1 Introduction
Jacob A. Bikker and Michiel van Leuvensteijn
The competition and efficiency of financial institutions have been investigated exhaustively in the economic literature, because of their important role in smoothing consumption of households over time and providing capital to enterprises. In principle, competition is expected to (i) enhance efficiency, and thus to lower prices; (ii) stimulate new innovations; and (iii) open up new financial markets. However, the financial crisis of 2007â2008 has also revealed to policymakers worldwide that the relationship between competition and financial stability is of major importance. On the one hand, competition may enhance financial stability by pushing unstable banks out of the market. On the other, competition is regarded as one of the possible drivers of risk-taking behaviour of banks. Competition may encourage banks to take more risks in order to be more profitable. To be able to assess the impact of competition on banksâ risk-taking behaviour, a good measure of competition is needed. This book aims to contribute precisely to that.
In the aftermath of the collapse of Lehman Brothers, many banks worldwide have been rescued by their national authorities. Some were supported with capital injections, for instance Bank of America, Citigroup and Morgan Stanley in the US, Royal Bank of Scotland (RBS) in the UK, ING in the Netherlands and 1KB in Germany. Others were nationalized, including ABN AMRO in the Netherlands, Hypo VereinBank in Germany and Northern Rock and Bradford & Bingley in the UK. Furthermore, government guarantees were used to facilitate the funding of banks (OECD, 2011). These interventions have tilted the playing field and distorted competition among banks. Banks receiving state support can attract savings and deposits against lower interest rates and may grant loans against lower lending rates or become involved in riskier projects (because they have an implicit guarantee), compared with banks without government assistance. Bank bail-outs in the US have led to higher profit margins of the banks after the bailout of all surviving banks and competitive distortions (Koetter and Noth, 2012). In response to possible competitive distortions, the European Commission has âpunishedâ banks in the European Union (EU) that needed state support with a view to avoiding competitive disadvantages for banks surviving without support. State aid prevents strong banks from being able to acquire a large market share, which they would have been able to acquire in the absence of state support. Punitive measures include requirements to split up banks and financial conglomerates (as in the case of ING) and prohibition of aggressive price setting for banks with state support. Although conducive to competition in the long run, the latter measure may impair competition in the short run. Furthermore, the current crisis has shown that large banks are more likely to be rescued than small banks. This implicit state guarantee may provide large banks with an incentive to take on more risk than otherwise and compared with smaller banks. As a response to this âtoo big to failâ moral hazard problem, the EU has developed a bail-in policy, shifting the possible burden of bank defaults not only to the bankâs shareholders but also to junior debt holders and large deposit holders (and maybe even to senior debt holders). Finally, during the crisis we have also seen that banks reduced their foreign activities in favour of serving their domestic market. As a consequence, cross-border competition has often run dry. Now the immediate crisis is over and the banking sector is restabilizing, it is time to reconsider the trade-off between financial stability and competition. In a depressed economy, competition in the banking sector is crucial as an engine for accelerating economic growth. All in all, the crisis and its aftermath have shown that the need to measure banking competition, as a basis of anti-trust policy, is larger than ever.
It is difficult, if not impossible, to observe competition and efficiency of banks and other financial institutions directly, since precise information on both input and output is rare and public data on the costs of separate bank and insurance products are unavailable. This impairs the calculation of price-cost margins of separate products. The literature has tried to measure competition by means of many different methods, none of which, however, has been entirely conclusive or unchallenged. One stream of literature focuses on the structure of the market, particularly at the level of market concentration; examples are the so-termed HirschmanâHerfindahl index (HHI), the market share of the five largest banks (C5), and the structure-cost-performance (SCP) model. A second strand of literature estimates competition by focusing on market conduct. Examples of these approaches include the Cournot model, as used in the new empirical industrial organization (NEIO) measurements, the elasticity-adjusted Lerner index, the H-statistic of Panzar and Rosse, regarding the pass-through behaviour of input price increases to revenues, and the conjectural variation parameter λ of the Bresnahan or Lau model, which describes responsiveness on reactions of rival firms.
Apart from the many theoretical shortcomings of competition measures as discussed in the literature, a practical problem is that different methods yield diverging estimates. Evaluating a broad field of research, Bikker and Bos (2008) introduce a general framework to describe a profit-maximizing bank and demonstrate how widely-used types of competition measurement models as listed above can be fitted into this framework. In particular, the framework sets out the assumptions which are implicit in various competition and efficiency measurement approaches. More precisely, the Bikker and Bos model indicates three factors determining competition: market share, price elasticity of demand and conjectural variation. Many approaches contain a partial analysis and focus on a single factor only, or on a combination of two factors. This in part explains the theoretical shortcomings of the various methods and the great diversity in the empirical outcomes. The PCS framework adds a new element to this theoretical analysis by considering how differences in efficiency across banks have impact on either market shares or profits, or both. The stronger this impact is, the heavier is the competition in the respective market. Furthermore, it may introduce dynamics over time; changes in market shares may improve scale economies, so that further market share gains are possible. In order to take endogeneity into account, we need to estimate with instrumental variables.
