Why High Earners Work Longer Hours

Featured in print Digest

"Between 1979 and 2002, the frequency of long work hours increased by 14.4 percentage points among the top quintile of wage earners, but fell by 6.7 percentage points in the lowest quintile."

During most of the 1900s, the hours of work declined for most American men. But around 1970, the share of employed men regularly working more than 50 hours per week began to increase. In fact, the share of employed, 25-to-64-year-old men who usually work 50 or more hours per week on their main job rose from 14.7 percent in 1980 to 18.5 percent in 2001.

This shift was especially pronounced among highly educated, high-wage, salaried, and older men. For college-educated men, the proportion working 50 hours or more climbed from 22.2 percent to 30.5 percent in these two decades. Between 1979 and 2002, the frequency of long work hours increased by 14.4 percentage points among the top quintile of wage earners, but fell by 6.7 percentage points in the lowest quintile. There was no increase at all in work hours among high-school dropouts.

As a result, there has been a reversal in the relationship between wages and hours. In 1983, the most poorly paid 20 percent of workers were more likely to put in long work hours than the top paid 20 percent. By 2002, the best-paid 20 percent were twice as likely to work long hours as the bottom 20 percent. In other words, the prosperous are more likely to be at work more than those earning little. This trend has been a puzzle for some economists.

In The Expanding Workweek? Understanding Trends in Long Work Hours Among U.S. Men, 1979-2004 (NBER Working Paper No. 11895), Peter Kuhn and Fernando Lozano attempt to explain why the century-long trend of shrinking work hours -- probably a reflection of rising prosperity -- reversed around 1970, essentially for the first time except during World War II. The authors also try to reconcile the trend towards longer workweeks for full-time workers with the fact of overall declining participation of men in the labor force. However, as Kuhn and Lozano note, highly educated men were not likely to leave the work force, but rather were much more likely to work longer hours; while high-school dropouts were more often leaving the work force or, if still at a job, working fewer hours.

After testing various possible causes for these trends, Kuhn and Lozano conclude that many salaried men work longer because of an increase in "marginal incentives" to supply hours beyond the standard 40 per week. These workers don't immediately get overtime pay for the "extra" hours. But over a longer time period, they get a substantial reward in the possibility of earning a bonus or a raise within their current position, or they may win a promotion to a better job, or simply signal to the labor market that they are productive and ambitious and thus suitable for a better job in another firm. Alternatively, the longer hours may enable them to acquire extra skills or to establish networks and contacts that could be rewarded in their current firm or in another one. In addition, the long hours may enhance their prospect of keeping their current job if the firm decides to lay off workers in the future. Studies suggest that perceived job insecurity has risen substantially among highly educated workers.

As evidence, the authors note that an extra hour beyond 40/week was associated with a 1.2 percent increase in earnings for male workers overall between 1983 and 1985, and with more than a 2 percent increase by 2000-2. For salaried workers, the man putting in 55 hours per week in the early 1980s earned a weekly salary of 10.5 percent more than an equivalent worker putting in normal hours. By the early twenty-first century, that gap had more than doubled, to 24.5 percent. Such pay gaps, or "long-hours premiums," were accommodated by a markedly wider dispersion of earnings within an occupation between 1983 and 2002.

In their research, the authors are able to rule out several factors as explanations for this change in work behavior. It is not the result of changing techniques in the Current Population Survey, a survey that provides the statistical base for their study. It is not a purely cyclical phenomenon. Nor is it attributable to a changing mix of occupations and industries in the male labor force. It cannot be attributed to rising education levels, an aging workforce, or decreasing unionization. Nor can it be explained by the declining economic fortunes of American men over the past two decades. Real earnings for 40-hour weeks remained essentially flat among hourly male workers in the years between 1983-5 and 2000-2, and increased only slightly for salaried workers. Nor, the authors find, is the change a consequence of increased self-employment. And, it is not related to an increase in multiple jobholding, or to advances in communication technology (such as the Internet) that facilitate additional work from home.

Rather, the authors note, U.S. firms have changed their methods of compensation for skilled, salaried workers over the past quarter century. It could be that longer-than-normal workweeks help firms to produce better products and services in "winner-take-all" type of markets

-- David R. Francis