Within neoclassical economic growth approaches there is no fully satisfactory explanation for regional asymmetric growth. More recent extensions of the neoclassical model (new endogenous growth theories) recognise the possible influence of a set of intangible factors, which, nevertheless, do not allow a treatable empirical formalisation. Neo- institutional and evolutionary theories highlight the complex nature of economic growth, recognising the difficulty of non-linear dynamic modelling; however, both theories highlight the influential role played by innovation and technology diffusion processes. Missing from these theories is any appreciation of the social network characteristics of regions and the interaction between social and economic systems in conditioning the growth and development processes. In the present PhD thesis, we take the institutional network in small towns as the unity for analysis where information, money and support flows among them are studied. We apply a social network analysis framework, which allow us to find numerical indicators for the whole network as for each individual institutional role in it. Applying this framework in two case studies (in Portugal and in the USA) we found a set of interesting differences about the institutional mediation roles. While in the US, the private sector leaded by banks plays the keystone sector role, in Portugal, we still have public institutions as the main players. After the identification of the significant variables and attending to the different results in the two studies, we built a testable cross-section model that, controlling for proximity from urban centre, transportation costs, local factor endowments, amenities and so forth, can be tested using a broader database, to correlate town economic performances with the network identified variables. The deep understanding of the relational structure within these small towns makes it possible to quantify a set of crucial intangible variables that, acknowledged in a regional policy design, will contribute to increase its efficiency. Without a complete understanding of these social network structures, it is possible that current EU criteria used to define regions eligible for assistance may be inadequate. At the present time, the criteria for receipt of regional development funds are based on relative (to EU average) per capita levels and do not explore differences in economic potential or the capacity for regions to develop. Part of this capacity is rooted in the social network structure. Further on-going research will contribute to the deepening of neoclassical growth models, complementing considerations that evolutionary economics uses to explain the complexity of modern socio-economic life.
