Reprinted with permission from Business Times (Singapore)

If economic forecasters are to be believed, an economic recovery in the United States is just around the corner. Almost all major forecasters predict that the recession that started last year will be over some time in the course of this year.

In other words, forecasters are echoing the words of Chauncey Gardner—the character played by Peter Sellers in the movie Being There—that "there will be growth in the spring." Or certainly by the summer.

Are such forecasts credible? The simple answer is `Yes'. Recessions in industrialized countries have tended to last under a year; so forecasting a recovery in the year following a recession is a pretty good bet.

The evidence is that private sector forecasters have done a reasonably good job of forecasting recoveries in industrialized countries over the 1990s. The forecast error (averaged over all the recoveries) has been about 0.3 percentage points for forecasts made at the start of the year of the recovery.

A word of caution, however. A small number of recessions did drag on for over a year—the record of forecasters in predicting which ones would do so has not been particularly good.

Consensus Forecasts has provided macroeconomic forecasts for industrialized countries on a monthly basis since October 1989. Each issue of the publication surveys a number of prominent financial and economic analysts, and reports their individual forecasts as well as simple statistics summarizing the distribution of forecasts. Attention is generally focused on the mean forecast (the `consensus').

Effect of Desert Storm

The `event' being forecast is annual average real GDP growth. Every month, a new forecast is made of this event. To take a concrete example, suppose that the event is 1991 real GDP growth. The sequence of forecasts that are available for this event are the 24 forecasts made between January 1990 and December 1991. The first 12 forecasts, the ones made during 1990, are referred to as year-ahead forecasts; the 12 forecasts made during 1991 are called current-year forecasts.

Consider first the evidence for the U.S. over that period. The forecast at the 24-month horizon (that is, in January 1990) was for about 2.5 percent growth. Following Iraq's invasion of Kuwait in August 1990, forecasts for U.S. growth started to be marked down substantially. By the start of 1991, the forecast was for a modest recession that year.

How did the recognition of the recession affect the forecasts for 1992, the year following the recession? Not very much initially. The year-ahead forecasts for 1992 (those made during 1991) remained virtually unchanged at about 2.5 percent. The current-year forecasts for 1992 showed somewhat greater variation but clearly never came close to forecasting a continuation of the recession of the previous year. The economy did indeed recover and the forecast error was small.

This pattern is quite typical of forecasts around turning points for other industrialized countries. Recessions typically arrive before they are forecast. The recognition that the country is in a recession does not generally lead to drastic markdowns of the forecast for the post-recession year.

In other words, forecasters act as though the recession is going to last under a year. And since most recessions are indeed over within a year, this behavior turns out to deliver reasonably good forecasts.

Some recessions do drag on for over a year. When this happens, forecasters make large forecast errors in the second year of a two-year recession. An example is the UK recession of 1991-92 where the year-ahead forecasts (made during 1991) were that it will be a normal year with growth of around 2 percent. It is only about half-way through the current-year forecasts that the realization set in that a multi-year recession was underway. In short, large forecast errors were made in 1992 as the recession broke the mould of ending within a year, taking forecasters by surprise.

Also of interest were the UK forecasts for 1993. The surprise that 1992 had in store for forecasters did not affect the forecasts they made for 1993.

This behavior of forecasts is typical of that exhibited in the course of other multi-year recessions in industrialized countries over the 1990s.

Why are recessions more difficult to forecast than recoveries? Victor Zarnowitz has some interesting speculations. He notes that during the boom years "various stresses and imbalances accumulate gradually as more and more industries approach high capacity operations.... Although these internal developments, if permitted to take their course, could alone bring about a downturn, it is also possible for some adverse shocks to speed up this outcome."

Zarnowitz goes on to say that "the forecaster faces an extremely difficult problem in that (a) it is very difficult to anticipate just when the stresses and imbalances will do their work and (b) the timing of true random shocks that matter is always unpredictable, even if their consequences are not."

A Goldman Sachs 2001 report contains a more recent discussion of why private sector economists tend to avoid forecasting recessions. First, recessions are relatively rare events, which make forecasting positive growth a pretty good bet in most years. Second, as a related point, macroeconomic models are built to capture normal relationships among variables and are hence not always suited to predicting recessions unless there is a clear large exogenous shock. Third, there is 'herding' tendency among forecasters.

Breaking from the pack

The report notes that "to forecast a recession substantially ahead of the pack, a forecaster must be willing to deviate from the consensus in an extremely transparent manner, knowing very clearly that in any given year a recession is a low probability outcome."

Once a recession is underway, however, predicting its end is an easier task because forecasters can closely monitor the behavior of indicators that tend to herald a recovery. For example, forecasters can observe how quickly inventories are being adjusted in key industries. Actions taken by policymakers, such as the central bank, can also be studied for clues on how quickly the recovery might arrive. So, under this explanation, recessions are triggered by a combination of accumulated stresses and an adverse (unpredictable) shock, making them difficult to forecast. However, recovery does not require a favorable (unpredictable) shock; it comes about as imbalances are addressed and through the actions of policymakers, all of which can be more easily monitored by forecasters.

Zarnowitz speculates that there may be, in addition, psychological reasons for why recoveries are forecast but not recessions: "Predicting a general downturn is always unpopular, and predicting it prematurely ahead of others may prove quite costly to the forecaster and his customers. On the other hand, most users are likely to await eagerly an upturn during a recognized recession, so forecasts of a recovery will be welcome and often accepted on the basis of early signs of improvement."

Conventional wisdom among forecasters is that the U.S. economy will start to grow again this year. What this forecast has going for it is the fact that most recent recessions have tended to last under a year. But a few recessions do end up lasting longer: when that happens, the evidence suggests that forecasters are caught flat-footed.