Sunday, September 20, 2009

Time to reinvent economics

Sep 16, 2009

By Robert J. Shiller

THE widespread failure of economists to forecast the financial crisis that erupted last year has much to do with faulty models. The lack of sound models meant that policymakers received no warning of what was to come.

The current financial crisis was driven by speculative bubbles in the housing market, the stock market and commodities markets. Bubbles are caused by feedback loops: rising speculative prices encourage optimism, which encourages more buying, and hence further speculative price increases - until the crash comes.

But you won't find the word 'bubble' in most economics treatises. A search of papers produced by central banks in recent years yields few instances of 'bubbles' even being mentioned. Indeed, the idea that bubbles exist has become so disreputable in much of the profession that bringing them up in an economics seminar is like bringing up astrology to a group of astronomers.

A generation of mainstream macroeconomic theorists has come to accept a theory that has an error at its very core: the axiom that people are fully rational. And as the statistician Leonard 'Jimmie' Savage showed in 1954, if people follow certain axioms of rationality, they must behave as if they knew all the probabilities and did all the appropriate calculations.

So economists assume that people do indeed use all publicly available information and know the probabilities of all conceivable future events. They update the probabilities as soon as new information becomes available. So any change in their behaviour must be attributable to their rational response to new information. And if economic actors are always rational, then no bubbles - irrational market responses - are allowed.

But abundant psychological evidence has shown that people do not satisfy Savage's axioms of rationality. This is the core element of the behavioural economics revolution that has begun to sweep economics over the last decade or so.

In fact, people almost never know the probabilities of future economic events. They live in a world where economic decisions are fundamentally ambiguous, because the future doesn't seem to be a mere repetition of a quantifiable past. For many people, it always seems that 'this time is different'.

Neuroscientists Scott Huettel and Michael Platt have shown, through functional magnetic resonance imaging, that 'decision-making under ambiguity does not represent a special, more complex case of risky decision-making; instead, these two forms of uncertainty are supported by distinct mechanisms'. In other words, different parts of the brain and emotional pathways are involved when ambiguity is present.

Economist Donald J. Brown and psychologist Laurie R. Santos are running experiments to try to understand how human tolerance for ambiguity in economic decision-making varies over time. They theorise that 'bull markets are characterised by ambiguity-seeking behaviour and bear markets by ambiguity-avoiding behaviour'. These behaviours are aspects of changing confidence, which we are only just beginning to understand.

To be sure, the purely rational theory remains useful for many things. It can be applied in areas where the consequences of violating Savage's axiom are not too severe. Economists have been right to apply the theory to microeconomic issues, such as why monopolists set higher prices.

But it has been overextended. For example, the Dynamic Stochastic General Equilibrium Model Of The Euro Area, developed by Frank Smets of the European Central Bank and Raf Wouters of the National Bank of Belgium, is very good at giving a precise list of external shocks that are presumed to drive the economy. But nowhere are bubbles modelled: the economy is assumed to do nothing more than respond in a completely rational way to external shocks.

Milton Friedman and Anna J. Schwartz, in their 1963 book A Monetary History Of The United States, showed that monetary policy anomalies were a significant factor in the Great Depression. Economists such as Barry Eichengreen, Jeffrey Sachs and Ben Bernanke have helped us to understand that these anomalies were the result of individual central banks' effort to stay on the gold standard, causing them to keep interest rates relatively high despite economic weakness.

To some, this revelation represented a culminating event for economic theory. The worst economic crisis of the 20th century was explained - and a way to correct, it suggested - with a theory that does not rely on bubbles. Yet the Great Depression, as well as the recent crisis, will never be fully understood without understanding bubbles. The fact that monetary policy mistakes were an important cause of the Great Depression does not mean that we completely understand that crisis, or that other crises fit that mould.

The failure of economists' models to forecast the current crisis will mark the beginning of their overhaul. This will happen as economists listen to scientists with different expertise. Only then will monetary authorities gain a better understanding of when and how bubbles can derail an economy, and what can be done to prevent that outcome.

The writer is Professor of Economics at Yale University.


No comments: