A Century of Cycles

September 8, 1958

In a study just issued by the National Bureau of Economic Research, a private organization, Geoffrey H. Moore finds that recent data support earlier indications that the current business contraction may be drawing to a close. Most measures of aggregate economic activity, he finds, have risen above their lows of April or May and personal income has regained its pre-recession level.

These immediate facts are largely known. What I find most interesting in the study is the elaborate comparisons of past business cycles, particularly the comparisons of 24 cycles over a period of a century, from 1854 to 1954. These disprove some hardy myths about business cycles.

Perhaps the hardiest of these myths is the belief in the “regularity” of the business cycle. This belief is found, for example, in Keynes’s General Theory of Employment, Interest and Money: “There is some recognizable degree of regularity in the time sequence and duration of the upward and downward movements. . . . The duration of the downward movement [has] an order of magnitude which is not fortuitous, which does not fluctuate between, say, one year this time and ten years next time, but which shows some regularity of habit between, let us say, three and five years.”

MEASURING RECESSIONS

Now measuring business cycles is an extremely difficult (and partly arbitrary) task. Different indicators of economic activity have different amplitudes of movement. All reach their relative peaks and troughs at different times. But the statistical studies of the National Bureau of Economic Research are the most complete and meticulous yet made. And we find from Dr. Moore’s tables that the average duration of the downward movement of the 24 cycles in the period from 1854 to 1954 was just twenty months.

But this statistical average conceals a wide range of duration. The contraction beginning in August 1918 lasted only seven months; that beginning in October 1873 lasted 65 months. In spite of Keynes’s impression of regularity, here is a difference in duration of almost ten times as much in one case as in another.

Had Keynes been discussing the average duration of the whole cycle, instead of merely the downward phase, his guess would have come near the mark. The expansion and contraction phase together, of the 24 cycles, add up to just 50 months, or slightly over four years. But this average again conceals wide differences. For whereas the average expansion phase of the 24 cycles lasted 30 months, the range was from as low as ten months to as long as 80 months.

PRICES VS. WAGE RATES

Statistics by themselves cannot provide either forecasts or remedies. We also need a study of the reasons behind the statistics. Moore points out that with the exception of the 1929 depression, the typical length of the contraction period seems to have shortened since 1920. I suggest that the main reason for this is that since 1933 we have followed an almost continuous policy of inflation. This has restored a workable relationship between costs and prices not by reducing costs but by further raising prices.

But it does not follow that we need the evils of inflation to cure the evils of recession. There is a striking and instructive comparison between 1920 and 1929. Judged by almost any standard, the crisis of 1920–21 was far more severe than the crisis of 1929–30. Here, for example, drawn from Moore’s tables, are the percentage declines in selected indicators in one year in the first period as compared with the second:

Why is it, then, that the recovery from 1921 was astonishingly rapid, whereas the depression of 1929 got steadily worse? The chief answer will be found in what happened in the two periods to wage rates. In the first they were still flexible downward as well as upward; they adjusted to lower prices, and allowed a quick recovery.