Controlling for observable characteristics such as the wage and demographic characteristics, workers in occupations where mean unemployment was three months in the year labored almost 2 hours less per day than workers in occupations with mean unemployment of 0. Given their choice of occupation, workers probably had little control over their daily hours.

Work Intensity

Changes in work intensity will affect the relationship between the hourly wage and the number of hours, particularly in a comparison of the 1890s with recent data. When 12 hour days were the norm in the steel industry, the idle time of an open hearth crew might be 54 percent for second helpers and 70 percent for steel pourers. Workers might even spend some of their idle time sleeping. Workers may have wanted idle time on the job and could take it if they enjoyed broad autonomy over the pace of their work, working as semi-independent businessmen within the firm, as was true of skilled workers in iron, glass, pottery, foundry, mining, and precision industries (Montgomery 1979: 11, 37-38, 41).

Although I cannot ascertain exactly whether the workers working the longest day were also those working less intensely, an examination of the 1890s data suggests that the observed relationship between the hourly wage and the number of hours is not determined by the intensity of work. Craftsmen had the greatest control over the pace of their work and therefore might be more likely to work in a less intensive, pre-industrial fashion, but their average hours were lower than those of operatives and they constituted only 9 percent of the bottom 20th wage percentile and 63 percent of the top 20th. Furthermore, craftsmen’s ratio of daily hour worked in the 90th to the 10th wage percentile was similar to that observed among operatives and laborers (see Table 4). Firms worried about the pre-industrial work habits of their employees sought to pay by the piece or by the hour (Montgomery 1979: 38), but among hourly workers the difference in daily hours between the lowest and highest paid workers was greater than among non-hourly workers (see Table 2).