Sectoral Shift and Labor Force Participation of Older Males

A Traditional explanation for the fall in the labor force participation of older males in the era of industrialization is that it was in part produced by the decline in agriculture. A number of recent studies rejected this view based on the result that farmers were no less likely to retire than were nonfarmers. An examination of a longer period, however, shows that farmers were less likely to retire than were nonfarmers, as the conventional view suggests. Only the first decade of the twentieth century, which the revisionist view drew evidence from, exhibits the opposite pattern. This peculiarity of the years between 1900 to 1910 is likely to have resulted from the unusually high appreciation of farm property during the same period that would have stimulated retirement of farmers. According to the counterfactual LFPR of older males that would have resulted had there been no decline in the relative size of agriculture since 1820, the decrease in the labor force employed in farming accounts for about 20 percent of the fall in the LFPR of men 60 and older between 1880 and 1940. The current concept of the labor force has been in use since the 1940 census; the concept of gainful 1 employment was used by the U.S. census through 1930. Therefore, the data on the labor force before and after 1940 are not directly comparable. Economic historians have been at odds over how to determine whether a person participated in the labor force using the censuses enumerated prior to 1940. The conventional view is that men who were gainfully employed should be counted in the labor force. The estimate of the labor force participation rate of older men based on this definition exhibits a downward trend from the late nineteenth century (Durand 1948, Long 1958, Moen 1997a). Ransom and Sutch (1986) have challenged this view, suggesting that a part of aged men who had gainful occupations may have been practically out of the labor force. For the 1900 and 1910 censuses, they removed from the calculation of the labor force those individuals reporting six or more months of unemployment (the long-term unemployed, hereafter), considering them "permanently unemployed." According to their estimate, the labor market involvement of older men remained stable between 1870 and 1937. See Lee (1998b) for a more detailed survey of this debate and its implications. A number of recent studies have attempted to resolve this debate by examining the labor market status of the long-term unemployed. The results from these studies suggest that it would be more reasonable to use the conventional definition by including the long-term unemployed in the labor force. Margo (1993) found that the long-term unemployed and those with no gainful occupation in the random sample of the 1900 and 1910 censuses had significantly different characteristics. After analyzing a longitudinal sample of older males, Lee (1998b) concluded that the long-term unemployed were more similar to aged male workers who were employed six months or more than those who were not gainfully employed in terms of the pattern of the transitions in the labor market status between 1900 and 1910. The definition proposed by Ransom and Sutch for 1900 and 1910, more importantly, cannot be applied to other years prior to 1940 because the length of unemployment is either incompletely reported or unavailable. For these reasons, I chose to follow the conventional definition of the labor force given in Moen (1987a) that excludes from the counting of the labor force the following males aged 50 and over: inmates of institutions, individuals with no occupation, and those whose occupation was reported as "capitalist," "landlord" or "retired." 1 1. The Issue The labor force participation rate (LFPR, hereafter) of older males in the United States fell dramatically over the last 120 years. In 1880, nearly four out of five men 65 and older were gainfully 1 employed. Today, only 15 percent of males at those ages participate in the labor market. Such a sharp decline in the labor market activity of older men has been regarded as one of the most marked changes in the U.S. labor market that the twentieth century witnessed, along with the rise in the LFPR of married women. Many economists has attributed this phenomenon to the impact of the implementation and expansion of the social insurance programs, especially social security (Boskin 1977, Parsons 1980, 1991,