Racial Minorities, Economic Scale, and the Geography of Self-Employment

Self-employment is receiving substantial attention in the United States and elsewhere. The notion that self-employment is beneficial for both individuals and for society as a whole has a long history in the United States; witness Horatio Alger's stories and references to them in modern culture.¹ Among the recent echoes is the notion that encouraging entrepreneurship would be a good strategy for improving the economic status of blacks. According to Washington Post journalist William Raspberry, Paul Pryde, president of Capital Access Group LLC, has argued that there "is a need to shift the black focus from jobs to ownership, from income to wealth, from political office to using politics to improve the climate for black business development."² This paper sheds light on the striking racial differences in the frequency of self-employment by examining how the characteristics of local populations affect the decision of minorities to enter self-employment.

Our research bridges several recent areas of interest. First, there is now an extensive econometric literature on self-employment. Much of the focus has been on two related questions: at a given point in time, what factors determine who is self-employed and who is not. Or, alternatively, over the course of, say, a year, who makes a transition from or into self-employment.³ A central point of this literature has been the degree to which access to capital limits the ability of individuals to enter self-employment, especially the role of such constraints in explaining racial differences in self-employment rates. Our research also draws on the literature in discrimination. In particular, Stephen G. Bronars and George J. Borjas (hereafter referred to as BB) and Dan Black construct simple equilibrium models of the decision to enter self-employment when consumers or employers may be prejudiced. Using data from the 1980 census, BB found that the likelihood of minority self-employment in a given area increased with the fraction of the minority population in that area. Black also shows that spatial concentration of minorities into enclaves serves to reduce the incidence of discrimination by employers.⁴

Finally, our analysis is also indirectly related to the literature on spatial mismatch. In that literature the suburbanization of jobs coupled with suburban housing (and labor) market discrimination against minorities reduces the employment opportunities of inner-city minorities.⁵ Although the implications of spatial mismatch for employment and compensation outcomes have been extensively studied, implications for minority self-employment rates have been largely overlooked. A possible response of minority groups to spatial mismatch, however, would be to seek out self-employment opportunities. Because racial segregation (and discrimination) is closely associated with concerns of spatial mismatch, this suggests that increased segregation could be associated with greater minority self-employment, other things being equal.

The starting point for our investigation is the considerable spatial variation in self-employment rates, in general, and race-specific self-employment rates, in particular. In all that follows, individuals are defined as self-employed if they report self-employment as their principal job on their 1990 census form. Based on that definition, table 1 shows regional rates of self-employment for men, 25 to 64 years old, living in urban areas; these individuals are the focus of our analysis. As shown, the overall U.S. self-employment rate of 11.4 percent embodies a range from 9.9 percent in the Northeast to 12.7 percent in the Pacific region; a difference of nearly 30 percent.

For race-specific rates of self-employment, two facts stand out. First, the overall rate of self-employment differs greatly across races, ranging from a low of 4.3 percent among blacks to 12.7 among whites. Second, within individual minority populations, the spatial variation in self-employment rates across census regions is considerable. For blacks, the lowest rate is 3.2 percent, while the highest rate (5.5 percent) is nearly 70 percent greater. Among Hispanics, the difference is an even more striking 166 percent (11.2 percent versus 4.2 percent), while among Asians, the...