Financial cycles and stability
Summary report by William R White, Economic Adviser, Bank for International Settlements, at the Bank of France's International Monetary Seminar, 31 May 2002.
I am reminded of a US comedian who once said, "We only have a minute, let us talk about the world". Financial cycles and central banks is a major topic and we have to be very modest as we begin approaching it. This does not mean that your week here has been wasted and I am sure you have learnt a considerable amount. However, we have to be modest. We do not know a significant amount about the interaction between the real side and the financial side. We do not know how the real has an impact on financial institutions. For example, no one is quite sure what might happen should we have a double dip. We are also not sure what happens when financial institutions run into difficulties or the extent to which that feeds back on the real side.
Furthermore, there are no magic or miracle solutions in this area. Instead, we are faced by trade-offs. For example, you might use policy to obtain more price stability but at the cost of less financial stability. You might achieve greater efficiency but less safety. Mark to market or fair value accounting might lead to quicker reaction to real problems but could also lead to inappropriate reactions to non-problems. Similarly, incentives are very subtle and psychological, and there are no right answers.
With that background, I will focus on four major points that were central to today's discussion: the underlying nature of the cyclical process; the various methods of crisis prevention; monetary policy; and other policies for getting out of crises that have arisen.
Let us first look at the underlying nature of the problem. Financial cycles would seem to evolve from excessive optimism. That is, credit expansion, which feeds into asset prices, which lowers the cost of capital, which leads to more investment, which leads to a boom. Eventually, the investments are found not to be sustainable in that they do not generate profits. Profits are the central factor in this cycle: if over-investment cannot generate profits, the whole structure collapses, moving from boom to bust. Furthermore, this kind of boom-bust cycle can occur without any overt inflation. In this context, Peter Praet identified a central issue: the very fact that the investment boom is increasing capacity and productivity growth, puts downward pressure on prices. If you begin from a position of very low inflation, where you have just enough capacity to keep prices in control when demand was high, you find yourself in a very difficult situation when demand collapses. The process can thus be summarised in three terms: boom, profit collapse, and subsequent bust.
Economic history books are full of historical examples of this process, and there is much available theory on the subject. For example, the major debate of the 1920s and 1930s between the Austrians (Hayek etc.) on the one hand, and Keynes on the other. In a nutshell, Keynes was focusing on the demand side and on how to get out of the bust, whereas Hayek was focusing on the supply side and on how we got into it in the first place.
A number of comments were made today about how this boom-bust financial cycle can be made better or worse. I would make three major points in this area. First, even if these cycles are inherent, "New Era" expectations can make them worse. The fact that there is a very good reason for being optimistic in the first place can subtly lead from justified optimism into unjustified optimism, and from rational exuberance to irrational exuberance. New Era concepts can thus be dangerous.
Second, opinion is still divided on whether these cycles are made better or worse by a more market oriented system as opposed to one that operates primarily through financial intermediaries. We currently have a much more market driven system than we did 20 years ago. Last year, the global economy had a huge number of shocks and yet we have come out unscathed to date. It could be argued that this was due to the fact that more complete and resilient financial markets were able to stem damage on the down side. In terms of more complete markets, a number of references were made to the fact that people can now obtain credit from more channels. For example, the commercial paper market virtually dried up and people fell back to the banks or went to the bond market. That is, they had the capacity to get money somewhere else. On the household side, the fact that mortgage markets are now more complete and people can collateralise their housing wealth, has kept things moving. The markets are also more resilient and there is less reliance on individual financial institutions. If the market rather than banks is providing credit, then it is the venture capitalists or pension funds that go under, and the whole system is not threatened. A number of references were also made to easier access to information. All of these factors have contributed to the system being more resilient.
However, when we look at the banking system as such, a very large number of these markets are now dominated by a very small number of financial institutions. The new instruments have been very helpful in many respects. However, we have not yet had a deep recession, and we do not know the extent to which these instruments will prove resilient in a deep recession.
Is a system of markets less or more stable than a system of financial intermediation? I believe that there are arguments on both sides. However, I was struck by comments made to me by Harold James in response to the question of whether our current situation resembles the 1920s or the much milder recessions pre-WWI. Harold James believes that the 1920s is the only case where the banking system was absolutely central to credit expansion. In all previous cases, we had a much more market based-system, similar to what we have today. He thus believes that the system is now more stable.
How do we reduce the likelihood of these financial crises? My first general point, which was raised by a number of speakers today, is that there is no correct answer, as in the Shumpeterian concept of creative destruction. On the one hand, many people were able to obtain credit in the past ten years who would not otherwise have been able to do so. Those people came up with new ideas which will improve our future standards of living. On the other hand, many people who obtained credit should never have done so. There is thus no correct answer, only a trade-off.
In this context, I would note that different people choose different trade-offs. The US seems to be more willing to drive for growth and efficiency, and to deal with the greater cyclical variance later. In contrast, Europeans seem to be more concerned with reducing the cyclical variance and less concerned with the mean. Furthermore, trade-offs change over time. After the Great Depression, there was a huge recognition of market excesses and much regulation was introduced. We have spent the past 30-40 years removing that regulation. Should we get out of our current situation unscathed, there may not be a swing back to re-regulation. However, if we have a serious double dip, the pendulum will likely move back in the direction of increased regulation.
