Geographic and sector externalities from highly qualified human capital: the importance of the business service sector

Abstract

This article develops a simple model where knowledge spillover is related to geography and sectors of the economy. Knowledge externality comes from those with college degree at sector level and the relative size of the business service sector at geographic level. We examine the model using data from Brazilian National Household Survey (PNAD) for the year 2008. We find evidence that sector concentration of highly qualified workers (with at least a college degree) generates knowledge externalities. Hence, workers benefit from the skills of their peers by learning from them. Also complementing the literature, we find a geographic knowledge spillover played by the business service sector. This last result holds even if we estimate rates of return to education by sectors or estimate wage regressions over sectors (not shown). This article also provides evidence of increasing returns to schooling in Brazil.

Keywords:

• knowledge externalities
• business service sector
• increasing returns to scale

JEL Classification:

• J31
• J24
• L80

Notes

¹ Moretti (2004) provides a review on the literature prior to 2004.

² Some results are dependent on the methodology utilized by the authors (Ciccone and Peri, 2006).

³ In the first case, our assumption is that highly educated individuals improve the overall organization and foster business with better ideas that are followed by others in the same sector. Workers in the same sector benefit from the skills of individuals with college education by learning from them. In the second case, the business service sector spreads its knowledge when supplying services to other companies or sectors. Services including consultancy, information technology and telecommunication, among others, contribute to improve organization and productivity in the economy. Therefore, it has sector and geographical knowledge spillovers.

⁴ Available at www.ibge.gov.br. The PNAD survey has a complex sample design and thus requires adequate treatment.

⁵ Sector classification details can be provided upon request.

⁶ Detailed information about the variables and can be provided upon request.

⁷ The variables included in the selection equation are family income, family member, member < 14, schooling, age, region, colour, married and female with child.