On the location of knowledge-intensive entrepreneurship in developing countries: lessons from São Paulo, Brazil

Abstract

This article empirically appraises the geographical distribution of knowledge-intensive entrepreneurship (KIE) in the settings of an emerging economy. We start from the typical agglomeration approach and then introduce a set of variables related to local market conditions, distance from the economic hub, and knowledge & innovation system to explain KIE location and density on the basis of city-level data in the State of São Paulo, Brazil. Findings indicate KIE concentration in and around a few urban areas, providing support to agglomeration economies concepts. There is strong evidence that the local presence of research-oriented universities, access to capital, and business concentration are correlated to KIE emergence and density. Results also indicate the moderating effect of agglomeration diseconomies mainly related to factors of rapid and anarchic expansion of urban centers and the consequences of extreme inequalities in income distribution. This challenges the usability of concepts of entrepreneurial ecosystems from advanced economies if not adapted to the realities of developing countries.

Keywords:

• Knowledge-intensive entrepreneurship
• geography of innovation
• developing countries
• entrepreneurial ecosystems

Acknowledgements

The authors acknowledge support by the Sao Paulo Research Foundation (FAPESP) in connection to the Sao Paulo Excellence Chair in innovation systems, strategy and policy established in the Department of Science and Technology Policy of the University of Campinas (UNICAMP). Bruno Fischer and Paola Schaeffer also recognize FAPESP support under the project The Economic Geography of Entrepreneurial Ecosystems in the State of Sao Paulo (2016/17801-4). Nicholas Vonortas acknowledges the infrastructural support of the Center for International Science and Technology Policy at the George Washington University. He also acknowledges the support of FAPESP through the Sao Paulo Excellence Chair in technology and innovation policy at Unicamp, Brazil. And, he acknowledges support from the Basic Research Program at the National Research University Higher School of Economics within the framework of the subsidy to the HSE by the Russian Academic Excellence Project ‘5–100’. None of these organizations is responsible for the contents of this paper. Remaining mistakes and misconceptions are solely the responsibility of the authors.

Notes

1. An additional shortcoming of the existing literature is the lack of approaches dealing with the geography of KIE in developing countries (e.g. Naude et al. 2008; Ghani, Kerr, and O’Connell 2014; González-Pernía, Jung, and Peña 2015).

2. However, some authors see high levels of income as a deterrent factor for entrepreneurship as it may signal higher labor costs (Li et al. 2016).

3. On the other hand, there is an extensive body of literature asserting that geographical proximity per se is not a determinant of network formation (Boschma 2005; Boschma and Martin 2010), thus not necessarily leading to the desired innovation outcomes.

4. Arauzo-Carod (2008) reaches a similar conclusion for location strategies of firms in general in the Catalonian context.

5. Last data available for comparison at OECD Statistics.

6. Data is not available for the entire state area.

7. http://www.doingbusiness.org/data/distance-to-frontier

8. Including countries and subnational units. Brazil ranks slightly better at the 196th position.

9. We have ran robustness tests excluding recurring winners – unless projects took place in different locations (the number of projects was reduced to 1030). Significance and coefficient of variables remained basically unaltered. As the focus of our analysis is on entrepreneurial activity (not individuals), findings are presented for the whole sample. Results for alternative specifications are available from the authors upon request.

10. This calls for a clarification on expectations of potential endogeneity disturbances (location of knowledge institutions related to local entrepreneurial profile and not vice versa). Most research-oriented institutions (particularly outside of the city of São Paulo) are public entities. Hence, they have been established by public authorities as possible backbones of local productive systems, neither relying substantially on private funds for research (Rapini et al. 2009) nor on ‘market seeking’ strategies for students. Hence the causality direction of this variable should run as specified in estimations.

11. Over the last decades, about a fourth of business incubators (7 out of 25) was created in areas without the presence of any research-oriented university. The situation is less problematic with S&T parks, since only one (out of a total of 14) was created away from these academic institutions.

12. We would like to thank an anonymous reviewer for raising this discussion.