WPS6604
Policy Research Working Paper 6604
Background Paper to the 2014 World Development Report
Lending Concentration, Bank Performance
and Systemic Risk
Exploring Cross-Country Variation
Thorsten Beck
Olivier De Jonghe
The World Bank
Development Economics
Office of the Senior Vice President and Chief Economist
September 2013
Policy Research Working Paper 6604
Abstract
Using both market-based and annual report-based might also enhance stability. This paper finds that sectoral
approaches to measure lending specialization for a broad specialization increases volatility and systemic risk
cross-section of banks and countries over the period exposures, while not leading to higher returns. The paper
2002 to 2011, this paper is the first to empirically gauge also documents important time, cross-bank, and cross-
the relationship between bank lending specialization county variation in this relationship, which is stronger
and bank performance and stability in an international post 2007, for richer countries, countries without
sample. Theory suggests that banks might benefit from regulatory requirements on diversification, banks with
specialization in the form of higher screening and lower market power, and banks with more traditional
monitoring efficiency, while a diversified loan portfolio intermediation models.
This paper—prepared as a background paper to the World Bank’s World Development Report 2014: Risk and Opportunity:
Managing Risk for Development—is a product of the Development Economics Vice Presidency. The views expressed in this
paper are those of the authors and do not reflect the views of the World Bank or its affiliated organizations. Policy Research
Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at O.deJonghe@
uvt.nl, and T.Beck@uvt.nl.
The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.
Produced by the Research Support Team
Lending Concentration, Bank Performance and Systemic Risk:
Exploring Cross-Country Variation
Thorsten Becky Olivier De Jonghez
September 16, 2013
Keywords: sectoral specialization, lending concentration, bank risk, systemic stability, herding
JEL Classiﬁcations: G01, G21, G28, L5
The authors would like to thank Martin Melecky and seminar participants at the University of Glasgow for helpful comments
and Maria Dimitraki, Janneke Driessen, Christian Hassert, Joni Koch, Pim Meulendijks, Stefﬁe Vandenbosch, Tim van Rijn and
Daan van Zonsbeek for excellent research assistance. This paper was prepared as background paper for the World Development
Report 2014 entitled Risk and Opportunity: Managing Risk for Development. This paper and its ﬁndings do not necessarily reﬂect
the opinions of the World Bank, their Executive Directors or the countries they represent.
y
Cass Business School, Tilburg University, CEPR and European Banking Center. t.beck@uvt.nl
z
CentER, European Banking Center, Tilburg University. o.dejonghe@tilburguniversity.edu
1
1 Introduction
Sectoral specialization of production and employment varies widely across countries; often, though not ex-
clusively, related to the level of development (Imbs and Wacziarg (2003)). Different degrees of specialization
and concentration also have an impact on the scope for and the extent of lending concentration by banks. In
turn, lending concentration affects bank risk as well as banking system stability via different and possibly
opposing channels. This paper tests these effects empirically and documents how lending specialization is
related to bank performance (valuation and returns), bank-speciﬁc risk (total volatility), as well as systemic
risk (the marginal expected shortfall). While lending specialization has often been ﬂagged by academics and
regulators as a critical dimension in banks’ performance and stability, research has been hampered by the
dearth of data. Using both market-based and annual report-based approaches to measure lending specializa-
tion for a broad cross-section of banks and countries over the period 2002 to 2011, this paper is the ﬁrst to
empirically gauge the relationship between bank lending specialization and bank performance and stability.
Lending concentration by banks has two opposing effects on banks’ risk-taking incentives and hence on
their stability. On the one hand, the traditional portfolio theory view posits that diversiﬁcation largely elimi-
nates the impact of idiosyncratic shocks on banks’ loan portfolio. On the other hand, lending specialization
can also result in better screening of potential borrowers and loan applications and more efﬁcient monitor-
ing, hence leading to lower default risk and higher (risk-adjusted) returns. Focused banks will gain expertise
in the sectors they lend to, and hence can detect a deterioration of the borrower’s business earlier and may
react in a timely manner by risk mitigation (for example, by requesting additional collateral). Moreover, a
credible threat of better monitoring skills might also prevent risk-shifting by borrowers, as in Stiglitz and
Weiss (1981). However, lending concentration or diversiﬁcation not only affects bank-speciﬁc risk, but also
the stability of the entire sector. In particular, from a systemic point of view, a third channel may play an
important role. In countries where the scope for lending diversiﬁcation is limited, banks’ loan portfolios
will be more similar to each other, leading to a more homogeneous ﬁnancial system. Furthermore, even
1
when diversiﬁcation is feasible, banks may have incentives to herd, thus causing a lack of diversity. This
lack of diversity is potentially more costly for society as it implies that similar institutions will more likely
face problems at the same time. On the other hand, recent work has pointed to risks from diversiﬁcation, if
all ﬁnancial institutions diversify into the same portfolio (Wagner (2010)).
