رقابت در بازار‌های بخش بندی شده: مدارک و شواهد جدید در مورد صنعت بانکداری آلمان با توجه به بحران مالی درجه دو

Of all of the EU member states, Germany has the largest banking market. However, not all German banking institutions necessarily face fierce competition. Because the industry is highly fragmented, strict separation of the three existing banking pillars may impede competition, with negative effects on financial stability. We assess the competitive stances of 1,888 universal banks from 2001 to 2009 by using the Panzar–Rosse revenue test. We find evidence that measuring competition at an average country level does not necessarily generate valid evaluations of fragmented markets. In addition, we find no clear indication that either the particular objectives of cooperative and savings banks or the legal protection of these institutions impedes competition or discriminates against private banks. Therefore, as long as the relationship between competition and financial stability is dubious, the overall effect and the social costs or benefits of political measures that influence the structure of the German banking market are at least questionable.

The current financial crisis has clearly demonstrated the importance of a stable financial system to economic growth and welfare. In this system, banks play a vital role because they attract short-term deposits from many small investors and grant long-term loans to borrowers. In addition, the banks’ role as intermediaries allows them to provide liquidity to depositors while also protecting borrowers from the liquidity needs of their investors (Diamond and Dybvig, 1983 and Diamond and Rajan, 2001). The banks’ abilities to diversify by financing independent projects further reduces information asymmetry and the costs of delegation (Diamond, 1984) and enables banks to transform indivisible, illiquid and risky assets into divisible, liquid and (nearly) riskless liabilities. These transformation activities traditionally explain the advantage of financial intermediation. However, at the same time, they expose banks to the risk of runs in that the expectation of a bank’s failure may, in the sense of a self-fulfilling prophecy, actually initiate the failure. Further, the failure of a single bank may have contagion effects on the whole banking system either by creating a panic among depositors or by affecting the interbank lending relationships. In this context, large banks are of particular importance because their failure could pose significant risks to other financial institutions and the financial system as a whole. Specifically, a large bank’s failure could trigger a systemic crisis that negatively affects the monetary system and real production (Diamond and Dybvig, 1983 and Stern and Feldman, 2004). To ensure financial stability, those institutions considered as ‘too-big-to-fail’ might implicitly or explicitly be protected by public guarantees or subsidies, as observed during the subprime crisis. The probability of default of a systemically important bank may depend on the degree of competition in the financial industry. According to the charter value hypothesis, banks facing a lower degree of competition can earn monopoly rents and will hold higher capital buffers to preserve those rents for the future. Consequently, increasing competition will erode rents, trigger excessive risk-taking ( Besanko and Thakor, 1993 and Boot and Greenbaum, 1993 and Hellman et al., 2000) and thus decrease financial stability. Though not undisputed (for a deeper investigation, see Carletti and Hartmann, 2002, Vives, 2010 and Jiménez et al., 2010), this theory is supported by a significant number of empirical studies. For example, Keeley (1990) analyses the US banking market and finds that competition is negatively related to bank solvency and positively related to the perceived bankruptcy risk (measured as the risk premiums for uninsured certificates of deposits). He concludes that decreasing charter values induced by liberalisation and deregulation encourage risk-taking. Thus, he argues that increasing competition at least partially contributes to the fragility of the banking sector. Additionally, Beck et al. (2006) analyse data for 69 countries from 1980 to 1997 and support this view by finding that concentrated banking systems are less likely to experience systemic crises. In contrast, Boyd and de Nicoló (2005) develop a theoretical model and claim that market power leads to higher loan rates, which, in turn, create moral hazards. Specifically, entrepreneurs will choose riskier projects and thereby lower the quality of banks’ loan portfolios while simultaneously increasing bankruptcy risk. However, both studies rely on structural measures of competition, as they assess competition based on concentration measures (Beck et al., 2006) or the number of institutions in the market (Boyd and de Nicoló, 2005). Those measures are, in fact, a weak proxy for competition because they do not explicitly account for the conduct of banks, as mentioned by Shaffer (1982b) and empirically verified by Claessens and Laeven (2004) and Schaeck et al. (2009). This shortcoming becomes even more important in fragmented banking systems, such as those in Germany or the US, where a large number of institutions operate in a specific local area with only a limited number of rivals. Those institutions may even operate as monopolists, a fact that remains undiscovered by those studies using structural measures as well as the multi-country studies using non-structural measures relying on the average degree of competition in a country. In particular, the German banking market is often criticised for being outdated and ‘overbanked’ by the Directorate-General for Competition of the European Commission and by large private banks. Especially savings and cooperative banks – also known as trustee savings banks and credit unions in other jurisdictions – often apply regional or territorial principles to their business models, which means that a specific local area, such as a city or administrative district, is reserved for an individual institution that is not directly competing with other similar institutions (through its network of branches) but is at least indirectly competing with other institutions of the respective pillar through other means, such as electronic banking (for a brief description of the German banking market, see Gischer and Stiele, 2009). The claim is that regional demarcation or a possible non-profit-maximising behaviour of the savings and cooperative banking group – in addition to the legal protection of these institutions provided by the rules governing their ownership structures – prevent a competitive equilibrium from emerging with negative effects on private banks. In this context, it should be noted that large private banks experienced significant losses during the subprime crisis and received partial governmental support, whereas small and medium-sized cooperative and savings banks seemed to play a stabilising role; these observations may be consistent with the charter value hypothesis. Following this line of reasoning might create conflicts for political and regulatory decision makers, as they will struggle with a possible trade-off between competition and financial stability. Although financial stability may benefit because of hypothetical competitive advantages of cooperative and savings banks, this might also increase the risk of having to bail out large private institutions that suffer from competitive disadvantages. As a result, these banks may be unable to build up adequate capital buffers or may have incentives to take excessive risks, which might negatively affect financial stability. The aim of this paper is to evaluate the degree of competition in the German banking industry and to reveal potential imbalances in competitive conduct. To that end, we complement and extend the existing literature on this issue in several respects. First, to the best of our knowledge, this study is the first since Lang (1997) to empirically investigate the competitive environment of universal banks in Germany. Analysing this issue is important because these banks provide a wide range of financial services to the entire German population and may therefore be more vulnerable to bank runs than more specialised institutions (e.g., savings and loan associations, business development banks, central institutions, specialised financing institutions or the so-called Landesbanken), whose inclusion in the sample could distort the results. Second, whereas previous empirical studies have assessed the average competitive stance of German banks within the context of multi-country studies, we will explicitly consider the high level of fragmentation in this industry by dividing the sample into characteristic groups according to the sector and size of the banks. In this manner, we will provide appropriate evidence addressing whether certain institutions show different competitive behaviours than others and whether the three-pillar structure of the market might generate competitive discrimination against private banks. Third, we will overcome the various methodological shortcomings of earlier studies. Unlike structural measures, the Panzar–Rosse revenue test allows us to explicitly assess the competitive conduct of banks. Further, because nearly all of the works applying this methodology to the German banking industry contain a critical misspecification or misinterpretation bias, the generalisability of their results may be limited. These problems create a need for further research to fill this gap. Fourth, this study is also the first to investigate how the experiences and various public rescue programs related to the subprime crisis have influenced the competitive stance within the German banking industry. The remainder of this paper is structured as follows. Section 2 summarises the existing research techniques used to assess the degree of competition in a market. Section 3 briefly describes the Panzar–Rosse methodology used in this study and summarises the existing literature on this issue. Sections 4 and 5 contain the empirical model and the data used, respectively. The empirical results are presented and discussed in Section 6, and Section 7 concludes.

