"...the decreases in evening and night work were sharpest among those workers earning the highest wages..."
According to a recent study by NBER Research Associate Daniel Hamermesh, "the incidence of evening and night work declined sharply in the United States between the early 1970s and the early 1990s, while the fraction of work performed at the fringes of the traditional working day grew." These changes are consistent with, and magnify, the increase in wage inequality in the United States that occurred during this period, he writes in The Timing of Work Time Over Time (NBER Working Paper No. 5855).
The data that Hamermesh uses come from the federal government: the May Multiple Jobholding supplements to the Current Population Surveys for 1973, 1978, 1985, and 1991. He reports that significantly less work is occurring between 7 p.m. and 5 a.m., and especially between midnight and 5 a.m. However, this has not been replaced by pronounced growth in the 9-to-5 workday, he finds. Rather, over the years there is more work being reported between 6 and 7 a.m., and between 5 and 6 p.m. Hamermesh points out that this does not mean that any individual worker is putting in longer days; rather, the definition of when a standard workday begins and ends seems to be changing.
The patterns that Hamermesh uncovers are similar for men and women, but the changes in men's schedules are more pronounced than those in the women's schedules. "In particular," he writes, "the growth in the fraction of work performed at 6 a.m. and 7 a.m. is significantly greater for men, as is the decline in the fraction performed between midnight and 5 a.m."
Hamermesh finds that the trend away from work in the evening and at night is not explained by changing industrial characteristics or by changing demographics. He believes that the pattern can be explained by two facts: First, work at these times of day is inferior, and so it diminishes with education level and is lower at peak work ages. In other words, people with the most earning power (economists call it "human capital") work at more desirable times.
Second, real wages of the median worker have grown over this quarter century. Thus, people with rising earning power increasingly have avoided working at undesirable times, including overnight. Indeed, most of the change in work timing that Hamermesh uncovers occurred after 1978, while real earnings rose least (or, fell most) in his sample during 1973-8.
Hamermesh finds that the decreases in evening and night work were sharpest among those workers earning the highest wages and lowest among those earning the least. The distribution of work at these undesirable times became less equal during this time period, paralleling the increasing inequality of wages in the United States. He concludes that while work in the evenings and at night declined sharply in the United States between the 1970s and the 1990s, and while the regular workday spread out somewhat from the archetypal 9-to-5 range, "none of these changes was due to the changing demographic structure of the American labor force" nor the result of a "shift in the mix of industries" away from manufacturing and toward retail trade and services.