Bank Supervision, Regulation and Efficiency:

Evidence from the European Union

† ‡
Georgios E. Chortareas , Claudia Girardone* and Alexia Ventouri

August 2010
Preliminary Draft: Please do not quote

Abstract

This paper investigates the dynamics between bank regulatory and supervisory
policies associated with Basel II’s three pillars and various aspects of banks’ cost
efficiency and performance for a sample of EU commercial banks over the period
2000-2006. We use frontier analysis and traditional accounting ratios to measure
efficiency. We also use a quasi-likelihood estimation method which is fully robust
and relatively more efficient than the OLS and Tobit approaches for models with
fractional dependent variables. Our findings suggest that interventionist supervisory
and regulatory policies such as empowering capital restrictions, fortifying official
supervisory powers, private sector monitoring and restricting bank activities, can
impede the efficient operation of banks. The produced evidence also suggests that
banks from countries with more open, competitive and democratic political systems
are more likely to benefit from higher operating efficiency levels.

JEL classification: G21, G28.

Keywords: Efficiency; Basel II; Capital Regulation; European Banks; Fractional
Data.

†
Department of Economics, University of Athens, 8 Pesmazoglou Street, Athens 10559, Greece.
Phone: +30-210-3689805. Email: [email protected].
* (Corresponding author): Essex Business School, University of Essex, Wivenhoe Park, Colchester,
CO4 3SQ (United Kingdom); Tel: +44-1206-874156; Fax: +44-1206-873429. Email:
[email protected]
‡
School of Business, Management and Economics, University of Sussex, Falmer, Brighton, BN1 9RH
(United Kingdom). Email: [email protected].
1. Introduction

Maintaining confidence and stability in the financial sector imply render capital

structure a key concern in the banking, industry. At the EU level for instance, the

structural and conduct deregulation of the 1990s has been accompanied by a parallel

increase in prudential rules, particularly capital adequacy regulation. Prior to the

recent crisis, a number of studies have emphasised the role of capital standards in

protecting banks from failure and in protecting their customers, as well as the whole

economy, from negative externalities (e.g. Hovakimian and Kane, 2000; Gorton and

Winton,1995; Rochet, 1992).

The ‘specialness’ of banks in the economy and the need to preserve financial

stability has resulted in a traditionally heavily regulated and supervised industry. Even

at the global level, international agencies including the IMF and the World Bank have

developed extensive checklists of “best practice” recommendations urging all

countries to adopt them. The most renowned example is the current accord on capital

adequacy (Basel II) – that revised and extended the first (1988) version, and is based

on three main pillars: minimum capital requirements, supervisory review and market

discipline. Basel II was expected to produce significant benefits in helping banks and

supervisors manage risks, improve stability and enable market participants to make

better risk assessments (Molyneux, 2003). Nonetheless, the recent financial turmoil

and bank insolvencies have revealed many shortcomings of the Basle II Accord that

prompted the urgency of a revised capital adequacy framework.1 Ironically, Basel II

was implemented in the EU in 2007, i.e. the year when the financial crisis began.

1
In 2009 the Basel committee responded to the lessons of the crisis by taking measures to strengthen
the Basel II framework and approved for consultation a package of proposal to strengthen global
capital and liquidity regulations with the goal of promoting a more resilient banking sector. See
http://www.bis.org/press/p091217.htm
 While regulation can be precise, through detailed and prescriptive rules, it is

not always accurate, however. Capital adequacy rules may specify how much capital

each bank should hold, but if such rules do not truly reflect the risks involved they

could unintentionally induce banks to hold either too much or not enough capital. Too

little capital increases the danger of bank failure whilst excessive capital imposes

unnecessary costs on banks and their customers and may reduce the efficiency of the

banking system. Furthermore, economic theory provides conflicting predictions about

the impact of regulatory and supervisory policies on bank performance (e.g. Barth et

al., 2004; 2007a).

Existing research on the relationship between different types of regulations,

supervisory practices and bank performance is rather limited and focuses on

individual countries (Barth et al., 2004; Beck et al., 2006; Berger et al., 2008).

Furthermore it typically relies on traditional measures of bank efficiency and

performance derived from simple accounting ratios (Barth et al., 2003a,b; and

Demirguc-Kunt, et al., 2004). Barth et al. (2006) investigate the impact of a broad

range of regulatory and supervisory practices on bank development, performance,

stability and the degree of corruption in bank lending. They provide a detailed account

of such practices for over 150 countries, including an examination of the three pillars

of Basel II, namely capital regulations, official supervision, and market discipline.

Another strand of literature relies on efficiency estimates measured using

frontier analysis as they are often deemed superior to traditional efficiency measures

based on accounting ratios (Berger and Humphrey, 1997). A handful of recent studies

focus on the relationship between regulation on bank efficiency measured using

alternative frontier methods (e.g. Fries and Taci, 2005; Grigorian and Manole, 2006;

Pasiouras et al., 2009). Nonetheless, the existing empirical evidence is scant

considering the timelines of the issue and in particular the significant increase in the

demand for regulation produced by the recent financial turmoil.

This paper contributes to the existing literature by first investigating the

impact of bank regulatory and supervisory approaches on bank cost efficiency

measured using both accounting ratios and a non-parametric frontier approach (Data

Envelopment Analysis or DEA). The data required to construct the indices that relate

to bank regulatory and supervisory practices were obtained from Barth et al. (2001b;

2006; 2007b) the World Bank (WB) database. The chosen period of study is 2000-

2006 and we focus on selected EU countries. In particular, we employ the net interest

margin and overhead costs (as in Barth et al. 2006) as well as technical inefficiency

measures to capture information about all aspects of a bank’s performance and

efficiency. Secondly, we estimate the relationship between alternative measures of

performance/efficiency and the bank regulation and supervision variables, using

Papke and Wooldridge (1996)’s method. This method is expected to yield better

estimates than the OLS and Tobit approaches used in the extant regulatory and

supervisory literature to model bank performance (as recently demonstrated by

McDonald, 2009).

Overall our results indicate that there is a strong link between various forms of

banking regulation and supervision and bank efficiency. In particular, more

demanding regulatory practices appear to decrease significantly the efficient

operations of banks.

The remaining of the paper is organized as follows. Section 2 describes the

literature on regulation and supervision and its potential relationship with efficiency.

Section 3 and 4 present the empirical methodology and the data. Section 5 discusses

the empirical results and, Section 6 concludes.

2. Regulation, Supervision and Efficiency in the Banking Sector

2.1 The Pros and Cons of Banking Regulation: An Overview

The financial sector is one of the most heavily regulated sectors in the economy and

banking is by far the most heavily regulated industry. Bank regulation typically refers

to the rules that govern the behaviour of banks, whereas supervision is the oversight

that takes place to ensure that banks comply with those rules.2 The issue of financial

regulation – particularly in relation to the banking sector – is often considered a

controversial issue. Regulation is costly and can give rise to moral hazard problems.

In addition distortions between regulated and unregulated institutions can occur

(Barth et al., 2006).

Yet the special role that banks play in the economic system implies that banks

should be regulated and supervised not only to protect investors and consumers but

also to ensure systemic stability. More specifically, bank regulations exist for

safeguarding the industry against systemic risk, protecting consumers from excessive

prices or opportunistic behaviour and finally to achieve some social objectives,

including stability (Llewellyn, 1999). Last but not least regulation is important for the

efficiency of the banking industry. In this respect, it is noticeable that whenever

regulation is implemented with the aim of restricting or limiting banking activities, the

banks’ conduct of business and the efficiency with which they operate will be

affected. This in turn could induce banks to engage in riskier activities and /or to

2
In the literature it is possible to identify three types of financial regulation: systemic, prudential and
the conduct of business regulations. For more discussion on the types of regulation, see among others,
Casu et al. (2006).
invest in ways to circumvent regulation. According to some studies, it could even

ultimately affect economic growth (Jalilian et al., 2007).

The financial crises that affected various countries around the world over the

last fifteen years or so (e.g. Latin America, South-East Asia, Russia) and the recent

US sub-prime crisis has ‘re-activated’ the long-lasting debate between those who

advocate more effective regulation and supervision and those who attribute those

crises to the failure of regulation (the so-called ‘free banking’). Most importantly, the

release of Basel II has generated a lively discussion and, while around 100 countries

are currently planning to adopt the new framework by 2015, there is still an on-going

debate as to its costs and benefits (Herring, 2005). The importance to assess the

effects of Basel II type of regulations was also recently highlighted by Raghuram

Rajan, the chief economist of the IMF who argued that rethinking bank regulation is a

big challenge and a thought- provoking issue, and criticised the current practices of

bank regulation and supervision. Moreover, under the latest proposals from the Basel

Committee (which have been dubbed ‘Basel III’) regulators are trying to make banks

better equipped against catastrophe. The increased demand for the reassessment of the

prudential rules of regulation has also been fortified, especially in light of the recent

credit crisis.