We gauge the relationship between lending specialization and bank performance and stability with two
novel data sets. First, we rely on stock return-based indicators of sectoral factor exposures. Using an
extended market model, we test whether banks are well diversiﬁed and are only exposed to the returns on
a broad market index, or whether they additionally exhibit signiﬁcant exposures to certain sector-speciﬁc
portfolios. This method is similar in spirit to returns-based style analysis, which is a statistical technique
mainly used to deconstruct mutual fund returns in exposures to investment strategies or asset classes (e.g.
with respect to large versus small stocks or value versus growth stocks, (see e.g. Sharpe (1992), Brown and
Goetzmann (1997) and ter Horst et al. (2004)). Second, we also construct a hand-collected database on the
sectoral exposures of the largest banks based on the information they report in the notes to their ﬁnancial
statements. We limit this analysis to listed banks with total assets in excess of US$ 10 billion as these are
more likely to publish a detailed report on their website, and ﬁnd useful information for 317 banks over the
period 2007-2011. Both the return-based and accounting-based sectoral exposures are then used to compute
several lending specialization indicators, such as focus of the portfolio (standard deviation or HHI of the
exposures) or herding with other banks (Euclidean distance measure).
Our main ﬁndings are that sectoral specialization ia positively related to bank risk, while not being
associated with higher returns. Moreover, it is associated with higher systemic risk exposures as well as
increases in total volatility. Hence, it seems that the portfolio diversiﬁcation gains outweigh the potential
beneﬁts of screening and monitoring efﬁciency in specialized lending. Furthermore, we ﬁnd that dissimi-
larity with respect to other banks is also associated with higher risk along all dimensions, in contrast with
the theoretical predictions. The results are similar using either the return-based or hand-collected sectoral
2
exposure indicators, notwithstanding differences in sample period (2002-2011 versus 2007-2011), sample
size (2030 versus 317 banks) and methodology (bank ﬁxed effects or not).
Sample splits show that our ﬁndings hold for the whole sample period, but are stronger for the post-2007
crisis period. A major advantage of our database compared to related studies is its international dimension,
which allows examining cross-country differences in the lending specialization-performance relationship.
We document that the relationships between sectoral specialization, valuation, volatility and systemic risk
contribution are driven by developed countries and countries with no regulatory requirements on diver-
siﬁcation, banks operating in more competitive markets and banks focusing on traditional intermediation
business. These ﬁndings suggest that the regulatory response to our ﬁndings cannot be "one size ﬁts all",
but rather has to be tailored to bank and country circumstances.
The relationship between lending concentration and bank performance and stability is not only interest-
ing for academics but has been central to policy and regulatory discussions. Historical experience shows
that concentration of credit risk in asset portfolios has been one of the major causes of bank distress (e.g. by
being overexposed to Enron, Worldcom and the likes). According to a 2004 Basel committee study, credit
concentration caused nine of the 13 major banking system crises around the world in the twentieth century,
resulting in calls for a revised regulatory approach to sectoral concentration to overcome one of the main
shortcomings of the ﬁrst Basel Accord (i.e. ignoring the potential consequences of this specialization within
banks’ credit portfolios). Consequently, the second Basel agreement incorporated adjustments regarding the
impact of bank lending specialization, though only in the second pillar on the supervisory review process
rather than in the ﬁrst pillar of capital requirements. In “Basel 2.5”1 and Basel III, there have been no major
adjustments regarding concentration risk.
Our paper is related to the literature on lending concentration. Concentration risk could stem from either
1
“Basel 2.5” is the intermediate change in capital requirements that came into force on December 31, 2011 (e.g. by means of
the second and third Capital Requirement Directive in the EU).
3
imperfect granularity (i.e. exposure to large single names) or imperfect sectoral diversiﬁcation. Both lead
to deviations from the asymptotic single risk factor framework. Empirical evidence for developed countries
indicates that the impact on economic capital is larger for sectoral concentration than for name concentration.
Using German data (but with a hint that the conclusions are generalizable to other continental European
banks), Duellmann and Masschelein (2007) ﬁnd that economic capital increases from 7.8% in the case of
the most diversiﬁed benchmark portfolio to 11.7% for a portfolio concentrated in one sector. However,
there is also theoretical and empirical evidence that shows how lending specialization may be beneﬁcial and
reduce risk or increase (risk-adjusted) returns. Winton (2000) shows theoretically that it is likely that the
bank’s monitoring effectiveness is lower in new sectors, with the effect that diversiﬁcation lowers average
returns on monitored loans, as banks are less likely to improve monitoring incentives, and is more likely
to increase the bank’s chance of failure. The aforementioned effects theoretically imply a reduction in the
probability of default. Empirical evidence by Acharya et al. (2006) for Italy and by Hayden et al. (2007)
for Germany documents that specialization in certain industries is indeed accompanied by lower loan loss
rates. Boeve et al. (2010) ﬁnd that German cooperative and saving banks exert more and better monitoring
if they are specialized rather than diversiﬁed. Empirical evidence from Brazil, by Tabak et al. (2011) also
hints to the fact that loan portfolio concentration seems to improve the performance of banks in both return
and risk of default. In addition, these authors also document that the loan portfolios of Brazilian banks are
more concentrated compared to e.g. Germany, Italy and the U.S. While the existing literature focuses either
on single countries or syndicated lending (Cai et al. (2013)), our paper is the ﬁrst cross-country study on
the relationship between lending specialization and bank performance and risk, which also allows testing
for cross-country variation in the relationship. While other studies have focussed on banks’ diversiﬁcation
in interest and non-interest business, we use unique and novel data to shed light on lending specialization.2
2
See De Jonghe (2010), Demirguc-Kunt and Huizinga (2010) and Stiroh and Rumble (2006), among others, for studies on
interest vs. non-interest business of banks.
