Bad year for forecasters
Like Christmas trees, economic forecasts should be discarded early in the new year. Anyone who retains a forecast too long will find it wilting more rapidly than an ageing spruce. At times of heightened economic uncertainty, such as we have recently been living through, forecasts have even shorter sell-by dates.
Forecasting errors for the last year were egregious. According to Consensus Economics, in December 2000 forecasters anticipated US growth of 3 per cent. As it turned out, the US economy went into recession some three months later and economic growth for 2001 will come in at around 1 per cent. Market research firms also failed miserably last year in the less ambitious task of predicting growth in their specialist areas. For instance, IDC, which under the motto "Analyse the Future" provides technology research, forecast personal computer sales growth of nearly 10 per cent in 2001. The outcome? The PC market contracted by around 15 per cent.
Why do such errors abound? Every year the world's computing power increases. Forecasters are evolving ever more sophisticated econometric techniques. In theory, these developments should result in more accurate prognosis. Perhaps the theory is wrong. Paul Ormerod is a well-known apostate of the forecasting world whose work will be familiar to many Prospect readers. In a slim and readable book, Butterfly Economics, he turned to chaos theory to explain why forecasting often fails. Conventional economics, explains Ormerod, assumes that individual agents have immutable preferences, regardless of how others behave. Such assumptions lend economics a scientific mien and are necessary for the building of forecasting models. But drawing on the behaviour of ant colonies, Ormerod suggests another model of economic behaviour.
An ant faced with two identical food sources, lying equidistant from the colony, has three choices: first, it may return to the source it last visited; secondly, it may choose at whim to visit the alternative food supply; thirdly, it may be influenced by the recent movements of other ants. The resulting behaviour of ants, switching from one food pile to another, appears to be random but can in fact be modelled using chaos and complexity theory, with its positive and negative feedback loops.
A similar process is at work in the stock market, where the actions of investors are constantly influencing the decision-making of others. As Keynes noted, most speculators are concerned not with the long-term prospects of the company whose shares they purchase, but with anticipating how short-term changes in the market psychology will affect the value of those shares.
How does this relate to recent events? During 1999 and 2000, there was a huge boom in technology spending, driven by grand visions for the application of the internet. This expenditure was funded with a vast issuance of debt and new equity. The boom continued as long as shares soared upwards. However, there came a point when market psychology shifted: shares started to decline, the debt markets baulked and the business models of most internet start-ups appeared increasingly flawed.
Falling bond and share prices in turn sent most dotcoms and new telecoms carriers to the wall, triggering a big reduction in capital spending in both information technology and telecoms equipment. Most business forecasts for 2001 were already redundant by the twelfth night of Christmas.
Central Bank Policy
Before arriving in Washington, Alan Greenspan earned a living as a business forecaster. It is said that he enjoys reading data on factory output while taking his morning bath. However, even the omniscient Greenspan was surprised by the rapidity of the US downturn last year. Now, the Fed Chairman is facing increasing criticism for his management of the economy during the late 1990s.
Over the last two decades, central bankers around the world have been obsessed with controlling consumer price inflation. They have not been as concerned with the inflation of either asset prices or debt. Yet, as Stephen King of HSBC observes, financial bubbles tend to occur during periods of low inflation. This was true of the US in the 1920s and 1990s and Japan in the 1980s.
When asset bubbles collapse, they leave behind both a mass of unproductive investment and a mountain of debt. Currently, US businesses have more unused capacity than at any time since the recession of the early 1980s. Furthermore, according to "debt anorak" Peter Warburton, outstanding liabilities in the US are currently around 170 per cent of GDP. By contrast, in 1990 Japanese debt was running at just over twice national income. When economies carry so much debt they are more than unusually vulnerable to any shock, whether internal or external.
It is more than a decade since Japan's economy hit the rocks and the condition of the world's second largest economy continues to worsen. The consensus forecast is for -0.6 per cent growth in 2001, but this may be over-optimistic. If the US rebounds in 2002-as most economists and both the stock and bond markets predict-then Greenspan's policy of ignoring asset prices will be vindicated. However, if the recovery falters and world economic growth continues at its sluggish pace, then calls for revising central bank priorities to focus on asset price inflation and credit creation will become overwhelming.
