ABSTRACT
This book derives nine approaches to measuring competition and efficiency from a single theoretical profit-maximizing framework, assuming that these models share the same features as our baseline model. The major conclusion is that all models focus on a single variable instead of a set of variables as theory prescribes. For this reason, all models may suffer from identification problems in the sense that they pick up market power when estimating efficiency and vice versa. Also contributing to this problem is the measurement of input and output prices in banking. These problems may explain why the various approaches result in such diverging outcomes. The banking landscape has changed considerably over the last decade. First, demand is affected by foreign competition and competition from non-bank financial firms. This calls into question the underlying assumption that the price elasticity of demand faced by all firms is the same and constant over time. Second, banks have reacted to changes in regulation and production technology. They have branched out into new products and behave less like the traditional intermediaries we model them after. Reaction curves may have shifted considerably, both on the market level and for individual banks. Although competition has intensified internationally, some banks may occupy dominant positions within national borders that allow them to react differently than their smaller competitors. Some of the models we reviewed are theoretically unable to cope with these changes, as they have traditionally assumed that all banks react similarly to each other. Third, the markets banks operate in have also changed as, for example, concentration has gone up, which may weaken competition. But foreign competition has intensified, so that it is uncertain what, on balance, the effect on individual banks has been. Most approaches ignore the fact that banks produce various products and operate in various markets, where competitive positions may differ per product or market.An exception is the Bresnahan model, which considers competition in one submarket (e.g. loans, deposits). Approaches based on bank observations (Iwata, Panzar-Rosse) can circumvent this problem distinguishing various bank-size classes linked to different markets, e.g. small banks on local or retail markets and large banks on international or wholesale markets (Bikker and Haaf, 2002a). Where ample observations are available, gradual effects on competition of the trends over time should be incorporated by using time (or
trend) dependent coefficients (Bikker and Haaf, 2002a). Structural changes in banking markets and the lack of reliability of the data (particularly interest rates for credit loans and deposits) reduce the reliability of the estimates of the Bresnahan approach. One notable problem for the efficiency models discussed here is the fact that their outcomes are very difficult to validate. There is no sound theory providing the correct distribution of the efficiency term, and we know little about the economic validity of the efficiency scores. Particularly with increasing internationalization, contestability and foreign competition, it is hazardous to transpose best practices in one country or market to another. All in all, we expect that the observed trends have similar consequences for most banks: increases in competition result in lower profit margins, higher cost efficiency and lower profit efficiency. In absolute levels, we also expect cost reductions. The dynamics of the consolidation process, however, may have increased the volatility of earnings.
