Does bank performance contribute to economic growth in the European Union?
Ferreira, Candida
INTRODUCTION
The recent protracted financial crisis has raised attention to the
importance of bank performance for economic growth all over the world
and very particularly in the European Union (EU). In the context of the
EU, the relevance of banking institutions in the process of financing
economic growth is generally accepted, although their specific roles are
controversial, as has been clearly recognised among others by Goddard et
al. (2007) and Molyneux (2009).
Banks and other financial institutions are usually supposed to
guarantee the financing of productive investments and activities as they
mobilise and allocate financial resources, and also by means of their
specific money-creation processes through bank credit. Moreover, there
is a general consensus that well-functioning markets and financial
institutions may decrease transaction costs and asymmetric information
problems.
The analysis of the importance of bank performance for economic
growth has been the object of discussion for decades and intensified
after the renowned contribution of King and Levine (1993). During the
last two decades there has been an increase in empirical studies at the
aggregate level, which explain output variables with financial ratios
and variables such as liquid liabilities, bank loans to the private
sector, and stock market capitalisation, which may be representative of
the development of financial systems and institutions.
Khan and Senhadji (2000), analysing the literature concerning the
empirical evidence of the relationship between financial development and
economic growth, concluded that the results of these studies indicate
that while the general effects of financial development on the outputs
are positive, the size of these effects varies with the different
variables considered, with the indicators of financial development and
with the estimation method, data frequency, or the defined functional
form of the relationship.
At the same time, authors such as Rajan and Zingales (1998) have
argued that there is no clear causality between financial development
and economic growth. Rather than adhering to the traditional explanation
of economic growth by the proxy of financial development, these authors
test the hypothesis that financial markets and banking institutions not
only reduce the cost of financing, but also help to combat the problems
provoked by asymmetrical information, concluding with their test that
the sectors most dependent on external financing will be the ones that
grow faster and in line with the development of the financial markets
and institutions to which these sectors have access.
More recently, Hassan et al. (2011) provided evidence from panel
data estimations and concluded that there is a general positive
relationship between financial development and economic growth in low-
and middle-income countries classified by economic region, although it
seems that well-functioning financial systems may be necessary for
mostly developing countries, but not a sufficient condition to reach
steady and sustainable economic growth.
Other authors analyse the importance of financial markets'
performance for real economic growth, but put particular emphasis on the
relations between business and financial cycles. For example, Borio
(2012) considers that it is not possible to understand macroeconomics,
business fluctuations, and policy challenges in recent decades without
understanding financial cycles. At the same time, Claessens et al.
(2012) empirically test the interactions between business and financial
cycles with monthly data for 44 countries, covering the time period
1960-2010, and mostly conclude that recessions associated with financial
disruptions tend to be longer and deeper, emphasising the importance of
financial market developments for the real economy.
Another strand of literature takes into account the development of
the global trend of bank consolidation, and increases the theoretical
debates and empirical tests analysing the particular relationship
between bank market concentration and bank performance. Until the 1990s,
there was a general belief that mergers did not clearly contribute to
bank performance improvements and several empirical findings were
consistent with the traditional structure-conduct-performance
statements, in particular with the 'quiet life hypothesis'
(among others, Hannan and Berger, 1991; Houston and Ryngaert, 1994;
Pilloff, 1996). But from the year 2000, this general consensus was
broken and particular attention was paid to specific characteristics of
the banking markets such as the presence of asymmetric information,
contagion phenomena, and imperfect competition, or the specific impacts
of bank concentration, competition, and regulation on bank performance
(eg De Bandt and Davis, 2000; Bikker and Haaf, 2002; Berger et al, 2004;
Hasan et al, 2009; Schaeck et al, 2009).
However, it is generally recognised that not many works have
addressed the possible relationship between economic growth and banking
market structure, and also between economic growth and bank performance,
namely measuring this bank performance through bank efficiency.
One of the few examples of these works can be found in Maudos and
Fernandez de Guevara (2009), who used different measures of bank market
competition for a sample of 21 countries and 53 economic sectors during
the time period 1993-2003, concluding not only that there is a positive
effect of financial development on economic growth, but also that the
exercise of bank market power promotes economic growth.
Different conclusions were obtained by Claessens and Laeven (2005),
who used industry-specific and country-specific data for 16 countries
for the time periods 1980-1990 and 1980-1997 to estimate a measure of
banking competition based on industrial organisation theory and then
related this competition measure to the growth of industries. Their
findings point to the evidence that greater competition in
countries' banking systems will contribute to the faster growth of
financially dependent industries, and so there is no support for the
hypothesis that market power is good for access to financing and
promoting economic growth.
Carbo Valverde et al. (2003) analyse the relationship between
financial market competition and economic growth in five large regions
in Spain and conclude that the differences in competition are not
associated with improved regional growth.
Regarding the measurement of the quality of the financial
development and its possible influence on economic growth, Hasan et al.
(2009) use a sample of 147 regions in 11 European countries between 1996
and 2004, and conclude that regional economic growth benefits
significantly from more efficient banks.
This paper contributes to the literature with the analysis of the
possible contribution of bank market structure and the performance of
the banking institutions to economic growth, here represented by per
capita gross national income. To our knowledge, not many authors have
addressed these issues in the particular context of all EU member states
during the last decade, taking into account the influence of the
international financial crisis, and considering the specific influence
of bank market structure and of bank performance on economic growth. To
represent bank market structure here, we use a bank market concentration
measure and bank performance is proxied both by the capital ratio of
bank equity to bank total assets and also a Data Envelopment Analysis
(DEA) bank cost-efficiency measure.
The main empirical results confirm the controversial influence of
bank market structure on economic growth. However, they clearly reveal
that the equity to total assets ratio had a significant negative
influence on economic growth before and after the beginning of the
recent financial crisis. Simultaneously, and in line with the findings
of Hasan et al. (2009), we conclude that DEA bank efficiency positively
contributes to economic growth, although not as statistically
significantly for the years after the beginning of the crisis.
The paper is organised as follows: The section 'Data and
methodology' presents the data used and the methodological
framework; The section 'Results obtained with the generalised
method of moments (GMM) estimates' reports the results of the
dynamic panel estimations; and the section 'Concluding
remarks' concludes.
DATA AND METHODOLOGY
This paper uses dynamic Generalised Method of Moments (GMM) panel
estimations in order to analyse mostly the effects on economic growth of
bank market concentration and bank performance, here represented both by
the ratio of bank equity to bank total assets and by a DEA bank
efficiency measure. The economic growth of each EU member state is
represented by the natural logarithm of the country's gross
national income at current prices per capita.
As control variables we include nominal short-term interest rates
(a proxy for monetary policy interest rates) and general government net
lending/ borrowing (taking into account the recognised importance of
public finances for economic growth and also the fact that some EU
countries recently faced a sovereign debt crisis).