The next section reviews in detail the existing recent literature on the

relationship between regulation and bank performance and efficiency.

2.2 A Review of the Literature on Bank Regulation and Performance/Efficiency

The literature on bank regulatory practices is copious. We focus on a few key papers

that motivate our research on the effects of regulatory and supervisory policies on

various aspects of bank performance and efficiency.

 Theoretical studies have emphasised the relative importance of capital

adequacy requirements in bank regulation (Dewatripont and Tirole, 1993). Capital

serves as a buffer against losses that can absorb the possibility of bank failure.

However researchers disagree as to whether the imposition of a minimum capital

requirement actually reduces risk-taking incentives (Blum, 1999). According to the

‘official supervision approach’, official supervision can reduce market failure by

monitoring and discipline banks thus weakening corruption in bank lending and

improving the functioning of banks as intermediaries (Beck et al., 2006).

Alternatively, powerful supervisors may exert a negative influence on bank

performance. Supervisors may use their powers to benefit their own private welfare

rather than the social welfare (Becker, 1983; Shleifer and Vishny, 1989). Finally,

many researchers over the years, have shifted their attention on market monitoring as

a more successfully type of regulation than official supervisions and capital

requirements (Herring, 2004). Again, researchers disagree as to the effectiveness of

this strategy especially in poor developed markets.

One of the first study to provide empirical evidence on each of the three pillars

associated with the Basel II capital accord is that of Barth et al. (2004). In their study

the authors examine the relationship between bank regulation and supervision and

bank development, performance and stability. More specifically, they designed and

implemented a survey funded by the World Bank to collect information on bank

regulations and supervisory practices for 107 countries. Their findings raise a

cautionary flag regarding Basel II’s first two pillars, as they find no statistically

significant relationship between capital stringency, official supervisory power and

bank performance and stability. In contrast, they provide evidence indicating that

encouraging and facilitating private monitoring can boost bank performance. In

accordance with the ‘activity restrictions approach’, Barth et al. (2004) put forward

various reasons for and against restricting bank activities. However, overall their

results indicate that restricting them may not only lower banking efficiency but also

increase the probability of a banking crisis.

Similarly, Demirguc-Kunt et al. (2004) investigate the impact of bank

regulations, market structure, and national institutions on the cost of intermediation as

measured by accounting ratios (net interest margin and overhead costs). Information

on commercial bank regulations is obtained from Barth et al. (2001b; 2003) databases

for a sample of 1400 banks operating in 72 countries. The time span considered is

1995 to 1999. Their evidence indicates that tighter regulation on banking services and

activities boosts the cost of financial intermediation. A result in line with the view that

by allowing banks to engage in a broad range of activities one might expect better

diversified and more stable banking institutions.

Using the same time span as in the above study and a sample of stock

exchange-listed banks, Fernandez and Gonzalez (2005) provide evidence to suggest

that in countries with low accounting and auditing requirements more power on

official supervisory authorities may reduce risk-taking behaviour from managers’

perspectives. The authors also reveal that higher restrictions on bank activities can

diminish the probability of a banking crisis. Beck et al. (2006) use firm-level data on

2500 firms across 37 non-transition economies to examine the relationship between

supervisory strategies and corporate financing obstacles. Their empirical findings

indicate that strengthening the power of supervisory agencies may actually reduce the

integrity of bank lending with adverse implications on the efficiency of credit

allocation. Their evidence lends support to the view that private monitoring will have
a positive impact on the banking industry in terms of efficient operations and sounder

banks.

The studies reviewed above have relied on accounting measures to measure

the efficiency of the banking sector. More recently, researchers have turned their

attention to the use of frontier analysis in investigating regulatory practices in

banking. In a cross-country analysis, Pasiouras (2008) investigates the impact of

several regulations on banks’ technical efficiency for a dataset of 715 commercial

banks operating in 95 countries during 2003. Using a non-parametric analysis 3 to

estimate technical efficiency and Tobit regressions in the second step, the results

provide evidence in favour of all three pillars of Basel II that strict capital adequacy,

powerful supervision and market discipline power promote technical efficiency.

However, only the latter one is significant in the estimated specifications. Using a

parametric frontier analysis, another study by Pasiouras et al. (2009) investigate the

link between bank regulations and cost and profit efficiency. The analysis is based on

a sample of 615 publicly quoted commercial banks from 74 countries covering the

period 2000-2004. Their findings are in favour of supervisory power and market

discipline mechanism, indicating that enhanced market discipline and supervisory

power increase both profit and cost efficiency. On the other hand the results on capital

requirements and restrictions on bank activities provide mixed results.

Overall, although there appears to be some evidence in favour of market

discipline, existing literature on the impact of different aspects of regulations on bank

efficiency provides mixed results. Moreover, there seems to be a lack of studies

focusing specifically on the European area, as existing ones are mainly using

international cross-country data. Lastly, while only a handful of these studies rely on

3
Namely the DEA or Data Envelopment Analysis (details can be found in Section 3.1).
frontier analysis in their regression models, none of these have a reliable

methodological framework that allows to consider the fractional nature of the

efficiency estimates in the second step analysis.

In this paper, we employ a Quasi-likelihood estimation method (see Section

3.2) to estimate the relationship between alternative measures of efficiency and the

bank regulatory and supervisory practices. We also allow for a better understanding of

bank regulation and supervision on the different aspects of a banks production

performance, by using both accounting ratios and a frontier method measure as

proxies for the level of bank efficiency. The next section illustrates the methodology

used for the empirical analysis.

3. Methodological Issues

3.1 Measuring Efficiency

As discussed in the previous section, the extant literature on bank regulations and

supervisions relies mostly on accounting ratios for measuring bank performance.

Recent literature however has highlighted the benefits of using frontier efficiency

measures as indicators of bank performance with respect to traditional accounting

ratios (see e.g. Berger and Humphrey, 1997),

The efficiency scores used in our model are constructed using an input-

oriented Data Envelopment Analysis (DEA) methodology. Originally developed by

Charnes et al. (1978), DEA employs a linear programming framework and makes

some fairly general assumptions about the production technology (Ray, 2004), in

order to provide an estimate of the Farrell (1957) efficiency measure for each bank in

the sample.4 In the generic situation of n banks, with each of them consuming m

4
For a systematic introduction to DEA methodology, see among others, Ray (2004).
different inputs to produce s different outputs and constant returns to scale, this

translates into the following linear programming problem being solved n times; each

time for a different bank in the sample.

Minθ ,λ θ ,
st θxi − Xλ ≥ 0,
(1)
− y i + Yλ ≥ 0,
λ ≥ 0,

where θ is a scalar, λ is a vector of ones and finally X and Y are the m × n input

and s × n output matrices respectively. In this context θ is the efficiency score of

each bank and is measured relative to an estimate of the true production frontier

which is known as the best practice frontier. When the value of θ is unity the bank

operates on the efficient frontier and is therefore deemed efficient.

The efficiency scores are estimated relative to a common best practice frontier

by pooling the data across countries. In particular, our sample comprises commercial

banks operating in selected European countries. A common frontier is computed

under the assumption that the banks operating in these countries share the same

technology. Following Berger and Humphrey (1997), we adopt the “intermediation

approach”, which views banks as institutions that employ labour, physical capital and

deposits to produce loans and other earning assets. Accordingly, we consider

personnel expenses, total fixed assets, and deposits and short term funding as inputs

and total loans and other earning assets as outputs. Capturing the non-traditional

activities of banks is essential, especially when dealing with banking institutions

across the European area that have a wider scope of activities than the traditional.

Hence, the fee-based financial services are included as a third output. The estimated
efficiency scores are then regressed against a set of regulatory, bank-specific and

macro variables as explained in Section 3.2.

3.2 The Determinants of Efficiency estimates

The analysis of the determinants of efficiency has evolved in recent years from using

simple OLS regressions to employing more sophisticated and improved models.

Some studies have employed censored regression models using Tobit estimators (e.g.

Pasiouras, 2008). More recent works employ Monte Carlo simulations to demonstrate

that a truncated regression is more appropriate in this context (Simar & Wilson,

2007).

Nonetheless, a typical characteristic of efficiency scores is that it takes values

between zero and one. The same logic applies for inefficiency (1-efficiency). Simar

and Wilson (2007) have recently argued that the efficiency variable follows truncated

distribution and therefore truncated regressions are more appropriate than the OLS

regressions in the second stage analysis. A recent development by McDonald (2009)

asserts that efficiency is not the outcome of a truncated process but rather the

outcome of a fractional logit process. That is, it takes values between zero and one

and is not a latent variable. For that reason Papke and Wooldridge (1996)’s Quasi-

Likelihood estimation methods have been proved more appropriate for this type of

data. In this study we employ the proposed methodology by Papke and Wooldridge

(1996) to assess the impact of bank regulatory and supervisory approaches on bank

efficiency, while controlling for bank-specific and country-specific characteristics.

