Response of Labor Productivity and Employment to Changes in Overtime Hours in US Manufacturing from 1982 to 2004

The intensity of use of overtime hours has risen markedly since the early 1990s. Recent research suggests that there has been a structural break in manufacturing productivity beginning in the 1990s in the US. This paper investigates how a given increase or decrease in hours - measured here as average weekly hours of overtime work - affects the subsequent adjustment of output, productivity and employment in the manufacturing sector. Impulse responses are constructed using parameter estimates obtained from vector auto regression (VAR) models consisting of hours, employment, output and wages in the manufacturing industries over the period 1982-2004. The period is subdivided into two periods, prior to and after a structural break identified in 1995. Impulse responses are used to contrast the relative size of productivity, output and employment adjustments to a given shock in average weekly overtime hours. Findings indicate that in recent years, a given increase in overtime hours suppresses average labor productivity for a more sustained period of months than it had in the past, at least for the non-durables sector of manufacturing. In addition, increases in overtime hours lead to somewhat smaller increases in employment than observed prior to 1995, somewhat in non-durables manufacturing but more so in durables, where evidence of a jobless recovery is more apparent. Thus, the more jobless the recovery the smaller is the negative impact on labor productivity. Implications are derived for future research regarding the jobless recovery, the pro-cyclicality and secular trend of productivity and the long-term escalation of average weekly overtime hours