https://macropaperwarehouse.com/authors/steven-j.-davis/Steven J. DavisHugo -- gohugo.ioen-ushttps://macropaperwarehouse.com/papers/gross-job-creation-and-destruction-microeconomic-evidence-and-macroeconomic-implications/https://macropaperwarehouse.com/papers/gross-job-creation-and-destruction-microeconomic-evidence-and-macroeconomic-implications/<p>Davis and Haltiwanger use a longitudinal establishment-level data set covering roughly 160,000 U.S. manufacturing establishments from 1972 to 1986 to document that gross job creation and destruction are large, persistent, and strongly countercyclical, and they argue that standard business-cycle analysis, which abstracts from this establishment-level heterogeneity, is for this reason incomplete. They calculate that annual gross job creation and destruction rates averaged 9.2 percent and 11.3 percent respectively (5.37 percent and 5.62 percent at quarterly frequency), that most of this job creation and destruction is persistent rather than transitory (73 percent of jobs created in 1974-75 still existed a year later, and 72 percent of jobs destroyed over the same interval remained destroyed), and that establishment births and deaths account for a substantial share of the totals (20 percent of gross job creation and 25 percent of gross job destruction). Decomposing the quarterly job reallocation rate (which swings from 6.9 percent to 15.4 percent over their sample) into idiosyncratic, sectoral, and aggregate-time components, they find the overwhelming bulk of its time variation is idiosyncratic rather than driven by differential sectoral responses to common shocks, and that this idiosyncratic component moves countercyclically with both own-sector and aggregate manufacturing employment growth. They then build a theoretical model in which both aggregate demand disturbances and allocative (reallocation-intensity) disturbances drive job flows, showing the two types of shocks have different dynamic signatures &ndash; adverse aggregate shocks raise job destruction and, because reallocation becomes relatively cheap in a downturn, accelerate the pace of reallocation itself, while a rise in the intensity of allocative disturbances raises job destruction immediately but raises job creation only with a lag. Empirically they find allocative disturbances account for a large share of the variance of job creation and destruction movements at short, medium, and long horizons, and that measures of job reallocation have a statistically significant positive relationship with unemployment.</p>