Determinants and effects of growth strategies in the banking industry

This dissertation examines the linkage between growth choice and changes in performance. Banks are assigned growth strategies of branching, bank acquisition, and/or product expansion. The "Any Activity Method" assigns banks a given strategy for one or more application(s) to the Federal Reserve for a particular growth activity. The Primary Activity Method assigns strategies only if the proportion of the activity is greater than 25 percent of all activity. Strategies are assigned for the 1983 to 1988 and the 1986 to 1991 time periods, and changes in performance calculated in the 1989 to 1991 and 1992 to 1994 periods respectively. Market model results indicate negative and significant cumulative average residuals for the entire sample of publicly traded banks. A significant difference is found between banks that incorporate branching and those that acquire banks and product expand, indicating the market values branching. Operating performance models indicate bank acquirers have significantly lower changes in performance than the average bank from 1989 to 1991 using the Any Activity Method. Using the Primary Activity Method, banks that branch have significantly lower changes in profit margin. In seven out of twelve models, size is positively related to performance. In every model, the change in the capital ratio is positive and significant. In eight out of twelve models, the change in demand deposits is positive and significantly related to performance. De novo banks are tested separately to mitigate the problem of ignoring strategic choices that occur in prior periods. De novo banks that branch have positive and significant changes in ROA and profit margin using the Primary Activity Method. The changes in asset size, loan to assets, and bank deposit concentration are significant and positively related to changes in performance. The change in charge-offs to assets is negative and significantly related to changes in performance. A multinomial logistic model is used to discover determinants of growth choice. The model allows for simultaneous consideration of the three growth choices. During both strategic determination periods, multi-bank holding companies are more likely to Bank Acquire than Branch, or separately to Product Expand. De novo banks are more likely to branch that any other growth strategy in both time periods. Federally chartered banks are more likely to Bank Acquire or Product Expand as compared to Branch. Banks with assets over $1 billion are more likely to Branch than to Bank Acquire in every model. Significant performance variables indicate that performance is a determinant of growth choice in the 1986 to 1991 period. The statewide branching binary variable is positive and significant in most multinomial models for the Bank Acquire and Product Expand relative to Branch equations.