The theory’s predictions accord well with my case study evidence. Do large-scale statistical results concur? Yes. Econometric evidence in chapter 7, covering all municipalities over the period 1994-2007, shows that where a large number of firms interacted through the political system with an organizationally rich civil society, local policy decisions were responsive to the objective needs and subjective preferences of voters. Firms and civic organizations proved to be important determinants of local decision-making, and our empirical strategy allows us to identify how. Both firms and GROs affected how local governments prioritize local needs – via lobbying, voter mobilization, or otherwise mediating information flows and helping to sustain political competition. They not only pressed local governments for the specific policies they prefer, often at cross purposes, but also interacted directly with each other in the policy-making process.

These interactions are independently significant not only in the narrow statistical sense, but substantively as well, in the sense that they resolved the competing priorities of different actors. For example the independent effect of GROs was to increase, and of firms to decrease, investment in education. The tension was resolved when firms and GROs interacted directly through the local political system; these interactions led to investment increases that were positively related to local need and huge. In urban development, by contrast, both firms and GROs worked to increase investment in places where it was needed less. But the effect of their mutual interactions went in the opposite direction, increasing investment where infrastructure was scarce and decreasing it where infrastructure was abundant.