4
The remainder of the paper is structured as follows. The next section introduces our sample and unique
(hand-collected) data. In Section 3, we outline the methodology and describe the estimated impact of two
different sets of specialization measures on bank performance. In Section 4, we exploit the multiple country
dimension of the database and examine whether the impact of lending specialization on performance is
conditional on certain country-speciﬁc or bank-speciﬁc characteristics. Section 5 concludes the paper with
policy implications and avenues for further research.
2 Data
In the analysis, we combine data from several sources. We obtain information on banks’ balance sheets and
income statements from Bankscope, which is a database compiled by Fitch/Bureau Van Dijck that contains
information on banks around the globe, based on publicly available data-sources. Bankscope contains infor-
mation for listed, delisted as well as privately held banks. While Bankscope does not contain stock market
information on a daily basis, it does contain information on the ticker as well as the ISIN number of (de)listed
banks’ equity, which enables matching Bankscope with Datastream. From Datastream, we retrieve infor-
mation on a bank’s stock price as well as its market capitalization. The combined Bankscope-Datastream
sample, cleaned for missing items on variables of interest, yields 12; 689 observations, on 2; 005 banks from
77 countries over the period 2002 2011. We include commercial banks, bank holding companies, as well
as saving banks and cooperatives.
Our independent variables of interest are several proxies of the degree of sectoral specialization of
a bank’s loan portfolio. These data are not directly available from (commercial) databases for a cross-
country sample of banks.3 Therefore, we take a two-pronged approach. First of all, we rely on return-
3
Authors of studies on lending concentration have either used conﬁdential data gathered by the central bank’s credit register (for
single country studies) or relied on syndicated loan exposures (e.g. Cai et al. (2013)). In the latter case, the sample is limited to the
subset of very large, internationally active ﬁnancial instutions (which are mainly located in the US). Moreover, the exposures are
5
based indicators of sectoral factor exposures (subsection 2.1). Secondly, we construct a hand-collected
database of the sectoral exposures reported by the largest banks in the notes to their ﬁnancial statements.
The procedure to hand-collect the accounting-based sectoral shares will be exposed in subsection 2.2. To
gauge the relationship between sectoral specialization and bank performance, we use various dimensions
of bank risk and return. The construction of these dependent variables is described in subsection 2.3. The
correlations between both types of sectoral specialization measures as well as their relationship with bank
performance measures are described in subsection 2.4.
2.1 A stock return-based approach to measuring banks’ sectoral exposures and lending
specialization
A bank’s stock price is inﬂuenced by exposures to systematic risk as well as idiosyncratic news. If a bank
holds a well-diversiﬁed loan portfolio, then its stock return should mainly co-move with returns on a broad
market-wide index. On the other hand, if a bank’s loan portfolio is (over)exposed to certain sectors, then
the bank’s stock return should not only react to economy-wide shocks, but also to sector-speciﬁc news.
Using an extended market model, we test whether banks are well diversiﬁed and are only exposed to the
market index, or whether they additionally exhibit signiﬁcant exposures to certain sector-speciﬁc portfolios
(and hence violate the assumption underlying the asymptotic single risk factor framework). This method is
similar in spirit to returns-based style analysis, which is a statistical technique mainly used to deconstruct
mutual fund returns in exposures to investment strategies or asset classes (see e.g. Sharpe (1992), Brown
and Goetzmann (1997) and ter Horst et al. (2004)). These exposures are then interpreted as a measure of
a fund or portfolio manager’s style (e.g. with respect to large versus small stocks or value versus growth
stocks). A similar approach is used by Acharya and Steffen (2013) to infer European banks’ sovereign risk
exposure from asset prices. They relate banks’ stock returns to yields on German government debt and
then limited to the syndicated loans, which may not be representative for the overall portfolio of commercial and industrial loans.
6
yields on GIIPS countries’ debt, to obtain market-based indicators of banks’ exposures to sovereign risk.
In particular, we estimate the following equation for each bank:
S
X
M s s
ri;t = c + rt + rt + "i
t (1)
s=1
M ) as well as on the
We regress a bank’s stock return (ri;t ) on the returns on a broad market index (rt
s ). The sectoral indices are based on the Industry Classiﬁca-
return to S (=10) different sectoral indices (rt
tion Benchmark (ICB). More speciﬁcally, we use the level 2 decomposition, which divides the total market
into 10 industries: oil & gas, basic materials, industrials, consumer goods, healthcare, consumer services,
telecommunications, utilities, technology, and ﬁnancials. As we are interested in exposures to sector-speciﬁc
news (and not the movement in sectoral indices due to economy-wide news), we ﬁrst orthogonalize each
s series with respect to market-wide returns (r M ) and the ﬁnancial sector returns.4 Doing so, we
of the rt t
clean the sectoral returns from market-wide news as well as their dependence on ﬁnancial sector (shocks).
