Abstract
This study uses detailed time diaries from household surveys for 1975 and 1981 to examine how changes in the use of time on the job affect earnings. Among nonunion workers, the marginal minute of break time apparently increases earnings, but not as much as does the marginal minute of work time. Among union workers, additional time in unscheduled breaks appears to be associated with significantly higher earnings, though other break time is not. The author concludes that further growth in on-the-job leisure would reduce productivity, that monitoring workers would yield returns to the firm, but that entirely eliminating breaks would be counterproductive.