Specifically, the estimated regressions can be written as follows:

BankInefficiencyi,k = α + β1Si + β 2 Bi,k + β 3Ci + ε i ,k (1a)

(IM i ,k = α + β1 S i + β 2 Bi ,k + β 3Ci + ε i ,k (1b)

OCi , k = α + β1S i + β 2 Bi , k + β 3Ci + ε i , k (1c)

where i indexes country i , k indexes bank k , S i is a vector of bank

regulatory and supervisory indicators in country i , Bi ,k is a vector of bank-specific

characteristics for each bank k in country i , Ci is a vector of country-specific

control variables in country i and ε i,k is the error term.

The dependent variable Bank Inefficiency in equation (1a) is the managerial

technical inefficiency measure INEFF, which measures how far the bank is from the

estimated efficient frontier. In other words, efficiency is the ability of a firm to fully

exploit the available production technology and relates to efficiencies brought about

by superior management or technologies. The dependent variables in equations (1b,

1c) are two accounting ratios that proxy for the level of bank inefficient

intermediation as in Barth et al. (2004; 2006). In particular, the NIM, in equation (1b)

is the net interest margins over interest bearing assets. As pointed out by Barth et al.

(2006) one should expect high net interest margins to signal inefficient intermediation

and greater market power that allows banks to charge higher margins. In other words,

the NIM measures the gap between what the bank pays savers and what the bank

receives from borrowers, thus it is focused on the traditional borrowing and lending

operations of the bank.

On the other hand, the dependent variable in equation (1c) is the overhead

costs over total assets variable, OC, used to capture cross-bank differences in the

efficiency with which banks are managed. It is argued that high overhead costs can
signal unwarranted managerial perquisites and market power that contradicts the

notion of efficient intermediation of banks (Barth et al. 2006). Cost inefficiencies and

market power may be reflected in high overhead costs. Therefore, the results using all

three performance measures can be used for comparative purposes to compare

different concepts of efficiency of banks and how technical estimates relate to

standard financial ratio measures of efficiency and performance.

To account for Basel II’s pillars on bank regulation and supervision we use

the variables in the vector Si of equations (1a-1c). Data for the regulatory and

supervisory variables were obtained from Barth et al. (2001b; 2006; 2007b) WB

database. Specifically the vector is defined as follows:

S i = ( SPOWERi , CAPRQi , PRMO(ITi , ACTRSi ) (2)

where SPOWER is Official Supervisory Power and measures the ability of

supervisory authorities to take specific action in banking decisions to prevent and

correct problems. It is calculated by adding 1 for each ‘yes’ and 0 for each ‘no’ to

questions relating to supervisory power in terms of prompt corrective power,

restructuring power and declaring insolvency power. Higher values of this variable

indicate greater power of supervisory authorities to get involved in banking decisions.

Strong official supervision may signal efficient banking institutions, preventing

managers from engaging in excessive risk-taking behaviors. On the other hand,

excessive supervision may be a signal of corruption or obstruct bank operations

resulting in lower efficiency. Yet, the bulk of empirical evidence on official

supervision provides mixed results.

 Capital Regulatory variable, CAPRQ, is an index of capital requirements that

accounts for both overall and initial capital stringency. Overall capital stringency

examines whether capital requirement reflects certain risk elements and deducts

certain market values losses from capital before minimum capital adequacy is

determined, whereas initial capital stringency indicates whether certain funds may be

used to initially capitalize a bank and whether they are officially verified. It is

calculated on the basis of nine questions with higher values indicating greater capital

stringency. Both official supervision and capital requirements constitute Basel II’s

first two pillars, yet again the empirical evidence is mixed. Barth et al. (2006) for

example find no strong association between power supervisions, capital requirements

and bank efficiency.

The third variable that proxies for Basel II’s third pillar on market discipline is

Private Monitoring, PRMONIT. This variable is calculated by adding 1 for each ‘yes’

and 0 for each ‘no’ to ten questions, with higher values indicating more informative

bank accounts. Those questions indicate the degree of information that is released to

officials, the public, credit audit requirements and whether credit ratings, either from

domestic or international credit rating agencies are required. Barth et al. (2006, 2004a)

provide evidence that regulations that enhances and facilitates private monitoring can

significant boost bank efficiency. More recently, Pasiouras (2008) shows that

encouraging and facilitating private monitoring of banks can boost efficiency.

Finally, Activity Restrictions, ACTRS, is an indicator of the degree of which

banks may engage in real estate investment, insurance underwriting and selling,

underwriting, brokering and dealing in securities and all aspects of the mutual fund

industry. This variable ranges from 0 to 4, with higher values indicating greater

restrictiveness. According to Barth et al. (2001a,b, 2003a,b), activity restrictions may

have an important impact on bank efficiency by reducing competition and limiting

economies of scope, which results in lower efficiency levels. We expect a negative

sign between efficiency and activity restrictions.

In the regression models described in equations (1a-1c) we also account for

bank specific Bi, k and country specific C i control variables. Specifically, the vectors

are defined as follows:

Bi , k = ( HERFi , k , L(TAi , k , LIQi , k , EQASi , k ) (3)

C i = ( ∆ GDP i , VOICE i , CORR i , GOVER( i , OPE(POL i ) (4)

The vector of bank specific variables, Bi,k, as described in equation (3),

includes measures for market share, bank size, liquidity and bank equity. To account

for market share we identify the Herfindahl index, which represents market share (in

terms of total assets) of every bank in the sample and gives a measure of market

concentration. Banks in concentrated markets are assumed to benefit from low

competitive pressure, allowing them to exercise market power and thus earn

monopolistic profit by e.g. offering lower deposits rates and charging higher loan

rates. These assumptions are derived from the traditional SCP theory in the industrial

structure of banks, which posits that banks in concentrated environments are more

likely to engage in collusive behaviour (Berger and Hannan, 1998). It has been

shown, however, that concentration is not necessarily synonymous of market power if

the banks operate efficiently. Specifying the expected sign of the HERF coefficient

can be fairly complicated due to different forces working in opposing directions.

These above arguments suggest that on balance the relationship between competition
and bank efficiency is by no means straightforward (see e.g. Casu and Girardone,

2006; Schaeck et al., 2009).

The Bank Size variable (LNTA) which is defined as the natural logarithm of

bank’s total assets is included in the regression to measure how bank size influences

efficiency level. Bank size may positively influence efficiency levels as big banks are

able to hold less capital compared to their smaller counterparts, and may also be able

to have greater portfolio and loan diversification and gain from size advantages

(Hughes et al., 2001; Yildirim and Philippatos, 2007; Altunbas et al., 2007).

We also include the loans to deposits ratio (LIQ) to account for the level of

liquidity, which proxies for differences in bank assets. Liquid assets reduce banks’

liquidity risk. Simultaneously, though, banks have to incur an opportunity cost for

holding them that could hamper their cost efficiency and adversely affect their

competitive viability. Therefore, a positive relationship between Liquidity and

efficiency is anticipated (Elyasiani et al., 1994).

The equity over total assets (EQAS) ratio is a measure of the degree of risk

taken by bank managers as higher leverage increases the risk of insolvency which

could result in greater borrowing costs (Berger and Mester, 1997; Dietsch and

Lozano-Vivas, 2000; Flannery et al., 2004). Moreover, higher levels of capitalisation

may reflect higher incentives from the stockholders to monitor management, thus

resulting in alleviating the efficiency problem caused by conflicts of interest

(Eisenbeis et al., 1999). In light of the above arguments, one would expect a positive

correlation between well capitalised banks and higher efficiency levels. Nevertheless,

financial capital also affects costs through its use as a resource for financing loans

(Berger and Mester, 1997). Since raising capital, however, through issuing shares

typically involves higher costs than raising deposits, a negative correlation between
EQAS and efficiency levels should be expected. In light of the above, predicting the

sign of the EQAS coefficient is a fairly complicated task due to different forces

working in opposing directions.

Additionally, the vector of control variables in equation (4), Ci contains

measures of economic conditions and institutional environment. The average annual

growth rate of per capita (∆GDP) is an environmental variable used to control for

local economic conditions. A high level of per capita GDP captures the cyclical

conditions of the macroeconomic environment. It is also expected to capture the

implications for bank efficiency stemming from operating in different economic

environment, as demand for financial products depends on the level of economic

activity. Empirical studies tend to find that countries with relatively high GDP growth

are characterised by more efficient banking institutions (e.g., Demirguc-Kunt and

Maksimovic, 1998; Schure et al., 2004; Yildirim and Philippatos, 2007).

The vector of institutional control variables in the efficiency equation includes

the following variables: voice and accountability (VOICE), control of corruption

(CORR), government-owned banks (GOVERN) and political openness (OPENPOL).

Voice and accountability is an indicator of the degree to which a country’s citizens are

able to participate in selecting their government, as well as freedom of expression,

freedom of association and a free media. This variable is taken from Kaufman et al.

