Industry cluster theory has been the predominant model guiding economic development policy throughout the world for nearly two decades. As appealing as the cluster approach has been to regional scientists and policy makers it suffers from a number of theoretical and empirical shortcomings, including an inability to explain economic dispersion and the presence of high-growing firms that thrive in non-clustered industries and locations. This dissertation tracks the growth and survival of a cohort of more than 300,000 establishments operating in Pennsylvania during the 1997-2007 period. It reveals that firm characteristics are 10-times more powerful than industry and cluster characteristics, and 50-times more powerful than location characteristics, in explaining and predicting establishment-level growth and survival. It also finds a Power Law is present in the distribution of establishment growth, indicating that a sub-set of businesses systematically accumulate a disproportionate share of employment growth. Roughly 1% of establishments created 169% of all net new jobs added in the state over a ten-year period. Growth is further concentrated among businesses that are able to sustain growth over multiple years. This suggests that the principal driver of regional growth is cumulative firm growth - the accumulation of a disproportionate amount of growth among a small number of firms through sustained expansion over multiple years. I conclude that the path to building better theory and more effective development policies is one that explicitly links regional growth to the growth of firms. Such an approach should focus on endogenous firm dynamics rather than exogenous heuristics such as industry and location.
