in the form of repercussions for the economy as a whole. These repercussions have to be carefully considered when we assess the benefits and costs of financial regulation. 

To be honest, I will not be able to provide a definite – sometimes not even a preliminary – answer. I will raise some issues concerning financial regulation that deserve consideration and perhaps further critical assessment. In particular, I want to argue by way of examples that macroeconomic aspects should be given more consideration in the design and the practice of financial regulation. 

Having said this I should point out that the regulatory community and academia have recently discussed numerous arguments and thoughts in this direction. I am therefore not raising new issues but rather trying to contribute to an ongoing discussion. I hope that this discussion will ultimately result in a better understanding of the possibilities as well as the limits of financial regulation.

2. What Is the Problem?

When we look at financial crises we experienced in the past, nearly all of them were caused by large macroeconomic shocks: huge hikes in interest rates, extreme exchange rate and stock market movements or recessionary dynamics in general. This observation is important because it suggests that the exposure to these risk factors lies at the heart of the problem of financial instability and systemic crises. 

Emphasizing the link between financial instability and adverse macroeconomic developments might seem obvious to most of you. However, it is worthwhile to point out that much of the economic literature on banking crises and bank runs drawing on an influential paper by Diamond and Dybvig (1983) has taken a different perspective. The focus in this literature has been on the dynamics of an event that starts somewhere in the banking system with the failure of a single institution and then affects other institutions through various contagion channels, such as information effects, or through direct interbank credit links. 

The macroeconomic perspective suggested above hints to the other source of financial instability. Financial institutions may have correlated exposures, and an adverse economic shock may directly result in simultaneous failures. Research recently undertaken at the Oesterreichische Nationalbank and the Bank of England has shown that quantitatively correlated exposures dominate the contagion problem by a huge margin as a source of systemic risk, providing   evidence for the casual observation made above.

Looking at the problem from this perspective does of course not provide us with an answer to the question how these exposures to macroeconomic risk factors are built up and whether this is something that regulatory policy can and should deal with. Here the story has to be augmented by further considerations in particular by a time dimension.

Borio (2003) described a stylized pattern of a financial cycle as it had repeatedly been observed in the past, notwithstanding the many differences between individual crises with respect to the different patterns of vulnerability as well as with respect to the triggering events. The build-up of imbalances that trigger a crisis usually starts under booming economic conditions. This boom is accompanied by a climate of overly optimistic risk assessment, the gradual weakening of financing and credit constraints and rising asset prices – in particular those of property and real estate. In this climate financial and real imbalances build up. At some point an essentially unpredictable trigger like an asset price drop or the interruption of an investment boom causes a sudden rundown of financial buffers, and once these buffers are exhausted and the contraction exceeds a certain threshold, a full-scale financial crisis occurs. 

This stylized story has some essential features which are worthwhile to emphasize here. First we see a dynamic interaction between the financial system and the real economy over time during the build-up of the crises. Therefore, rather than the cross-section perspective, the intertemporal perspective, which looks at a set of institutions at a single point in time, is essential. Common exposures to interest rate risk, exchange rate risk and the business cycle in general play a dominant role. Much of the build-up of the crises occurs on the asset side of a bank’s balance sheet, whereas the liability side plays its main role in the unfolding and triggering of the crises. 

Maybe less obvious but nevertheless important is the fact that the build-up of a crises through the dynamic, intertemporal interaction between the financial system and the real economy suggests that the risks are to a large extent endogenous. They both influence and are themselves influenced by the behavior of people in the financial system. This feedback mechanism should be taken into account when policy measures are designed. Again, it may be obvious to most of you familiar with the day-to-day reality of the financial system. I emphasize it because many of the quantitative state-of-the-art risk assessment models we use in analyzing financial risks rely on the assumption that financial risks are driven by an exogenous random process which influences the behavior of financial market participants but itself is not influenced by their behavior. These models therefore assume away the feedback mechanism between behavior and risk, which plays such an important role in the reality of financial markets, in particular in situations of crises and distress.

Having sketched a stylized picture of financial instability and having pointed out the main features characterizing it, we may ask what the role of a regulator in all of this is. We have learned from the past that the triggering event is essentially unpredictable and diverse. It is furthermore beyond the control of any particular individual, even of such institutions as financial regulators and central banks. Their potential role seems to be connected to the dynamic problem of the build-up of imbalances and the monitoring and early understanding of the emergence of a crisis situation rather than to the triggering event itself. 

