Behavioural financial regulation
All macroeconomic actions - monetary, fiscal or regulatory - are ultimately intended to influence behaviour. Monetary policy changes interest rates in order to affect people's preference to hold or spend money. Fiscal stimulus increases aggregate demand, giving people more income in the short term which we hope they'll spend, and directly brings unemployed resources into use by paying them to work. Regulatory changes restrict or encourage some behaviour more directly, but generally aim to indirectly affect the behaviour of consumers by directly restricting the actions of financial institutions (or more properly, their staff).
All of these approaches to managing the economy share a common assumption: that people, on the whole, respond rationally to price incentives. If the price of immediate consumption goes up, people will consume less immediately. If the demand for labour rises, purchasers of labour will pay more and suppliers of labour will work more. If bank capital requirements increase, banks will pay more for capital, raise interest rates and lend less money.
But we know from recent experience that this is not always an accurate assumption.
Lord Turner, writing the review of financial regulation for the UK's Financial Services Authority, identified "self-reinforcing irrational exuberance followed by confidence collapse". Martin Wolf has said that global finance "is prone to wild swings of mood, from euphoria to panic". Robert Shiller has written on animal spirits. And even arch-neoclassicist Alan Greenspan pointed out the phenomenon of irrational exuberance twelve years before Lord Turner picked up on it.
So if people are irrational, will policy and regulation based on rationality successfully keep the economy stable? Of course, to some extent this does work. In most behavioural experiments, subjects are influenced in the expected way by incremental changes in price, supply or demand. And thus, a reduction in interest rates will, ceteris paribus, lead to an increase in consumption.
But superimposed on this are powerful behavioural phenomena which are not explained by conventional rational price and utility models. A well-known experiment (Smith, Suchanek and Williamson 1988) demonstrates that people in experimental conditions will still bid up the price of financial instruments to levels well above their fundamental value, even when that value is exactly predictable and is made known to them at all times.
The problems for financial markets arise not only from irrationality, but from correlated irrationality. The recent proposal from Hart and Zingales, for example, which aims to use CDS pricing as a regulatory signal and require banks to raise additional equity if their CDS spread exceeds a known level, would work well for a single institution but not when a large proportion of all banking assets are under- or over-valued.
Can regulation address the issue of correlated irrationality? Fortunately economics has started to develop tools to be able to do just that.
Behavioural and experimental economics now provide objective ways to measure and monitor such variables as risk appetite, monetary value expectations, and cognitive capabilities of investors when evaluating financial products. Using known expected values, it is possible to calibrate investors against the hypothetical rational agent, and find out whether they are irrationally exuberant - or irrationally pessimistic.
While it is rarely possible to determine with certainty the expected value of a single financial instrument, it is much easier to calculate the value of certain portfolios. For instance, it is easy to see if the stockmarket as a whole is priced for aggregate profits that represent an unsustainable share of future GDP, but harder to know whether it's Exxon Mobil or Amazon.com that is overvalued. However, by sampling a suitable range of investors and testing against diverse assets, it should be possible to gain an insight into correlated irrationality.
In other cases, behavioural tests can give insight into accurate asset pricing. A sharp rise in house prices, for example, might originate partly in irrational buying decisions, and partly in by a genuine increase in the expected long-term return on property. While the long-term return cannot be tested today, the irrationality of consumers can, through specific experiments. Thus the change in expected returns can be deduced by subtracting the irrationality effect from the actual price rise.
I propose that regulators develop a small set of measures of irrationality that can be calculated and published at least monthly. These might include measures related to expected personal income, job security and asset values; measures of expectations about the performance of the economy as a whole; and measures of hyperbolic discount rates and other specific observable cognitive biases.
Having ascertained the population's divergence from classical economic rationality, what should regulators do about it? As well as offering measurement, behavioural economics also offers a range of tools to influence behaviour. A well-known example is Benartzi and Thaler's "Save More Tomorrow" programme which overcomes the effect of hyperbolic discounting by asking consumers to precommit to saving a proportion of future salary increases.
There are many other tools which have been experimentally tested at the micro level but not yet applied to macroeconomics. Price anchoring, framing, endowment effects, confirmation bias and various social and peer effects all demonstrably allow us to influence behaviour in the lab; they should have applications to what we might call "macrorationality".
In many cases, simply releasing the results of these measurements may correct asset prices to some extent. If a regulator's report had announced in 2006 that consumers' estimates of future house prices were unsustainable and that it was investigating methods to correct the overvaluation, it is very likely that prices would have been lower in 2007 than they actually were.
But this will not work in all cases. Announcing that consumer confidence is irrationally low might provoke a self-fulfilling fall in consumer spending. Thus it will be important to conduct research on how to extend these influencing tools from the laboratory to the macroeconomic theatre.
There are important subtleties to this argument. The definition of "irrationality" is one. By irrationality, I mean departures from the model of classical utility-maximising rational agents. Models of the mind sometimes provide reasons and systematic description of this kind of irrationality; philosophically it could be described as rational, but it does not conform to economic orthodoxy. To properly understand this area, explicit models of decision-making need to be built that more closely match real behaviour but are still mathematically tractable.
Another is the distinction between influencing choices that are otherwise freely made, and coercion. This difference, familiar from Thaler and Sunstein's "Nudge" concept, is important in understanding the trade-off between liberty and security that is present in all decisions about regulation.
A final question is whether, if this approach does resolve or mitigate the problems of correlated irrationality, a new problem will surface which leads to new forms of crisis. Perhaps this will happen. We can call that progress.
But the problems that confront us now and in the recent past are all closely related to a mass phenomenon of over- or underestimates of asset values and income. Developing tools to directly attack this problem would be one of the key steps that a national - or a putative international - regulator could take to end this crisis and prevent the next one.
- Turner, Adair. "The Turner Review" (2009) pp25, 39-40: http://www.fsa.gov.uk/pubs/other/turner_review.pdf
- Wolf, Martin. Interview transcript (2009): http://www.atkearney.com/main.taf?p=5,13,2,3
- Shiller, Robert J. and George A. Akerlof (2009). "Animal Spirits: How Human Psychology Drives the Economy, and Why It Matters for Global Capitalism", Princeton University Press
- Smith, V.L., G.L. Suchanek and A.A. Williams. "Bubbles, Crashes,and Endogenous Expectations in Experimental Spot Asset Markets." Econometrica, 56, (1988), pp.1119-1151
- Hart, Oliver and Luigi Zingales (2009): "To regulate finance, try the market" Foreign Policy: http://experts.foreignpolicy.com/posts/2009/03/30/to_regulate_finance_try_the_market
- Benartzi, Shlomo and Richard Thaler (2004): Journal of Political Economy, Vol. 112, No. 1, pp. S164-S187, February 2004
- Thaler, Richard H. and Cass R. Sunstein (2008), "Nudge: Improving Decisions About Health, Wealth, and Happiness", Yale University Press