The Journal of Economic Asymmetries 18 (2018) e00099

Contents lists available at ScienceDirect

The Journal of Economic Asymmetries

journal homepage: www.elsevier.com/locate/jeca

The eurozone financial crisis and bank efficiency asymmetries:

Peripheral versus core economies
Grigorios Asimakopoulos a, Georgios Chortareas b, c, *, Michail Xanthopoulos d
a
Polytechnic School, Universidad Carlos III deMadrid, 126 Madrid Street, 28903, Getafe, Madrid, Spain
b
King's Business School, King's College London, Bush House, 30 Aldwych, London, WC2 B4BG, UK
c
Department of Economics, University of Athens, 10559, Athens, Greece
d
Piraeus Bank, Athens, Greece

A R T I C L E I N F O A B S T R A C T

JEL classification: This paper assesses the impact of the financial crisis on the levels of banking efficiency within the
G21 Eurozone. We examine if the crisis had asymmetric effects on bank efficiency across different
C33 regions of the Eurozone, comparing the banks in the financially stressed European periphery to
those in the surplus economies of the core during 2005–2012. We use Data Envelopment Analysis
Keywords:
(DEA) to measure bank efficiency and the Brockett and Golany (1996) test to identify group-based
Bank efficiency
differences in efficiency. Our results indicate a gradual convergence process in efficiency between
Convergence
Financial crisis the banks of the core and periphery countries up to 2008. This process is reversed with the
Core-periphery escalation of the financial crisis from 2009 to 2012 and the pattern of bank performance becomes
Data envelopment analysis asymmetric. Moreover, our findings suggest a more benign impact of the crisis on the core banks,
Mann-Whitney rank statistic which anyway outperform peripheral banks throughout the period considered.

1. Introduction

After the establishment of the Economic and Monetary Union (EMU) in Europe the level of banking competition increased both in
terms of width (Goddard, Molyneux, & Wilson, 2001) and volume (Cetorelli, 2004) on the back of a bold pro-competitive deregulation
process. Key elements of this competition and efficiency enhancing process included the removal of entry barriers and the creation of a
level-playing-field in banking services provision (Claessen & Laeven, 2004). Along with its obvious economic advantages, however, the
process of financial integration within the EMU entailed more pronounced contagion risks, which were not fully perceptible before the
burst of the financial crisis. The euro-zone financial crisis emerged against this background with a dual nature, involving both insolvent
sovereigns and banks. An extensive empirical literature considers various aspects of the interactions between bank efficiency and
economic activity. Limited research exists, however, examining, the potentially asymmetric effects of the euro-zone crisis on bank
efficiency and performance. This paper assesses the implications of the euro-area crisis for banking efficiency on a comparative basis. We
consider whether significant asymmetries exist in the performance of banks in core and peripheral economies in the run-up and the
aftermath of the crisis. We examine if the crisis had unbalanced bank efficiency implications on the banks in the financially stressed
European periphery (including Greece, Ireland, Italy, Portugal, and Spain) as compared to those in the surplus economies of the core
(including Austria, Belgium, Germany, France, and the Netherlands). Our data spans from 2005 to 2012. We use the Data Envelopment

* Corresponding author. King's Business School, King's College London, Bush House, 30 Aldwych, London, WC2 B4BG, UK.
E-mail addresses: gasimako@ing.uc3m.es (G. Asimakopoulos), georgios.chortareas@kcl.ac.uk, gchortar@econ.uoa.gr (G. Chortareas), michxanth@yahoo.gr
(M. Xanthopoulos).

https://doi.org/10.1016/j.jeca.2018.e00099
Received 7 May 2018; Received in revised form 7 July 2018; Accepted 10 July 2018
1703-4949/© 2018 Published by Elsevier B.V.
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

Analysis (DEA) approach to obtain measures of bank efficiency and then utilize the Brockett and Golany (1996) test to identify
group-based differences in efficiency. Our results indicate a gradual convergence process in efficiency between the banks of the core and
periphery countries up to 2008. This process is reversed with the escalation of the financial crisis from 2009 to 2012. Our findings also
suggest more benign impact of the crisis on the core banks which anyway outperform peripheral banks throughout the period
considered.
The remaining of the paper is organized as follows. Section 2 reviews the relevant literature in the area of the DEA application in
measuring banking efficiency. Section 3 discusses data and methodology issues. In Section 4 the results from the empirical analysis are
presented and finally Section 5 discusses the results and presents directions for future research.

2. Background and literature review

2.1. Background

From the early years of the EMU, commercial banks engaged in rapid credit growth, taking advantage of the declining interest rates,
especially in the peripheral euro-zone economies, which experienced very low, and occasionally, negative rates. The low interest rate
environment under the single currency framework founded a false perception that credit risk is similar across the different regions,
leading some EMU sovereigns to increased levels of borrowing. The peripheral economies traditionally characterized by structural
weakness, an underlying loss of competitiveness, and external imbalances, continued enjoying an overrated credit status, attributed to
their EMU membership, and an unobstructed access to capital markets, relying on both domestic and foreign bank's appetite towards
sovereign bonds. The Euro-zone bank exposure to both private and sovereign credit risk increased in the wake of the new economic
environment. Thus, along with its obvious economic advantages, the process of financial integration within the EMU entailed more
pronounced contagion risks, which have not fully perceptible before the burst of the financial crisis.
Increasing risk premia in the sovereign debt and money markets adversely affected the cost and the composition of some euro area
banks' funding within the first decade of the euro-zone. Sovereign debt downgrades spilled over to bank's credit ratings, balance sheets
weakened due to losses on government debt holdings and collateral values declined imposing restrictions on liquidity provision from the
interbank market and the ECB gradually inhibited access to markets. Banks in Greece, Ireland and Portugal experienced severe diffi-
culties in raising wholesale debt and vanishing access to capital markets, becoming increasingly more reliant on central bank liquidity.
Risks spread to periphery banks, witnessed a large portion of their customer deposits moving to the EMU core (e.g., Germany, France,
Netherlands) banks in a flight to safety. To hinder this trend periphery banks have been compelled to raise the interest rates paid on their
deposits, a strategy which weighed on their profitability. On the credit supply side, the awareness of risks related to the euro-zone crisis
along with the demand for compliance with the Basel III regulatory framework, restricted peripheral banks' appetite for new lending.
In addition to the deepening recession, austerity programs and fiscal adjustment in the peripheral economies resulted in lower in-
vestment, lower demand for loans and, more importantly, increased loan delinquencies due to the erosion of the corporate and
households' repayment ability (IMF report, 2012). Banks in core economies experienced declining wholesale funding costs which were
efficiently transmitted to the implicit required return on their loans portfolio and were reflected on tightening margins in retail banking
activities.
The financial crisis generated distortions in the euro-area banking system and the fragmentation between the core and the periphery,
highlighted the emerging challenges for financial integration within the EMU. The interbank market was constructed on top of national
banking systems with the powers of oversight and supervision of commercial banks being assigned to the national central banks. In the
absence of uniform and transnational resolution mechanisms, governments had to back commercial banks ‘creditworthiness against
default risk and to extend guarantees on their national banks. The last, in turn, started providing liquidity to banks with non-eligible
assets for ECB refinancing operations. Thus, while the euro-area banks used to operate in a highly intergraded interbank market
their monitoring, creditworthiness, and solvency remained highly country-specific and asymmetric across the eurozone.

2.2. Literature review

In addition to academic concerns, the evaluation of bank performance has direct and practical implications for various stakeholders
in the real economy, including investors, bank managers, and regulators. An extensive literature explores the bank incentives for
reducing operating costs, utilizing resources more efficiently, and improving management to achieve competitiveness gains (e.g. Hsieh,
Shen, & Lee, 2010; Miguel-Davila, Cabeza-García, Valdunciel, & Fl orez, 2010). Two approaches dominate the literature on bank per-
formance, namely the financial ratio approach and the frontier approach. The financial ratio approach is based on the evaluation of
financial indices measuring profitability, liquidity and credit quality. One of its key advantages is the ability to effectively distinguish the
outperforming banks while providing sufficient controls for size effects (Samad, 2004). Nevertheless, the financial ratio approach suffer
from the lack of consensus, at least among academics, about the most representative combination of financial ratios and their respective
weightings in the analysis of bank efficiency (Yang, 2012).
The efficient frontier approach is based on both statistical and mathematical programming techniques which produce efficiency
frontiers and assess bank performance by assigning relative efficiency scores, with the higher score indicating always higher relative
performance. Bauer, Berger, Ferrier, and Humphrey (1998) provide a comprehensive review of frontier approaches including the Data
Envelopment Analysis (DEA), the Stochastic Frontier Analysis (SFA), the Distribution Free Analysis (DFA) and the Thick Frontier
Analysis (TFA). They apply all competing frontier approach to a data set of 683 US banks for the period from 1977 to 1988, and conclude
that there is no single approach dominates is allowing to define an efficient frontier. Instead, each technique seems to react to varying

2
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

characteristics of the data, highlighting the importance of methodological cross-checks of outcomes derived by alternative methods.
Bauer et al. (1998) is the only cross-checking study that quantifies differences in large and heterogeneous samples consisting of dis-
similar financial institutions. More recent studies, (e.g., Casu & Girardone, 2002) find higher consistency in the efficiency estimates
derived with the application of DEA and DFA. Another typical finding in the literature is that SFA produces higher efficiency scores as
compared to DEA (e.g., Beccalli, Casu, & Girardone, 2006) and that small samples of banks with similar characteristics tend to un-
derestimate differences between DEA and SFA. Weill (2004), however, finds that efficiency scores do not differ significantly across
techniques and display positive correlation between SFA and DFA.
DEA studies can follow either an input-oriented or output-oriented approach. The former quantifies the proportion of the inputs that
can be reduced without having output losses while the latter quantifies how much can the output increase by keeping inputs constant.
Fethi and Pasiouras (2010) provide a comprehensive review of 196 studies assessing efficiency in the financial sector. The survey
identifies DEA as the most commonly used technique in examining bank efficiency. DEA can be related with different measures of
efficiency depending on the orientation of the modeling approach. According to Fethi and Pasiouras (2010), the majority of studies
apply the input-oriented approach probably because of the assumption that bank managers control inputs rather than outputs. Other
studies use price information for the input and output variables and approach the efficiency measures in terms of cost efficiency (Isik &
Hassan, 2002; Tortosa-Ausina, 2002) and/or profit efficiency, which is limited in the literature due to the lack of reliable output price
data (Kirkwood & Nahm, 2006). The major advantage of DEA is that it functions sufficiently well with small samples and it does not
require assumptions about the distribution of inefficiency. Nevertheless, the DEA is not immune to outliers and tends to produce
misleading and biased results when there are few observations and many heterogeneous input/output variables (Rao, O'Donnell, &
Battese, 2005) while assuming that the data is free of measurement error.
A substantial academic literature considers the factors that impact on bank efficiency focusing on particular bank-specific charac-
teristics such as size, profitability, capital adequacy, ownership structure, and risk profile (Ataullah & Le, 2006; Casu & Girardone, 2004;
Casu & Molyneux, 2003). Other studies examine the impact of country specific factors such as the market concentration and interna-
tionalization, the GDP growth and the public deficit ratio (Ataullah & Le, 2006; Hauner, 2005). The empirical evidence is mixed in many
cases, however. Studies focusing on bank size produce evidence indicating that larger banks are more efficient (Grigorian & Manole,
2006) while other studies suggest that no link exists between size and cost efficiency (Cebenoyan, Cooperman, Register, & Hudgins,
1993). A number of studies find that larger banks are less profit-efficient (Nikiel & Opiela, 2008) but more cost-efficient (Kaparakis,
Miller, & Noulas, 1994). The evidence is mixed as well on ownership structure, with papers suggesting that private banks display su-
perior efficiency when compared to government banks (Fries & Taci, 2005; Fu & Heffernan, 2007) and papers refuting this result by
producing evidence that state-owned banks do not underperform their private competitors (Altunbas at al., 2001; Denizer, Dinc, &
Tarimicilar, 2000). Several studies explore the effects of deregulation process in financial markets finding that it enhances bank effi-
ciency (Canhoto & Dermine, 2003; Casu & Molyneux, 2003; Lin & Zhang, 2009) while other authors find that inhibits bank efficiency
(Denizer at al., 2007; Grifell-Tatje & Lovell, 1996). Finally, a number of papers do not find significant links between deregulation and
bank efficiency (Hao, Hunter, & Yang, 2001; Havrylchyk, 2006).
An issue of increasing importance, especially in the aftermath of the financial crisis, is the relationship between bank efficiency and
managerial risk-taking. Evidence is again mixed with Berger, Leusner, and Mingo (1997) and Kwan and Eisenbeis (1997) suggesting that
poorly performing US banks are more sensitive to excessive risk-taking and rising nonperforming loans while Sturm and Williams (2004)
show that mismanagement tends to generate loan portfolios of poor quality and performance. Fiordelisi, Marques-Ibanez, and Molyneux
(2011) provide evidence suggesting that underperformance abets higher managerial risk taking in the near term. In contrast,
cross-country studies focusing on European data indicate that more efficient banks tend to have higher loan-to-asset ratios (Maudos,
Pastor, Perez, & Quesada, 2002) or loan-loss reserves (Altunbas, Carbo, Gardener, & Molyneux, 2007). Other studies incorporate risk
measures such as nonperforming loans (Berg, Førsund, & Jansen, 1992), equity capital (Altunbas, Evans, & Molyneux, 2001) or pro-
visions on loans losses (Pastor & Serrano, 2005) to control for risk. Finally, another group of studies (Carvallo & Kasman, 2005;
Koutsomanoli-Filippaki & Mamatzakis, 2011; Semih Yildirim & Philippatos, 2007) analyze risk as an external factor that impacts on
bank efficiency, applying two-stage DEA or SFA models.
DEA studies have been applied almost in all the banking industries around the world and only few studies provide cross-country
comparison (Beccalli et al., 2006; Casu & Molyneux, 2003; Lozano-Vivas, Pastor, & Pastor, 2002). The empirical research conducted
on the European banking industry is relatively very small in comparison to the balk of studies on the US banking industry (Berger &
Humphrey, 1997; Mitchell & Onvural, 1996). However, the rapid economic developments and the challenging structure of the Eurozone
have attracted increasing interest on the European banking industry. The clear majority of studies in Europe is country-specific and
concerns mainly banks from Germany, France, the UK and Spain while there has been increasing interest for the transition economies.
The existing cross-country comparisons do not provide compellingly robust evidence of superior bank efficiency in any particular
country. There are trends in the empirical research, however, indicating that German banks are relatively more efficient (Allen & Rai,
1996; Altunbas et al., 2001; Casu & Molyneux, 2003; Cavallo & Rossi, 2002), followed by Austrian banks (Allen & Rai, 1996; Altunbas
et al., 2001). The worst performers are the banks from the UK (Allen & Rai, 1996; Altunbas et al., 2001; Cavallo & Rossi, 2002), followed
by the French banks (Allen & Rai, 1996; Lozano-Vivas et al., 2002; Weill, 2004) and the Italian banks (Allen & Rai, 1996; Casu &
Molyneux, 2003; Lozano-Vivas et al., 2002).
In this study we perform a cross-country evaluation of bank efficiency with respect to the fragmentation of the banking industry after
the eruption of the financial crisis, which split the euro area into two groups: the first group consists of financially stressed countries
whose solvency has been questioned by the recent crisis, and adopted painful adjustment measures; the second group consists of surplus
countries with top-rated credit outlook and a “safe haven” status ascribed by the financial markets.

3
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

3. Data and methodology

3.1. Data sources and description

The dataset is covers commercial banks from seventeen Eurozone countries and has been acquired from the Bankscope database
(Bureau van Dijk). Moreover, it is unbalanced and covers an eight-year period from 2005 to 2012 to capture the financial crisis from its
very beginning up to its full swing. We exclude from the sample investment banks, national central banks, and other type of banks which
are not involved in retail banking activities. We also exclude from the sample banks which display missing values and inconsistent
observations. We treat banks which were engaged in mergers or acquisitions as separate operational units only for the period prior to the
merger or acquisition event.
A typical disadvantage of frontier models and specifically the DEA is its sensitivity to outliers (Fethi & Pasiouras, 2010). Thus we use a
test introduced by Wilson (1993) to detect the presence of outliers based on Wilson. The method is particularly useful when using frontier
models to measure efficiency and distortions from individual or groups of observations with extreme values may emerge. Applying the
Wilson (1993) routine, allows to exclude a small group of banks from the initial dataset and the final number of banks included in the
sample, categorized by year and type of group, is presented in Table 1. We split the banks into three groups based on the degree of
involvement in the debt crisis of the origin country. This is done in order to investigate the relative efficiency differences between groups
of banks that operate in different countries based on our introductory narration which suggests that the financial crisis had asymmetric
effects and generated distortions and fragmentation in a monetary union which was thought to have achieved a high degree of capital
market integration (as evidenced by declining country risk premia in the early millennium). The first group consists of banks from
Austria, Belgium, Germany, France and the Netherlands, the so called “safe havens” in the financial market's perception during the period
of the crisis. The second group includes banks from Greece, Ireland, Italy, Portugal and Spain, the so-called PIIGS which were in the eye of
the cyclone during the financial turmoil. Finally, banks operating in the rest of the euro area are categorized as a third group. In total, our
sample includes 3519 observations with an average number of around 440 banks per year. Banks from the core and the periphery
dominate the entire sample, displaying an average number of 243 and 147 banks per year respectively. Table 1 presents a breakdown of
the dataset with the number of banks per year and group, along with additional information for the size of the banks included in the
analysis. The bank assets mean varies from 36.272 to 88.401 million euros with an average of 57.371 million euros.

3.2. Methodology

To measure I study bank efficiency we use DEA which is a non-parametric technique based on that relies on mathematical pro-
gramming to estimate the relative efficiency of decision making units (DMU) by determining a production frontier which is made up by
the most efficient of them. DEA's major advantage is that it does not require dealing with assumptions on the distribution of the variables
included as inputs and outputs. Furthermore, DEA allows for the inclusion of inputs and outputs which are not supposed to have a pre-
specified relationship and the measurement of inputs and outputs could be in different units.
The model of Charnes, Cooper, and Rhodes (1978), despite the numerous modifications over the years, remains the most widely
applied DEA model in banking. The model of Charnes at al. (1978) assumes that the decision-making units operate under constant returns
to scale (CRS). A key limitation of this model is the assumption that all firms operate at the optimal scale (Banker, Charnes, & Cooper,
1984), which is somewhat strong, especially in regulated industries such as the banking industry. Banker et al. (1984) develop a model
which allows for variable returns to scale (VRS). That is, namely firms can exhibit constant, increasing or decreasing returns to scale.
The use of DEA permits identifying the most efficient units operating on the frontier and constitute the benchmark for the relatively
inefficient units which are enclosed by the envelope that make up the efficient firms. The inefficient units should adjust their inputs or
outputs, depending on the model orientation, in order to improve their operational efficiency. DEA orientation depends on the modeler's
choice and on the inputs and outputs used. Output-oriented DEA maximizes output for a given level of the inputs used, while input-oriented
DEA minimize inputs for a given level of outputs. The initial input-oriented CRS model (Cooper, Seiford, & Tone, 2007) is given by:

minθ;λ θ
s:t:

yi þ Yλ  0
θxi
Xλ  0
λ0

Table 1
Descriptive statistics of the Euro-zone banking industry.

Number of banks

Country/Region 2005 2006 2007 2008 2009 2010 2011 2012

Core economies 207 256 263 270 269 288 284 111
Peripheral economies 157 169 159 163 160 153 146 74
Rest 42 45 51 56 56 57 55 28
Total 406 470 473 489 485 498 485 213
Assets Mean 36272.43 61345.01 66776.02 57077.20 55522.91 45052.99 48525.90 88401.72

Notes: Core economies include Austria, Belgium, Germany, France, and the Netherlands. Peripheral economies include Portugal, Ireland, Italy, Greece, and Spain. Assets
mean in millions.

4
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

where Y is a M  N matrix of output quantities and X is a K  N matrix of input quantities assuming that the firms uses M outputs and K
inputs, yi is a M  1 vector of output quantities for the i-th bank, xi is a K  1 vector of input quantities for the i-th bank, λ is a N  1
vector of constants and θ is the efficiency score of the bank i. The efficiency score obtained varies between 0 and 1. The solution of this
linear programming problem for each bank provides the efficiency score, with θ ¼ 1 indicating that it lies on the frontier and thus is
perfectly efficient, while firms with θ < 1 exhibit inefficiencies compared with the firms on the frontier. The DEA input-oriented VRS
linear programming problem is the same as the one under CRS but with the difference of one additional constraint that the lambda
P
weights should add to one, λ ¼ 1, that is:

minθ;λ θ
s:t:

yi þ Yλ  0
θxi
Xλ  0
λ0
X
λ¼1

As suggested above, a significant number of developments have been proposed based on DEA. The present study focuses on
comparing differences in efficiency among groups of banks depending on the country they belong to. To do so we use a methodology
developed by Brockett and Golany (1996) and allows determining group-based rather than individual efficiency differences after
eliminating managerial inefficiency differences. In other words, the routine proposed basically determines if there exist programmatic
efficiency differences between the two groups and tests if the two groups of banks share the same efficiency frontier.
The Brockett and Golany (1996) routine consists of the following steps:

I. Split all banks (i ¼ 1, …n) into two groups of banks n1 and n2 such as n1þn2 ¼ n and run DEA separately for each group.
II. In each of the two groups, adjust inefficient banks to their “level if efficient” by projecting each bank onto the efficient frontier of
its group.
III. Run a pooled DEA with all n banks at their adjusted efficient levels.

Table 2
Bank inputs and outputs descriptive statistics.

Year Inputs Outputs

Personnel expenses Fixed assets Total customer deposits Loans Other Earning Assets

2005 Mean 260.02 259.36 11300.00 16900.00 17400.00

Std. dev. 571.30 542.79 25300.00 36600.00 47700.00
Min 0.10 0.10 0.10 0.20 2.42
Max 5638.00 4812.00 193000.00 305000.00 360000.00
2006 Mean 410.99 444.86 19000.00 27200.00 30700.00
Std. dev. 1134.44 1327.41 54100.00 66500.00 97100.00
Min 0.10 0.10 0.10 0.10 0.40
Max 10300.00 12500.00 497000.00 523000.00 946000.00
2007 Mean 414.44 434.23 20200.00 29800.00 32900.00
Std. dev. 1184.01 1354.19 57800.00 74400.00 107000.00
Min 0.10 0.10 0.87 0.20 2.57
Max 11100.00 13200.00 528000.00 574000.00 1150000.00
2008 Mean 343.19 337.07 16900.00 28000.00 25500.00
Std. dev. 917.32 909.56 46400.00 70300.00 72900.00
Min 0.20 0.10 0.50 0.10 0.20
Max 10000.00 9260.50 523000.00 622000.00 609000.00
2009 Mean 347.76 348.37 18100.00 28200.00 23600.00
Std. dev. 927.52 1005.93 51500.00 72400.00 65400.00
Min 0.30 0.10 0.20 0.20 0.30
Max 9344.50 12100.00 478000.00 669000.00 502000.00
2010 Mean 283.89 259.22 14100.00 22400.00 20100.00
Std. dev. 643.20 643.54 31900.00 49700.00 56700.00
Min 0.10 0.05 0.40 0.10 0.40
Max 5665.00 6006.00 299000.00 456000.00 482000.00
2011 Mean 310.14 279.81 15000.00 23500.00 21500.00
Std. dev. 785.25 753.81 35700.00 53400.00 63200.00
Min 0.02 0.01 0.10 0.50 0.90
Max 9401.70 8635.20 356000.00 556000.00 609000.00
2012 Mean 535.57 489.92 26400.00 40500.00 40400.00
Std. dev. 1072.02 1033.61 50300.00 71400.00 87900.00
Min 2.99 0.09 0.10 13.00 33.81
Max 9137.80 8469.50 369000.00 547000.00 697000.00

Notes: values in millions.

5
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

IV. Apply a statistical test to the results of III to determine if the two groups of banks have the same distribution of efficiency values
within the pooled DEA. The nonparametric statistical test suggested by Brockett and Golany (1996) is the Mann-Whitney rank
test which is the one adopted in the present study.

The bulk of studies that assess bank efficiency seem to be dominated by two main underlying theoretical approaches (Berger &
Humphrey, 1997). The first is the production approach which assumes that banks use a number of production factors such as labor and
capital to produce loans and deposits, and the second is the intermediation approach which perceives banks as intermediaries between
savers and investors. Our analysis adopts the intermediation approach which appears as the most appropriate for evaluating financial
institutions as a whole (Fethi & Pasiouras, 2010). In line with this approach, personnel expenses, fixed assets and total customer deposits
(Maudos et al., 2002, Casu & Molyneux, 2003, Havrylchyk, 2006, Chortareas, Girardone, & Ventouri, 2013) are used as inputs, and loans
and other earning assets are used as outputs (Casu & Girardone, 2004, 2006; Chortareas et al., 2013). The descriptive statistics for each
year of the aforementioned inputs and outputs used in the DEA model are presented in Table 2. The statistics shown refer to the entire
number of banks of the seventeen Eurozone countries during the period considered and all the variables are expressed in million euros.

4. Empirical results

4.1. Bank efficiency in the euro-zone

Table 3 presents the results obtained from the input-oriented DEA with variable returns to scale for the whole dataset. The figures in
Table 3 represent the average scores by country and year and provide a straightforward characterization of the efficiency distribution.
The results indicate an average efficiency score of around 0.19 for the banks in Belgium, Estonia, Finland, France, Germany, Ireland,
Malta and Netherlands, which is above the sample's average efficiency score during the whole period. It should be noted that Irish banks
consistently outperform during the entire period, displaying an efficiency score which varies from 0.42 to 0.56, and they are followed by
banks from Estonia and Malta. The rest of the banks belonging mainly to peripheral economies underperform, showing an average
efficiency score which lies below the sample's average. Specifically, Cyprus, Greece and Slovakia exhibit the worst efficiency scores with
a maximum of 0.08 and a minimum of 0.03, both of them observed in Greece. The time trend of the efficiency scores shown in Table 3,
indicates that when comparing the first and the last year of the dataset Finland, Ireland, Italy, Germany and Spain show major im-
provements. In contrast, Estonia, Malta and Netherlands exhibit a decrease in their average efficiency scores between 2005 and 2012. In
the remaining countries, banks show relatively stable efficiency scores.
As far as individual banks efficiency scores are concerned, the majority of the banks on the frontier for each year are French, German,
Dutch, Italian, and Irish.1 On the other extreme, the banks with the lowest efficiency scores are located mainly in Italy, Germany,
Slovenia, Slovakia, Spain, Netherlands, Austria, and Portugal. Interestingly, the number of inefficient banks increases dramatically after
2008. The next section compares the banks efficiency estimates of core and peripheral economies within the Euro-zone.

4.2. Core versus peripheral economies

The current section explores efficiency differences between the core and the peripheral banks in the Eurozone. We apply the Brockett

Table 3
DEA efficiency scores: Euro zone 17.

Country 2005 2006 2007 2008 2009 2010 2011 2012

AUSTRIA 0.1709851 0.1207232 0.1228215 0.1328533 0.1464162 0.1844682 0.1444263 0.1967809

BELGIUM 0.2750921 0.2741274 0.2249139 0.1155252 0.1484074 0.1624486 0.1433912 0.2954854
CYPRUS 0.0480403 0.0552909 0.0576327 0.0472361 0.043948 0.0700322 0.0438589 0.0496865
ESTONIA 0.5392558 0.3284877 0.3048622 0.2253759 0.2606166 0.324433 0.0560678 0.2756997
FINLAND 0.1114113 0.2068267 0.1706116 0.1333238 0.1574786 0.2202202 0.283566 0.3629702
FRANCE 0.2536211 0.2642247 0.262116 0.1893614 0.1725034 0.2186856 0.2046917 0.2888006
GERMANY 0.2794085 0.2871903 0.2630715 0.195666 0.1936216 0.2351331 0.2421581 0.3641863
GREECE 0.066674 0.0493352 0.0544511 0.0359215 0.0331528 0.0668109 0.0676845 0.0862702
IRELAND 0.4522239 0.4292534 0.4478031 0.4392759 0.4636524 0.4951314 0.5655242 0.5692769
ITALY 0.1647386 0.2020864 0.1856734 0.1899575 0.1644529 0.1901969 0.1693467 0.2357441
LUXEMBOURG 0.0856365 0.0603883 0.1253982 0.1356652 0.0770247 0.1117112 0.1475123 0.10354
MALTA 0.3725364 0.3767733 0.3726365 0.3138545 0.2046579 0.2143648 0.1754017 0.2480966
NETHERLANDS 0.3174092 0.3739111 0.3036431 0.1926449 0.1933633 0.1600354 0.1519132 0.2354244
PORTUGAL 0.0861341 0.090944 0.0921866 0.0632884 0.055983 0.1133791 0.0744419 0.0950536
SLOVAKIA 0.0592027 0.0724382 0.0699775 0.0427495 0.046291 0.0716145 0.0350817 0.0705635
SLOVENIA 0.1232025 0.15797 0.0914992 0.0870197 0.0533809 0.1051034 0.0713703 0.0896484
SPAIN 0.1117838 0.1518791 0.1609229 0.1011029 0.1011174 0.124425 0.1449588 0.2244588

Notes: Average efficiency based on output oriented DEA model with variable returns to scale (VRS).

1
Detailed results are available by the authors. For the external imbalances within the Eurozone see Arghyrou and Chortareas (2008).

6
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

and Golany (1996) routine to identify programmatic efficiency differences between the two groups. The DEA modeling options are the
same as in the first model, namely input-oriented with VRS. The efficiency scores for each country and the medians by group of countries
for each year are shown in Table 4.
In general, the group of the core banks performs better than the group of the peripheral banks, except for the banks in Ireland which
constantly outperform every single country in the two groups. Moreover, we can observe a general declining tendency of the average
efficiency score in the core countries from 2005 to 2009 a gradual stabilization after 2009 and an increasing trend after 2011. Country-
wise, German and Dutch banks display the highest efficiency scores and Austrian banks exhibit the lowest efficiency scores in the core
banks group. In the group of the peripheral economies, the Irish banks outperform consistently followed by the Italian banks, while the
Greek and Portuguese banks appear to be the worst performers in both the two groups.
Fig. 1 shows the evolution of the median efficiency scores by group of countries and their respective difference. It emerges that bank
efficiency in the periphery follows a slightly upward trend from 2005 to 2007, thereafter a decline until 2009, and finally an
improvement after 2010. A similar pattern is observed in the group of the core banks where 2009 is a more obvious turning point in
terms of median efficiency score. In 2012, bank efficiency improves with both groups displaying a parallel pattern. Nevertheless, pe-
ripheral banks are lagging again behind their core competitors. Moreover, we detect contradictory trends even within the same group
with some countries (e.g., Germany, Spain, Ireland and Greece) improving their efficiency scores, others (e.g., the Netherlands)
recording a declining efficiency score after the onset of the crisis, and others (France, Italy and Portugal) exhibiting unclear patterns with
ups-and-downs.
In order to analyze programmatic efficiency differences between the core and the peripheral countries we need to use a non-
parametric statistic since the DEA efficiency scores have unknown distribution (Brockett & Golany, 1996). Table 5 shows the z-score
statistics of the Wilcoxon-Mann-Whitney rank test. Based on the results, the “null hypothesis”, suggesting that there is no difference in
the distribution of the bank efficiency scores between the two groups of countries, can be rejected. This result is valid for all years in the
sample except for 2008, where the null hypothesis is marginally rejected at a 0.05 level. This finding suggests that there was a significant
programmatic difference in the banking efficiency scores between the core and the peripheral banks for all the years except for 2008. In
other words, the Eurozone financial crisis, which started in 2009 and escalated dramatically in 2012, interrupted a convergence process
and restored the programmatic efficiency differences between the core and the periphery.

5. Conclusion

The paper first analyses the evolution of asymmetries in the efficiency of the European banking sector around the eurozone financial
crisis. We apply the DEA approach to calculate efficiency and obtain the data from the Bankscope database for the years from 2005 to
2012. An additional objective of the paper is to evaluate whether there exist asymmetries between the banks of particular groups of
countries, namely, core and peripheral economies. To answer this question, we use the Brockett and Golany (1996) procedure which
allows comparing programs within the DEA efficiency framework. The results show that banks in the EMU core economies consistently
outperform peripheral banks throughout the period examined. Moreover, we find that the financial crisis interrupted a convergence
process between the two groups after 2009. This is in line with the finding of Tsionas, Assaf, and Matousek (2015) who show that most
banks experienced declining efficiency levels during financial crisis. Efficiency asymmetries between the core and the peripheral banks
reach their minimum level in 2008, just before the outburst of the EU crisis and thereafter increases again. In terms of individual country
performance, our results reveal considerable efficiency asymmetries between fully intergraded Euro zone members.
The results suggest that core banks have benefited from the crisis more than the peripheral banks, possibly due to their perceived role
as “safe havens”. In contrast to previous research our findings indicate that the financial crisis caused asymmetric, in terms of magnitude
and location, distortions on the quality of the bank inputs and outputs and thus on bank efficiency. The results also indicate that, in many
cases, efficiency improved after the burst of the crisis, suggesting positive effects from shifts in risk preferences, declining operational
and funding costs, extensive monitoring and regulation in the post crisis period. One should note, however, that interpreting efficiency
scores can be misleading when one does not control for possible risk-efficiency trade-offs for banks which displaying very high efficiency

Table 4
DEA efficiency scores: Core and Peripheral Euro-area economies.

Country 2005 2006 2007 2008 2009 2010 2011 2012

Core AUSTRIA 0.221469 0.154262 0.1718126 0.1999112 0.1926872 0.2024106 0.1540509 0.2250295
BELGIUM 0.3537928 0.3208427 0.2692009 0.184099 0.2022578 0.1830705 0.1545225 0.3027313
FRANCE 0.2676463 0.2741881 0.2733513 0.2251001 0.1964481 0.2291301 0.2096694 0.3101304
GERMANY 0.3008995 0.2986691 0.2783645 0.23314 0.2210001 0.2466318 0.2529374 0.3740644
NETHERLANDS 0.345414 0.409963 0.3463956 0.2487865 0.2417771 0.1865702 0.1645586 0.2528768

Median 0.1361309 0.1301378 0.1354147 0.1040993 0.0876356 0.1265052 0.1032842 0.1660949

Peripheral GREECE 0.0843373 0.0653428 0.0702625 0.0638643 0.0551281 0.0742787 0.0731756 0.0977155
IRELAND 0.4532008 0.4363485 0.4536569 0.4602561 0.4816729 0.5057359 0.5672894 0.5772429
ITALY 0.1745154 0.2130238 0.1948724 0.2243465 0.1917395 0.1969349 0.1730783 0.2462452
PORTUGAL 0.1261769 0.1280192 0.1298771 0.1188014 0.0998112 0.1257249 0.078247 0.1112161
SPAIN 0.1380134 0.1758119 0.1883399 0.1571299 0.1332724 0.1335844 0.1499284 0.2317046

Median 0.0817659 0.0868384 0.0990397 0.0870346 0.0679949 0.0941798 0.0853209 0.1179138

Notes: Average efficiency based on output oriented DEA model with variable returns to scale (VRS).

7
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

Fig. 1. Median efficiency scores by group of countries and their difference.

Table 5
Efficiency difference test between banks in Core and Peripheral Euro-area economies.

2005 2006 2007 2008 2009 2010 2011 2012

Wilcoxon–Mann–Whitney test Z ¼ 5.0635 Z ¼ 3.756 Z ¼ 2.8394 Z ¼ 1.8978 Z ¼ 3.0666 Z ¼ 2.7659 Z ¼ 2.2404 Z ¼ 2.4782
p-value 0.0000 0.0001 0.0044 0.0576 0.0021 0.0055 0.0249 0.0129

scores (for instance in Ireland). Finally, we find that in many cases efficiency improved after the burst of the crisis, possibly due to shifts
in risk preferences, declining operational and funding costs, enhanced monitoring, and redesigned regulation in the post crisis period.
One should note, however, that interpreting efficiency scores can be misleading when one does not control for possible risk-
efficiency trade-offs for banks that display very high efficiency scores (for instance in Ireland). Future work should try to identify
possible determinants of efficiency controlling for the above factors by including a second stage regression analysis. Additionally, ac-
cording to Aiello and Bonanno (2018) using the intermediation approach, as in the current paper, yields possibly lower efficiency scores
and this should be taken into account when interpreting the results.

Appendix A. Supplementary data

Supplementary data related to this article can be found at https://doi.org/10.1016/j.jeca.2018.e00099.

References

Aiello, F., & Bonanno, G. (2018). On the sources of heterogeneity in banking efficiency literature. Journal of Economic Surveys, 32(1), 194–225.
Allen, L., & Rai, A. (1996). Operational efficiency in banking: An international comparison. Journal of Banking and Finance, 20(4), 655–672.
Altunbas, Y., Carbo, S., Gardener, E. P., & Molyneux, P. (2007). Examining the relationships between capital, risk and efficiency in European banking. European
Financial Management, 13(1), 49–70.
Altunbas, Y., Evans, L., & Molyneux, P. (2001). Bank ownership and efficiency. Journal of Money, Credit and Banking, 926–954.
Arghyrou, M. G., & Chortareas, G. (2008). Current account imbalances and real exchange rates in the Euro area. Review of International Economics, 16(4), 747–764.
Ataullah, A., & Le, H. (2006). Economic reforms and Bank efficiency in developing countries: The case of the indian banking industry. Applied Financial Economics,
16(9), 653–663.
Banker, R. D., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale inefficiencies in data envelopment analysis. Management Science,
30(9), 1078–1092.
Bauer, P. W., Berger, A. N., Ferrier, G. D., & Humphrey, D. B. (1998). Consistency conditions for regulatory analysis of financial institutions: A comparison of frontier
efficiency methods. Journal of Economics and Business, 50(2), 85–114.
Beccalli, E., Casu, B., & Girardone, C. (2006). Efficiency and stock performance in European banking. Journal of Business Finance and Accounting, 33(1–2), 245–262.
Berger, A. N., & Humphrey, D. B. (1997). Efficiency of financial institutions: International survey and directions for future research. European Journal of Operational
Research, 98(2), 175–212.
Berger, A. N., Leusner, J. H., & Mingo, J. J. (1997). The efficiency of bank branches. Journal of Monetary Economics, 40(1), 141–162.
Berg, S. A., Førsund, F. R., & Jansen, E. S. (1992). Malmquist indices of productivity growth during the deregulation of Norwegian banking, 1980-89. The Scandinavian
Journal of Economics, S211–S228.
Brockett, P. L., & Golany, B. (1996). Using rank statistics for determining programmatic efficiency differences in data envelopment analysis. Management Science, 42(3),
466–472.
Canhoto, A., & Dermine, J. (2003). A note on banking efficiency in Portugal, New vs. Old banks. Journal of Banking and Finance, 27(11), 2087–2098.
Carvallo, O., & Kasman, A. (2005). Cost efficiency in the Latin American and Caribbean banking systems. Journal of International Financial Markets, Institutions and
Money, 15(1), 55–72.
Casu, B., & Girardone, C. (2002). A comparative study of the cost efficiency of Italian bank conglomerates. Managerial Finance, 28(9), 3–23.
Casu, B., & Girardone, C. (2004). Financial conglomeration: Efficiency, productivity and strategic drive. Applied Financial Economics, 14(10), 687–696.
Casu, B., & Girardone, C. (2006). Bank competition, concentration and efficiency in the single European Market. The Manchester School, 74(4), 441–468.
Casu, B., & Molyneux, P. (2003). A comparative study of efficiency in European banking. Applied Economics, 35(17), 1865–1876.
Cavallo, L., & Rossi, S. P. (2002). Do environmental variables affect the performance and technical efficiency of the European banking systems? A parametric analysis
using the stochastic frontier approach. The European Journal of Finance, 8(1), 123–146.

8
G. Asimakopoulos et al. The Journal of Economic Asymmetries 18 (2018) e00099

Cebenoyan, A. S., Cooperman, E. S., Register, C. A., & Hudgins, S. C. (1993). The relative efficiency of stock versus mutual S&Ls: A stochastic cost frontier approach.
Journal of Financial Services Research, 7(2), 151–170.
Cetorelli, N. (2004). Real effects of bank competition. Journal of Money, Credit, and Banking, 36(3), 543–558.
Charnes, A., Cooper, W. W., & Rhodes, E. (1978). Measuring the efficiency of decision making units. European Journal of Operational Research, 2(6), 429–444.
Chortareas, G. E., Girardone, C., & Ventouri, A. (2013). Financial freedom and Bank efficiency: Evidence from the European union. Journal of Banking and Finance,
37(4), 1223–1231.
Claessen, S., & Laeven, L. (2004). What drives bank competition? Some international evidence. Journal of Money, Credit and Banking, 563–583.
Cooper, W. W., Seiford, L. M., & Tone, K. (2007). Data envelopment analysis: Acomprehensive text with models, applications, references and DEA-solver software. Springer.
Denizer, C. A., Dinc, M., & Tarimcilar, M. (2007). Financial liberalization and banking efficiency: Evidence from Turkey. Journal of Productivity Analysis, 27(3),
177–195.
Denizer, C. A., Dinc, M., & Tarimicilar, M. (2000). Measuring banking efficiency in the pre-and post-liberalization environment: Evidence from the Turkish banking system (Vol
2476). World Bank Publications.
Fethi, M. D., & Pasiouras, F. (2010). Assessing Bank efficiency and performance with operational research and artificial intelligence techniques: A survey. European
Journal of Operational Research, 204(2), 189–198.
Fiordelisi, F., Marques-Ibanez, D., & Molyneux, P. (2011). Efficiency and risk in European banking. Journal of Banking and Finance, 35(5), 1315–1326.
Fries, S., & Taci, A. (2005). Cost efficiency of banks in transition: Evidence from 289 banks in 15 post-communist countries. Journal of Banking and Finance, 29(1),
55–81.
Fu, X. M., & Heffernan, S. (2007). Cost X-efficiency in China's banking sector. China Economic Review, 18(1), 35–53.
Goddard, J. A., Molyneux, P., & Wilson, J. O. (2001). European Banking: Efficiency, technology, and growth. John Wiley.
Grifell-Tatje, E., & Lovell, C. K. (1996). Deregulation and productivity decline: The case of Spanish savings banks. European Economic Review, 40(6), 1281–1303.
Grigorian, D. A., & Manole, V. (2006). Determinants of commercial bank performance in transition: An application of data envelopment analysis. Comparative Economic
Studies, 48(3), 497–522.
Hao, J., Hunter, W. C., & Yang, W. K. (2001). Deregulation and efficiency: The case of private Korean banks. Journal of Economics and Business, 53(2), 237–254.
Hauner, D. (2005). Explaining efficiency differences among large German and Austrian banks. Applied Economics, 37(9), 969–980.
Havrylchyk, O. (2006). Efficiency of the Polish banking industry: Foreign versus domestic banks. Journal of Banking and Finance, 30(7), 1975–1996.
Hsieh, M. F., Shen, C. H., & Lee, J. S. (2010). Factors influencing the foreign entry mode of Asian and Latin-American banks. The Service Industries Journal, 30(14),
2351–2365.
Isik, I., & Hassan, M. K. (2002). Technical, scale and allocative efficiencies of Turkish banking industry. Journal of Banking and Finance, 26(4), 719–766.
Kaparakis, E. I., Miller, S. M., & Noulas, A. G. (1994). Short-run cost inefficiency of commercial banks: A flexible stochastic frontier approach. Journal of Money, Credit
and Banking, 26(4), 875–893.
Kirkwood, J., & Nahm, D. (2006). Australian banking efficiency and its relation to stock returns. Economic Record, 82(258), 253–267.
Koutsomanoli-Filippaki, A. I., & Mamatzakis, E. C. (2011). Efficiency under quantile regression: What is the relationship with risk in the EU banking industry? Review of
Financial Economics, 20(2), 84–95.
Kwan, S., & Eisenbeis, R. A. (1997). Bank risk, capitalization, and operating efficiency. Journal of Financial Services Research, 12(2–3), 117–131.
Lin, X., & Zhang, Y. (2009). Bank ownership reform and bank performance in China. Journal of Banking and Finance, 33(1), 20–29.
Lozano-Vivas, A., Pastor, J. T., & Pastor, J. M. (2002). An efficiency comparison of European banking systems operating under different environmental conditions.
Journal of Productivity Analysis, 18(1), 59–77.
Maudos, J., Pastor, J. M., Perez, F., & Quesada, J. (2002). Cost and profit efficiency in European banks. Journal of International Financial Markets, Institutions and Money,
12(1), 33–58.
Miguel-D  Cabeza-García, L., Valdunciel, L., & Fl
avila, J.A., orez, M. (2010). Operations in banking: The service quality and effects on satisfaction and loyalty. The Service
Industries Journal, 30(13), 2163–2182.
Mitchell, K., & Onvural, N. M. (1996). Economies of scale and scope at large commercial banks: Evidence from the Fourier flexible functional form. Journal of Money,
Credit, and Banking, 28(2), 178–199.
Nikiel, E. M., & Opiela, T. P. (2008). Customer type and bank efficiency in Poland: Implications for emerging market banking. Contemporary Economic Policy, 20(3),
255–271.
Pastor, J. M., & Serrano, L. (2005). Efficiency, endogenous and exogenous credit risk in the banking systems of the Euro area. Applied Financial Economics, 15(9),
631–649.
Rao, D. P., O'Donnell, C. J., & Battese, G. E. (2005). In T. J. Coelli (Ed.), An introduction to efficiency and productivity analysis. Springer.
Samad, A. (2004). Performance of Interest-free Islamic banks vis-a-vis Interest-based Conventional Banks of Bahrain. International Journal of Economics, Management and
Accounting, 12(2).
Semih Yildirim, H., & Philippatos, G. C. (2007). Efficiency of banks: Recent evidence from the transition economies of Europe, 1993–2000. European Journal of Finance,
13(2), 123–143.
Sturm, J. E., & Williams, B. (2004). Foreign bank entry, deregulation and bank efficiency: Lessons from the Australian experience. Journal of Banking and Finance, 28(7),
1775–1799.
Tortosa-Ausina, E. (2002). Bank cost efficiency and output specification. Journal of Productivity Analysis, 18(3), 199–222.
Tsionas, E. G., Assaf, A. G., & Matousek, R. (2015). Dynamic technical and allocative efficiencies in European banking. Journal of Banking & Finance, 52, 130–139.
Weill, L. (2004). Measuring cost efficiency in European banking: A comparison of frontier techniques. Journal of Productivity Analysis, 21(2), 133–152.
Wilson, P. W. (1993). Detecting outliers in deterministic nonparametric frontier models with multiple outputs. Journal of Business and Economic Statistics, 11(3),
319–323.
Yang, C. C. (2012). Service, investment, and risk management performance in commercial banks. The Service Industries Journal, 32(12), 2005–2025.