Still the question is what the exact nature of the problem in the build-up of crises is. I think we can identify two different problems. One is a fairly straightforward externality problem. Financial market participants may have an incentive to take more risks than a socially optimal risk allocation would require because in their plans they take into account only the individual costs of their decisions but not the social costs that occur in a situation of financial distress. Yet it is of course not fully clear what the exact reason may be that market prices do not reflect the full costs of taking risky decisions. Given one believes the externality story the problem of the regulator is to implement an optimal risk allocation by suitable policy instruments. In practice this is of course not exactly a neat textbook problem and requires considerable judgment and practical expertise.   The second problem seems to be a coordination problem. The feedback from behavior to the build-up of risk and back to behavior can lead to a coordination problem: it is individually rational to act on the overly optimistic risk assessment as long as everybody else does but again the social optimum might lie elsewhere.   Many such situations are discussed abstractly in game theory. Here the role of the regulator becomes one of supporting coordination on a sustainable equilibrium of behavior.

The challenging problem in all cases is to deal with the characteristic feedback loop where actions of financial market participants influence financial risks and these risks feed back on actions and so on. This feedback loop lies at the heart of the problem why regulatory instruments may have macroeconomic consequences. 

3.     Regulatory Policy and Macroeconomic Consequences

Financial Regulation encompasses both markets and financial institutions, most importantly banks. In the area of markets, financial regulation mainly concentrates on providing a legal framework (a set of rules) that facilitates the market exchange of financial instruments. From an abstract viewpoint, financial contracts are promises to exchange goods or money between different points in time and across different possible future states of the world. This very nature of financial contracts brings along a host of incentive, enforcement and information problems, which creates a direct need for this legal infrastructure. If we pass on to financial institutions, regulation often interferes directly with the decisions that can be taken by economic agents, like restrictions on asset holdings and competition or restrictions on the capital structure of banks. Thus, these regulatory measures interfere with the feedback loop between behavior and financial risks discussed above and are therefore notoriously hard to assess.

The regulatory instrument that has been given most attention during the past decade certainly is capital adequacy for banks. It is the regulatory instrument most frequently discussed in the context of financial stability as well as concerning its macroeconomic consequences. It thus perhaps serves as a good example to discuss some of the more general issues that I want to raise here.

Usually, the following two arguments are given in favor of capital adequacy: First, it is seen as an instrument limiting excessive risk taking by bank owners with limited liability and, thus, promoting optimal risk sharing between bank owners and depositors. Second, capital adequacy regulation is often viewed as a buffer against insolvency crises, limiting the costs of financial distress by reducing the probability of insolvency of banks. It is the hope of regulators that capital adequacy provides a safeguard against a systemic crisis, the widespread breakdown of financial intermediation.

In the context of the stylized facts of the build-up of crises and the dynamics sketched above, these arguments focus on the cross-sectional rather than on the intertemporal aspect of risk. The idea that a bank that holds more equity has larger buffers against the risks it is exposed to does not relate directly to the dynamic story of how equity may be optimally adjusted over time. 

The cross-sectional view has also shaped the big recent changes to capital adequacy in the Basel II framework. In an effort to improve risk sensitivity, the new framework stipulates that minimum capital for a given portfolio changes with its perceived riskiness as measured by external or internal ratings. This is a big reform step that leads to a much better relative risk assessment across institutions at a given point in time. Without doubt, this reform is a milestone that will substantially improve risk management and risk awareness. 

One of the discussions in the run-up to the New Basel Capital Accord concerned the macroeconomic implications of capital requirements if they are set independently of macroeconomic conditions. The arguments that have been raised have been based mostly on an intertemporal perspective and mainly concerned the endogenous nature of financial risks. 

The basic argument is plain and straightforward. When the economy is hit by a recession enterprises’ debt service is impaired. The reduction in bank profits – in absence of compensating issues of outside equity or reduction in dividends – leads to a reduction in bank equity. No matter how sophisticated the cross-sectional capital regulation is calibrated this decline in bank equity must result in reductions in bank lending. This in turn reduces business investments and exacerbates the shock to aggregate demand. Thus, in an intertemporal context the regulation that boosts individual buffers at the bank level acts as a potential crises amplifier at the system level through the macroeconomic consequences of the regulatory instrument. Clearly, if we take the stylized story of financial crises and crisis dynamics seriously the intertemporal aspects of risk are perhaps not yet sufficiently taken into account. The previous analysis of crisis dynamics implies that the adjustment of capital over time requires as much attention as the cross-sectional aspects that are now taken into account in a very detailed and sophisticated way.