These macroeconomic data, that is, the two control variables and
also the dependent variable (the natural logarithm of the gross national
income at current prices per capita), are all sourced from the European
Commission AMECO database.
The data used to obtain the bank market concentration measure and
to represent bank performance (more precisely, the equity to total
assets ratio and the inputs and outputs used to obtain the DEA bank
efficiency measure) are sourced from the privately owned financial
database maintained by the Bureau van Dijk: Bankscope.
Bank market concentration
Bank market concentration is measured through one of the most
popular indicators: the percentage share of the total assets held by the
three largest banking institutions (C3) of each EU member state.
The C3 results are presented in Table 1 and clearly show that, with
some exceptions, there is a general increase in bank market
concentration. The exceptions are to be found in the Netherlands and
Greece and most particularly in certain new EU member states, namely
Bulgaria, Ireland, Latvia, Lithuania, Malta, Poland, Romania, and the
Slovak Republic.
On the other hand, and in spite of the general increase in EU bank
market concentration during the considered time period, the levels of
concentration continue to be relatively low in the relevantly developed
EU countries, namely France, Germany, Spain, and the United Kingdom.
Ratio of bank equity to bank total assets
The ratio of equity to total assets is one of the most important
capital ratios, representing the book value of equity divided by total
assets. Taking into account that equity represents a cushion against
asset malfunction, this ratio measures the amount of protection afforded
to the bank by the equity invested in the bank: the higher this ratio
is, the more protected the bank is. The equity to total assets ratio
also measures bank leverage levels and reflects the differences in risk
preferences across banks.
Table 2 reports the values of the ratio of bank equity to bank
total assets. The values reveal that in spite of some clear country
differences and important oscillations, there is a general tendency of
this ratio increasing during the considered time period, particularly as
a response to the international financial crisis that began in 2007.
Nevertheless, exceptions showing a decrease of this ratio and thus the
decrease of bank protection are to be found in some of the new EU member
states (such as Bulgaria and Romania) and, very particularly, in some of
the developed EU countries such as Denmark, France, the Netherlands, and
the United Kingdom.
Data envelopment analysis (DEA) bank efficiency
The analysis of efficiency is usually based on the estimation of
efficiency frontiers with the best combinations of the different inputs
and outputs of the production process and then on the analysis of the
deviations from the frontier that correspond to the losses of
efficiency.
Most of the empirical studies on the measurement of bank efficiency
adopt either parametric methods, such as the Stochastic Frontier
Analysis, or non-parametric methods, in particular Data Envelopment
Analysis (DEA).
Here, we will adopt the DEA methodology, which was originally
presented in Charnes et al. (1978), assuming constant returns to scale,
which can be accepted as optimal but only in the long run. Later, Banker
et al. (1984) introduced an additional convexity constraint ([lambda])
and allowed for variable returns to scale. Following also Coelli et al.
(1998), Thanassoulis (2001), and Thanassoulis et al. (2007), we can
assume that at any time t, there are N decision-making units that use a
set of X inputs (X = [x.sub.1], [x.sub.2], ..., [x.sub.k]) to produce a
set of Y outputs (Y = [y.sub.1], [y.sub.2], thus obtaining the DEA
input-oriented efficiency measure of every i DMU, solving the following
optimisation problem: (1)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]
The DEA approach provides, for every i DMU (here every
country's banking sector), a scalar efficiency score
([[theta].sub.i] [less than or equal to] 1). If [[theta].sub.i] = 1; the
DMU lies on the efficient frontier and will be considered an efficient
unit. On the contrary, if [[theta].sub.i] < 1 the DMU lies below the
efficient frontier and will be considered an inefficient unit; moreover,
(1 - [[theta].sub.i]) will always be the measure of its inefficiency.
Here we follow the intermediation approach, considering that the
banks' total costs will depend on three bank outputs (total loans,
total securities, and other earning assets), and also on three bank
inputs (borrowed funds, physical capital, and labour).
More specifically, using the Bankscope database we define the
outputs and the inputs of the cost function with the following
variables:
Outputs:
1. Total loans = the natural logarithm of the loans
2. Total securities = the natural logarithm of the total securities
3. Other earning assets = the natural logarithm of the difference
between the total earning assets and the total loans.
Inputs:
1. Price of borrowed funds = the natural logarithm of the ratio of
interest expenses to the sum of deposits
2. Price of physical capital = the natural logarithm of the ratio
of non-interest expenses to fixed asset
3. Price of labour = the natural logarithm of the ratio of
personnel expenses to total assets (2)
The results obtained are provided in Table 3 and reveal not only
some year-on-year oscillations, but also a clear general tendency
towards the increase of DEA bank efficiency between 1999 and 2013.
The good bank performance of the economically more developed EU
member states is proven by the frequency of the achievement and
maintenance before and after the beginning of the international
financial crisis in the DEA efficiency frontier (represented by the
value 1) of countries such as Belgium, France, Germany, Luxembourg, the
Netherlands, and the United Kingdom.
Using the presented information, and also two control variables
(the interest rate and the government net lending/borrowing) the basic
model to be estimated in this paper is:
[GNP.sub.i,t] = [a.sub.0] + [[alpha].sub.1] interest [rate.sub.i,t]
+ [[alpha].sub.2] government net lending/[borrowing.sub.i,t] +
[[alpha].sub.3] bank market [concentration.sub.i,t] + [[alpha].sub.4]
ratio of equity to total [assets.sub.i,t] + [[alpha].sub.5] bank
[efficiency.sub.i,t] + [[epsilon].sub.i,t] (1)
where
GNP=the natural logarithm of gross national income, at current
prices, per capita;
1 = EU country (i = 1, ..., 28 in Panel 1; i=1, ..., 22 in Panel
2);
t = year (t = 1999, 2013 in interval A; t = 1999, ..., 2007 in
interval B; t = 2008, ..., 2013 in interval C);
interest rate = the nominal short-term interest rate;
government net lending/borrowing = general government annual net
lending or net borrowing;
bank market concentration = the percentage share of total assets
held in each country by the three largest banking institutions (C3);
ratio of equity to total assets = the book value of equity divided
by total assets;
bank efficiency = Data Envelopment Analysis (DEAJ bank
cost-efficiency measure; and
[epsilon] = error term.
Before proceeding with the estimations of the presented equation
some panel properties of the variables are investigated. First, we test
the stationarity of the series using the Levin-Lin-Chu panel unit root
test (Levin et al., 2002), and then we test the cointegration between
the series by applying the Westerlund (2007) cointegration tests.
Stationarity of the series
We test the stationarity of the series with the Levin-Lin-Chu panel
unit root test (Levin et al, 2002), which may be viewed as a pooled
Dickey-Fuller test or as an augmented Dickey-Fuller test, when lags are
included to account for serial correlation in the errors and the null
hypothesis is the existence of non-stationarity. The basic augmented
Dickey-Fuller equation is [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN
ASCII].