(2009) dataset on institutional development. Control of corruption is an index of the

overall corruption in the country’s government, where larger values indicate that

government officials are less likely to demand illegal payments. This variable is also

taken from Kaufman et al. (2009), measuring perceptions of the extent to which

public power is exercised for private gain, including both petty and grand forms of

corruption, as well as “capture” of the state by elites and private interests. The
government-owned banks variable is used as proxy for the degree of state-owned

banks. It is calculated as the fraction of the banking system’s assets that is held by

banks that are more that 50 percent owned by the government. Finally, political

openness measures the openness, competitiveness and the level of democracy of the

country. This variable is from Marshall and Jaggers’s Polity IV database5. Higher

values indicate greater openness in the political environment.

4. Data

The dataset used in this study is composed of individual bank data sourced from

unconsolidated statements of banks operating in selected European countries area

available in the BankScope database of Bureau van Dijk. The banks we consider are

those operating in 11 selected EU countries: Austria, Belgium, Bulgaria, Czech

Republic, Denmark, Italy, Latvia, Luxembourg, Netherlands, Portugal and Sweden6.

The chosen time span is 2000 to 2006.

The sample used in this study is comprised of institutions classified as

commercial banks. The data have undergone substantial editing to void

inconsistencies, reporting errors and double counting of institutions. Moreover in

order to obtain a relatively homogenous dataset and further detect and remove the

potential outliers from the sample, we apply the following screening methods: (1)

eliminate bank-year observations with total-loans-to-assets ratios below the 20

percent, (2) apply the Jackstrap methodology as described in the Appendix.

5
The type of political system was examined using data (Polity2) from the Polity IV data set, and the
variable Polity ranges from –10 to +10, where a higher score indicates a more democratic system
(Marshall & Jaggers, 2002). Available at http://www.systemicpeace.org/polity/polity4.htm
6
Due to unavailability of data or/and missing values for significant number of banks we excluded
Estonia, Greece, Hungary, Ireland, Lithuania, Poland, Slovakia, Slovenia from our EU dataset. For
similar reasons we exclude from the analysis Cyprus, France, Germany, Malta, Romania, Spain and
United Kingdom. The main reason the unavailability of data for Private Monitoring Index variable,
from Version II for the period 2001-2003.
 Therefore, after implementing the aforementioned screening methods, we

obtain an unbalanced panel consisting of 1,382 commercial bank observations7. Table

1 contains the number of bank observations by year, as well as their total assets,

which are expressed in million Euros.

Table 1
Time and size distribution of banks included in the estimation of the frontier
Europe
Asset size
2umber of
obs.
Year Mean Min Median Max
2000 190 5,635.37 47.73 673.95 101,083.40
2001 197 5,258.71 41.03 516.64 132,285.55
2002 190 5,515.59 29.22 455.76 153,447.10
2003 180 8,065.36 30.88 544.72 337,247.59
2004 200 7,597.19 27.05 605.85 370,663.69
2005 234 10,990.88 43.56 871.62 427,192.50
2006 191 15,122.11 50.71 1,258.00 451,832.41
Source: Bankscope.

Data for regulatory and supervisory variables were collected form Barth et al.

(2001b; 2006; 2007b) databases. In particular we have specified three groups of

variables. The first group contains bank regulatory and supervisory indicators,

focusing on Official Supervisory Power, Capital Regulatory Index, Private

Monitoring and Activity Restrictions. The second group includes bank-specific

characteristics and the third group contains country-specific factors that are expected

to influence banks’ (in)efficiency. Information on bank regulation and supervision is

obtained from the Wold Bank database developed by Barth et al. (2001b) Version I,

and updated by Barth et al. (2006, 2007b) with Versions II & III.

7
The banks we consider are those of Austria, Belgium, Bulgaria, Czech Republic, Denmark, Italy,
Latvia, Luxembourg, Netherlands, Portugal and Sweden.
 Descriptive statistics for the inputs and outputs used in the DEA efficiency

measurement and the explanatory variables employed in the model are available in

Tables 2 and 3, respectively.

Table 2
Descriptive Statistics of Banks Inputs and Outputs
Variable Mean Std.dev Minimum Maximum Median
Technical Efficiency
Inputs
Personnel expenses 58.45 205.47 0.20 2,409.50 7.68
Total Fixed Assets 42.46 149.55 0.03 1,842.00 5.96
Deposits & Short-term
Funding 6,432.01 25,985.82 17.38 369,000.69 547.06

Outputs
Total Loans 3,674.22 12,837.99 7.71 130,221.70 362.80
Total Other Earning Assets
4,016.59 18,009.04 0.10 273,663.41 223.15
Fee-based Income 55.40 198.22 0.04 2,545.70 6.60
a
Figures are in mil. Euros.
Source: Bankscope and own calculations.
 Table 3
Descriptive Statistics for the Variables Employed in the Cross Sectional Regressionsa
Symbol Definition Mean St.dev. Minimum Maximum Median
Dependent Variables
INEFF Technical Inefficiency measure using the Data Envelopment 0.15 0.14 0.00 0.56 0.13
Analysis (DEA) methodology (VRS)
NIM
0.02 0.04 0.00 0.57 0.01
(Interest Income - Interest Expenses) / Interest-Bearing Assets
OC Overhead Costs / Total Assets 0.03 0.02 0.00 0.21 0.02
Regulatory and Supervisory Variables
SPOWER Official Supervisory Power 9.78 2.39 5.00 14.00 10.00
CAPRQ Capital Regulatory Index 6.04 1.92 3.00 10.00 6.00
PRMONIT Private Monitoring 6.76 1.66 3.00 9.00 6.00
ACTRS Activity Restrictions 7.52 2.50 3.00 12.00 7.00

Bank-Specific Control Variables

HERF Herfindahl index of local market concentration 0.24 0.17 0.03 0.88 0.18
LNTA Logarithm of Total Assets 6.82 1.79 3.30 13.02 6.52
LIQ Total Loans / Total Deposits 0.72 0.33 0.22 4.49 0.71
EQAS Shareholder's Equity / Total Assets 9.67 6.69 0.74 77.37 8.29

Country-Specific and Institutional Variables

∆GDP Annual Growth rate of per capita GDP 0.02 0.02 -0.02 0.09 0.02
VOICE Voice and Accountability 1.31 0.33 0.48 1.83 1.39
CORR Control of Corruption 1.49 0.80 -0.14 2.35 1.91
GOVERN Government-Owned Banks 3.28 4.51 0.00 25.00 0.16
OPENPOL Political Openness 9.79 0.58 8.00 10.00 10.00
a
All financial variables measured in millions Euros. Annual GDP growth is measured at constant 1995 market prices.
Sources: WB (Barth et al. 2001b; 2006; 2007b), Governance Matters VIII (Kaufman et al., 2009), Polity IV database (Marshall and Jaggers, 2002), AMECO, Bankscope and own
calculations.
5. Results
The banks’ efficiency scores are estimated relative to a common best practice frontier

by pooling the data across countries. This approach allows for estimating efficiency

differentials not only between banks within a country but across countries as well

using the same benchmark technology.

Figure 1 illustrates the average DEA efficiency scores by year and by country.

Overall, the results show relatively high average technical inefficiency levels of about

15%, which is broadly in accordance with previous bank efficiency studies covering

Europe (e.g. Goddard et al., 2001; Lozano-Vivas et al., 2002; Casu and Molyneux,

2003). Technical efficiency displays a decreasing trend (see panel a), peaking in 2003

and then weakening over the following years. This trend is probably driven by the

anticipated changes that took place during the studied period (e.g. the new member

states that joined the EU in 2004) which probably took their toll on the EU banking

sector of some specific countries included in our sample. The (commercial) banking

sectors that achieved better operating efficiency during the early millennium recession

have been the Luxembourg and the Italian ones, as displayed in panel (b).

Figure 1
Technical efficiency scores by year and country
in selected European countries (2000-2006)

Panel (a) Panel (b)

 In this section, we analyse how regulatory and supervisory practises, based on

Basel II accord, affect a bank’s production performance. First, we regress our DEA

inefficiency scores against bank regulatory and supervisory practices and then we

advance the literature by comparing the results with those of Barth’s et al. (2006),

using NIM and overhead costs as dependent variables. To this aim we regress net

interest margins, overhead costs and a frontier efficiency measure, used as proxies for

a bank’s profit, cost and inefficiency, respectively, on a set of regulatory and

supervisory variables, while controlling for country and bank specific effects.

The results of the regressions are reported in Tables 4-6. Each table presents

six regressions. The first column presents the basic regression that includes bank-

specific control variables, economic growth, and the bank regulatory and supervisory

variables (model 1). The next four columns include country-specific control variables

one at a time (models 2-5). The last column corresponds to a re-estimation of the

basic model with year dummies to control for potential time effects (model 6). Due to

the fact that the institutional country variables are highly correlated we do not run a

regression of all these variables together, to avoid problems of multicollinearity. The

reported estimates of equations (1a,b,c) result from applying the Papke and

Wooldridge (1996) Quasi-Likelihood estimation method to the full sample of banks in

a cross sectional regression over the period 2000-2006.

