The Onset and Persistence of Secular Stagnation in the U.S. Economy: 1910-1990

Beginning of article

Examination of the rate of total output growth over the last century reveals a surprising break in trend around 1910. Before this time, the annual rate substantially exceeded 4 percent a year. However, using John Kendrick's real GNP data reported in the Historical Statistics of the United States, we find real growth from 1913 to 1929 was only 2.8 percent. For the entire period 1909-1970, the real GNP growth rate was only 3 percent, and from 1970 to 1990 it was 2.7 percent. The economy has never recovered secularly from the disastrous drop in the growth rate that occurred around 1910.

Retardation in the total output trend sometime in the twentieth century has long been recognized by such scholars as Kendrick [1961, 59-601, Gailman [1966, 8-91, and Denison [1962, 30]. However, the time of the trend fall has not been pinpointed.

Four secular concerns--the timing of the retardation onset, the main relevant economic accompaniments of that trend break, the possible explanations for the break suggested by those concomitant changes, and the consequences for the economy and economic policy--are the focuses of this discussion. We presume to have some probable but not definitive answers to these questions. But the phenomenon has been bypassed, and even modest contributions should be worthwhile.

Retardation is not stagnation per se. However, the post-1910 record can be interpreted as stagnation. To do this, it is necessary to introduce the concept of potential growth.

There have been a large number of revisions of the estimates of real GNP/GDP or NNP for the period 1870-1930. The most widely used are those of Kendrick [1961 59-60], who modified the work of Simon Kuznets and brought it into conformity with the standard national income accounting. Five such series calculated over the last 35 years are by: Kendrick, Berry [1978, 451, A. Schwartz [1975], Balke and Gordon [1989, 84-85], and C. Romer [1989, 22-23]. Although they show some disagreement concerning levels for some periods, the long-term patterns are remarkably consistent.

The trends in Christina Romer's data are essentially the same as in other studies, although she finds cycle differences. They all show that laissez-faire capitalism in the United States began a significant slowdown of growth about 1909-10.

In the two longer periods, 1878-1908 and 1908-1928, the difference in growth rates averages 1.14 percent per year among the five estimators. No one thinks the decline in the growth rate is less than the 0.84 percent obtained by Schwartz. Balke and Gordon agree with Romer that the real GNP in 1982 dollars was $384 billion in 1910.

If real GNP subsequently grew at the rate of 3.16, obtained by Schwartz for 1908-1928, the real 1990 GNP would have been $4,626 billion. Had it grown at the 4.01 she found for the earlier period (the lowest of the early period estimates), it would have been $8,919 billion by 1990. We can then say that the decline in the growth rate reduced 1990 GNP by $4,293 billion. Measured real GNP in 1982 dollars was $4,156 billion, so the 3.16 growth rate of 1908-1928 was above the average for the rest of the century. On a three-year average basis, GNP grew at 3.17 percent, almost exactly the Schwartz, Berry, and Romer estimates, for …