Hard times ahead Depressions caused by loss of economic confidence might be predictable. 15 August 2001
PHILIP BALLLibrary of Congress
A new model of economic fluctuations suggests that a global recession is indeed imminent - probably worse than that of 1990-91 - and that forecasters should have spotted it a year ago1.
Depressions in the US economy become a near-inevitability at least a year before they happen, says Elliott Middleton of Placemark Investments in Dallas, Texas. His model predicts, a year in advance, all five of the slumps between 1969 and 1998, and generates no false alarms.
With this track record, the model's forecast of a recession in the second half of 2001 seems worth taking very seriously. Last May, professional economists put the chance of such a slump at about one in three; now many businesses are bracing themselves for it. Middleton expects the coming recession to be less serious than those of the 1970s and 1980s.
Slumps seem to be an inevitable aspect of any economy; the business cycle of boom and bust has become a standard part of economic lore. But these changes are not really cyclic at all. Although bad times inevitably follow good, there is no regularity to the rise and fall of the economy.
This makes recessions embarrassingly hard for economists to predict. In the words of John Kay of the London Business School: "Economic forecasters all say more or less the same thing at the same time, but what they say is almost always wrong."
Middleton believes that there is something of a self-fulfilling prophecy about recessions. Businesses and investors lose confidence in the buoyancy of the market; their timidity causes a slowing and eventually a turnaround of economic growth. The British economist John Maynard Keynes dubbed these subjective levels of market confidence 'animal spirits'.
Nothing, it seems, deflates these spirits like unemployment. The moment the unemployment rate ceases to improve, confidence plummets and the economy heads for a turnaround, transforming a boom into a slump. Because unemployment levels cannot get better for ever, a boom, in effect, carries the seed of its own demise.
Loss of confidence can be detected in the variations of an economic index before the slump itself takes hold, Middleton says. Crucially, he asserts, market agents are adaptive: they take their norms from recent events rather than from any absolute judgement about what is a good or bad unemployment rate.
So Middleton's model derives confidence changes from adaptive changes in unemployment, and uses the resulting prediction to forecast slumps.
For the post-war US economy, the model works well. But in other countries - where the memory of the 1930s' Great Depression is less haunting - unemployment might not be the main influence on confidence levels, Middleton cautions. This might explain why previous attempts at forecasting have met with varying success in different countries.
References
1. Middleton, E.'Animal spirits' and expectations in nonlinear US recession forecasting. Preprint, (August 2001).
-------------- next part -------------- An HTML attachment was scrubbed... URL: <../attachments/20010815/d1d5f51c/attachment.htm>