 Table 4
Results on Bank Inefficiency and Regulation using equation (1a)
Years: 2000-2006 Models
Dep.Var.: I2EFF (1) (2) (3) (4) (5) (6)

Regulation and Supervision

CAPRQ (Pillar I) 0.041* 0.060*** 0.041* -0.002 0.016 0.083***
SPOWER (Pillar II) 0.017 0.024 0.046** 0.067*** 0.045** -0.022
PRMONIT (Pillar III) 0.042* 0.059** 0.091*** 0.119*** 0.038 0.108***
ACTRS (Pillar III) 0.096*** 0.139*** 0.187*** 0.168*** 0.077*** 0.031

Bank Specific variables

HERF 1.371*** 0.949*** 0.883*** 0.882*** 0.910*** 2.015***
LNTA -0.231*** -0.239*** -0.237*** -0.217*** -0.211*** -0.244***
LIQ -0.172* -0.139 -0.143 -0.051 -0.385*** -0.245**
EQAS -0.025*** -0.026*** -0.027*** -0.029*** -0.028*** -0.029***

Country Specific variables

∆GDP 5.318*** 9.088*** 10.168*** 2.715* -5.743** 4.543**
VOICE - -0.596*** - - - -
CORR - - 0.362*** - - -
GOVERN - - - -0.091*** - -
OPENPOL - - - - -0.1 -

Constant 0.041* -3.122*** -3.582*** -2.748*** -2.434** -1.937***

Year Dummies No No No No No Yes

Observations 1382 1382 1382 1382 1190 1382
Number of Countries 11 11 11 11 10 11
*p<0.1; **p<0.05; ***p<0.01
Note: SPOWER= Official Supervisory Power, CAPRQ= Capital Regulatory Index, PRMONIT=
Private Monitoring, ACTRS= Activity Restrictions, HERF= Herfindahl Index, LNTA= LN of Total
Assets, LIQ= Total Loans/Total Deposits, EQAS= equity/assets, ∆GDP= Real GDP Growth, VOICE=
Voice & Accountability, CORR= Control of Corruption, GOVERN= Government-Owned Banks,
OPENPOL= Political Openness, Constant= constant term.
Estimated using Papke and Wooldridge (1996) Quasi-Likelihood estimation method.
 Table 5
Results on Bank Inefficient Intermediation and Supervision
using equation (1b)
Years: 2000-2006 Models
Dep.Var.: 2IM (1) (2) (3) (4) (5) (6)

Regulation and Supervision

CAPRQ (Pillar I) 0.062*** 0.064*** 0.062*** 0.032** 0.034** 0.051***
SPOWER (Pillar II) -0.029 -0.028 -0.024 -0.014 -0.009 -0.018
PRMONIT (Pillar III) 0.089*** 0.089*** 0.093*** 0.112*** 0.091*** 0.065***
ACTRS (Pillar III) 0.048*** 0.050*** 0.060*** 0.063*** 0.019 0.069***

Bank Specific variables

HERF 0.368*** 0.337** 0.276* 0.163 0.194 0.455***
LNTA -1.159*** -1.160*** -1.160*** -1.157*** -1.154*** -1.158***
LIQ -0.769*** -0.772*** -0.781*** -0.818*** -0.826*** -0.767***
EQAS 0.020*** 0.020*** 0.021*** 0.022*** 0.017*** 0.020***

Country Specific variables

∆GDP 6.186*** 6.428*** 6.976*** 4.968*** 9.719*** 7.577***
VOICE - 0.031 - - - -
CORR - - 0.048 - - -
GOVERN - - - -0.028*** - -
OPENPOL - - - - 0.180** -

Constant 1.760*** 1.695*** 1.529*** 1.691*** 0.244 1.674***

Year Dummies No No No No No Yes

Observations 1382 1382 1382 1382 1190 1382
Number of Countries 11 11 11 11 10 11
*p<0.1; **p<0.05; ***p<0.01
Note: SPOWER= Official Supervisory Power, CAPRQ= Capital Regulatory Index, PRMONIT=
Private Monitoring, ACTRS= Activity Restrictions, HERF= Herfindahl Index, LNTA= LN of Total
Assets, LIQ= Total Loans/Total Deposits, EQAS= equity/assets, ∆GDP= Real GDP Growth, VOICE=
Voice & Accountability, CORR= Control of Corruption, GOVERN= Government-Owned Banks,
OPENPOL= Political Openness, Constant= constant term.
Estimated using Papke and Wooldridge (1996) Quasi-Likelihood estimation method.
 Table 6
Results on Bank Inefficient Intermediation and Supervision
using equation (1c)
Years: 2000-2006 Models

Dep.Var.: OC (1) (2) (3) (4) (5) (6)

Regulation and Supervision

CAPRQ (Pillar I) 0.007 0.002 0.007 -0.003 -0.006 -0.012
SPOWER (Pillar II) 0.032*** 0.030*** 0.028** 0.042*** 0.049*** 0.047***
PRMONIT (Pillar III) 0.085*** 0.081*** 0.078*** 0.101*** 0.087*** 0.052***
ACTRS (Pillar III) 0.073*** 0.062*** 0.060*** 0.087*** 0.059*** 0.099***

Bank Specific variables

HERF 0.310*** 0.435*** 0.388*** 0.225*** 0.122 0.371***
LNTA -0.167*** -0.166*** -0.168*** -0.164*** -0.142*** -0.170***
LIQ 0.149** 0.144** 0.148** 0.168*** 0.097 0.137**
EQAS 0.019*** 0.018*** 0.018*** 0.019*** 0.020*** 0.019***

Country Specific variables

∆GDP 0.228 -0.837 -0.478 -0.408 1.553 1.028
VOICE - -0.161*** - - - -
CORR - - -0.051** - - -
GOVERN - - - -0.015*** - -
OPENPOL - - - - 0.091 -

Constant -4.444*** -4.081*** -4.184*** -4.647*** -5.392*** -4.306***

Year Dummies No No No No No Yes

Observations 1382 1382 1382 1382 1190 1382
Number of Countries 11 11 11 11 10 11
*p<0.1; **p<0.05; ***p<0.01
Note: SPOWER= Official Supervisory Power, CAPRQ= Capital Regulatory Index, PRMONIT=
Private Monitoring, ACTRS= Activity Restrictions, HERF= Herfindahl Index, LNTA= LN of Total
Assets, LIQ= Total Loans/Total Deposits, EQAS= equity/assets, ∆GDP= Real GDP Growth, VOICE=
Voice & Accountability, CORR= Control of Corruption, GOVERN= Government-Owned Banks,
OPENPOL= Political Openness, Constant= constant term.
Estimated using Papke and Wooldridge (1996) Quasi-Likelihood estimation method.
 Our results indicate that the major two pillars of Basel II (pillar 1 & 2)

uncover a significant impact on bank inefficiency, indicating that higher capital

requirements and more powerful supervisors can adversely affect the efficiency of

banks. In particular, CAPRQ has usually a positive and statistically significant

coefficient in all regression models. This finding appears to be particularly strong

when NIM is the dependent variable (eq. 1b), and in most of the cases when using

INEFF (eq. 1a). Concerning pillar 2, the coefficient for the variable SPOWER is

positive and significant in three cases when DEA is the dependent variable (see

models 3-5 in Table 4). The relationship appears more unambiguously important

when overhead costs are included as dependent variable (eq. 1c).

Official capital adequacy regulations play a crucial role in aligning the

incentives of bank owners with depositors and other creditors. Theoretical models,

however, disagree over whether the imposition of capital requirements actually

reduces risk-taking incentives. As emphasized in Barth et al. (2004), it is

extraordinarily difficult for regulators and supervisors to set capital standards that

mimic those that would be demanded by well informed, undistorted private-market

participants. This view is consistent with theories that suggest that if regulators have

high screening ability, then looser regulatory policy with lower capital requirements

can improve efficiency (VanHoose, 2007). In an event of a banking crisis, weaker-

ability regulators will tend to tighten capital requirements (thus increase inefficiency),

while stronger-ability regulators will tend to loosen capital requirements, which

improves efficiency.

On the other hand, governments with powerful supervisors may use this power

to benefit favoured constituents or their own private welfare, which in turn results in

inefficient banking institutions and overcoming market failures. Moreover, onerous

capital regulations and supervisions may reflect excessive government involvement,

which in turn distort bank incentives leading to the undertaking of risky investments

(Barth et al., 2004).

Turning to the variable explaining market discipline (pillar 3), our results

uncover a strong and statistically significant relationship with inefficiency. To recall,

higher values for PRMONIT imply more informative and transparent banks’ balance

sheets. As shown in Tables 2-5, we find a positive and significant impact of

PRMONIT on inefficiency, across all specifications and for all chosen measures of

performance. In other words, our evidence suggests that fortifying private sector

monitoring actually impedes the efficient operation of banks. This result is in contrast

with previous empirical findings that provide evidence in favour of private monitoring

practices (Barth et al., 2004, 2006; Pasiouras, 2008).

