Two tales on human capital and knowledge spillovers: the case of the US and Brazil

Abstract

This article uses a quasi-Mincerian approach to verify whether the concentration of college-educated individuals employed in the business support services sector and in the own sector contributes to increased productivity in other sectors of the economy. We estimate the returns to education using data from the 2008 US Current Population Survey (March supplement) and from the 2008 Brazilian household survey. This article finds evidence of a positive and significant human capital sectorial spillover effect, which is consistent with Acemoglu’s (1996) conjecture. The sectorial concentration of highly educated workers contributes to increase wages for all workers. This study also finds evidence of increasing returns to education in Brazil and diminishing returns to education in the United States. This finding may be explained by differences in supply of skilled workers in both economies. In addition, the short supply of highly skilled workers in Brazil likely explains the importance of the spillover effect from the business supporting sector.

Keywords:

• human capital
• knowledge externalities
• returns to scale
• economic development

JEL Classification:

• O30
• O57

Funding

This work was supported by CNPq – Brazilian National Research Council [471213-2012-8].

Notes

¹ Pecuniary externalities result from changes in relative prices and are different from technological externalities, which result from changes in resource endowments.

² The selection equation is estimated including the following variables: family income, number of family members, number of children under 14 years old in the family, years of schooling, age, region, skin colour, marital status and female with at least one child under 14 years old. To obtain the estimates omitted in the article, please contact the authors.

³ The microdata can be downloaded from http://www.ibge.gov.br.

⁴ More precisely, for Brazil, experience is computed as the individual’s age minus the age at which he/she informed to have started working, divided by 10. For individuals who did not provide information about when they started working, experience is computed as their age minus their years of schooling, divided by 10. For the US, potential experience is computed as age less years of schooling less six. We could have used potential experience for both countries in order to gain compatibility, but this would lead to a serious underestimation for Brazilian workers, who often start working long before they leave school.

⁵ The codes used for matching the industries are omitted in the article. We will be happy to provide a table showing the industry definitions and matching upon request.

⁶ Trostel modifies the Mincer (1974) equation to allow for a nonlinear relation between education and income, making it possible to identify the return to scale in schooling.