THE LENGTH OF THE WORK DAY: Implications

I have shown that the distribution of hours worked was much less egalitarian in the past than it is today and that much of the change in the inequality of the length of the work day could be explained by declines in the number of hours workers were willing to supply. In the past, in contrast to recent times, the labor supply curve was strongly backwards bending. The backwards bending labor supply curve of the 1890s implies that circa 1890 increases in wages should have led to a decrease in the length of the work day. In fact, between 1890 and 1919 when real wages increased by 43 percent, the work day fell from 10 to 8 hours. The elasticities estimated for 1973 and 1991 are small and between 1973 and 1991 the length of the work day barely changed. If the elasticity of labor supply in 1920 was small as well then that might explain why the length of the work day has remained unchanged since 1920. Recent estimates of the elasticity of daily hours supplied suggest that the average work day is likely to remain constant or even increase.

w6504-14
The changing labor supply curve has implications for earnings inequality. Table 10 shows that between 1973 and 1991 26 percent of the earnings inequality of men between the 90th and the 10th wage deciles could be attributed to differences in hours worked. For women more than all of the earnings inequality could be attributed to differences in hours worked. Table 10 also shows that had the 1991 pattern of hours worked prevailed in the past (but the number of days worked per week had remained unchanged) weekly earnings inequality would have been much greater in the past than it actually was. Because this calculation does not account for the fact that higher paid workers may have worked fewer hours on Saturday, the extent to which the inegalitarian distribution of daily work hours at the end of the last century equalized income is probably underestimated. Nonetheless, the results suggest that an examination of income alone would underestimate the extent of inequality in the past.
w6504-15
In the 1890s when few wives worked in the labor force earnings inequality among married couples was determined by the distribution of husbands’ hourly wages and hours worked. But, this was not true in 1973 and in 1991. In 1973 and in 1991 the trend in combined hours of work of husbands and wives by the sum of their weekly earnings was very different from that of own hours by own wage decile (see Table 11). Although between 1973 and 1991 the hours of husbands in the top household earnings decile increased relative to the hours of husbands in the bottom of the household earnings decile, labor force participation rates of wives in the bottom of the household earnings decile rose relative to the labor force participation rates of wives in the top of the household earnings decile. Among wives in the labor force, the hours of wives in the bottom decile rose relative to the hours of wives in the top decile.
w6504-16