Thursday, 19 March 2009

Towards a rational exuberance

Lord Turner's report "A regulatory response to the banking crisis" was released yesterday.

One of the key findings is that markets, and market participants, can be irrational. This can lead to problematic outcomes such as bubbles in asset prices and massive mispricing of derivatives such as CDOs and CDSs.

However, Turner's recommendations barely address this problem. There are several valid recommendations about procyclical reserves and a hint at a power to intervene in momentum trading (such as short-selling which feeds on itself). But this only dances around the edges of the problem.

He also recommends technical training or qualifications for bank executives. But that's not where the irrationality is. Indeed, many bankers have been all too rational throughout this crisis - extracting rents for themselves at the expense of shareholders and creditors. Turner does recognise this principal-agent problem and some of his recommendations deal with it. However it is not a problem of irrationality; the irrationality that matters is that of borrowers and investors, not of banks.

The corollary of this is a more radical proposal.

Regulators - central banks or financial regulators such as the FSA - need to monitor aggregate irrationality. There are two ways to do this:
  • measures based on the behaviour of individuals, aggregated across a whole market
  • measures based on the aggregate risk taken and valuation implied by market asset prices
The second of these can be measured using (fairly) standard financial measures - projections of economic growth, share of output accruing to capital or to profits, discounted and compared with asset prices. The answers won't be exact, but measures that are wildly outside of economic rationality will be apparent.

There will certainly be challenges in measuring across different asset classes - typically a single regulator does not cover equity and debt markets, for example, and if there is an aggregate overvaluation, it will be very hard to gauge whether equity holders and debt holders are both expecting a too-high share of overall returns, and who is wrong. But with unified or at least coordinated regulation, it should be possible to determine that overall prices are too high.

The former requires insights from behavioural economics and particularly from behavioural finance. In many situations it is possible to measure objectively what the rational outcome or decision is. And it is possible to see where an individual investor diverges from this. Sometimes this does not matter for the system as a whole; and sometimes it does. But as Turner points out, markets are not always self-correcting and certainly not in the short term. Thus, it is likely that a dedicated regulator could identify substantial deviations from rationality in large populations.

On the border between these is the collective action problem: where individual rationality leads to an outcome which is not a rational one for the group. Turner does mention this but offers no solution. Again it is clearly measurable: at present, for example, we have an excess of desired savings leading to a reduction in overall output - clearly not a desirable outcome for the population as a whole, but perhaps rational for each individual saver. Somehow the interest rate mechanism is not resolving this problem (the zero bound is one reason; sticky prices, wages and investments are another).

So our hypothetical regulator has identified irrationality of some kind - irrational exuberance or irrational depression. What should it do?

Here we turn again to behavioural economics. There are clear ways in which irrational behaviour can be guided or corrected by specific stimuli. Specific examples:
  1. Framing of choices. The right kind of framing can influence people to take more or fewer risks, to consider the future more or less, and to put a higher or lower value on assets. Framing methods include choice of language, the range of pricing and risk choices available to buyers, subconscious signalling such as branding, and sensory cues such as colour and music. The mechanisms for regulators to communicate their decisions downward to the framing of commercial decisions do not yet exist, but could certainly be put in place. After all, they currently influence interest rates, so the argument for commercial freedom is not an absolute barrier to this.
  2. Visibility of irrationality. People tend to become more rational either if they have more time to reflect, or if their irrationality is made visible to them. A regulator could certainly have a role in making this happen. Simplistically, one could ask why not make people more rational all the time. But there are transaction costs and diminishing returns involved. It's not realistic for a regulator to step into all transactions everywhere; but if they are able to measure who is irrational, when and where, they can focus their efforts where they'll have an effect.
  3. Lengthened time horizons. Buyers generally make less rational decisions when they consider the consequences over a shorter time period. The longer a period that is considered, the better the decision will be. Interest rates are one way of influencing the time periods that investors consider; others include the credibility of inflation targets or the availability of externally anchored future events. For instance, subjects who are asked questions about their age and retirement dates will subsequently make a different kind of investment decision than those who are not.
  4. Increased scope of social contract. Buyers who act purely as individuals will make different choices to those who also consider the effects on their family, their social groups or their society. It is possible to measure the divergences between individual and group interest, and to influence the degree to which people consider the consequences for a group when making decisions.
This proposal undoubtedly needs further research: we don't yet know the most accurate measures of irrationality or the most effective influences on it. And there are obvious limits on rationality in all circumstances - predictions of the future are not perfect, information is not always available, and people cannot always know their own preferences with regard to time discounting.

But I am confident that direct methods to help people and groups to be more rational will have a more powerful, and more timely, effect than relying on banks' capital cushions to make the corrections for them. This isn't all about correcting for overconfidence; it will also work on the risk aversion and underconfidence we're seeing now. Exuberance can be rational, and growth will be more steady and reliable when it is.

Update: Two VoxEU articles I've written on this subject: Models of bounded rationality and the credit environment and Behavioural financial regulation

No comments: