Say ”R”

Economists have a dismal record in predicting recession.

Print

As recently as February, 95% of American economists said it wouldn’t happen, but it has. America is now in recession, according to — don’t laugh — the Business Cycle Dating Committee at the National Bureau of Economic Research (NBER), the official arbiter of American business cycles. This group of six economists reckons that the recession began in March, after ten years of expansion, the longest in American history.

To followers of The Economist‘s R-word index (which counts the number of times that the word “recession” appears in American newspapers), this comes as no surprise. The index, unrigorous as it is, has accurately pinpointed the start of previous recessions, and it started to flash red in the first quarter of this year. But most American economists thought otherwise. Last January The Economist‘s poll of forecasters predicted, on average, that GDP growth would be 2.3% in 2001. Now the country will be lucky if it sees growth of 1%. Even in early September few economists were forecasting a recession. Now it appears that one had already been under way for almost six months.

How should you define recession? The most popular rule of thumb is that it means two consecutive quarters of declining GDP, and so far America has seen only one quarter’s decline. But the NBER committee rejects this criterion as neither necessary nor sufficient.

It is possible to have a recession without two consecutive quarterly contractions, as America and others have had. GDP could fall sharply in one quarter, for example, rise slightly in the next quarter, and then plunge again in a third. That certainly ought to count as a recession. On the other hand, if output fell only slightly in each of two consecutive quarters, that might not be enough to warrant the label. Most newspapers were quick last week to declare the German economy in recession after German GDP fell by 0.03% in the second quarter and by 0.1% in the third. That would probably not pass the NBER recession test.

The NBER committee defines recession as a significant decline in activity, spread across the economy and lasting more than a few months, and also visible in industrial production, employment, real income, and business and retail sales. The declared starting-point of a recession is then based on a compromise between the different times at which all these indicators start turning down.

The committee does not pay much attention to GDP figures themselves, since these are published only quarterly and are subject to large revisions. The first estimate of GDP for the third quarter of 1990, the official start of the previous recession, showed that the economy was still growing. It was later revised down to reveal that the economy had already started to contract. There is a good chance that America’s slim reported growth in the second quarter of this year will also be revised away.

It does not help that “recession” means different things in different countries. Conventionally, it is seen as a fall in output, yet that is too crude. For instance, very low GDP growth in an economy with rapid population growth implies falling GDP per head. Likewise, low growth that is below trend in an economy with a high trend rate of growth (perhaps because of a fast-growing labour force or high productivity growth) will feel like a recession and will cause unemployment to rise.

This problem is particularly acute in emerging economies. For instance, South Korea’s GDP has risen by 1.8% over the past year. By European standards, that sounds fairly respectable, but compared with South Korea’s average annual growth of 6% over the past ten years, it should be counted as a recession. A more sensible definition of recession might be when growth falls significantly below its long-term potential, causing unemployment to rise. In practice, potential growth rates are hard to estimate.

Depth Gauge

Why are recessions so hard to forecast? A study published last year by the International Monetary Fund looked at the economists’ record. It is bad. Of 60 recessions in developed and developing economies during the 1990s, two-thirds remained undetected by consensus forecasts as late as April of the year in which the recessions occurred. In one-quarter of cases, the consensus forecast in October of that year still expected positive growth.

Are economists any better at predicting when recessions will end? Most now forecast that America will start to recover early next year, following one of the mildest recessions in its history. The consensus forecast is for a peak-to-trough decline of just over 1%, compared with an average decline during recessions in the past half-century of over 2%.

The average recession since the second world war has lasted for 11 months, which suggests that the recession might be over by next February. On the other hand, between 1854 and 1945 the average recession lasted for 21 months. The present economic cycle may have more in common with those before the second world war than with those after. In particular, this recession, unlike earlier post-war recessions, was not caused by America’s Federal Reserve raising interest rates sharply in response to a burst of inflation. Rather, a financial bubble popped and an investment boom turned to bust. Such recessions, with the deflationary forces they generate, tend to be more prolonged — witness Britain in the early 1990s and Japan today.

Copyright © 2001 The Economist Newspaper and The Economist Group. All rights reserved.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>