Racial Diversity and Macroeconomic Productivity across US States and Cities

Abstract

Sparber C. Racial diversity and macroeconomic productivity across US states and cities, Regional Studies. Racial diversity in the USA continues to rise. Past analyses have argued that diversity can have both positive and negative consequences. The overall macroeconomic effects of diversity within the USA require further examination. This paper exploits variation across US regions from 1980 to 2000 to determine whether racial heterogeneity creates gains or losses for states and cities. Fixed-effects analysis indicates that diversity enhances the productivity of cities. Evidence at the state level is more ambiguous, as significant results only appear in random-effects specifications.

Sparber C. La diversité ethnique et la productivité macroéconomique à travers les états et les grandes villes aux E-U, Regional Studies. Aux Etats-Unis, la diversité ethnique est toujours à la hausse. Des études antérieures ont cherché à affirmer que les conséquences de la diversité ethnique peuvent s'avérer à la fois positives et négatives. L'effet global macroéconomique de la diversité aux Etats-Unis nécessite un examen plus approfondi. Cet article analyse la variation à travers les régions des Etats-Unis entre 1980 et l'an 2000 afin de déterminer si, oui ou non, l'hétérogénéité ethnique a des retombées positives ou négatives sur les états et les grandes villes. Une analyse par effets fixes laisse voir que la diversité augmente la productivité des grandes villes. Les preuves qui proviennent des états sont plus ambigues, parce que d'importants résultats ne sont évidents qu'à partir des spécifications des effets aléatoires.

Diversité ethnique Productivité macroéconomique

Sparber C. Ethnische Vielfalt und makroökonomische Produktivität in Bundesstaaten und Städten der USA, Regional Studies. In den USA nimmt die ethnische Vielfalt weiter zu. In den Analysen der Vergangenheit wurde argumentiert, dass sich eine Vielfalt sowohl positiv als auch negativ auswirken kann. Die generellen makroökonomischen Auswirkungen der Vielfalt innerhalb der USA bedürfen einer genaueren Untersuchung. In diesem Beitrag werden die Schwankungen innerhalb der Regionen der USA im Zeitraum von 1980 bis 2000 untersucht, um festzustellen, ob eine ethnische Heterogenität für die Bundesstaaten und Städte Gewinne oder Verluste erzeugt. Eine Festeffekt-Analyse weist darauf hin, dass sich durch Vielfalt die Produktivität der Städte erhöht. Die Belege auf der Ebene der Bundesstaaten fallen weniger eindeutig aus, da signifikante Ergebnisse nur in den Spezifikationen willkürlicher Effekte auftreten.

Ethnische Vielfalt Makroökonomische Produktivität

Sparber C. Diversidad racial y productividad macroeconómica en estados y ciudades de EE.UU., Regional Studies. La diversidad racial en los Estados Unidos sigue creciendo. En análisis previos se ha postulado que la diversidad puede tener consecuencias tanto positivas como negativas. Es necesario examinar mejor los efectos macroeconómicos generales de la diversidad en los Estados Unidos. En este artículo analizo la variación entre las regiones estadounidenses de 1980 a 2000 para determinar si la heterogeneidad racial crea ganancias o pérdidas a los estados y las ciudades. Los análisis de efectos fijos indican que la diversidad aumenta la productividad de las ciudades. La evidencia a nivel estatal es más ambigua porque los resultados significativos sólo aparecen en las especificaciones de efectos aleatorios.

Diversidad racial Productividad macroeconómica

Keywords:

• Racial diversity
• Macroeconomic productivity

JEL classifications:

• J24
• O40
• O51
• R11

Acknowledgements

The author would like to acknowledge the guidance and advice from three anonymous referees as well as from the faculty and students of Colgate University, Union College, and the University of California – Davis. The author especially thanks Giovanni Peri, Hilary W. Hoynes, Alan M. Taylor, Florence Bouvet, Takao Kato, and Ahmed S. Rahman. A fellowship from the University of California Office of the President provided support for this research.

Notes

Estimates are based on standard metropolitan areas – not incorporated city limits.

For the earliest analyses, see Mauro (1995) and Knack and Keefer (1997).

Also see Greif's (1993) account of eleventh-century Maghribi traders.

In addition to previously cited literature, see also Caselli and Coleman (2002), Alesina et al. (1999), and Poterba (1997). Putnam (2007) provides a more extensive survey of this literature.

Individual minorities also chose to cooperate more frequently than did individual Whites, but these rates fell between the cooperation rates of White teams and diverse teams.

For similar objections, see Posner (2004) and Fearon (2003).

Hispanics complicate the race and ethnicity dichotomy. According to the US Census, as well as the National Research Council, Hispanics compose an ethnic group. However, Hispanics often see themselves as belonging to a separate race. The National Research Council (2004, p. 34) writes, ‘In the 2000 Census, 97 percent of people reporting “some other race” were of Hispanic origin.’ Rather than subscribing to a traditionally defined race, ‘about one-half of Hispanics either marked “some other race” or marked “two or more races”’ on the Census form. This motivates the National Research Council to argue that ‘Hispanic’ is an ethnicity and not a race in the traditional definition of the term, but that analyses of race in the USA should include Hispanics as a distinct group.

‘Hispanics’ includes all those who claimed Hispanic origin on the Census form. Therefore, the ‘White’ variable is equivalent to ‘White, Non-Hispanic’ (and similarly for Asians, Blacks, and Others). In 2000, respondents were allowed to select ‘Two or more races’ on the Census form. The analysis in this paper categorizes individuals as those who chose this option as ‘Others’, so long as they did not also mark ‘Hispanic’ on the form.

For a list of the 103 metropolitan regions for which the IPUMS provides necessary demographic data in each decade, see Appendix A1.

Unlike for subsequent decades, the 1970 Census does not record the state (or metropolitan area) of a person's employment. Instead, this paper assumes that an employee worked in the same state (and metropolitan area) in which he or she lived in 1970.

Employment figures come from the IPUMS. The US Department of Commerce, Bureau of Economic Analysis (2004) provided real gross state product data for each of 1980, 1990, and 2000. As of 26 October 2006, the BEA renamed the gross state product series ‘gross domestic product by state’. This revision created a discontinuity in 1997, when data changed from Standard Industrial Classification (SIC) to North American Industry Classification System (NAICS) industry codes. The BEA now recommends against appending data before and after this date. However, the data in the present paper were obtained before the BEA revisions occurred, and are available upon request.

Appendix A2 offers an alternative methodology involving growth accounting and total factor productivity regressions.

The average years of education for the workforce are estimated using the IPUMS education recode (EDUCREC) variable.

The BEA identifies eight economic regions: New England, Middle Atlantic Coast, Great Lakes, Plains, Southeast, Southwest, Rocky Mountains, and Far West.

The Hausman test maintains the null hypothesis that random effects are sufficient and fixed effects are unnecessary. The test statistic comes from the F-distribution, and the null is rejected if the corresponding p-value is less than 0.05. The regression in column 6 of Table 3 delivers a p-value of 0.20, suggesting that the random-effects specification is valid.

This assumption will be relaxed later in the analysis.

This is the IPUMS variable RENT divided by ROOMS.

Appendix A3 illustrates that the wage effects identified in this section do not vary across regions of the USA.

As with all other share variables, the employment share value is recorded in whole-number terms.