Subsequently, we standardize the orthogonalized exposures, which facilitates comparing the exposures to
s
different industries. The estimated coefﬁcients then reﬂect both the exposure to as well as the riskiness
(volatility) of the sectoral shocks. The residual, "i
t , captures the idiosyncratic or bank-speciﬁc news compo-
nent. We estimate Equation (1) for each bank and for each year using daily returns, such that we end up with
a panel database on sectoral exposures that varies at the bank-year frequency. The panel dataset of estimated
exposures consists of 12; 689 bank-year observations, covering 2; 005 banks from 77 countries over a ten
year period starting in 2002. We do not impose constraints on the coefﬁcients and hence allow that a bank
has a negative exposure to, and hence is short in, a speciﬁc industry. Information on the estimated exposures
is reported in Table 1.
4
The returns on the ﬁnancial sector index are orthogonal with respect to the market.
7
It is important to note that there is an asymmetry in the interpretation of signiﬁcant and insigniﬁcant
factor loadings. While signiﬁcant factor loadings can be interpreted as implying (over)exposure to a speciﬁc
sector, ﬁnding a zero (or non-signiﬁcant) exposure on average can be due to three different reasons. First,
banks are opaque and stock market participants are not able to make an accurate assessment (hence imprecise
and insigniﬁcant estimates). Second, banks are transparent (to stock market investors) but do not have an
imbalanced loan portfolio (precise, but zero, estimates). Third, banks may specialize in certain sectors, but
could use derivative contracts to hedge these (over)exposures (precise zero estimates, but different from
sectoral composition).
Panel A of Table 1 reports for each estimated factor loading the mean and standard deviation across
12; 689 observations, as well as the 5th , 50th ; and 95th percentile of the panel of estimated factor loadings.
In panel B, we report for each factor the relative frequency of observing t-statistics in ﬁve groups. We
consider both conventional signiﬁcance levels (95%, absolute value of t-statistic above 1:645) as well as a
weaker threshold (absolute value of t-statistic in excess of one) to account for the fact that the exposures
are estimated over relatively short windows (one year) of daily return data (which can be noisy) and have
been orthogonalized (hence estimation error) in a previous step. As illustrated in Panel A, the average and
median exposure is close to zero for all but two sectors (i.e. utilities and ﬁnancials). This indicates that the
stock market believes that banks are, on average, not exposed to shocks to these sectors. Unsurprisingly, the
exposure to the ﬁnancial sector is larger and it is more likely that the t-statistic with the associated exposure
will exceed 1 or even 1:645 (for 22% of the banks). As indicated in Panel B, for each of the sectoral factor
loadings, we ﬁnd substantial heterogeneity across bank-year pairs and many of these sector exposures are
often statistically signiﬁcant. Speciﬁcally, the share of signiﬁcant t-statistics (at the 5% level) ranges from
15% in the oil&gas and industrial sectors to 28% in the ﬁnancial sector.
Based on the estimated coefﬁcients of Equation (1), we compute several time-varying bank-speciﬁc
measures of the intensity of sectoral specialization. More speciﬁcally, for each bank and for each year, we
8
calculate the following measures. First, we count the number of sectoral exposures with a t-statistic (in
absolute value) larger than one, thus ranging from zero to ten. We label this measure: Signiﬁcant Sectoral
Factors. Second, we compute the contribution, of the sectoral factors (excluding the contribution of the
ﬁnancial sector) to the R-squared of the return-generating model (Sectoral Contribution to R2 ). A larger
value indicates a larger exposure to sector-speciﬁc news that is not created by economy-wide or ﬁnancial
events. The ﬁrst two indicators also indicate the extent to which the asymptotic single risk factor assumption
is valid for a given bank in a given year. Third, we construct the measure labelled Dispersion (factors) which
s
captures the dispersion in the estimated sectoral exposures (standard deviation in the ten coefﬁcients).
Fourth, we compute a measure of differentiation (or its opposite: similarity or herding) by banks within a
country. For each bank, we compute the Euclidean distance between a bank’s estimated sectoral exposures
and the country-average (excluding that bank) of the sectoral exposures. The Euclidean distance is computed
as follows:
v 0 12
u
uXS Ij
u
@ s 1 X s A
Dis tan cei;j;t =t i;j;t i;j;t (2)
Ij
s=1 i=A
where Ij is the number of banks in country j . The measure, labelled Differentiation (factors), will
be larger the more the bank’s sectoral exposures deviate from the average bank in the country. A similar
measure has also been used by Cai et al. (2013) to measure bank herding based on syndicated loan exposures.
We report summary statistics on these measures in panel C of Table 1. We ﬁnd that the average bank has
four signiﬁcant sectoral factor loadings in a given year, that differ substantially from each other (as indicated
by the specialization measures) and which lead to a substantial increase in R-square of 7:6% compared to
a model that only includes the market factor which has an average R-square of 9% (the ﬁnancial sector
contributes an additional 1%). The average dispersion is 31% and the average bank’s differentiation from
the country-average is 1.5 factors. More importantly, all measures exhibit substantial variation, which will
enable us to assess how these measures are related with our proxies for bank performance and stability.
9
2.2 Constructing a hand-collected database of reported sectoral exposures
Detailed information on banks’ loan composition is hard to obtain from publicly available or commercial
databases. Typically, one can ﬁnd a breakdown in real estate, consumer or business loans.5 However, in
general, there is no information on the sectoral composition of the business loan portfolio. Two exceptions
are the credit registers maintained by some central banks and syndicated loan exposures. The former is,
however, conﬁdential, only available for few countries and does not allow cross-country comparisons. The
latter is limited to very large loans by very large banks. Nevertheless, many banks provide information on
their sectoral exposures in the notes to their ﬁnancial statements.
