Relationship between hours and employment in manufacturing: Does work sharing make sense?

Over the business cycle, hours of work are a known to complement employment, decreasing when the economy is slowing down and rising when the macroeconomy is recovering or expanding. During a recession, both hours and employment decline. However, in a recovery period, though the length of the average workweek increases, there may be no commensurate increase in employment, particularly in sectors such as manufacturing. This potential asymmetry in the pattern and lag times between hours and employment change is one of the factors behind the phenomenon known as a “jobless recovery.”

The purpose of this paper is to investigate whether reducing hours in an expansionary period would further augment employment and whether reducing hours during a contraction would reduce layoffs. We develop a 4-equation threshold VAR system consisting of manufacturing employment, the length of the average workweek, manufacturing output, and real manufacturing wages.  The purpose of the threshold VAR system is to distinguish between expansionary and contractionary periods with the idea that the two periods are not mirror images of each other. VAR analysis permits us to observe the nature of the pattern of employment following a “shock” (adjustment) of a given (“unit”) size in both output and hours. For example, we will be able to compute how much employment would respond following a shock to employment in both contractionary and expansionary periods.  We would then compare this response to how employment would respond if in addition to the output shock, there had also been a negative hours shock.

We examine monthly manufacturing data between 1978 through 2011. The data source for employment, hours and wages is the BLS’s Current Employment Statistics (CES) of the U.S.. Data on manufacturing output will come from industrial production data provided by the Federal Reserve.