The Formation of Network and Public Intervention: Theory and Evidence from the Chilean Experience

The first part of the paper deals with the theoretical foundations of new industrial policy tools aimed at promoting a process of interacting learning among firms. I discuss the issue at three different levels: first, I define the theoretical boundaries of my research interest within the considerable economic literature dealing with industrial networks; secondly, I concentrate on some endogenous growth and development models, in order to analytically define the existing relationship between firm interactions, knowledge flows, and productivity. Then, I discuss the relationship between knowledge diffusion and productivity, with particular emphasis on the fundamental concept of network multiplier. Finally, I carry out a microeconomic analysis of the motivations that bring firms to interact with each other, and look for a role for public institutions in promoting such interaction. I discuss in which cases public intervention promoting the formation of a knowledge-sharing network is justified by the existence of a sort of “market failure”, and identify which variables are involved. In the second part of the paper I analyze the most important Chilean networking program, the PROFO program. The availability of relational data on a significant number of firm networks allows me to investigate in detail the relationship between network structure, public intervention and firm competitiveness. The econometric analysis confirms a strong correlation between PROFO firms’ innovativeness and industrial cooperation, proving the existence of an interactive learning process among participant firms. I used sociometric data to refine my analysis of the impact of the program on the network multiplier: not only do participant firms also achieve better performance in terms of productivity, but this performance is quite strongly correlated with firm centrality and network density, which are the two variables best representing the structure and function of the network multiplier and that, as I previously mentioned, are strongly affected by PROFO.