The one question you must never ask an economist

The one question you must never ask an economist
Mark Thornton

Economists are everywhere. Steve Levitt, Tim Harford and Steven Landsburg use newspaper columns and best-selling books to show how economics can account for why drug dealers live with their mums, why you can’t find space to park, why school teachers cheat, why people share umbrellas and why sexually transmitted diseases are so rife. Simple economics, it seems, can explain everything.

Everything, that is, except the economy. Although orthodox economics can do a good job of explaining why people get a divorce or the clap, it does a much worse job of accounting for what people think it should explain.

Take last week’s tumbles in world stock markets. These raise four puzzles.

Take last week’s tumbles in world stock markets. These raise four puzzles. First, why should the threat of default on some US sub-prime mortgages – loans to high-risk, poorer customers – be such a big deal? Conventional economics says risk should be split up into small bits and sold off to those people most willing and able to take it; all that jargon about collateralised debt obligations describes how this is done. Spread over trillions of dollars of assets, losses of even billions of dollars should be little problem. As recently as March, Ben Bernanke, the chairman of the Federal Reserve, said: “The impact on the broader economy and financial markets of the problems in the sub-prime market seems likely to be contained.” Last week at least, stock markets thought he was wrong. Why?

Secondly, there’s a timing problem: why should stock markets worry about the problem now? The risks of sub-prime lending have been known for months; stock markets fell (albeit temporarily) in February for just this reason. Basic economics – the idea that the market is efficient – says that information should be immediately embodied in share prices. It shouldn’t take so long for sub-prime problems to hit prices.

The third puzzle came on Thursday, in the market’s reaction to the injection of €95 billion into the banking system by the European Central Bank. Investors could have thought: “There’s more money around. This is great for shares.” But they didn’t. They thought: “The ECB’s bailing out banks – things must be even worse than we knew” and shares fell as a result. Why did the latter reaction dominate?

The fourth puzzle, deepened by yesterday’s recovery in prices, is: why are shares so volatile? The past few weeks have reminded us of what Robert Shiller, of Yale University, established back in 1981, that shares move much more than their “value” – the discounted present value of future dividends – would warrant.

These four puzzles all have a common root. Orthodox economics assumes that people know roughly what they are doing, that they are rational, and that rationality is unambiguous. Such assumptions are often fair enough in everyday life – hence the justified success of Levitt, Landsburg and Harford. But in financial markets, people often don’t know what they are doing. Recognising this helps to solve our puzzles.

So, risk can’t be allocated efficiently because it can’t always be measured accurately and priced. In particular, we just can’t know the small probability of one-off events such as default. These are, in Donald Rumsfeld’s useful phrase, unknown unknowns.

This means that many hedge fund managers, for all their fancy jargon and maths PhDs, do what Nassim Nicholas Taleb accused them of in his book The Black Swan: they are just picking up pennies in front of a steamroller. And sometimes the steamroller accelerates.

And if risk can’t be measured precisely, that leaves a role for sentiment; sometimes people are happy to lend, other times not. There can therefore be systematic financial crises and bubbles which conventional economics, with its emphasis on well-functioning markets, denies.

Secondly, facts aren’t always embodied immediately in prices because we just don’t know what these facts are. Economists use the word “data” – Latin for “givens” – as a synonym for “facts”. But facts aren’t given. They have to be found. One way we do this is to take cues from what others believe, because only an arrogant charlatan thinks he knows it all.

So, in February – when the stock market dipped – we worried about the sub-prime problem because others worried. Then others stopped worrying and the market rose, so we stopped worrying. And in the past few weeks they started worrying again. These are examples of information cascades – people believe things because others do.

Thirdly, rationality is ambiguous. It would have been reasonable for the market to react with optimism to the ECB’s cash injection last week – “there’s more money around”. If so, it would have been an example of what the American philosopher Robert Nozick called – apologies for the jargon – causal expected utility. The pessimistic reaction – “things must be bad” – is what he called evidential expected utility. Both principles, he explained, are rational. Both are also contradictory and it is utterly arbitrary which way the market will jump. The idea that the rational response leads to a clear course of action is wrong.

Fourthly, markets are volatile partly because we can’t quantify the probabilities of booms or disasters. As these probabilities change even slightly, prices will swing a lot even though the actual economy stays quite stable. In these ways, the gyrations in stock markets are genuinely interesting. They’re not just about wide boys losing money, but raise questions about human nature. What can we know? How can we know it? What exactly does it mean to be rational?

But these gyrations also invites every idiot to repeat the cliché, “it’s all about greed and fear”, as if this were anything other than vacuous. And abandoning the elegant discipline of conventional economics opens the door to every crank. To paraphrase G. K. Chesterton, when people stop believing in orthodox economics, they start believing not in nothing but in anything.

But perhaps we should believe in nothing. Maybe the lesson here is that suggested by Alasdair MacIntyre in his classic After Virtue – that there are no law-like generalisations in the social sciences that we can use to predict or control the world.

The “experts” not only know less than they pretend, but cannot possibly know enough. The best we can do is tell nice stories after the fact.

Chris Dillow is a columnist for Investors Chronicle who blogs at stumblingandmumbling.typepad.com