Focusing on the variable explaining activity restrictions, the results of a

positive and statistically significant relationship with all measures of performance

indicate that restricting banks from engaging in security activities is strongly

associated with lower bank efficiency and performance. This result is in accordance

with previous findings in the banking literature (see among others, Barth et al., 2004,

2006; Demirguc-Kunt et al, 2004). Specifically, the literature suggests that restrictions

on bank activities and general impediments to bankers in their business conduct

reduce the efficiency of bank operations without a corresponding benefit in terms of

other measures of bank performance.

The overriding message that emerges from these analyses is that there appears

to be a strong link between the three Basel II pillars and bank efficiency and

performance. In particular results suggest that heavier capital requirements, powerful

supervisions by monetary authorities, excessive private monitoring and regulatory

restrictions on bank activities are associated with greater banking system inefficiency,

not enhanced efficiency. These findings have important policy implications,

particularly in light of the existing debate on the validity of Basel II’s pillars, as for

which of these regulations really work.

Turning to the vector of bank specific variables, the first notable results is the

evidence of a positive and significant relationship between market share and

inefficiency. In particular, the coefficient for Herfindahl index (HERF) enters positive

in all of our specifications. This result is in line with previous European evidence and

can be interpreted as if the lack of market discipline in concentrated markets results in

a reduction of a bank’s costs, profits and efficiency (Berger and Hannan, 1998;

Yildirim and Philippatos, 2007). In the absence of market discipline mechanisms,

managers of firms with more market power may be allowed to pursue their own

objectives, thus causing firms’ profitability to decline. This view is consistent with

studies that focus on structural market imperfections arising from imperfect

competition which may cause market power and lax market discipline in concentrated

markets.

As in most cases, bank size appears to be an important significant factor that

drives the differences in efficiency across banks, which is documented by a negative

and statistically significant sign for the LNTA coefficient. This finding is consistent

with previous studies European countries (e.g. Stavarek, 2004; Altunbas et al. 2007;

Yildirim and Philippatos, 2007). The Liquidity variable (LIQ) has been included in

the analysis to capture the degree of association between the level of liquidity risk

measured as total loans over total deposits, and their efficiency. For this variable we

find mixed results. The coefficient is negative and statistically significant when

INEFF and NIM are dependent variables, thus indicating that the most cost efficient
and profitable banks are also less liquid (Elyasiani et al., 1994). In contrast, the

positive sign for the variable LIQ in Table 5.6 reveals that this can also be costly in

terms of overhead costs.

The results for Bank Equity (EQAS) reveal a significantly negative

relationship with INEFF and EQAS. This evidence suggests that higher capital ratios

are related with greater efficiency. This finding most likely reflects the benefits from

higher capitalisation on alleviating agency problems between managers and

shareholders. Shareholders in this case have greater incentives to monitor

management’s performance and ensure that the bank is run efficiently (Eisenbeis et

al., 1999). On the other hand, our results reveal that higher capital ratios are also

associated with greater costs in terms of interest margins and overhead costs. This

finding is in line with Barth et al.(2006), thus one could say that while higher

capitalisation leads to more efficient banking institutions, this also comes with a cost

in terms of accounting ratios.

As far as the macro environment is concerned, the GDP growth (∆GDP)

variable enters with a positive sign in all regressions and for all specifications with

INEFF and NIM. This result indicates that banks in expanding markets may be less

efficient in controlling their costs, thereby results in lower efficiency levels. Turning

to the country-specific control variables, we observe that banks operating under more

open frameworks are more likely to achieve higher efficiency levels. In particular, the

variable VOICE, that measures the degree of freedom of expressions and free media

in a country’s system, is negatively associated with both technical inefficiency and

overhead costs, indicating that more developed and democratic systems are associated

with more efficient banking sectors.

 Unsurprisingly, the control of corruption in lending (CORR) is positively

associated with inefficiency. However, the same variable is also found to be

significantly negatively related with costs, a result that probably reflects the incentives

of bank managers to focus on increasing operational efficiency rest assured that the

market is not corrupted or experience less corruption levels.

On the other hand, our results on GOVERN variable that equals the fraction of the

banking system’s assets that is held by banks that are more than 50% owned by the

government, are in line with previous studies that argue that in the absence of

corruption, governments have adequate information to promote socially desirable

investments thus increase the efficiency in the industry and economy welfare. We find

that greater government ownership is generally associated with less inefficient

banking industries, a result in line with the view that government-owned banks

contribute to economic development (Stiglitz, 1994).

Finally, the coefficient for OPENPOL (that takes values of -10 if the level of

democracy, the openness and the competitiveness is poor and 10 otherwise) shows a

positive and significant sign only when NIM is the dependent variable, implying that

the quality of democratic authority in governing institutions has an impact mainly on

bank margins.

6. Conclusions

This paper contributes to the existing literature by empirically examining the impact

of regulatory and supervisory policies and the three pillars of Basel II on bank

efficiency and performance. We focus on a sample of banks operating in selected

European countries and the period under study is 2000 to 2006. We obtain efficiency

scores using an input-oriented Data Envelopment Analysis methodology. We also

consider performance measures calculated using traditional accounting ratio, namely

net interest margin and overhead costs. We use both types of performance measures

not only to enhance the robustness of our analysis but also to make our results directly

comparable with those of previous studies.

The results raise a cautionary flag regarding the efficacy of Basel II pillars

regarding capital regulations and official supervision on the banking sector

performance In particular, we produce evidence that tightening the official

supervisory power or increasing capital requirements can have a discernible negative

impact on bank efficiency, possibly by increasing agency problems, enhancing market

power and minimising operational efficiencies. Moreover, our findings on pillar 3 of

Basel II, market discipline, indicate that excessive private monitoring and regulatory

restrictions on bank activities can affect the efficient operation of banks. A message

that emerges from this analysis is that there is a strong link between bank efficiency

and bank regulatory and supervisory policies that obstructs private sector monitoring,

bank activities.

Considering the determinants of bank supervisory and regulatory policies, our

findings also suggest that larger banks operating in countries with more competitive

systems tend to have relatively higher levels of efficiency. Moreover, our results are

consistent with the view that the functioning of national political systems may affect

the efficient operation of banks. Controlling for these broader, national characteristics,

can explain cross-bank differences in terms of efficiency. From a public policy

perspective, we acknowledge that these findings confront strong prejudices by

supervisory agencies and central banks that typically use the phrase “strengthen

official supervision and regulation” as synonymous with policies that promote the

safety and soundness of the financial sector. Although we do not include explicitly
stability in our analysis, our evidence provides a strong indication of the perils that

heavy-handed regulation policies can have in the efficient functioning of the banking

sector. Yet there is no doubt that the recent financial markets turmoil puts this

discussion on a new basis. The next challenge is to consider the role of political, legal

and regulatory systems in facilitating the well-functioning of banks in the post-crisis

period.
References

Altunbas, Y., Carbo, S., Gardener, E.P.M. and P. Molyneux, (2007), “Examining the

Relationship between Capital, Risk and Efficiency in European Banking”,

European Financial Management, 13(1), pp. 49-70.

Altunbas, Y., Gardener, E.P.M., Molyneux, P., and B. Moore, (2001), “Efficiency in

European banking”, European Economic Review, 45, pp. 1931-1955.

Andersen, P. and N.C. Petersen, (1993), “A Procedure for Ranking Efficient Units in

Data Envelopment Analysis”, Management Science, 39, pp. 1261-1264.

Barth, J.R., Caprio, G. and R. Levine, (2007b), “Bank Regulations are Changing: But

for Better or Worse?”, July, World Bank, available at:

http://go.worldbank.org/SNUSW978P0.

Barth, J.R., Caprio, G. and R. Levine, (2007a), “The Microeconomic Effects of

Different Approaches to Bank Supervision”, In Haber, S., North, D. and

B.Weingast, (Eds), The Politics of Financial Development, Stanford University

Press.

Barth, J.R., Caprio, G. and R. Levine, (2006), Rethinking Bank Regulation: Till

Angels Govern, Cambridge: Cambridge University Press.

Barth, J.R., Caprio, G. and R. Levine, (2004), “Bank Regulation and Supervision:

What Works Best?”, Journal of Financial Intermediation, 13, pp. 205-248.

Barth, J.R., Caprio, G. and R. Levine, (2003b), Bank Regulation and Supervision:

Lessons from a New Database, In Garza J.A.M. (Ed), Macroeconomic Stability,

Financial Markets, and Economic Development, Mexico City: Banco de

Mexico.

Barth, J.R., Caprio, G. and R. Levine, (2001b), The Regulation and Supervision of

Banks around the World: A New Database, In Litan, R.E. Herring, R., (Eds),
 Integrating Emerging Market Counties into the Global Financial System,

Brookings-Wharton Papers in Financial Services, Brooking Institution Press,

pp. 183-240.

Barth, J.R., Caprio, G. and R. Levine, (2001a), Banking Systems around the Globe:

Do Regulations and Ownership Affect Performance and Stability?, In Mishkin

F.S. (ed), Prudential Supervision: What Works and What Doesn’t, University of

Chicago Press, pp. 31-88.

Barth, J.R., Nolle, D.E, Phumiwasana, T. and G. Yago, (2003a), “Across-Country

Analysis of the Bank Supervisory Framework and Bank Performance”,

Financial Markets, Institutions and Instruments, 12, pp. 67-120.

Basel Committee on Bank Supervision, (2009), “Strengthening the Reliance of the

Banking Sector – Consultative Document”,

available at: http://www.bis.org/publ/bcbs164.htm.

Beck, T., Demirguc-Kunt, A. and R. Levine, (2006), “Bank Supervision and

Corruption in Lending”, Journal of Monetary Economics, 53, pp. 2131-2163.

Becker, G., (1983), “A Theory of Competition among Pressure Groups for Political

Influence”, Quarterly Journal of Economics, 98, pp. 371-400.

Berger, A.N. (1995), “The Profit-Structure Relationship in Banking – Tests of

Market-Power and Efficiency-Structure Hypotheses”, Journal of Money, Credit

and Banking, 27, pp. 404-431.

Berger, A.N. and E. Bonaccorsi di Patti, (2006), “Capital Structure and Firm

Performance: A New Approach to Testing Agency Theory and an Application

to the Banking Industry”, Journal of Banking and Finance, 30(4), pp. 1065-

1102.
Berger, A.N. and T. Hannan, (1998), “The Efficiency Cost of Market Power in the

Banking Industry: A Test of the ‘Quiet life’ and Related Hypotheses”, Review of

Economics and Statistics, 80(3), pp. 454-465.

Berger, A.N., Herring R.J. and G.P. Szego, (1995), “The Role of Capital in Financial

Institutions”, Journal of Banking and Finance, 19, pp. 393-430.

Berger, A.N. and D.B. Humphrey, (1997), “Efficiency of Financial Institutions:

International Survey and Directions for Further Research”, European Journal of

Operational Research, 98, pp. 175-212.

Berger, A.N., Hunter, W.C., and S.G Timme, (1993), “The efficiency of financial

institutions: A review and preview of past, present and future”, Journal of

Banking and Finance, 17, pp. 221-249.

Berger, A.N., Klapper, L.F. and R. Turk-Ariss, (2008), “Banking Structures and

Financial Stability”, Mimeo.

Berger, A.N. and L.J. Mester, (1997), “Inside the Black Box: What Explains

Differences in the Efficiencies of Financial Institutions?”, Journal of Banking

and Finance, 21, pp. 895-947.

Blum, J., (1999), “Do Bank Capital Adequacy Requirements Reduce Risks”, Journal

of Banking and Finance, 23, pp. 755-771.

Calem, P. and R. Rob, (1999), “The Impact of Capital-Based Regulation on Bank

Risk-Taking,” Journal of Financial Intermediation, 8, pp. 317-352.

Casu, B. and C. Girardone, (2006), “Bank Competition, Concentration and Efficiency

in the Single European Market”, The Manchester School, 74(4), pp. 441-468.

Casu, B., Girardone, C. and P. Molyneyx, (2006), Introduction to Banking, FT

Prentice Hall, Pearson Education, 1st Edition, Harlow, England.

Casu, B. and P. Molyneux, (2003), “A Comparative Study of Efficiency in European

Banking”, Applied Economics, 35(17), pp. 1865-1876.

Cazals, C., Florence, J.P. and L. Simar, (2002), “Non Parametric Frontier Estimation:

A Robust Approach”, Journal of Econometrics, 106, pp. 1-25.

Charnes, A., Cooper, W.W. and E. Rhodes, (1978), “Measuring the Efficiency of

Decision Making Units”, European Journal of Operational Research, 2(6), pp.

429-444.

Claessens, S. and L. Laeven, (2002), “Financial Development, Property Rights and

Growth”, CEPR Discussion Papers, No. 3295.

De Sousa, M.C. and B. Stosic, (2005), “Technical Efficiency of the Brazilian

Municipalities: Correcting Nonparametric Frontier Measurements for Outliers”,

Journal of Productivity Analysis, 24, pp. 157-181.

Demirguc-Kunt, A., Detragiache, E. and T. Tressel, (2008), “Banking on the

Principles: Compliance with Basel Core Principles and Bank Soundness”,

Journal of Financial Intermediation, 17, pp. 511-542.

Demirguc-Kunt, A., Laeven, L. and R. Levine, (2004), “Regulations, Market

Structure, Institutions, and the Cost of Financial Intermediation”, Journal of

Money, Credit and Banking, 36, pp. 593-622.

Demirguc-Kunt, A. and V. Maksimovic, (1998), “Law, Finance and Firm Growth”,

Journal of Finance, 53(6), pp. 2107-2137.

Dewatripont, M. and J. Tirole, (1993), The Prudential Regulation of Banks,

Cambridge: Cambridge University Press.

Dietsch, M. and A. Lozano-Vivas, (2000), “How the Environment Determines

Banking Efficiency: A Comparison Between French and Spanish Industries”,

Journal of Banking and Finance, 24, 985-1004.

Eisenbeis, R.A., Ferrier, G.D. and S.H. Kwan, (1999), “The Informativeness of

Stochastic Frontier and Programming Frontier Efficiency Scores: Cost

Efficiency and Other Measures of Bank Holding Company Performance”,

Federal Reserve Bank of Atlanta, Working Paper, No. 99-23.

Elyasiani, E., Mehdian, S. and R. Rezvanian, (1994), “An empirical test of

association between production and financial performance: The Case of the

Commercial Banking Industry”, Applied Financial Economics, 4, pp. 55-59.

Fernandez, A.I. and F. Gonzalez, (2005), “How Accounting and Auditing Systems

Can Counteract Risk-Shifting of Safety-Nets in Banking: Some International

Evidence”, Journal of Financial Stability, 1, pp. 466-500.

Flannery, M.J., S.H. Kwan, and M. Nimalendran, (2004), “Market Evidence on the

Opaqueness of Banking Firms’ Assets”, Journal of Financial Economics, 71(3),

pp. 419-460.

Fries, S., and A. Taci, (2005), “Cost efficiency of banks in transition: Evidence from

289 banks in 15 post-communist countries” Journal of Banking and Finance,

29, pp. 55-81

Goddard, J.A., Molyneux, P. and J.O.S. Wilson, (2001), European Banking:

Efficiency, Technology, and Growth. Chichester: John Wiley.

Gorton, G. and R. Rosen, (1995), “Corporate Control, Portfolio Choice and the

Decline of Banking”, Journal of Finance, 50, pp. 1377-1420.

Gorton, G. and A. Winton, (1995), “Bank Capital Regulation in General

Equilibrium”, Wharton Financial Institutions Centre, Working Paper.

Grigorian, D., and V. Manole, (2006), “Determinants of commercial bank

performance in transition: An application of Data Envelopment Analysis”,

Comparative Economic Studies 48, pp. 497-522.

Herring, R., (2004), “How Can the Invisible Hand Prudential Regulation? And How

Can Prudential Supervision Strengthen the Invisible Hand?”, in Borio, C.,

Hunter, W., Kaufman, G. and KJ. Tsatsaronis, eds. Market Discipline across

Countries and Industries, Cambridge, MA: MIT Press, pp. 363-380.

Herring, R., (2005), “Implementing Basel II: Is the Game Worth the Candle?”,

Financial Markets, Institutions and Instruments, 14, pp. 267-287.

Hovakimian, A, and E.J. Kane, (2000), “Effectiveness of capital regulation at U.S.

commercial banks, 1985 to 1994”, Journal of Finance, 55, pp. 451–468.

Hughes, J.P., Mester, L. and C. Moon, (2001), “Are Scale Economies in Banking

Elusive or Illusive. Evidence Obtained by Incorporating Capital Structure and

Risk-Taking into models of Bank Production”, Journal of Banking and Finance,

25(12), pp. 2169-2208.

Jeon, Y. and S.M. Miller, (2001), “Deregulation and Structural Change in the US

Commercial Banking Industry”, University of Connecticut, Department of

Economics, Working Papers, No. 7.

Jalilian, H., Kirkpatrick, C., and D. Parker, “The Impact of Regulation on Economic

Growth in Developing Countries: A Cross-Country Analysis”, World

Development, 35(1), pp. 87-103.

Jeon, Y. and S.M. Miller, (2002), “Bank Concentration and Performance”, University

of Connecticut, Department of Economics, Working Papers, No. 25.

Kaufmann, D., Kraay, A. and P. Zoido-Lobaton, (1999), “Governance Matters”,

Policy Research Working Paper, No. 2196, The World Bank.

Kaufmann, D., Kraay, A. and M. Mastruzzi, (2009), “Governance Matters VIII:

Aggregate and Individual Governance Indicators, 1996-2008”, Policy Research

Working Paper, No. 4978, The World Bank.

Kwan, S. and R. Eisenbeis, (1997), “Bank Risk, Capitalization and Operating

Efficiency”, Journal of Financial Services Research, 12(2 and 3), pp. 117-131.

Lang, G., and P. Welzel, (1998), “Technology and cost efficiency in universal

banking: A “Thick Frontier”analysis of the German banking industry” Journal

of Productivity Analysis, 10, pp. 63-84.

Llewellyn, D., (1999), “The Economic Rational for Financial Regulation”, FSA

Occasional Papers in Financial Regulation, Occasional Paper Series 1.

Lozano-Vivas, A., Pastor, J.T. and J.M. Pastor, (2002), “An Efficiency Comparison

of European Banking Systems Operating Under Different Environmental

Conditions”, Journal of Productivity Analysis, 18, pp. 59-77.

Marshall, M. and K. Jaggers, “Data Users’ Manual”, Polity IV Project, Centre for

International Development and Conflict Management, University of Maryland.

http://www.systemicpeace.org/polity/polity4.htm

McDonald, J., (2009), “Using Least Squares and Tobit in Second Stage DEA

Efficiency Analyses”, European Journal of Operational Research, 197, pp.

792-798.

Mester, L.J., (1996), “A study of Bank Efficiency taking into account Risk-

Preferences”, Journal of Banking and Finance, 20, 1025-1045.

Molyneux, P., (2003), “Regulation and Financial Innovation Trends in European

Banking and the Impact on the Supply and Demand for Financial Services in

Europe”, United Nations University, Institute for New Technologies, EIFC -

Technology and Finance Working Papers, No. 16.

Papke, L.E. and J.M. Wooldridge, (1996), “Econometric Methods for Fractional

Response Variables with an Application to 401(K) Plan Participation Rates”,

Journal of Applied Econometrics, 11, pp. 619-632.

Pasiouras, F., Tanna, S. and C. Zopounidis, (2009), “The Impact of Banking

Regulations on Banks’ Cost and Profit Efficiency: Cross-Country Evidence”,

International Review of Financial Analysis, 18, pp. 294-302.

Pasiouras, F., (2008), “International Evidence on the Impact of Regulations and

Supervision on Banks’ Technical Efficiency: An Application of two-stage Data

Envelopment Analysis”, Review of Quantitative Finance and Accounting, 30

(2), pp. 187-223.

Pasiouras, F., Gaganis, C. and C. Zopounidis, (2006), “The Impact of Bank

Regulations, Supervision, Market Structure, and Bank Characteristics on

Individual Bank Ratings: A Cross-Country Analysis”, Review of Quantitative

Finance and Accounting, 27 (4), pp. 403-438.

Rochet, J., (1992), “Capital Requirements and the Behaviour of Commercial Banks”,

European Economic Review, 36(5), pp. 1137-1170.

Schaeck, K., Cihak, M. and S. Wolfe, (2009), “Are Competitive Banking Systems More

Stable?”, Journal of Money, Credit and Banking, 41(4), pp. 711-734.

Schure, P., Wagenvoort, R. and D. O’ Brien, (2004), “The Efficiency and the

Conduct of European Banks: Developments after 1992”, Review of Financial

Economics, 13, pp. 371-396.

Sealey, C.W. and J.T. Lindley, (1977), “Inputs, Outputs and Theory of Production

and Costs at Depository Financial Institutions”, Journal of Finance, 34, pp.

1251-1266.

Seaver, B. and K. Triantis, (1992), “A Fuzzy Clustering Approach Used in Evaluating

Technical Efficiency Measures in Manufacturing”, Journal of Productivity

Analysis, 3, pp. 337-363.

Seaver, B. and K. Triantis, (1995), “The Impact of Outliers and Leverage Points for

Technical Efficiency Measurement using High Breakdown Procedures”,

Management Science, 41, pp. 937-956.

Shleifer, A., and R.W. Vishny, (1989), “Management entrenchment: The case of

manager-specific investments”, Journal of Financial Economics, 25, pp. 123-

139.

Simar, L., and P.W. Wilson, (2007), “Estimation and Inference in two-stage, semi-

parametric models of production processes”, Journal of Econometrics, 136, pp.

31-64.

Stavarek, D., (2004), “Banking efficiency in Visegrad countries before joining the

European Union” European Review of Economics and Finance, 3, pp. 129-167.

Stiglitz, J.E., (1994), “The Role of the Statein Financial Markets”, Proceedings of the

World Bank Annual Conference on Development Economics, The World Bank,

pp.19-52.

Stosic, B. and M.C. Sampaio de Sousa, (2003), “Jackstrapping DEA Scores for

Robust Efficiency Measurement”, Anais do XXV Encontro Brasileiro de

Econometria SBE, pp. 1525-1540.

VanHoose, D., (2007), “Theories of Bank Behaviour under Capital Regulation”,

Journal of Banking and Finance, 31, pp. 3680-3697.

Yildirim, H.S. and G.C. Philippatos, (2007), “Efficiency of Banks: Recent Evidence

from the Transition Economies of Europe (1993-2000)”, European Journal of

Finance,13 (2), pp. 123-143.

Appendix A:

A.1 JackStrap Methodology

DEA efficiency estimates are measured relative to estimated frontier and therefore are

sensitive to outliers, which could be due to measurement or other errors. It is therefore

imperative to identify these outliers and remove them from the sample in order to

make DEA efficiency estimate more robust. Various approaches for detecting outliers

have been proposed in the literature with some of them being based on “descriptive

methods” (see for example, Wilson, 1993; 1995) and others being based on non-

parametric methods (see for instance, super-efficiency score).

In order to deal with the sensitivity of DEA measurements to the presence of

potential outliers and data errors, especially when dealing with large datasets like the

one used here, we estimate the DEA efficiency scores by applying the recent proposed

“Jackstrap” methodology (Stosic and De Sousa, 2003). Jackstrap is based on the idea

of identifying those banks that have the highest influence on the rest of the sample. In

doing so, we re-estimate the efficiency scores and then compare the new scores with

those estimated before removing any bank from the sample, we square the differences

and then add them up altogether. This gives us the leverage of each bank in the

sample

This methodology combines Bootstrap and Jackknife re-sampling techniques,

to reduce the effect of outliers and possible errors in the dataset. The method is based

on calculating the effect of the presence of each bank in the sample on the efficiency

scores of all the other banks using:

∑ (θ
k =1, k ≠ j
*
kj −θ k )2
lj = (A.1)
K −1
 where θ κ are the original efficiency scores; θ kj* are the recalculated efficiency

scores after removing each bank one by one; and j = 1,..., K represents the removed

bank. The basic idea of this methodology is to quantify, for each bank, the effect the

removal of a bank would have on the efficiency scores of the remaining banks in the

dataset. In that way outliers and banks with errors in data can be detected and

removed from the dataset, as those are expected to exert the greatest influence of all.

Following Stosic and De Sousa (2003, 2005) the steps following this

procedure are as follows:

1. One selects randomly a subset of L banks (which is typically 10-20% of K) and

~
performs the above procedure to obtain subset leverages l j , where index j takes on L

different values from the set {1,…,K}.

~
2. Repeat the above step B times, accumulating the bubble leverage information l jb for

all randomly selected DMUs (each DMU is selected n j times, where B large enough

we should have roughly n j ≈ BL / K .

3. Calculate mean leverage for each DMU as

nj
~
~
∑
b =1
l jb
lj = (A.2)
nj

And the global mean leverage as

K
~
~
∑
j =1
lj
l = (A.3)
K
 It should be mentioned that omitting a bank and re-estimating the DEA

tantamount to a jackknife since we create a new sample from the original sample by

dropping a single observation and then estimate the statistic of interest (in this case

the “leverage score”. As far as the bootstrap part of the procedure, this is not exactly a

bootstrap but it is done to reduce the computational burden of the procedure, however,

although we acknowledge the misuse of the term, we have to be consistent with the

terminology used by the authors for the sake of consistency.

The next step regarding the procedure is how leverage information can be used

to identify potential outliers. For that purpose, we first order the DMUs according to
~ ~
their leverage values, such that li ≥ l j for i p j , and then proceed with removing

sequentially one by one the DMUs with the highest leverage, simultaneously applying
~
the threshold leverage value l0 to quantify the difference between the efficiency

distributions before and after the removal. For setting the leverage threshold value in

the case of large datasets (like the ones used in this thesis) we apply the proposed

function by Stosic and De Sousa (2003, 2005), taking into account the leverage

information and the sample size K. Therefore, for the leverage threshold value we use
~ ~
the product l0 = l log K .