Bank Office Outreach, Structure and Performance in Regional Banking Markets

Abstract

Burgstaller J. Bank office outreach, structure and performance in regional banking markets, Regional Studies. This paper studies bank penetration, market structure and banking conduct across Austrian districts. Neither differences in market concentration nor in the activity of out-of-market banks can explain the regional dispersion of interest rates, bank profitability and efficiency. A higher bank density is connected to more competition and lower mark-ups, and indicates more efficient banking markets as well. As efficiency is the main determinant of bank profits and competition seems to be (mildly) indicative of stronger regional growth, the preservation of a fair regional bank outreach appears to be an important policy target.

Burgstaller J. 区域性银行市场中银行的外部性拓展、结构以及绩效，区域研究。本文研究了澳洲银行的渗入、市场结构以及银行行为。 市场集中或者市场外的银行活动都无法解释利率、银行收益以及效率的区域性差异。 较高的银行密度与激烈的竞争及较低的加价相关。同时表明存在更加有效的银行市场。正如效率是决定银行获益的主要因素, 竞争反映出增强的区域增长，维持银行进行公平的区域性拓展似乎是重要的政策目标。

银行拓展 地区性市场 市场结构 竞争 银行绩效 区域发展

Burgstaller J. Le rayon d'action des guichets bancaires, la structure et les résultats des marchés bancaires régionaux, Regional Studies. Cet article cherche à étudier la pénétration bancaire, la structure du marché et le comportement des banques à travers les districts autrichiens. Ni les différences de la concentration du marché, ni la variation de l'activité bancaire hors marché n'expliquent ni la dispersion des taux d'intérêt, ni la profitabilité, ni l'efficacité des banques. Une densité des banques plus élevée se rapporte à une augmentation de la concurrence et aux marges moins élevées, et indique aussi des marchés bancaires plus efficaces. Etant donné que l'efficacité est le déterminant sine qua non des profits bancaires et que la concurrence semble (dans une certaine mesure) indiquer une croissance régionale plus forte, la pérennité d'un rayon d'action bancaire régional loyal semble être un objectif important de la politique.

Rayon d'action bancaire Marchés locaux Structure du marché Concurrence Résultats des banques Aménagement du territoire

Burgstaller J. Bankstellendichte, Marktstruktur und Performance in regionalen Bankenmärkten, Regional Studies. Die vorliegende Arbeit untersucht Bankendispersion, Marktstrukturen und Bankverhalten auf der Ebene von Bezirken in Österreich. Weder Unterschiede in der Marktkonzentration noch in der Dichte nicht bezirksansässiger Banken tragen zur Erklärung der Variation in Bankzinsen, Profitabilität und Effizienz bei. Eine höhere Bankendichte steht in Verbindung mit verstärkter Konkurrenz und niedrigeren Aufschlägen und ist auch ein Indiz für höhere Bankeneffizienz. Da Effizienz die Hauptdeterminante für die Gewinne der Banken darstellt und stärkerer Wettbewerb zwischen Banken positiv mit regionaler Entwicklung verbunden scheint, ist die Erhaltung einer gewissen Bankendichte durchaus ein erstrebenswertes wirtschaftspolitisches Ziel.

Bankstellendichte Lokale Märkte Marktstruktur Wettbewerb Bankenperformance Regionalentwicklung

Burgstaller J. Densidad, estructura y rendimiento de filiales bancarias en los mercados bancarios regionales, Regional Studies. En este artículo analizamos la penetración bancaria, la estructura del mercado y las operaciones bancarias en los distritos de Austria. Ni las diferencias en la concentración del mercado ni la actividad de los bancos fuera del mercado explican la dispersión regional de los tipos de interés, la rentabilidad y la eficacia de los bancos. Una densidad más alta de los bancos está relacionada con más competencia y menos márgenes e indica también mercados bancarios más eficientes. Puesto que la eficiencia es el principal determinante de los beneficios bancarios y la competencia parece ser un indicio (moderado) de un crecimiento regional más fuerte, el mantenimiento de una densidad bancaria regional justa parece ser un objetivo político importante.

Densidad bancaria Mercados locales Estructura del mercado Competencia Desempeño bancario Desarrollo regional

Keywords:

• Bank outreach
• Local markets
• Market structure
• Competition
• Bank performance
• Regional development

JEL classifications:

• G21
• L11
• R32

Notes

For spatial approaches to the evaluation of the European Union's cohesion policy, see, for example, Dall'erba and Le Gallo (2008) and Mohl and Hagen (2010).

The following paragraphs focus on business finance through the banking system, but much of the argument applies to household financing as well. The same is true for the funding of firms through equity and venture capital markets, which is discussed by Klagge and Martin (2005) alongside their evaluation of respective policy interventions in the UK and Germany.

The role of soft information in the lending process is extensively discussed by Petersen and Rajan (1995), Berger and Udell (2002) and Stein (2002). For a recent study on lending to small and medium enterprises, which also lists many relevant references on the topic, see Ely and Robinson (2009).

Berger et al. (2007) and Cerqueiro et al. (2009), however, cast some doubts on this proposition.

The portrayals of this assertion range from ‘wide acceptance’ (Fischer and Pfeil, 2004) to ‘strong evidence’ (Cohen and Mazzeo, 2007).