We all know that we need good internal governance but there are various reasons why it has not been achieved. The Enron incident illustrates the complexity of the issues involved. There were at least ten levels of governance in the Enron case, and they all failed. This relates to the issue of excessive optimism: hard questions are not asked when things are going well. The micro problem in the Enron case comes back to questions of conflict of interests. People at virtually every level of governance in Enron had a conflict of interest.
A number of speakers also referred to the fact that, as situations degrade, people will take on more risks and gamble for resurrection. In periods of radical change, internal governance almost always has a tendency to break down due to an internal conflict of interest. In times of rapid change, profits are under pressure but shareholders increasingly demand greater returns. The only way to resolve this dilemma is to take on more risk. This is why internal governance will not work by itself.
Much progress has been made, as illustrated by Christine M. Cumming's reference to quantification. We become more focused when we begin to consider how to better measure risk. It should be noted that we also have to use our own judgement on all these matters. Thus, if VaR is going up solely because the market has panicked, reacting to the increase in VaR by selling will simply make the situation worse. However, we can distinguish between measurement and response, and much progress has been made in terms of risk management in recent years.
Self governance will not be able to reduce the likelihood of crises on its own, and regulatory oversight will be needed. Many issues were raised in this area today. BIS staff were most concerned about this issue early on in the context of Basel II and pro-cyclicality. If risk weights can change to reflect relative risk, they can also change absolutely over the cycle in a way that makes the cycle worse. However, I find solace in the approach that considers Basel II as a process. Christine M. Cumming referred to the on-going discussions between supervisors and the private sector, raising consciousness about risk, and the necessity of measuring and responding to risk. That on-going cultural change could play an important role in reducing the problem of pro-cyclicality.
Many references were also made as to the appropriate body to carry out supervision. Henk J. Brouwer is concerned with the issue of where supervisory authorities should be located. This raises some profound issues due to the sectoral breakdown: who should have oversight responsibility given that insurance companies are becoming banks etc. Bob Merton has recently pointed out that many non-financial companies are actually financial companies. For example, General Motors has a pension fund worth USD 80 billion, whereas its market capitalisation is USD 30 billion. Do our methods of accounting for pensions provide us with a clear understanding of the exposure of these institutions?
Many references were also made to accounting issues. In general terms, in particular in relation to credit risk, people assume that good times will last forever and that there is no need to worry about any subsequent downturn. Many people believe that provisioning, or preparing for potentially poor losses, should be much more forward looking. Christine M. Cumming noted that fair value accounting basically amounts to pricing risks properly. If the risk is priced properly from the start and you have a higher rate of return, you do not need to take a provision at that point because the future interest rate payments will be coming in. If, at a certain point, risk is materialising unexpectedly quickly, a provision should be made against it. In a fundamental way, fair value accounting is thus forward looking and represents better risk pricing.
Christine M. Cumming pointed out that market discipline did not play much of a role in the 19th century. I would add that they have not played much of a role in recent years either. To reduce the prevalence of financial cycle crises we need a combination of better internal governance, better oversight and supervision, and better market discipline. They will each prove inadequate on their own.
What can monetary policy do about credit cycles, both on the way up and the way down? When asset prices are increasing, monetary policy is, unfortunately, not very useful. This is due to the fact that we cannot know the fair value of an asset. Nor can we know which asset price to target. Furthermore, if we are too concerned by asset prices, we will be less concerned by CPI inflation. It is also very difficult to raise interest rates when things are going well, particularly when the public has been told that price stability is the fundamental issue and it has been achieved. However, price stability may not be sufficient in avoiding crises. Furthermore, interest rates would have to be increased quite significantly to deal with an asset price bubble. Finally, if monetary policy is to be used to respond to asset prices, it has to be done in a symmetrical way.
Alastair Clark reminded us that much information is now available in relation to monetary policy, but we still need more information, in particular, on credit exposures. We now have various ways of obtaining market-based assessments of where credit exposure lies. For example, risk differentials in credit market, credit derivatives, or KMV type estimates of default probabilities.
It may be difficult to know what to do when confronted by the upside of these financial asset bubbles. However, when faced with the combination of fast credit growth, heavy levels of investment, and rising asset prices, we should be concerned that this will feed back into the financial system in a very negative way.
In so far as the way down is concerned, let us suppose that the bust has already occurred. One school of thought argues that, at this point, interest rates should be reduced in a slow and sustainable way. Otherwise, people may assume that the situation will be fine and the long bond rates will rise, cutting off any expansion that might otherwise have occurred. Another school of thought argues that interest rates should be reduced as quickly as possible. However, simply printing money to respond to the bust following successive periods of credit excess is akin to offering alcohol as a cure for a hangover.
This brings us to the essence of the Hayek-Keynes debate. Keynes focused on the demand side, and Hayek on the supply side. Hayek lost the debate because his philosophy seemed to offer no hope. His basic message was that the only answer to major over-capacity in terms of real capital was to liquidate. While this is not a very attractive argument, it does contain an element of truth. For example, Japan's basic problem is its failure to get rid of excess supply.
I would conclude that both a demand side and a supply side response is needed. This involves highly expansionary monetary policy, and fiscal policy, if needed. As well, the supply side is very important. The US is set up at the structural level to ensure the kind of supply side response the Hayek School would have recommended. The various actors involved will ensure (through the courts if needs be) that the necessary supply side adjustments are made.
I hope you have learned a considerable amount about financial cycles and stability in the past week. The main lesson to be learnt is to try and do the right things so that the situation does not arise in the first place.