The breakdown can be very detailed, but the level of detail can vary by bank and country as there is
no required ﬁnancial reporting format for these exposures. We hand-collect information on these exposures
according to the following procedure. Starting from the universe of banks covered by Bankscope, we impose
the following constraints: (i) banks need to be active in 2013, i.e. not have failed during the recent crisis; (ii)
banks need to have publicly traded equity; (iii) banks need to have total assets in excess of 10 billion US$ in
2011; (iv) we only keep commercial banks, savings banks, cooperative banks and bank holding companies;
and (v) information on basic characteristics, such as: common equity, total assets, the net interest margin,
loan loss provisions as well as a liquidity ratio are non-missing for the period 2009, 2010, and 2011.
This selection results in a sample of 435 banks. We focus on large, listed banks as these are more likely
to publish a detailed report on their website. However, this is not the case for all selected banks. The ﬁnal
database therefore consists only of banks for which the reports published on their website contain useful and
detailed information on the sectoral exposures (317 from the 435 banks). To harmonize the heterogeneity in
5
Liu (2011) investigates herding behaviour in bank lending by US commercial banks and looks at similarities in banks’ loan
exposures to ﬁve categories (commercial real estate, residential real estate, consumer and industrial loans, individual loans and
all remaining loans. He uses the Lakonishok et al. (1992) herding measure, which is initially developed to analyse herding by
institutional investors their buy and sell signals.
10
the sectoral breakdown across banks, we categorize each reported exposure in ten economic sectors based on
the one-digit Standard Industrial Classiﬁcation. Personal/consumer loans, loans to central governments and
interbank loans were excluded. The data are collected as meticulous as possible, but nevertheless subject to
some researcher-speciﬁc choices. For example, if the reported information is at a coarser level than the SIC
one-digit level (e.g. ‘Agriculture and Mining’), we divide the reported amount over the two separate sectors
(i.e. half of the exposure to ‘Agriculture’ and the other half to ‘Mining’).
The data collection yields a panel of accounting-based sectoral exposures at the bank level for the years
2007 2011. Summary statistics on these exposures are reported in panel A of Table 2. For each sector,
we report the mean, standard deviation, 5th , 50th ; and 95th percentile. There is variation in the average
exposure across the ten sectors, with the lowest average for the sector “Agriculture, forestry and ﬁshing”
and the largest one for “manufacturing”. Within each sector, there is substantial heterogeneity. The value
of the 5th percentile is almost always zero, whereas the exposure to manufacturing for the bank at the 95th
percentile is 37%.
Based on these hand-collected exposures, we construct indicators of industrial specialization in lending
by banks. These measures are reported in Panel B of Table 2. In particular, we capture various aspects of
lending specialization by (i) the dispersion in the reported sectoral exposures (standard deviation of the ten
shares), labelled Dispersion (accounting), (ii) the cumulative share of the three largest sectoral exposures
(Sectoral CR3), (iii) a Hirschmann-Herﬁndahl index (Sectoral HHI ) of industrial specialization6 , and (iv)
a proxy for the amount of differentiation (herding) in sectoral exposures at the country level. This measure,
Differentiation (accounting), is computed as the Euclidean distance between a bank’s sectoral loan portfolio
6
The HHI is measured as the sum of the squared sectoral shares. A higher value of the HHI indicates more concentration or
inequality. We also compute modiﬁed version of the HHI in which we exclude the “Others” category, as this might range from
nearly granular to a single exposure. The results are similar for both measures.
11
and the average bank’s sectoral composition (as in Equation 2, but replacing the estimated factors with
reported shares). The more similar the exposures, the lower the value of the measure and the higher the
exposure to common shocks.
The summary statistics of these measures in Panel B of Table 2 indicate that there is considerable het-
erogeneity across banks. Speciﬁcally, the standard deviation of sectoral exposures, Dispersion (accounting),
varies largely with a value of 0:068 at the 5th percentile and a value of 0:184 at the 95th percentile (with
a mean of 0:113). The cumulative exposure of the largest three sectors varies from 53% (5th percentile)
to 96% (95th percentile), with a mean of 70%. The HHI concentration ratio has an average of 0:227 and
a standard deviation of 0:088. Finally, Differentiation (accounting) also exhibits substantial cross-sectional
variation. The Euclidean distance between a bank’s exposure and the country’s average exposure ranges
from 0:09 to 0:55 (5th and 95th percentile), with a mean of 0:26.
2.3 Measures of bank performance and stability
Using stock return-based measures, we gauge several aspect of bank performance.7 In particular, we will
look at bank valuation, bank risk as well as exposure to systemic risk. More speciﬁcally, we will employ the
following dependent variables in our analysis. First, bank performance is gauged by the annualized average
daily return over a calendar year, thus measuring proﬁtability for shareholders. Second, volatility, measured
as the annualized standard deviation of a bank’s daily stock returns over the span of a calendar year, captures
7
We prefer capital market data to accounting data because equity prices are forward-looking and hence better identiﬁers of
prospective performance and risks associated with different strategic choices. In addition, accounting proﬁts reﬂect short-run
performance, rather than capturing long-run equilibrium behavior. Furthermore, accounting-based proﬁt (such as return on assets
or return on equity) adn risk measures may be noisy measures of ﬁrm performance as a result of differences in tax treatment and
(discretion over) accounting practices across countries, or different provisioning and depreciation practices. Noise and biases in the
dependent variable may result in low values of goodness-of-ﬁt tests in basically all empirical setups (Smirlock et al. (1984), Stevens
(1990)).
