We estimate the effects of manufacturers' use of employment services—comprised primarily of temporary help and professional employer organizations—on measured employment and labor productivity in manufacturing between 1989 and 2004. A major contribution of the paper is the construction of panel data on employment by occupation and industry from the Occupational Employment Statistics program. We use these data to document the dramatic rise of production and other manual occupations within the employment services sector and, in conjunction with information from the Contingent Worker Supplements, to estimate the number of employment services workers assigned to manufacturing over the period. Although measured employment in manufacturing declined by 4.1 percent from 1989 to 2000, counting employment services workers assigned to manufacturing, employment in that sector actually rose by an estimated 1.4 percent. Factoring in manufacturers' use of employment services workers does not erase the large declines in manufacturing employment since 2000, but a growing share of manufacturing work in the United States is being performed by employees of staffing agencies. In 2004, employment services workers added an estimated 8.7 percent to direct-hire manufacturing employment, compared to just 2.3 percent in 1989. In addition, we estimate that manufacturers' outsourcing to employment services significantly inflated manufacturing labor productivity measures, accounting for 0.5 percentage points of the annual growth rate from 1989 to 2000 and from 2001 to 2004. Although multifactor productivity measures should adjust for such outsourcing, available evidence suggests that KLEMS, the multifactor productivity measure for manufacturing, does not fully capture the relatively large effects that outsourcing to staffing services has on manufacturing productivity.
