LOCAL ENTROPY, BANKS` DENSITY AND FIRMS` INVESTMENTS: EVIDENCE FROM ITALIAN PROVINCES

Starting with the seminal contribution by Fazzari, Hubbard and Peterson (1988), the role of financial structure in firms’ investment decisions has been investigated widely in the economics literatures. Most studies show that financially-constrained firms have a greater investment sensitivity to internally generated financial resources, as generally measured through cash-flow indicators. Financial market imperfections, based on information asymmetries between borrowers and lenders, are generally seen as the main cause of financing constraints. This stream of the literature also suggests that small-sized firms tend to be more financially-constrained then larger ones since they are younger, they have less assets, and they face a higher entrepreneurial risk. The role of credit constraints and of transmission mechanisms for financial shocks appear to be particularly relevant in the case of Italian manufacturing industry, which is characterised by the strong presence and a wide spread of industrial districts and local clusters of SMEs. In these contexts, the action of spatial agglomeration forces and relationship banking, by affecting information asymmetries and credit access conditions, might overcome the negative effects related to small size (Dei Ottati, 1994; Alessandrini, Presbitero and Zazzaro, 2006). In this paper we analyse if the composition of the local system in which firms are located does affect their investment decisions. In particular, we ask if the level of firms’ investments in new equipment can depend on local banks availability within clusters characterized by a high level of sectoral specialization or within clusters where many different industries operate. While in the former case banks may benefit from specialization economies and higher knowledge of sector performance, in the latter case banks may find more profitable to diversify their credit portfolio and minimize financial risks. However, such a benefit may arise if the level of sectoral diversification is not too high, i.e. when the local system is characterized by an optimum degree of related variety (Frenken, Van Oort and Verburg, 2007; Boschma and Iammarino, 2010). We measure these effects by interacting the bank density with local entropy variables capturing the degree of risk diversification in each Italian administrative province. Operationally, for the period 2000-2007 we estimate an Error Correction Model for investments, using an unbalanced panel of 13,000 Italian firms. Data are drawn from AIDA, a dataset collected by Bureau Van Dijk Electronic Publishing which gathers information on balance sheet data for more than 200,000 Italian joint stock companies. We estimate dynamic investment equations using the one-step system GMM estimator developed by Arellano and Bond (1991, 1998) and Blundell and Bond (1998), which allows to control for potential bias due to non-observable fixed effects and endogenous dependent variables. Our preliminary estimates seem to support the existence of strong interactions between the local availability of banks and a moderately diversified local system of production which affect the level of investments of manufacturing firms. Since investments in new machinery and equipment can be considered as a proxy for capital-embodied technical change, our evidence not only confirms previous results (Benfratello, Schiantarelli and Sembenelli, 2008) stressing the role of bank development in increasing the probability of process innovation and in reducing the cash flow sensitivity of fixed investments spending, but also integrates it by emphasising the role of related variety in driving firm-level innovation.