Nada Kobeissi, Iftekhar Hasan, Bo Wang, Haizhi Wang, and Desheng Yin Regional Studies, Vol. 57 No. 1, pp. 57-71

DOI: https://doi.org/10.1080/00343404.2022.2030053

What the paper is about

Most research on what drives regional innovation focuses on firms that trade publicly, even though private firms make up the bulk of American businesses and tend to pursue a broader, more exploratory kind of innovation. This paper turns to a factor that has drawn little attention in that context, social capital, defined as the civic norms and social networks that let a community act collectively. The idea is that in places where people trust each other and interact often through civic and social organizations, private firms find it easier to get support for financing risky, hard-to-verify inventive projects, since that trust substitutes for some of the due diligence outside capital would otherwise demand.

Using a panel of 3,042 US counties from 1990 to 2004, the paper builds a county-level social capital index from voter turnout, census response rates, and the density of civic associations and nonprofits in a county, following a method used in earlier work on social capital. Patent data come from private-firm assignees in the NBER patent database, aggregated to the county level and measured four ways, how many patents a county produces, how often those patents get cited, and two separate measures of how broad or novel the underlying ideas are.

Counties with more social capital produce more patents, patents that draw more citations, and patents that are broader in scope, and the relation holds up when the analysis is instrumented using a county’s distance from the Canadian border, a variable Putnam had earlier identified as a strong predictor of social capital in the US. Splitting the index into its two components, civic norms and social networks, both show the same positive pattern on their own.

The relation is not unlimited. A squared term on the social capital index turns up negative, suggesting the benefit tapers off past some point, plausibly because dense, high-trust networks can start to crowd out exposure to new ideas rather than helping spread them. The benefit is also uneven across counties. It does more for innovation where the local supply of financial capital is thin, and it works alongside R&D spending rather than in its place, so the two appear complementary rather than substitutes for one another.

Using a spatial Durbin model, the paper also finds that a county’s social capital reaches beyond its own borders. Innovation in neighbouring counties rises along with a focal county’s social capital, and the estimated spillover to neighbours turns out to be larger than the effect within the county itself.

Highlights

  • County-level social capital is positively associated with the quantity, quality, and novelty of private-firm patents.
  • An instrumental variable approach using distance to the Canadian border supports a causal reading.
  • The benefit tapers off at high levels of social capital, consistent with a diminishing return.
  • Social capital matters more where financial capital is scarce, and it complements R&D spending rather than substituting for it.
  • Social capital in a county also boosts innovation in neighbouring counties, with spillovers larger than the effect at home.