Bank value added, risk, and mergers

This dissertation integrates the finance and production approaches in studying bank operations. Chapter I develops a unified theory of banking that combines insights from production, financial intermediation, and asset pricing. It shows that banking comprises three qualitatively distinct functions: (1) resolving asymmetric information, (2) providing depositor services, and (3) financing. The first two create bank value added, whereas the last only contributes to gross output. Next it shows how these components interact within a consistent framework of dynamic value maximization, which takes risk into account. The model establishes the separability of a bank's production function for value added. It resolves some long-standing debates in the literature, and provides a theoretical basis for measuring banking output in the National Income Accounts.; The second chapter then constructs the new flow measure of output for Bank Holding Companies (BHCs). It differs noticeably from the existing measures, carrying significant implications for the measurement of banking GDP. Chapter II also estimates various specifications of value-added production and cost functions. Returns to scale estimates using the new measure range from 0.77 for the single-output Cobb-Douglas production function to 2.5 for its dual cost function. If possible measurement error in the imputed output is the major source of error, then the production function estimate of 0.77 is less biased than the higher estimates from the cost function. By comparison, existing studies typically find constant returns to scale.; The third chapter utilizes the new output measure to analyze the impact of mergers on BHC cost and productivity. It shows that conceptual problems in existing measures may explain previous studies' paradoxical findings—significant profit increases following mergers, without cost savings or increases in market power. Chapter III finds that the most commonly-used output measure shows the least improvement in cost productivity. By contrast, the new measure shows the most improvement, although it is still insignificant, partly because the sample size is small. The gap widens further when one corrects for possible biases in the new measure. Thus, the new measure of bank output may potentially resolve the existing paradox, by showing that mergers do lead to cost savings.