The structure of this book is as follows. Chapter 2 discusses why competition in the financial sector is of crucial importance and sets out the major structural problems undermining competition on financial markets. It also addresses the ambiguous relationship between competition and financial stability. Chapter 3 and 4 present the seminal introductions of the PCS indicator by Hay and Liu (1995) and Boone (2004), respectively. The foundations for this indicator were laid by Hay and Liu, who provide both its theoretical underpinning under Cournot oligopoly and its first application. They focus on the relationship between efficiency and performance as an indicator of the level of competition, where performance is either profitability or gaining a larger market share. Boone (2004) shows theoretically that the impact of efficiency on profits is a robust measure of competition, using a fairly general model that include the possibilities that firms play Bertrand or Cournot competition. He calls the PCS measure the relative profit differences (RPD) and shows that under the conditions of homogenous goods and equal profits for firms that are equally efficient, the RPD measures competition correctly and does not suffer from the deficiencies of the (elasticity-adjusted) Lerner index, or price-cost margin (PCM). When competition increases, production reallocation may be expected from less efficient banks to more efficient banks, so that the average PCM may rise (Boone et al., 2007). Boone emphasizes the importance of profit elasticity with regard to marginal costs and does not investigate the impact of efficiency on market shares. At the same time, his models provide a theoretical basis for the relationship between market shares and efficiency (as also shown in Chapters 7 and 8). However, using market shares instead of profits may make the indicator theoretically less robust, because there is no relationship when efficiency gains are not translated into lower output prices, in which case market shares remain unchanged, but only in higher profits. Until early 2012, we referred in our papers to the PCS indicator as the Boone indicator, being unaware of the founding work of Hay and Liu (1995, 1997). In this book we introduce a new, more neutral name, reflecting that this approach is based on the efficiency hypothesis; in competitive markets, more efficient banks are expected to gain a larger market share and earn more profit. It is thereby an opposite view to the well-known structure-conduct-performance (SCP) model, which inspired us to come up with our name: the performance-conduct-structure (PCS) indicator.
Many different approaches to measure competition provide disappointingly divergent outcomes. Chapter 5 applies the many approaches from the literature, including the PCS indicator, to one specific dataset and searches for an explanation of the diversity in the results, which cannot be attributed to different datasets. Varying approaches appear to focus on different aspects of competition. Correlations between different approaches are fairly low (which is disappointing), but where they are significant, they all have the correct sign (which is encouraging). These proxies of competition also function quite well, that is, in line with expectations, as explanatory variables; apparently they make sense (which is encouraging too). This chapter develops an index of competition which appears to function better than the underlying components.
Chapter 6 provides an empirical underpinning to the PCS indicator, particularly examining whether this indicator is able to distinguish differences between regimes of competition over time with respect to the American Sugar Refining Company (ASRC). This analysis is based on testimonies at the Congressional hearing on the American sugar industry and data on the ASRC during the period 1890â1914. These old data were used because there was no regulation which hampered competition, ASRCâs marginal costs were known and different regimes of competitions can be verified on the basis of testimonies given during the Congressional hearings. This chapter shows that the PCS indicator can indeed properly identify different regimes of competition, like cartels and price wars, at least in the sugar industry case in that period. Furthermore, in this case, the PCS indicator works equally as well as the elasticity-adjusted Lerner index in identifying the different regimes of competition.
Chapters 7 and 8 present empirical studies on the measurement of banking competition in the major economies. In the euro area, banks are the main suppliers of external funds for companies, particularly small and medium-sized enterprises and, hence, play a more important role in this sense than capital markets. For consumers, the mortgage markets are crucial. Chapter 7 concentrates on the measurement of competition in the loan markets of the euro area and the US, the UK and Japan. Our findings indicate that in the period 1994â2004 the US had the most competitive loan market, whereas overall loan markets in Germany and Spain were among the most competitive in the EU. The Netherlands occupied an intermediate position, whereas in Italy competition declined significantly over time. The French, Japanese and UK loan markets were generally less competitive. Chapter 8 uses these results to investigate the role of competition in the functioning of financial markets from a monetary policy perspective across eight euro area countries. The European Central Bank (ECB) has defined an objective of price stability and sets its policy rate to achieve this aim. The effectiveness of the monetary policy instruments in reaching the ECBâs objective relies on the degree to which policy rate changes are translated into bank lending rates. We investigate four loan markets and two deposit markets separately. The analysis shows that the strength of monetary transmission depends significantly on the level of competition in loan product markets, particularly mortgage loans, consumer loans and short-term loans to enterprises. Furthermore, banks act as price makers in the deposit market and as price takers in the loan market. Stronger competition causes both lower bank interest rates and a stronger pass-through of market rate changes into bank rates.
Life insurance companies offer, among other things, insurance against old age by providing an annuity, and improve the possibilities to save throughout the life cycle. Chapter 9 investigates the competition in this market, by estimating scale economies, X-efficiency (indication of management shortcomings) and the PCS indicator. In general, given that information on profits is not always available, it is important for the PCS indicator method to have a reliable alternative in âmarket sharesâ. Both scale economies and X-inefficiencies are high, suggesting limited competition; indeed, stronger competition would force insurers to become more cost efficient. We show for the PCS indicator that analyses based on profits and analyses based on market shares produce comparable outcomes. Both point to rather limited competition on the life insurance market, compared with PCS values for other sectors. We use marginal cost derived from estimated translog cost models. A simpler alternative is the use of average costs (defined as operational cost divided by premiums), but it is less accurate because average costs incorrectly include fixed costs. We observe that the marginal cost and average cost results have a fairly similar pattern. Marginal costs refer to inefficiency caused by either economies of scale or X-inefficiencies (or managerial inability). In this chapter, we also derive marginal costs adjusted for scale economies to see the impact of pure X-inefficiencies on market shares. The results show that each type of inefficiency contributes separately to the dynamics in the life insurance market shares and that variation in X-inefficiency is the main source of variation in the PCS indicator over time.
In addition to the chapters in this book, we have recently applied the PCS model to the Chinese loan markets (Xu et al., 2013), an analysis of the effects of securitization on the relationship between competition and bank risk (Altunbas et al., 2013), life insurance (Bikker, 2012) and non-life insurance (Bikker and Popescu, 2014). Although these papers are not yet published in edited journals, they are valuable for further reading. Furthermore, a line of publications...