12
a bank’s total risk exposure. Third, to capture the return-risk trade-off in one metric, we will also employ
a measure of a bank’s franchise value, proxied by the ratio of market capitalization to the book value of
common equity. Finally, we estimate a bank’s systemic risk exposure using the Marginal Expected Shortfall
(Acharya et al. (2012)). Mathematically, the MES of bank i at time t is given by the following formula:
Q
M ESi;t (Q) = E [ri;t jrm;t < V aRm;t ] (3)
where ri;t denotes the daily stock return of bank i at time t, rm;t the return on a stock index at time
t and Q is an extreme percentile, such that we look at systemic events. Following common practice in
the literature, we compute MES using the opposite of the returns such that a higher MES means a larger
systemic risk exposure. In this paper, we measure MES for each bank-year combination and follow common
practice by setting Q at 5%. Doing so, M ESi;t corresponds with bank i’s expected equity loss per dollar in
year t conditional on the market experiencing one of its 5% lowest returns in that given year. Conceptually,
MES measures the increase in the risk of the system induced by a marginal increase in the weight of bank i
in the system.8 The higher a bank’s MES (in absolute value), the higher is the contribution of bank i to the
risk of the banking system. We deﬁne the banking system as the local, country-speciﬁc banking market. In
addition, the bank for which we compute the MES is excluded from the banking sector index.9
Summary statistics on these variables are reported in Table 3. The upper panel contains the summary
statistics for the full sample. The lower panel reports the summary statistics on the smaller sample of banks
X
N h i
8 Q
The Expected Shortfall of the market portfolio is given by: wi;t E ri;t rm;t < V aRm;t and is hence equal to the
i=1
weighted sum of the MES of all banks in the system. The ﬁrst derivative of the Expected Shortfall of the market portfolio with
respect to wi;t equals the MES of bank i at time t.
9
We also compute the MES when the global, rahter than country-speciﬁc, banking sector experiences distress. All results in
the paper reported in the paper are robust to using either the local or the global banking sector as the conditioning variable in the
marginal expected shortfall measure.
13
(and shorter period) for which we hand-collect the sectoral exposures. Information on the countries included
in the sample as well as the number of bank-year observations by country is reported in Appendix A. Over
our larger sample over the period 2002 to 2011, the annualized average stock return was 0:29%, with a
volatility of 39:3%. Both variables, however, show a large variation across banks and years. The market-to-
bank value of equity shows an average of 1:11; but ranges from 0:01 to 3:10. The average MES with respect
to the local market is 2:00. The dependent variables in the smaller and shorter (post-)crisis sample (2007 to
2011) show, on average, worse performance, higher volatility, lower market-to-book value (below one!) and
a higher MES. The descriptive statistics in Panel B show a much lower average return of 5:83%, reﬂecting
that the sample period of the smaller sample is dominated by the crisis, a similar average volatility of 42:2%,
a low average franchise value of only 0:58 and a MES of 3:94, reﬂecting the fact that our smaller sample is
dominated by large banks.
2.4 Relating performance to specialization: Exploring pairwise correlations
In the three previous subsections, we introduced the performance metrics as well as two sets of sectoral
specialization indicators, respectively based on return-based sectoral factor exposures and accounting-based
sectoral lending shares. In this subsection, we present a ﬁrst exploration of the relationship between these
variables by means of pairwise correlations. We repeat, for convenience, the deﬁnitions of the above-
mentioned measures in Table 4.
Table 5 contains three panels of correlation matrices. In the upper panel A, we report the correlation
coefﬁcients among and between the performance measures and the sectoral specialization measures based
on factor loadings (large sample). In the middle panel B, we report the correlation coefﬁcients among and
between the performance measures and the sectoral specialization measures based on accounting exposures
(small sample). In panel C, we report the correlation coefﬁcients among and between all sectoral specializa-
14
tion measures (for the small sample). The number-letter combination in the column headers (which differ
by panel) correspond with a variable in the rows. In each panel the correlation coefﬁcient as well as the
p-value are reported.
Focussing ﬁrst on the pairwise correlations between factor-based specialization measures and bank per-
formance, we ﬁnd that Dispersion and Differentiation correlate negatively with average stock returns and
positively with total volatility. Sectoral Contribution to R2 correlates negatively and signiﬁcantly with re-
turns, while Sectoral Factors correlates positively and signiﬁcantly with volatility. The franchise value
(market value or net worth scaled by book value of equity) is a market-based risk-adjusted return indica-
tor. The negative (positive) relation of sectoral specialization and average return (volatility) translates into
a negative relationship between sectoral specialization, as measured by Sectoral Contribution to R2 , Dis-
persion and Differentation and the franchise value. Finally, signiﬁcant sectoral exposures, dispersion and
differentiation are all correlated with a higher exposure to systemic risk, as measured by the marginal ex-
pected shortfall. All pairwise correlation coefﬁcients between the four return-based sectoral specialization
measures are positive (except one) and signiﬁcant. However, the correlation is far from perfect, indicating
that the information content of each of these measures is slightly different.
