Regional spillovers of knowledge in Brazil: evidence from science and technology municipal indicators

ABSTRACT

This paper evaluates the role of regional spillovers of knowledge in Brazilian municipalities. The main hypothesis here is that regional inequalities restrict the knowledge spillovers in developing economies. We use a random-effects spatial panel model with data for two years, 2010 and 2015. Estimations indicate that regional spillovers of knowledge are important for innovation in Brazilian municipalities. The local scientific structure and economic specialization are also relevant for municipal innovativeness. The findings herein also provide evidence suggesting that spillovers are spatially restricted and will tend to reinforce the long-term regional imbalances in the Brazilian innovation system in the foreseeable future.

KEYWORDS:

• Regional knowledge spillovers
• regional development
• spatial econometrics
• science and technology indicators
• Brazil

JEL:

• Codes: R58
• O30

Acknowledgments

The authors acknowledge to Rodrigo Simões for initially encouraging this research, as well as research assistants Thalita Freitas, Carlos Miranda and Paulo Alves da Costa Jr. The anonymous reviewers are also acknowledged for their valuable comments and suggestions on the earlier version of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 In 2010, there were 5,565 municipalities in Brazil. This number increased to 5,570 in 2015. The methodological option was to maintain the number of municipalities set in 2010, transforming the data for 2015 to make it compatible with the structure prior to the creation of the five new municipalities. In this way, the database is composed of 11,130 observations.

2 Patents per 1,000 inhabitants.

3 The numbers of patents per municipality were collected for 3-year intervals (2008–2010 and 2013–2015). The other data employed here cover the years of 2010 and 2015.

4 The calculation of the productive specialization index is based on Glaeser et al. (1992). Its adaptation for getting productive concentration according to technological intensity aims to observe how places specialized in low-, middle–low-, middle–high-, and high-tech activities perform in knowledge production.

5 Based on Marshall (1982), Arrow (1962), and Romer’s (1986) developments, Glaeser et al. (1992) suggest that technological overflows occur among firms in the same industry (productive specialization) and positively affect growth. This process became known as Marshall-Arrow-Romer externalities (specialization externality-MAR).

6 This index is widely used in regional and urban studies to evaluate industrial diversity. It compares productive structure in a locality with the productive structure of the entire country. Here the modified Hirschman–Herfindahl is employed following Simões and Freitas (2014), who took it to evaluate industrial diversification in Brazilian Micro-regions. Regional economics literature stresses that the most diversified a locality is more creative and innovative it must be (Jacobs 1969; Storper and Venables 2004).

Additional information

Funding

This work was supported by Fundação de Amparo à Pesquisa do Estado de Minas Gerais – FAPEMIG/BR [Grant number APQ-00577-17] and by Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPQ/BR [Grant numbers 428190/2018-9; 152859/2018-7].