Using data for 1888 universal banks operating in Germany from 2001 to 2009, this paper provides answers to three important questions concerning the competitive conduct in the banking industry. First, we find evidence that measuring competition at an average country level (as previous studies have done) not necessarily generates valid evaluations of fragmented markets, as a sophisticated investigation reveals significant differences among the estimated H-statistics of the banks in the three existing banking pillars and the different size classes. Second, although there are differences among the particular subsamples, we find no clear indication that either the particular objectives of cooperative and savings banks or the legal protection of those institutions impedes competition because monopoly power is ruled out for all of our subsamples. Additionally, we cannot be certain about significant differences in competitive conduct among German universal banks without further information about cost structures or market demand elasticity. Third, according to our data, estimation results do not confirm that there was a change in competitive conduct among German universal banks during the subprime crisis. However, this issue requires further investigation and analysis in the future using additional data from the entire crisis period. With respect to the policy implications arising from these main findings, it is difficult to identify clear indications of discrimination against (large) private banks. Neither cooperative nor savings banks possess monopoly power, and differences in estimated H-values – which do not necessarily indicate differences in competitive conduct – might be related to diverging corporate objectives instead of legal protection. From this perspective, we find no distinct empirical evidence supporting the argument that the German banking system is outdated or overbanked and no evidence that large private institutions might suffer from competitive disadvantages or are at increased risk of being bailed out with negative consequences for financial stability. Consequently, the effects of initiatives aiming to equalise the competitive conditions of private banks and cooperative or savings banks – such as increasing the permeability of the three-pillar structure – are unpredictable from a competitive point of view if we cannot unambiguously confirm differences in competitive conduct. With respect to financial stability, the consequences of such initiatives are even more ambiguous under the charter value hypothesis. According to this hypothesis, lowering the level of competition for systemically important financial institutions, allowing them to earn monopoly rents and build up capital buffers, might improve financial stability. However, this causes multiple problems. First, reducing the degree of competition, if possible at all, is exactly the opposite of the aim of the Directorate-General for Competition of the European Commission. Second, strengthening the competitive position of (large) private banks or systemically important financial institutions to (possibly) foster financial stability would certainly have negative effects on the competitive position and solvency of (small) cooperative and savings banks. Based on the stabilising role of those institutions during the subprime crisis, policy makers must ensure that increasing the stability of (large) private banks is not achieved by decreasing the stability of cooperative and savings banks. Third, the relationship between competition and financial stability – and the validity of the charter value hypothesis itself – is hotly debated in the economic literature. Thus, questions continue about the remaining uncertainties about differences in competitive conduct among German universal banks, the effectiveness – in terms of overall effects and social costs or benefits – of political measures on banking sector competition and the relationship between competition and financial stability in general. These questions demonstrate the need for future research in this field and highlight the conflicts facing political and regulatory decision makers.