Martin (1999) gives a description of many trends that are at work in this respect in the wake of deregulation and technological advances.

In the following, the focus is on evidence for small business customers. For a review of market size as perceived by households (with similar results to be deduced), see, for example, Amel et al. (2008) and Brevoort and Wolken (2009).

Attempts to estimate the scope of banking markets are mainly pursued by antitrust authorities in their evaluation of potential mergers. Thereby, the extent of competition is evaluated by using bank-level data to assess, for example, cross-price elasticities. Other approaches based on price data apply time-series methods (studying, for example, the co-integration properties of price series) or look at the uniformity of prices within markets of a perceived scope (as, for example, Hannan and Prager, 2004).

According to Rajan and Zingales (2003), interest group politics are an important factor in explaining differences in financial development across countries as well.

Banks from the following sectors are not regarded: special-purpose banks (including severance funds, investment companies and real estate funds), building and loan associations, and European Member State credit institutions.

For the vast similarity of the German banking system related to these characteristics, see, for example, Fischer and Pfeil (2004) and Schmidt and Tyrell (2004). The described ownership and network structures, as well as market segregation practices, are assessed to hinder competition (European Commission, 2007).

The Raiffeisen sector is the only one with a three-tier structure of primary banks, regional head institutions and one central organization.

Furthermore, the (regional) head institutions provide the primary banks with clearing services, product developments, marketing, research, training, and so on (for a description of corresponding networks in German banking, see, for example, Hackethal, 2004).

In general, the used data refer to 2006 and were obtained from the Austrian National Bank (Oesterreichische Nationalbank, OeNB). Financial figures for the full period (1998–2006) are applied with the dynamic calculation of competition measures (H-statistics and Lerner indices).

An Austrian district therefore is of comparable size with an average German ‘Kreis’, which is rated a useful market approximation by Fischer and Pfeil (2004).

Branch penetration was calculated by use of the data in European Central Bank (2010); population figures for 2008 were taken from Eurostat.

Population, as the income level, is frequently used as a proxy of the demand for banking services in the empirical literature (for instance, Cohen and Mazzeo, 2007; Richards et al., 2008; Coccorese, 2009).

In the following, the term ‘branch density’ is used interchangeably with ‘bank office density’, although the analysis is for bank offices (headquarters and subsidiaries) throughout.

In Germany, savings banks exhibit the most even regional distribution according to Wengler (2006) and Bresler et al. (2007).

Districts are defined non-neighbours (with zero entries for w _ij ) if the Euclidean distance between their centroids exceeds a certain threshold. The cut-off distance ensures that each district has at least one neighbour, and the weights are based on actual distance. According to Stakhovych and Bijmolt (2009), however, binary weights matrices based on shared borders, on average, perform better than those based on the n nearest neighbours or on distance in terms of estimation precision and a higher probability of detecting the true (spatial) model.

The presented correlation coefficients all are significantly different from zero. If not, this is mentioned explicitly.

Demographic bank office density and population density are negatively correlated (–0.27). A similar result for Germany is reported by Conrad et al. (2009).

The corresponding figures are very much alike those prevalent for Germany. Conrad et al. (2009) report a demographic branch penetration for German rural areas in 2003 of 0.76 per 1000 capita (based on 102 districts). The average for the forty-two Austrian districts in the Palme regions 6, 8 and 9 in 2006 is 0.78.

Both the mean and the median of the HHI distribution for Austria are quite similar to the corresponding statistics reported for local US markets by Park and Pennacchi (2009), which refer to the concentration of deposits. However, the average district-level bank concentration based on Austrian branching data is larger than similarly obtained figures for Spanish (Fernández de Guevara and Maudos, 2009), Italian (Coccorese, 2008, 2009) and US regions (Fischer, 2005).

Fischer (2005) reports such a pattern for the United States, where rural markets are more concentrated. By a simple inspection of an Austrian map with the districts' HHI, the only apparent pattern to be observed is that many districts in Tyrol and Salzburg (as well as in South-West Upper Austria) tend to have a rather low bank office concentration. In these districts, however, there often are very few banks present at the municipality level, which are small and possibly serve markets that are smaller than districts. The (alpine) form of the landscape and issues of traffic and accessibility therefore also affect the possibility of drawing conclusions about bank competition from the concentration level and other indicators of market structure.

The share of OMB branches is, at best, only weakly correlated with other covariates. Thus, its non-significance (as in other regressions described in the following) is not due to multicollinearity.

The efficiency level is highest for Vienna, followed by other urban (though not the suburban) and the intensive industrial regions. Intensive touristic districts exhibit the lowest efficiency scores.

A passable (and available) instrumental variable for competition measures is the number of distinct active banks (that operate at least one branch) in the local market. Instrument eligibility was assessed based on an F-test on excluded instruments in the first stage of two-stage least squares, and by the criterion that the instrument itself is not a significant factor in the growth regression (because there is no second suitable instrument, an over-identification test on instrument exogeneity cannot be performed). An attempt was made to instrument the HHI as well, with the density of single-branch banks as the most promising instrument. When applying Durbin–Wu–Hausman endogeneity tests, the null hypotheses of exogeneity is not rejected in these exercises.