Contact Us



Forecasts at Turning Points

US economists have grossly and chronically underestimated the US economy's weakness this year, in a fashion typically found during a turning point toward a new recession.

Over the past two months, forecasters have consistently predicted a resumption of the recovery and underpredicted big drops seen over the past two months in employment, production, retail sales and home purchases - even though most of those figures lag by a month, released by the time most of the developments of the reported month are already widely known.

For instance, economists, on average, underestimated the loss of payroll jobs in March by an average 80,000, or 80 per cent, following their prediction of a slight gain in February - when job losses, in fact, totalled more than 350,000.

The gap between their predictions and the actual numbers is now the widest in more than seven years, according to recent research by Goldman Sachs. That, in itself, is worrisome, say historians of the data.

"The consensus forecast typically makes enormous errors at turning points of the economy, when things change," said Anirvan Banerji, research director for the Economic Cycle Research Institute. "Most economists use forecasting models that in some sophisticated way just extrapolate from the recent past. These models fail to account for any possible change in the dynamic or any possible cascading effect."

Prakash Loungani, an analyst with the International Monetary Fund, who has also closely researched the historical accuracy of private-sector forecasts in the US and abroad, warns in a paper to be released this summer that forecasts for a recovery may be "good in cases where recessions end in about a year", since US recessions generally last only a year or less, but "recessions which drag on for a couple of years or which do a double-dip are poorly forecast".

"We are in a window of vulnerability - at a juncture where we could be tipped in recession or away from it," said Mr Banerji, "and the longer this period of vulnerability lasts, the more likely we are to be tipped into recession..."

The uncertainty about the economy's current plight and its future course has discouraged the National Bureau of Economic Research, the academic group seen as the official historian of US business cycles, from declaring an outright end to the recession, which it said began in 2001.

It has also prompted the Federal Reserve policymakers to abstain, for the first time ever, from making statements about the balance of risks between economic weakness and higher inflation.