Excess Labor and the Business Cycle

In a remarkable empirical study of 168 U.S. manufacturing plants, James Medoff and Jon Fay (1985) have examined the magnitude of labor hoarding during economic contractions. They found that during its most recent trough quarter, the typical plant paid for about 8 percent more blue-co&t hours than were needed for regular production work. Some of these hours were used for other worthwhile work, and after taking account of this, 5 percent of the bluecollar hours was estimated to be hoarded for the typical plant. The hypothesis that firms may hold “excess labor” during contractions was explored in my 1969 study, using monthly three-digit industry data. A model of labor demand was developed that is based on the idea that firms may at times hold excess labor. This model was originally estimated using the monthly three-digit industry data, and it was later estimated using aggregate quarterly data. The aggregate labor demand equations are part of my U.S. macro model. The latest discussion of the aggregate equations is in chapter 4 in my 1984 study. Both the monthly industry estimates and the quarterly macro estimates support the excess labor hypothe sis. The purpose of this paper is to see if the quantitative estimates of Medoff and Fay are consistent with the aggregate estimates. If this is the case, which the results in this paper show, it provides a strong argument in favor of the excess labor hypothesis. Essentially the same conclusion has been reached using two very different data sets. This is one of the few examples in macroeconomics where a hypothesis has been so strongly contirmed using detailed micro data.