Let me now – as a second example – turn to the problem of coordinating expectations. The coordination problem arises when in a climate of overly optimistic risk assessment it might be optimal for each individual to respond by equally overly optimistic assessments as long as everybody else holds excessively optimistic views about the economic environment. Here the dynamics of expectations have a potentially huge impact on financial cycles with repercussions for the macroeconomic environment. Expert risk assessments by institutions in charge of financial stability, such as central banks and supervisors, act as potentially important coordination devices and can contribute to coordinating expectations to realistic and well-founded risk assessments. Exactly these aspects are the motivation for many central banks to publish financial stability reports.

We have seen that led by the Baselprocess, impressive progress in quantitative risk assessment models and in the skillful use of these techniques by supervisors and regulatory institutions in charge of financial stability has been taking place. The difficulty of many of these techniques so far is that they are devised for individual banking institutions and not for the system as a whole. To capture systemic risk, it is however decisive to take a system perspective. In a recent paper, Elsinger, Lehar, Summer and Wells demonstrate that an individual institution approach that ignores correlations and interlinkages – as many quantitative risk assessment models do – does indeed underestimate systemic events of bank insolvencies by a considerable margin. 

I think that regulatory institutions in charge of maintaining financial stability can contribute a great deal to coordinate expectations on realistic risk assessments by devising credible and good quantitative models that are able to capture risk at the level of the banking system. Such contributions can establish a beneficial link between regulatory supervision and financial stability. If they smoothen expectation-driven financial cycles, a positive impact on macroeconomic stabilization can be established. If such models can be run at regular intervals with a continuous updating of information risk assessment over time may also be improved.

4. Conclusions

So let me conclude: I have taken the viewpoint of a regulatory skeptic to point out some of the particular challenges financial regulation has to face.   In an attempt to pin down the major reasons why we might want regulatory intervention in the first place I have identified an externality problem and a coordination problem. Both problems are intimately connected to the dynamic nature of financial instability. Both problems mainly materialize in the dynamics and the build-up of financial and real imbalances, which then suddenly unfold, triggered by an essentially unpredictable event. The dynamics are so intricate because the major risks are endogenously driven by a feedback from macroeconomic conditions to behavior and back to macroeconomic conditions. 

I have argued using the example of capital adequacy regulation that this instrument is very finely tuned to considerations of relative risk at a particular point in time and less so to the intertemporal aspects, which is where the externality problem kicks in. Thus further explorations to include the intertemporal aspects and therefore to design the instrument with a stronger focus on its macroeconomic consequences may be useful.

I have also pointed out the usefulness of quantitative models of risk assessment for coordinating expectations about financial developments. A shortcoming of many of these models is that they are geared very much towards individual institutions rather than to the system as a whole and may therefore perform poorly in the assessment of systemic events. If these improvements can be successfully accomplished a reform of supervisory monitoring may be achieved that has a beneficial role in coordinating expectations and in supporting financial and macroeconomic stability.

Overall I think we have to strive for progress with respect to both the regulatory instruments that we use as well as the quantitative models we rely on in making our assessments.   Thus with respect to financial stability we have to aspire to a state that has already been reached by monetary analysis. Here the policy instrument is clearly defined. The decisions are supported by a host of canonical quantitative models which are built on a combination of theory, practical relevance and empirical fit. The impact and feedback mechanisms of policy are fairly well understood and relatively uncontroversial.

There is no principal reason to assume that financial stability analysis and financial regulation will not arrive at a similar state of theoretical and practical knowledge in the future. The tremendous progress that has been made in the past two decades gives reason for such optimism. To foster this progress we will however need considerable further effort in research and a constant attempt to review current policy practice. I think that this conference and the activities of SUERF provide an excellent and valuable contribution to this ambitious goal.



References
 
Borio, C., 2003, Towards a macroprudential framework for financial market regulation and supervision, BIS Working Paper Nr. 128.

Elsinger, H., Lehar, A., Summer, M., 2002, Risk assessment for banking systems, OeNB Working Paper Nr. 79.

Elsinger, H., Lehar, A., Summer, M., Wells, S., 2004, Using Market Information for Banking System Risk Assessment, mimeo. 

Diamond, D., Dybvig, P., 1983, Bank Runs, Deposit Insurance and Liquidity, Journal of Political Economy, 91(3), pp. 401-19.


1) Acknowledgements: I want to thank Martin Summer for helpful discussions and comments.