The null hypothesis is that the series is non-stationary or
integrated of order 1. The Levin-Lin-Chu panel unit root test derives a
statistic (t-star), which is distributed as a standard normal under the
null hypothesis of non-stationarity. The test accounts for individual
effects, time effects and a possible time trend, but it assumes that
each cross-section in the panel shares the same auto-regressive
coefficient, meaning that all series in the panel have the same degree
of mean-reversion.
This test is adequate for heterogeneous panels of moderate size
such as the panels used in this paper. The main results obtained are
reported in Table 4 and all Levin-Lin-Chu t-star statistics are clearly
significant at all the usual testing levels, allowing us to reject the
existence of the null hypothesis and to conclude that the considered
series are stationary.
Cointegration between the series
Cointegration is tested with the implementation of the four panel
tests developed by Westerlund (2007), which test for the absence of
cointegration by determining whether individual panel members are error
correcting. These tests are flexible and work well in unbalanced,
heterogeneous, and/or relatively small panels, allowing for dependence
both between and within cross-panel units.
The application of these panel cointegration tests to the i series
included in one panel considers for each moment t (during the time
interval t = 0, the following error-correction model:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].
The Westerlund cointegration tests provide four test statistics:
Gt, Ga, Pt, and Pa. The Gt and Ga statistics test HO: [a.sub.i] = 0 for
all i versus HI: [a.sub.i] < 0 for at least one of the series, i,
starting from a weighted average of the individually estimated
coefficients [a.sub.i] and their respective t-ratios.
The Pt and Pa test statistics consider the pooled information of
all panel cross-section units to test HO: [a.sub.i] = 0 for all i versus
H1: [a.sub.i] < 0 for all cross-section units. Thus, the rejection of
HO must always be taken as the rejection of cointegration for the entire
panel. Any single cross-section unit can cause the rejection of HO and
it is not possible to identify which cross-section unit is responsible
for this rejection.
Table 5 presents the p-values obtained with the Westerlund
cointegration tests (Westerlund, 2007) for both of the considered panels
of EU countries (the values of the statistics and the Z-values are also
available and will be provided on request).
These results reveal that, in both panels, cointegration cannot be
rejected between the series of the natural logarithm of the per capita
national income and bank market concentration, and also to some extent
(according to the presented [G.sub.t] and [P.sub.t] results) between the
series of the natural logarithm of the per capita national income and
the nominal short-term interest rate, indicating that these series tend
to move together in the long run.
On the other hand, there is clear evidence that the series of the
natural logarithm of the per capita national income do not move together
with the series of the government net lending/borrowing, and also,
although not as clearly (because of the [G.sub.t] p-values), with the
series of the ratio of bank equity to bank total assets.
In what concerns the cointegration between the series of the
natural logarithm of the per capita national income and the DEA bank
efficiency, the results are not completely unanimous. Nevertheless, the
[G.sub.t] and [P.sub.t] p-values clearly point to the non-rejection of
their cointegration in both considered panels of EU countries.
Summarising, we may say that these results do not provide clear
evidence of the existence or non-existence of cointegration between the
dependent variable (the natural logarithm of the per capita GDP) and
each of the considered explanatory variables (in particular those
representing bank market structure and bank performance). Under these
conditions, in this paper we opt not to develop a cointegration
approach, namely testing the relevant cointegrating vectors and
analysing the long-run relationships between the variables.
RESULTS OBTAINED WITH THE GENERALISED METHOD OF MOMENTS (GMM)
ESTIMATES
In our estimation we chose to apply the GMM system. GMM uses
cross-country information and jointly estimates the equations in first
differences and in levels, with first differences instrumented by lagged
levels of the dependent and independent variables and levels
instrumented by the first differences of the regressors.
In order to test the consistency of the GMM estimates, namely the
validity of the additional instruments, we follow the tests proposed by
Arellano and Bond (1991). They are used to test autocorrelation, that
is, the assumption that the error term is not serially correlated using
the differenced error term; so, by construction, the autocorrelation of
the first order, AR(1), is supposed to be validated, but not the
autocorrelation of the second order, AR(2), or autocorrelation of a
higher order. In addition, the validity of the instruments is tested
through the Sargan statistic, which is robust to heteroscedasticity and
autocorrelation.
In the estimations we take into account the particular challenges
that the EU countries had to face during the last decade, namely
adapting to the implementation of the European Monetary Union and to the
enormous enlargement of the EU with new member states, facing the
consequences of the international financial crisis, and also facing the
possible interactions between business and financial cycles. (3) We also
take into account the availability of data and we consider the following
panels;
* Panel 1: EU 28, including all current EU member states, and
* Panel 2: EU 22, excluding countries that were under the
'Troika's' financial assistance (Greece, Ireland, and
Portugal) and also countries that recently faced growth problems and/or
troubles in their financial systems (Cyprus, Italy, and Spain).
For both panels we also consider three time periods:
* Period A: 1999-2013, using all available data in the Bankscope
database and taking into account that the European Monetary Union was
established in 1999;
* Period B: 1999-2007, including the years before the international
financial crisis that deeply affected Europe in 2008; and
* Period C: 2008-2013, covering the period after the beginning of
the global financial crisis.
Table 6 presents the results obtained for all panels using GMM
one-step estimates. In all situations the Wald tests clearly confirm the
overall fit of the considered model.
The quality of these estimates in both panels is corroborated by
the results obtained with the Arellano-Bond tests (Arellano and Bond,
1991) as they always reject the null hypothesis of no autocorrelation of
the first order and (with the exception of only Panel 2 - B) do not
reject the hypothesis of no autocorrelation of the second order.
Moreover, the Sargan test results for the overidentifying
restrictions allow us to consider that the included instruments in our
estimations are valid.
In general the results are very similar for both panels, revealing
that concerning the issues addressed in this paper, the differences
across EU member states are not very relevant, namely the differences
between the six countries that more clearly suffered the consequences of
the international crisis and the other countries.
In what relates to the control variables, we confirm the same kind
of results for both panels: the nominal short-term interest rate exerts
in all situations a statistically very significant negative influence on
the per capita national income growth, underlining the importance of
monetary policy for economic growth. More precisely, and as expected,
our results suggest that an increase of the nominal short-term interest
rate, which is often taken as a proxy of the monetary policy interest
rate, will contribute to the decrease of the growth of the per capita
GDP (more precisely the natural logarithm of gross national income, at
current prices, per capita taken from the AMECO database).
On the other hand, the other control variable, government net
lending/ borrowing, always contributes positively to per capita national
income growth and the results are statistically significant at 1 % for
the entire time period (1999-2013) and for the years after the beginning
of the global financial crisis (2008-2013).
As for the influence on economic growth of the bank market
conditions (here represented by the bank market concentration C3
measure), the results confirm the recognised ambiguity of these market
(competition) conditions. In general, the results obtained are not
statistically significant. Nevertheless, considering Panel 1 with all EU
28 countries, we obtain a 10% statistically positive influence, but only
for the sub-interval before the international crisis (1999-2007),
revealing that in those years bank concentration, which can also be
identified with less bank market competition, was in line with per
capita income growth.
However, if we exclude the six EU countries that had clear and
strong problems with the financial crisis and look at the results of
Panel 2 - EU 22, it looks as though that bank market concentration had a
negative influence (statistically significant at 10%) during the entire
time period (1999-2013), revealing that for those 22 EU countries bank
competition can promote economic growth. And this result is not
surprising, taking into account what we said before: in general the bank
market concentration is lower (meaning that bank market competition is
higher) in the more developed EU countries.
Regarding the chosen variables representing bank performance, on
the one hand the results obtained for the capital ratio (bank equity to
bank total assets), which is a recognised good measure of bank
protection, reveal that in both panels, and during all considered time
periods, more protected banks, that is, those with lower leverage
levels, had a negative influence on per capita income growth. This was
true before and after the beginning of the international financial
crisis, meaning that bank risk preferences were important in explaining
economic growth and that very cautious banks did not provide the
necessary financial support to guarantee the increase of the per capita
gross national income of EU countries.
On the other hand, if bank performance is proxied by the DEA bank
efficiency measure, our results confirm the assumption that
well-functioning bank institutions will contribute positively to
economic growth, although they did not do so as strongly for the years
after the beginning of the international financial crisis (here the time
period 2008-2013). The explanations for this statistically less
significant relevance of bank efficiency for economic growth after the
crisis is surely connected with the consequences of the crisis in the EU
financial system as well as with strict requirements, namely in terms of
capital ratios and legal regulations that EU banks were obliged to
respect as a response to the financial crisis.
CONCLUDING REMARKS
This paper uses dynamic GMM panel estimations and includes two
control variables to explain economic growth in all current EU member
states for the time period 1999-2013 and the sub-intervals before and
after the beginning of the recent international financial crisis.
Regarding the control variables, one represents the influence of
the monetary policy rate [proxied by the nominal short-term interest
rate) and the second is the government net lending/borrowing (having in
mind the recognised relevance of public finances for economic growth and
the recent sovereign debt crisis faced by many EU countries).
As expected, in all considered situations, the increase of the
nominal short-term interest rates clearly contributes to the decrease of
per capita national income growth. The results obtained for the other
control variable (government net lending/borrowing) are also not
surprising, as they always exert a positive influence on economic
growth. Moreover, this influence is statistically very significant when
we consider the entire time period (1999-2013) and for the years after
the beginning of the global financial crisis (2008-2013), when many EU
countries had to pay particular attention to their public finances and
to their possible consequences for national income growth.
In what concerns the importance for economic growth of the bank
market concentration C3 measure, the increase of which may be accepted
as synonymous with less market competition, the obtained results are
neither statistically very significant nor unanimous, confirming that
this issue deserves to be the object of further empirical analysis and
discussion. Nevertheless, as the results that we obtained during the
entire interval (1999-2013) are statistically relevant at 10%, but only
for Panel 2 (including only the 22 EU countries that were less clearly
affected by the international crisis), we confirm the 'quiet life
hypothesis', meaning that the increase in bank market concentration
(meaning less market competition) negatively contributes to per capita
national income growth. On the other hand, and also with results that
are statistically relevant at 10%, but now for Panel 1 (including all 28
current EU member states) and only for the subinterval including the
years before the international crisis (1999-2007), the increase of bank
market concentration can be considered as promoting economic growth.
These findings are mostly in line with the concerns of those authors
that, particularly after the year 2000, contradict the 'quiet life
hypothesis' and emphasise that some specific characteristics of the
banking markets contain the presence of asymmetric information,
contagion phenomena, and imperfect competition that surely affected many
EU banking markets before the beginning of the international financial
crisis.
Bank performance is represented in this paper by two specific
variables: a financial capital ratio (bank assets to bank total assets)
and a DEA bank cost-efficiency measure.
According to the results obtained for the bank equity to bank total
assets ratio, in all considered situations the increase of this ratio
had a statistically very strong negative influence on per capita income
growth. These findings confirm that this ratio is a good measure of bank
protection, as when this protection collapses with the increase of the
bank leverage levels it will be associated with the increase of
financial resources promoting economic growth, and so more bank
protection will contribute to less per capita income growth. The
recognition of these facts is one of the justifications for the
theoretical and political discussions associated with the relevance of
bank market regulations and the importance of bank behaviour not only
for promoting economic growth but also for avoiding financial crisis.
Clearly, on the one hand high bank leverage levels will provide more
financial resources to promote economic growth, but on the other hand
these high leverage levels (here identified with low bank equity to bank
total assets ratios) will be synonymous with high risks, and less
protected banks may become problematic and contribute to financial and
economic crises.
Finally, in what concerns the influence of DEA bank efficiency on
economic growth, our findings always point to a clear positive
contribution of the increase of bank efficiency to economic growth.
Nevertheless, the results are statistically not as significant for the
sub-interval 2008-2013, which includes the years after the beginning of
the international financial crisis. So, in spite of the specific
challenges, namely in terms of capital ratios, that the EU banking
institutions had to face with the crisis, our findings clearly confirm
the general assumption that well-functioning bank institutions will
contribute positively to economic growth.
Acknowledgements
The author would like to thank the participants at the 15th Annual
Conference of the International Network for Economic Research (INFER),
Orleans, 29 May-1 June 2013, and especially to Professor Camelia Turcu
and to three anonymous referees for their most helpful comments,
criticisms, and suggestions. The usual disclaimer remains.
REFERENCES
Arellano, M and Bond, S. 1991: Some tests of specification for
panel data: Monte Carlo evidence and an application to employment
equations. The Review of Economic Studies 58(2): 277-297.
Banker, RD, Charnes, A and Cooper, WW. 1984: Some models for the
estimating technical and scale inefficiencies in data envelopment
analysis. Management Science 30(9): 1078-1092.
Berger, AN, Demirguc-Kunt, A, Levine, R and Haubrich, JG. 2004:
Bank concentration and competition: An evolution in the making. Journal
of Money, Credit, and Banking 36(3): 433-451.
Bikker, JA and Haaf, K. 2002: Competition, concentration and their
relationship: An empirical analysis of the banking industry. Journal of
Banking & Finance 26(11): 2191-2214.
Borio, C. 2012: The financial cycle and macroeconomics: What have
we learnt?, Bank for International Settlements Working Paper No. 395,
www.bis.org.
Carbo Valverde, S, Humphrey, DB and Rodriguez Fernandez, F. 2003:
Deregulation, bank competition and regional growth. Regional Studies
37(3): 227-237.
Charnes, A, Cooper, WW and Rhodes, E. 1978: Measuring the
efficiency of decision making units. European Journal of Operational
Research 2(6): 429-444.
Claessens, S and Laeven, L. 2005: Financial dependence, banking
sector competition, and economic growth. Journal of the European
Economic Association 3 (1): 179-207.
Claessens, S, Rose, MA and Terrones, ME. 2012: How do business and
financial cycles interact? Journal of International Economics 87(1):
178-190.
Coelli, TJ, Prasada Rao, DS and Battese, GE. 1998: An introduction
to efficiency and productivity analysis. Kluwer Academic Publishers:
Norwell, MA.
De Bandt, O and Davis, EP. 2000: Competition, contestability and
market structure in European banking sectors on the eve of EMU. Journal
of Banking & Finance 24(6): 1045-1066.
Goddard, J, Molyneux, P, Wilson, JOS and Tavakoli, M. 2007:
European banking: An overview. Journal of Banking & Finance 31 (7):
1911-1935.
Hannan, TH and Berger, AN. 1991: The rigidity of prices: Evidence
from the banking industry. The American Economic Review 81 (4): 938-945.
Hasan, I, Koetter, M and Wedow, M. 2009: Regional growth and
finance in Europe: Is there a quality effect of bank efficiency? Journal
of Banking & Finance 33(8): 1446-1453.
Hassan, MR, Sanchez, B and Yu, J-S. 2011: Financial development and
economic growth: New evidence from panel data. The Quarterly Review of
Economics and Finance 51(1): 88-104.
Houston, JF and Ryngaert, MD. 1994: The overall gains from large
bank mergers. Journal of Banking & Finance 18(6): 1155-1176.
Rhan, MS and Senhadji, AS. 2000: Financial development and economic
growth: An overview, IMF Working Paper No. 209. Available at SSRN:
http://ssrn.com/abstract=880870.
King, RG and Levine, R. 1993: Finance and growth: Schumpeter might
be right. The Quarterly Journal of Economics 108(3): 717-737.
Levin, A, Lin, C-F and Chu, C-SJ. 2002: Unit root tests in panel
data: Asymptotic and finite-sample properties. Journal of Econometrics
108(1): 1-24.
Maudos, J and Fernandez de Guevara, J. 2009: Banking Competition,
Financial Dependence and Economic Growth, Munich Personal RePEc Archive
Paper No. 15254, https://mpra.ub .uni-muenchen.de/15254/.
Molyneux, P. 2009: Do mergers improve bank productivity and
performance? In: Balling, M, Gnan, E, Lierman, F and Schoder, JP (eds).
Productivity in the Financial Services Sector. SUERF: Vienna, pp. 23-43.
Pilloff, SJ. 1996: Performance changes and shareholder wealth
creation associated with mergers of publicly traded banking
institutions. Journal of Money, Credit and Banking 28(3): 294-310.
Rajan, RG and Zingales, L. 1998: Financial dependence and growth.
The American Economic Review 88(3): 559-586.
Schaeck, K, Cihak, M and Wolfe, S. 2009: Are competitive banking
systems more stable? Journal of Money, Credit and Banking 41(4):
711-734.
Thanassoulis, E. 2001: Introduction to the Theory and Application
of Data Envelopment Analysis. A Foundation Text with Integrated
Software. Kluwer Academic Publishers: Norwell, MA.
Thanassoulis, E, Portela, MCS and Despic, O. 2007: Data envelopment
analysis: The mathematical programming approach to efficiency analysis.
In: Fried, HO, Knox Lovell, CA and Schmidt, SS (eds). The Measurement of
Productive Efficiency and Productivity Growth. Oxford University Press:
Oxford, pp. 251-420.
Weill, L. 2004: Measuring cost efficiency in European banking: A
comparison of frontier techniques. Journal of Productivity Analysis
21(2): 133-152.
Westerlund, J. 2007: Testing for error correction in panel data.
Oxford Bulletin of Economics and Statistics 69(6): 709-748.
(1) For more details on this problem, see, among others, Coelli et
al. (1998) and Thanassoulis et al. (2007).
(2) The Bankscope database does not provide the number of employees
for the bank sector of all EU countries. So, following the approach
adopted, among others, by Weill [2004), we use the ratio of personnel
expenses to total assets as a proxy for the price of labour.
(3) Concerning these issues, we are grateful to three anonymous
referees for their suggestions, namely to the referee who gave the Borio
(2012) and Claessens et al. (2012) references.
CANDIDA FERREIRA
Economics, ISEG, Universidade de Lisboa and UECE, R Miguel Lupi,
20, Lisboa 1249-078, Portugal.
E-mail: candidaf@iseg.ulisboa.pt
Table 1: Concentration measure (C3)
1999 2000 2001 2002 2003 2004
Austria 85.09 80.4 83.71 82.14 80.57 70.89
Belgium 43.19 41.51 38.51 37.86 38.7 38.59
Bulgaria 73.79 68.57 65.1 60.31 68.69 59.82
Croatia 69.6 68.7 67.49 66.92 71.79 74.87
Cyprus 75.7 76.26 70.04 69.66 72.24 68.94
Czech Republic 65.97 76.65 81.09 79.2 80.36 72.29
Denmark 73.17 71.23 33.21 31.94 31.82 88.57
Estonia 98.9 98.74 99.94 99.93 99.92 99.82
Finland 99.12 100 100 100 100 98.07
France 33.84 33.75 45.86 65.57 66.16 72.15
Germany 48.98 56.88 57.88 50.29 48.63 48.44
Greece 100 100 100 100 100 59.51
Hungary 58 52.29 58.43 55.93 55.98 53.84
Ireland 77.13 76.78 82.08 75.99 100 95.28
Italy 58.42 67.89 51.42 57.17 58.56 87.34
Latvia 79.87 80.45 80.76 75.97 68.09 61.92
Lithuania 92.64 86.43 86.74 85.13 81.33 78.97
Luxembourg 27.34 29.51 37.23 34.35 34.95 33.88
Malta 100 95.21 100 100 92.41 98.07
The Netherlands 55.24 53.33 56.41 58.87 55.34 83.52
Poland 70.69 77.42 53.18 60.07 84.28 48.14
Portugal 66.65 78.68 93.05 99.38 99.58 98.18
Romania 76.94 70.69 69.16 63.34 61.08 58.34
Slovak Republic 79.62 78.68 76.86 76.81 76.38 74.72
Slovenia 64.7 62.96 66.89 66.69 64.33 62.96
Spain 47.67 63.42 66.78 73.77 88.45 70.75
Sweden 96.58 97.6 88.16 86.92 87.07 46.88
The United Kingdom 34.6 41.44 43.32 44.36 40.04 64.81
2005 2006 2007 2008 2009 2010
Austria 73.09 65.65 67.65 60.44 56.48 63.23
Belgium 81.47 80.99 81.86 79.14 74.7 71.2
Bulgaria 53.78 50.04 57.81 52.57 52.25 42.93
Croatia 67.73 55.41 55.71 56.09 56.87 57.25
Cyprus 66.93 75.66 82.47 80.64 81.19 89.67
Czech Republic 73.4 73.56 71.41 70.53 66.55 67.41
Denmark 86.44 84.25 84.46 84.17 82.71 83.58
Estonia 99.8 98.68 98.54 99.24 99.18 98.51
Finland 98.09 98.44 97.22 93.15 92.51 93.47
France 44.66 45.03 45.01 47.45 54.33 54.42
Germany 46.96 54.89 56.66 59.05 61.15 63.33
Greece 60.64 60.9 58.89 58.89 59 59.83
Hungary 53.99 54.47 55.9 54.41 57.16 58.2
Ireland 68.34 65.44 65.94 70.96 69.95 67.17
Italy 55.64 55.39 60.54 61.94 59.89 64.73
Latvia 63.98 63.14 60.75 62.77 59.74 51.98
Lithuania 72.06 72.56 72.67 73.59 71.26 66.21
Luxembourg 41.48 39.51 30.14 29.25 29.82 30.5
Malta 98.02 92.3 97.76 91.76 90.22 87.2
The Netherlands 89.25 91.28 91.67 89.5 81.31 79.91
Poland 46.22 44.62 47 41.06 42.53 41.46
Portugal 82.83 80.65 79.67 82.83 82.11 81.45
Romania 59.44 59.21 52.38 49.68 45.94 45.76
Slovak Republic 68.02 73.54 68.34 68.23 63.37 62.06
Slovenia 59.83 60.14 61.24 58.85 57.7 57.12
Spain 57.63 53.63 53.11 48.98 48.88 43.42
Sweden 95.58 95.12 95.46 95.6 94.23 91.93
The United Kingdom 49.09 49.04 56.21 62.92 53.69 55.07
2011 2012 2013
Austria 63.74 62.8 64.99
Belgium 68.47 64.84 65.56
Bulgaria 41.51 41.3 50.01
Croatia 58.72 60.14 61.36
Cyprus 89 93.14 93.62
Czech Republic 66.79 64.63 66.77
Denmark 83.35 83.81 84.58
Estonia 95.84 94.5 93.99
Finland 94.4 93.21 94.57
France 53.93 52.18 52.88
Germany 64.17 63.55 68.53
Greece 72.78 72.02 78.06
Hungary 58.44 58.38 61.38
Ireland 66.83 74.17 77.81
Italy 64.88 63.39 70.17
Latvia 49.98 49.97 49.99
Lithuania 79.04 83.93 81.47
Luxembourg 31.5 32.57 35.94
Malta 87.07 86.25 83.91
The Netherlands 82.64 83.9 86.2
Poland 40.91 39.93 41.36
Portugal 78.5 79.79 79.8
Romania 46.53 47.04 47.42
Slovak Republic 62.85 59.69 63.6
Slovenia 56.24 54.24 57.85
Spain 48.41 52.19 53.36
Sweden 91.75 90.96 90.67
The United Kingdom 55.16 53.5 50.55
Source: Author's calculations using data sourced
from the Bureau van Dijk Bankscope database.
Table 2: Ratio of equity to total assets
1999 2000 2001 2002 2003 2004
Austria 3.97 3.75 4.3 4.39 5.52 5.51
Belgium 4.76 5.3 5.05 5.23 5.8 4.99
Bulgaria 17.99 16.12 13.5 13.31 14.24 11.84
Croatia 12.6 11.06 8.94 8.82 8.36 8.54
Cyprus 8.07 9.93 9.89 9.99 7.26 6.25
Czech Republic 6.58 6.87 5.91 6.49 7.21 7.79
Denmark 7.11 8.28 12.62 12.2 13.62 4.7
Estonia 13.79 12.06 10.5 10.98 10.64 10.26
Finland 5.07 4.88 7.86 85.37 14.37 4.72
France 4.71 5.03 4.45 4.33 4.67 4.55
Germany 3.65 3.67 4.02 3.54 3.85 3.77
Greece 16.66 21.37 12.77 15.85 7.76 5.89
Hungary 7.33 7.96 8.29 9.06 8.79 9.11
Ireland 5.33 4.47 3.6 9.01 7.9 5.28
Italy 5.12 6 6.67 6.33 8.22 6.4
Latvia 10.52 9.02 9.04 8.8 9.14 8.6
Lithuania 11.19 12.19 11.39 11.85 9.77 8.63
Luxembourg 3.26 3.42 3.56 4.13 4.12 4.54
Malta 6.66 6.99 7.46 7.32 23.71 20.84
The Netherlands 5.23 5.09 5.34 5.02 5.58 2.95
Poland 9.28 9.72 10.87 14.64 8.21 11.11
Portugal 4.86 4.87 4.09 3.37 3.66 4.94
Romania 16.96 16.31 17.17 15.6 13.97 12.15
Slovak Republic 5.19 6.7 7.52 8.33 8.69 8.3
Slovenia 9.78 9.87 9.35 9.41 9.39 9.31
Spain 7.27 8.22 8.62 9.45 10.8 6.81
Sweden 4.07 4.07 5.36 5.73 5.94 13.04
The United Kingdom 7.44 8.1 10.48 11.23 12.47 4.66
2005 2006 2007 2008 2009 2010
Austria 5.71 6.82 8.06 6.68 7.65 7.62
Belgium 3.13 3.3 4.13 3.17 4.28 4.75
Bulgaria 11.24 10.8 10.99 11.64 13.21 13.73
Croatia 9.19 10 12.23 13.22 13.66 13.77
Cyprus 5.64 8.63 8.32 5.87 7.23 8.55
Czech Republic 7.78 7.35 6.84 8.26 8.98 9.08
Denmark 4.24 4.65 4.22 3.75 4.24 4.34
Estonia 8.66 7.6 7.68 9.27 8.55 9.77
Finland 9.03 9.42 7.2 5.25 5.44 4.57
France 3.68 3.87 3.51 2.98 4 4.3
Germany 4.31 3.55 3.88 3.39 4.23 4.43
Greece 7.07 7.9 7.62 5.76 5.76 7.02
Hungary 8.9 9.12 8.99 8.57 9.5 9.81
Ireland 4.52 4.55 4.47 4.34 5.24 4.79
Italy 6.71 6.77 6.98 6.44 7.1 7.31
Latvia 8.13 7.88 8.29 7.84 8.86 8.77
Lithuania 7.6 7.05 7.19 7.56 6.63 7.75
Luxembourg 4.63 4.64 4.43 4.51 5.47 5.66
Malta 19.41 20.78 16.58 12.61 16.35 15.46
The Netherlands 2.83 2.9 3.83 2.56 3.15 3.58
Poland 10.66 10.49 10.27 9.18 10.41 10.75
Portugal 5.74 6.44 6.09 5.66 6.61 5.97
Romania 10.85 9.77 8.84 9.13 9.62 10.62
Slovak Republic 8.4 7.92 8.24 8.32 8.76 9.18
Slovenia 8.47 8.04 7.79 7.9 7.62 7.56
Spain 6.3 6.17 6.41 5.72 6.2 5.46
Sweden 4.21 4.32 4.31 4.2 4.88 5.16
The United Kingdom 3.59 3.59 3.83 2.66 3.98 4.54
2011 2012 2013
Austria 7.32 7.68 7.59
Belgium 4.43 5.51 6.21
Bulgaria 13.84 13.84 13.25
Croatia 13.52 14.33 13.98
Cyprus 5.83 8.55 9.71
Czech Republic 9.25 10.35 10.2
Denmark 4.5 4.67 5.06
Estonia 16.69 18.97 19.48
Finland 3.54 3.55 4.1
France 4.17 4.33 4.68
Germany 4.73 5.34 5.78
Greece 0.93 -0.43 8.31
Hungary 9.63 10.88 11.35
Ireland 8.59 9.66 8.83
Italy 6.45 6.62 6.39
Latvia 10.02 10.11 9.93
Lithuania 9.96 11.46 11.84
Luxembourg 5.68 6.79 7.11
Malta 15.3 16.46 16.23
The Netherlands 3.32 4.08 4.2
Poland 10.47 11.65 11.56
Portugal 5.52 6.74 6.58
Romania 11 10.21 10.6
Slovak Republic 9.96 10.77 10.75
Slovenia 7.58 7.76 9.21
Spain 5.92 5 6.59
Sweden 4.92 5.04 5.65
The United Kingdom 4.58 4.76 4.88
Source: Bureau van Dijk Bankscope database.
Table 3: Yearly Data Envelopment Analysis (DEA)
cost-efficiency measures obtained for all EU
member states
1999 2000 2001 2002 2003 2004
Austria 0.579 0.753 0.515 3.754 0.820 0.639
Belgium 0.930 0.921 0.804 1.000 1.000 0.937
Bulgaria 0.892 1.000 0.855 1.000 1.000 1.000
Croatia 0.853 0.886 0.728 3.793 0.620 0.596
Cyprus 0.308 0.415 0.377 3.431 0.401 0.425
Czech Republic 0.706 0.663 0.571 3.790 0.833 0.733
Denmark 1.000 1.000 0.443 3.735 1.000 0.706
Estonia 0.416 0.581 0.489 3.477 0.480 0.543
Finland 0.276 0.337 0.148 1.000 1.000 0.672
France 1.000 1.000 1.000 1.000 0.935 1.000
Germany 1.000 1.000 1.000 1.000 1.000 1.000
Greece 0.508 0.119 0.362 0.261 0.423 0.998
Hungary 0.380 0.522 0.431 0.493 0.512 0.413
Ireland 1.000 1.000 0.959 0.491 0.458 0.515
Italy 1.000 1.000 0.704 0.750 0.501 0.705
Latvia 0.601 0.636 0.466 0.53 0.563 0.807
Lithuania 0.781 0.749 0.644 0.887 0.875 0.947
Luxembourg 1.000 1.000 1.000 1.000 1.000 1.000
Malta 0.292 0.545 0.475 0.640 0.683 0.712
The Netherlands 0.541 1.000 1.000 1.000 1.000 1.000
Poland 1.000 0.752 0.622 1.000 1.000 0.541
Portugal 0.244 1.000 0.624 1.000 0.965 0.410
Romania 0.786 0.676 0.940 0.986 0.756 0.731
Slovak Republic 0.606 1.000 0.840 0.866 0.82 0.819
Slovenia 0.597 0.688 0.701 0.626 0.537 0.486
Spain 1.000 1.000 0.845 0.750 0.493 1.000
Sweden 1.000 1.000 1.000 1.000 1.000 0.610
The United Kingdom 1.000 1.000 1.000 1.000 1.000 1.000
2005 2006 2007 2008 2009 2010
Austria 0.655 0.666 0.669 0.735 0.682 0.622
Belgium 0.702 0.851 0.628 0.755 0.859 0.891
Bulgaria 1.000 1.000 1.000 0.880 0.818 0.734
Croatia 0.615 0.662 0.740 0.759 0.719 0.762
Cyprus 0.442 0.638 0.622 0.689 0.612 0.503
Czech Republic 0.751 1.000 0.957 0.942 0.893 0.784
Denmark 0.731 0.922 0.687 0.790 0.667 0.794
Estonia 0.772 0.858 0.729 1.000 0.655 1.000
Finland 0.637 0.730 0.611 0.934 0.855 1.000
France 1.000 1.000 1.000 1.000 1.000 1.000
Germany 0.853 0.886 1.000 0.842 1.000 1.000
Greece 0.991 0.969 0.885 0.892 0.951 1.000
Hungary 0.451 0.551 0.575 0.509 0.441 0.494
Ireland 0.804 0.883 0.623 0.801 1.000 1.000
Italy 1.000 1.000 0.925 0.834 0.864 0.948
Latvia 0.788 0.673 0.680 0.639 0.645 0.660
Lithuania 1.000 0.957 0.921 0.772 0.688 0.678
Luxembourg 1.000 1.000 0.964 1.000 1.000 1.000
Malta 0.768 0.932 0.869 0.903 1.000 0.921
The Netherlands 1.000 1.000 1.000 1.000 1.000 1.000
Poland 0.498 0.964 0.942 0.856 0.630 0.583
Portugal 0.511 0.598 0.568 0.580 0.607 0.623
Romania 0.668 0.646 0.726 0.592 0.581 0.611
Slovak Republic 1.000 0.976 1.000 1.000 1.000 1.000
Slovenia 0.585 0.669 0.707 0.688 0.705 0.692
Spain 1.000 1.000 1.000 1.000 1.000 1.000
Sweden 0.641 0.723 0.489 0.624 0.568 0.748
The United Kingdom 1.000 1.000 1.000 1.000 1.000 1.000
2011 2012 2013
Austria 0.788 0.610 0.975
Belgium 1.000 0.755 1.000
Bulgaria 0.873 0.708 0.952
Croatia 0.882 0.795 0.985
Cyprus 0.577 0.553 0.619
Czech Republic 0.994 0.924 1.000
Denmark 0.824 0.716 0.916
Estonia 0.918 1.000 1.000
Finland 1.000 1.000 1.000
France 1.000 1.000 1.000
Germany 1.000 1.000 0.935
Greece 1.000 1.000 1.000
Hungary 0.598 0.515 0.577
Ireland 0.911 0.578 0.886
Italy 0.927 0.994 0.997
Latvia 1.000 0.912 1.000
Lithuania 0.928 0.702 0.873
Luxembourg 1.000 0.676 1.000
Malta 1.000 1.000 1.000
The Netherlands 1.000 1.000 1.000
Poland 0.739 0.482 0.660
Portugal 0.706 0.575 0.862
Romania 0.678 0.560 0.677
Slovak Republic 1.000 0.937 0.840
Slovenia 0.813 0.711 0.837
Spain 1.000 1.000 1.000
Sweden 0.758 0.688 0.748
The United Kingdom 1.000 1.000 1.000
Source: Author's calculations using data sourced
from the Bureau van Dijk Bankscope database.
Table 4: Levin-Lin-Chu panel unit root test
Variables Panel 1 - EU 28
(P1 A: 1999-2013)
t-star P>t
Natural log of the per capita -5.26517 0.0000
gross national income
Nominal short-term interest rate -19.38331 0.0000
Government net lending/borrowing -3.74542 0.0001
Bank market concentration -4.55126 0.0000
Ratio of bank equity to bank total assets -6.95812 0.0000
DEA bank efficiency -8.40442 0.0000
Variables Panel 2 - EU 22
(P2 A: 1999-2013)
t-star P>t
Natural log of the per capita -5.28534 0.0000
gross national income
Nominal short-term interest rate -15.82860 0.0000
Government net lending/borrowing -3.56866 0.0002
Bank market concentration -4.02627 0.0000
Ratio of bank equity to bank total assets -6.39360 0.0000
DEA bank efficiency -7.92980 0.0000
Table 5: Westerlund panel cointegration test (p-values)
Cointegration between the Panel 1 - EU 28 (Pl-A:
series of variables 1999-2013)
Gt Ga Pt Pa
Natural log of the per capita gross 0.000 0.996 0.007 0.461
national income and nominal short-
term interest rate
Natural log of the per capita gross 0.207 1.000 0.739 0.976
national income and government net
lending/borrowing
Natural log of the per capita gross 0.006 0.169 0.000 0.000
national income and bank market
concentration
Natural log of the per capita gross 0.072 0.941 0.609 0.782
national income and ratio of bank
equity to bank total assets
Natural log of the per capita gross 0.008 0.770 0.016 0.158
national income and DEA bank
efficiency
Cointegration between the Panel 2 - EU 22 (P2-A:
series of variables 1999-2013)
Gt Ga Pt Pa
Natural log of the per capita gross 0.000 0.991 0.020 0.479
national income and nominal short-
term interest rate
Natural log of the per capita gross 0.709 0.999 0.816 0.971
national income and government net
lending/borrowing
Natural log of the per capita gross 0.050 0.324 0.002 0.000
national income and bank market
concentration
Natural log of the per capita gross 0.076 0.942 0.716 0.836
national income and ratio of bank
equity to bank total assets
Natural log of the per capita gross 0.064 0.957 0.039 0.284
national income and DEA bank
efficiency
Table 6: Results obtained with dynamic GMM one-step system estimations
Panel 1 - EU 28
PANEL 1-A (1999-2013)
P>
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.102199 -8.67 0.000
Government net lending/borrowing 0.0361498 3.21 0.001
Bank market concentration -0.0069661 -1.09 0.278
Ratio of bank equity to bank -0.093506 -7.32 0.000
total assets
DEA bank efficiency 1.394033 3.58 0.000
Constant 3.404441 5.53 0.000
Wald [chi square] (5) = 264.49
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -3.65
in first differences Pr > z = 0.000
Arellano-Bond test for AR(2) z = -1.04
in first differences Pr > z = 0.298
Sargan test of overid. [chi square] (23) = 171.71
restrictions Prob. > [chi square] = 0.000
Number of observations 420
PANEL 1-B (1999-2007)
P>
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.1170908 -6.23 0.000
Government net lending/borrowing 0.0100156 0.25 0.806
Bank market concentration 0.01993 1.81 0.070
Ratio of bank equity to bank -0.0562786 -2.42 0.016
total assets
DEA bank efficiency 3.064111 4.72 0.000
Constant -0.0279268 -0.03 0.980
Wald [chi square] (5) = 117.39
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -1.74
in first differences Pr > z = 0.083
Arellano-Bond test for AR(2) z = -1.39
in first differences Pr > z = 0.165
Sargan test of overid. [chi square] (11) = 86.51
restrictions Prob. > [chi square] = 0.000
Number of observations 252
PANEL 1-Cl (2008-2013)
P>
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.0757305 -4.23 0.000
Government net lending/borrowing 0.0533712 3.43 0.001
Bank market concentration -.0073907 -0.53 0.596
Ratio of bank equity to bank -0.1264874 -7.24 0.000
total assets
DEA bank efficiency 0.0328352 0.09 0.926
Constant 4.83344 4.52 0.000
Wald [chi square] (5) = 86.2
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -1.78
in first differences Pr > z = 0.075
Arellano-Bond test for AR(2) z = -1.40
in first differences Pr > z = 0.160
Sargan test of overid. [chi square] (5) = 5.94
restrictions Prob. > [chi square] = 0.312
Number of observations 168
Panel 2 - EU 22
PANEL2-A (1999-2013)
P >
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.0742289 -5.00 0.000
Government net lending/borrowing 0.0258307 1.43 0.153
Bank market concentration -.0121716 -1.83 0.067
Ratio of bank equity to bank -0.1006469 -6.85 0.000
total assets
DEA bank efficiency 2.576677 5.03 0.000
Constant 2.65505 4.05 0.000
Wald [chi square] (5) = 217.50
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -4.04
in first differences Pr > z = 0.000
Arellano-Bond test for AR(2) z = -1.36
in first differences Pr > z = 0.173
Sargan test of overid. Restrictions [chi square] (23) = 124.72
Prob. > [chi square] = 0.000
Number of observations 330
PANEL2-B (1999-2007)
P >
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.0654121 -2.61 0.009
Government net lending/borrowing 0.0487936 1.05 0.292
Bank market concentration 0.0129485 1.16 0.245
Ratio of bank equity to bank -0.0731273 -2.75 0.006
total assets
DEA bank efficiency 4.298334 5.36 0.000
Constant -0.681047 -0.60 0.546
Wald [chi square] (5) = 102.48
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -2.37
in first differences Pr > z = 0.018
Arellano-Bond test for AR(2) z = -2.89
in first differences Pr > z = 0.004
Sargan test of overid. Restrictions [chi square] (11) = 62.33
Prob. > [chi square] = 0.000
Number of observations 198
PANEL2-C (2008-2013)
P >
[absolute
coef. Z value of z]
Nominal short-term interest rate -0.1093013 -5.55 0.000
Government net lending/borrowing 0.0803335 3.99 0.000
Bank market concentration 0.0017324 0.21 0.834
Ratio of bank equity to bank -0.1620307 -9.14 0.000
total assets
DEA bank efficiency -0.7542794 -1.52 0.129
Constant 5.341475 7.58 0.000
Wald [chi square] (5) = 99.41
Prob. > [chi square] = 0.000
Arellano-Bond test for AR(1) z = -2.30
in first differences Pr > z = 0.021
Arellano-Bond test for AR(2) z = 0.72.
in first differences Pr > z = 0.472
Sargan test of overid. Restrictions [chi square] (5) = 11.12
Prob. > [chi square] 2 = 0.049
Number of observations 132
Dependent variable: natural logarithm
of the per capita national income.
