The problem with forecastsby Stephen Nickell / February 24, 2010 / Leave a comment
Published in March 2010 issue of Prospect Magazine
Since the financial crisis started, economists have been lambasted for failing to predict it. But foretelling the future is always tricky, as any climate scientist will tell you. Foretelling the future of the economy is especially tricky because the very act of publishing forecasts will, if they are believed, affect the future path of the economy. At least the weather doesn’t respond to weather forecasts.
Suppose that in early 2007 economists had predicted that 2008 was going to be the year of financial collapse. Assuming that they were believed, people would have become a lot more cautious, and wary investors would have started to sell assets to ensure that they got out in time. In the blink of an eye, sellers would have been rushing into the market, asset prices would have plummeted and 2007 would have been the year of the financial collapse, not 2008. The economists would have been wrong yet again.
Of course, in practice, people who attempt to foretell the economic future are not believed, not least because there is always a wide variety of forecasts. Look at the British housing market. Every year since 2002, some pundits have predicted a housing market crash the next year. In 2007, they were right. But in 2002, 2003, 2004, 2005, and 2006, not to mention 2008, they were wrong.
This is an excellent example of why it is not a good idea to believe forecasts. If you need somewhere to live and have the income, then buy a house. The pundits may be right—in 2007 you would have been better off if you had postponed your purchase. But if you had believed them in 2002 and held off, it would have been a bad mistake. Since then, the real price of housing has been above its 2002 level. Indeed, from 2004 the real price of housing has exceeded its level in 2002 by at least 20 per cent.
Forecasting turning points in an asset market is a mug’s game and an investment strategy based on such predictions is not, on average, going to work. A better plan is…