Small Business and Job Creation: Dissecting the Myth and Reassessing the Facts

Article excerpt

FEW IDEAS about the U.S. economy reap greater homage in public discourse than the belief that small businesses are the fountainhead of job creation. Such claims about small business's ability to generate jobs are frequently presented as justification for tax incentives, regulatory policies and other government programs that favor the small business sector.

But it is not only the public discourse that motivates our interest in the relationship between employer size and job creation. Previous academic research convincingly establishes strong connections between employer size and important economic outcomes like the level and inequality of wages, the incidence of fringe benefits, workforce quality, the pace of technological innovation, and the likelihood of unionization.(1) These findings prompt us to ask how job creation and destruction behavior varies by employer size. In this paper, we address that question for the U.S. manufacturing sector. We also evaluate the empirical basis for conventional claims about the job-creating prowess of small business.

Our analysis of job creation and destruction behavior in the manufacturing sector relies upon the Longitudinal Research Database (LRD) housed at the Center for Economic Studies in the U.S. Bureau of the Census. The LRD contains plant-level data at annual sampling intervals for the U.S. manufacturing sector from 1972 to 1988. Information in the LRD permits classification of employers by various characteristics, including plant and firm size. Among U.S. data sets that have been used to study job creation and destruction, the LRD contains the most detailed information on plant characteristics, the most careful treatment of the statistical sampling frame, and the best treatment of plant entry and exit. We exploit the LRD to deepen our understanding of job creation and destruction in the U.S. manufacturing sector and, by extension, the entire U.S. economy.(2)


Although the concept of a job is easy to understand, measuring and interpreting job creation and destruction requires careful definitions. In this study, a job means an employment position filled by a worker. Our data do not distinguish among part time, full-time and overtime employment positions; all count equally as a single job. We do not measure the number of vacancies (i.e., unfilled positions) at a point in time or the change in vacancies over time. Rather, we measure plant-level changes in the number of filled employment positions.

The basic observational unit underlying our job creation and destruction measures is the plant -- a physical location where production takes place. In contrast to a plant, a company or firm is an economic and legal entity that encompasses one or more plants and, possibly, administrative offices specializing in nonproduction activities. While we provide tabulations sorted by plant and firm size, all job creation and destruction measures are cumulated from plant-level employment changes and are calculated over twelve-month intervals.

Because plants represent the observational unit in the LRD, our calculations capture the effects of firms that shift employment between plants. By the same token, however, our calculations do not capture the effects of job shifts between plants. Because of the point-in-time nature of LRD employment data, our calculations also do not record plant-level employment changes that are reversed within the sampling interval. For both reasons -- the failure to capture within-plant job shifts and the point-in-time nature of the employment data -- our job creation and destruction measures understate the true magnitudes.

With these remarks as background, we supply the following definitions: (1) Gross job creation at time t equals employment gains summed over all plants that expand and start up between t-1 and t; (2) gross job destruction at time t equals employment losses summed over all plants that contract or shut down between t-1 and t (In line with these definitions, plants with unchanged employment contribute to neither job creation nor job destruction. …