Union-nonunion compensation differentials across plant sizes: Evidence from CPS 1983

Previous empirical work on the relationship between plant size and union-nonunion differentials has focused either on wages (Podgursky, 1986) or benefits (Bramley, Wunnava, and Robinson, 1989; Freeman, 1981). This note extends this research by simultaneously focusing on both wages and benefits. There are several arguments that can be made in explaining union-nonunion differentials across plant sizes: (1) union threat effects, (2) efficiency wage effects, and (3) wage dispersion effects. Our study focuses on measuring the union-nonunion differential in total compensation. For this end, estimates of total compensation are obtained using Mellow's suggestion (1982) for combining hourly wage information with qualitative data on pensions and health insurance, and Ross's imputations (1989) for holidays and vacations benefits. Our results, based on the May 1983 CPS supplemented by BLS data on pensions, health insurance, holidays, and vacation benefits indicate significant union-nonunion total compensation differentials exist only for workers in establishments with less than 500 workers. These results are consistent with Podgursky's wage differential findings.