The abstract goes like this:
Entrepreneurial Risk and the Geographical Concentration of Industries: Evidence from the UK Manufacturing Sector.This paper provides an analysis of the effects that conditions of imperfect information and irreversibility of investment exert on location decisions, and subsequently produce some empirical evidence in support of those theoretical results, by looking into the concentration of manufacturing industries in the UK. We analyse the location decision of a firm that has to undertake a non-recoverable investment in order to enter a new market. Localities are characterised by their level of efficiency. The potential entrant faces two types of uncertainty: 1) industry profits are aleatory and 2) the characteristics of the single localities are unknown. Due to the existence of sunk costs and uncertainty over future profits the potential entrant faces the risk of producing losses from the venture (entrepreneurial risk). Our main hypothesis is that agents will use the location of existing plants in the industry to infer information on the characteristics of the single localities: the existence of an industry in a locality signals “good” local endowments. As a consequence, the expected level of profits conditional on the locality being “active” will be higher than the unconditional. This difference is bigger the higher the variability of industry profits. Our main prediction is that high-risk industries will be more concentrated geographically, ceteris paribus. We test this hypothesis on the geographical distribution of manufacturing activity in the UK. We do so by running a series of linear regressions of NACE Rev.1 4-digit industries’ Ellison and Glaeser’s gammas on a measure of entrepreneurial risk, after controlling for transport costs, natural resource intensity and Marshallian economies of agglomeration. Our results confirm a significant positive effect of entrepreneurial risk on geographical concentration. As additional contributions, a new proxy for vertical linkages is introduced, which performs consistently better than the alternatives used in previous empirical studies, and a discussion on how to proxy for transport costs in studies of the geographical concentration of industries are carried out.