The correlation coefﬁcients in the middle panel B indicate that the accounting-based lending dispersion,
concentration or differentiation measures are also positively and signiﬁcantly correlated with total volatil-
ity and MES, even though the data source is different, the sample period is shorter and the set of banks is
smaller. None of the specialization measures is signiﬁcantly correlated with average return. In contrast to
panel A, we ﬁnd in panel B that the franchise value is positively correlated with lending specialization. Note
that this opposite result is most likely caused by the franchise value measure. The signs of the correlation
between the franchise value and each of the three other performance metrics is also reversed in panel B
compared to panel A (while the sign of the correlation coefﬁcients between these three is the same in both
15
panels). We conjecture that this is driven by the fact that the smaller sample is driven by the crisis experi-
ence. Furthermore, each of the four accounting-based specialization measures is positively and signiﬁcantly
related with one another. The correlation coefﬁcients for the ﬁrst three indicators (dispersion, sectoral CR3
and sectoral HHI) are high and above 85%. Only the accounting differentiation measure seems to capture a
somewhat different aspect.
Finally, in the lower panel C of Table 5, we report the pairwise correlation coefﬁcients between all the
specialization indicators. While the correlation is almost always strongly positive and signiﬁcant within a
group, it is low and mostly insigniﬁcant across both groups. The market-based measures Signiﬁcant Sectoral
Factors and Dispersion (factors) are uncorrelated with each of the four accounting measures. The measures
Differentiation (factors) is positively and signiﬁcantly related with three accounting-based lending special-
ization measures. In contrast, the Sectoral Contribution to R2 is negatively and signiﬁcantly related to three
accounting-based sectoral specialization measures. This suggests that the market-based and accounting-
based measures capture different dimensions of lending specialization, as we already alluded to above. The
market-based measures have the advantage that they can also capture exposures of banks through market
and hedging operations in addition to asset-based exposures. On the other hand, these measures might
include noise if stock market participants are not able to make an accurate assessment. The accounting-
based measures are direct indicators of loan portfolio exposures, but do not capture hedging operations that
banks might be able to undertake. Recognizing the differences between these two approaches, we gauge
the relationship between sectoral specialization and bank performance across these two different groups of
indicators and samples.
3 Results
The contribution of this paper is to assess how lending specialization affects bank performance. To that
end, we will relate both the return-based and the accounting-based measures of sectoral specialization to
16
the variables that capture the various dimensions of bank performance, while controlling for other bank
characteristics that may affect bank performance. More speciﬁcally, we will estimate regressions of the
following form:
P erf ormancei;t = 1 Specializationi;t 1 + Xi;t 1 controls + ui + vt + "i;t (4)
The independent variables are lagged one year to mitigate concerns of reverse causality. We winsorize
all variables at the 1 and 99 percentile level to mitigate the impact of outliers. Next to the set of control
variables, discussed below, we also include year ﬁxed effects as well as bank ﬁxed effects. The standard
errors are clustered at the bank level. Before gauging the relationship between sectoral specialization and
bank performance, we will ﬁrst present the results of a regression of each dependent variable on the set of
control variables only (i.e. without a specialization measure). This serves two purposes. First, it further
describes the data and facilitates comparability with other papers. Second, in a subsequent subsection, we
will add the various sectoral specialization indicators one-by-one and will for the sake of space omit the
reporting of the results on the control variables.
3.1 Initial regressions: Leveling the playing ﬁeld
In Table 6, we show the results of regressing each dimension of bank performance on the set of control
variables. The control variables, of which the summary statistics are reported in Table 3, capture various
dimensions of a bank’s business model that might inﬂuence performance and stability. First of all, we
include bank size and non-interest income. The former is computed as the natural logarithm of total assets
expressed in 2007 US dollars.10 We measure a bank’s share of non-interest income to total operating income,
by dividing other operating income (which comprises trading income, commissions and fees as well as all
10
While most of the bank-speciﬁc variables are ratios, variables in levels (such as size) are expressed in 2007 US dollars (mil-
lions). Furthermore, bank size is highly correlated with many other bank characteristics, affecting the point estimates and the
magnitude of the standard errors. We therefore orthogonalize bank size with respect to all other control variables.
17
other non-interest income) by the sum of interest income and other operating income. The other bank-
speciﬁc variables are proxies for leverage (capital-to-asset ratio), the funding structure (share of deposits
in sum of deposits and money market funding)11 , asset mix (loans to assets ratio), proﬁtability (return-on-
equity), annual growth in total assets as well as expected credit risk (Loan Loss Provision to Total Assets).
These variables are often used in other studies that gauge the performance and stability of banks, including
Demirguc-Kunt and Huizinga (2010), Laeven and Levine (2009) or Beck et al. (2013).
In the left hand side part of Table 6, we report the regression results of the full sample, that will be
used when analyzing the impact of the factor-based specialization measures. We ﬁnd that larger banks
have lower returns on average but also less volatile returns, have lower market-to-book values, and higher
MES. Similarly, banks with higher non-interest income shares have lower but also less volatile returns,
have lower market-to-book values, and higher MES, although the last two results are not signiﬁcant. Better
capitalized banks have higher returns, higher market-to-book values and lower MES, while the relationship
with volatility is not signiﬁcant. Banks with higher loan-to-asset ratios have higher average but also more
volatile returns, have higher market-to-book values and lower MES. Banks with higher returns on equity in
the previous period have lower average but also less volatile returns, have lower market-to-book values and
higher MES. Banks with higher asset growth in the previous period experience higher returns and higher
market-to-book values. Banks with higher loan loss provisions, ﬁnally, have lower average returns, lower
market-to-book values and higher MES. The R2 vary substantially across the four regressions, with our
control variables explaining the highest share of variation in the market-to-book value regression (76:8%),
while they only explain 34% in the average return regression. The reported ﬁndings are, in general, as
11
Using several proxies for access to deposits and the use of bank deposits, Han and Melecky (2013) ﬁnd that greater access to
bank deposits can make the deposit funding base of banks more resilient in times of ﬁnancial stress and will hence also affect bank
performance and banks’ exposure to systemic risk.
18
expected and in line with other studies.
In the right hand side part of Table 6, we report the regression results for the smaller sample and shorter
(crisis and post-crisis) period. The results are mostly different, except for total volatility. Recall that the
sample now only spans ﬁve years, of which a large part is dominated by the global ﬁnancial crisis. Moreover,
the combination of a short sample period and bank ﬁxed effects gives less power to solely rely on the within
variation to identify signiﬁcant relationships. This is also apparent from the fact that the R2 stays high
(due to the bank and year ﬁxed effects), irrespective of the lack of signiﬁcant bank-speciﬁc characteristics.
Hence, irrespective of the quality of the return-based versus accounting-based measures, it will be harder to
identify signiﬁcant relationships for the latter compared to the former, given the shorter and smaller sample.
3.2 Lending specialization and bank performance
To gauge the relationship between lending specialization and bank performance and fragility, we add each of
the four return-based sectoral specialization indicators, discussed above, one-by-one to the regression setup
reported in the previous subsection. This results in 16 different speciﬁcations. We summarize the results in
Table 7, reporting only the information of interest. More speciﬁcally, for each regression, we only report
the coefﬁcient and t-statistic on the return-based sectoral specialization indicator as well as the number of
observations and the adjusted R2 of the regression. In each regression, we add the control variables and
bank and year ﬁxed effects as in Table 6. The standard errors are clustered at the bank level. The table is
constructed so that each of the four sectoral specialization indicators is reported in a different column. The
dependent variable varies by block of rows.
We do not ﬁnd any signiﬁcant relationship between annualized average returns and lending specializa-
tion of banks. In all four regressions with returns as dependent variable is the standard error of the lending
specialization variable higher than the coefﬁcient, leading to t-statistics well below one (in absolute value).
19
Specializing in speciﬁc sectors is therefore not reﬂected in stock returns. There is a signiﬁcant relationship
between the volatility of stock returns and lending specialization, as indicated in the second block of results.
Speciﬁcally, banks whose returns react more to sectoral indices (i.e. where sectoral factors contribute more
to the R2 ), for which we ﬁnd a higher dispersion of sectoral betas in the return regression, and which show a
higher factor-based differentiation (i.e. Euclidian distance from the country’s average exposure) have more
volatile stock returns. Lending specialization is thus associated with higher stock volatility. There is a
signiﬁcant, negative relationship between banks’ market-to-book values and lending specialization, as doc-
umented in the third row of results. Speciﬁcally, banks with a higher number of signiﬁcant sectoral betas,
whose returns react more to sectoral indices (i.e. where sectoral factors contribute more to the R2 ), for which
we ﬁnd a higher dispersion of sectoral betas in the return regression, and which show a higher Euclidian
distance from the country’s average exposure have lower market-to-book values. Lending specialization
thus seems to undermine market value.
Finally, we ﬁnd a signiﬁcant relationship between the MES and our different measures of lending spe-
cialization. Speciﬁcally, banks with more specialized lending portfolios according to our four indices have
a higher MES, thus contribute more to systemic fragility than other banks. In contrast to the theoretical pre-
dictions, we ﬁnd that differentiation (a larger distance, hence less herding) leads to more realized tail risk,
a more volatile stock and a lower market-to-book value. One potential explanation for this ﬁnding is in the
information content of this speciﬁc application of the distance measure. It measures the extent to which the
estimated sectoral factor exposures differ from the average of the estimated factor exposures in the country.
It misses hence a large part of herding or similarity to common shocks (which will be in the exposure to the
market factor). In addition, ﬁnding that more similar exposures lead to lower risk might already reﬂect the
idea the government will be more likely to step in if the likelihood of multiple banks facing distress is higher
(Acharya and Yorulmazer (2007)).
20
The economic signiﬁcance of our ﬁndings varies across the different regressions. A standard deviation
in the three signiﬁcant sectoral specialization indicators is associated with a three to 16% increase in the
standard deviation of total volatility. A standard deviation in the four sectoral specialization indicators is
associated with one to seven percent decrease in the standard deviation of the franchise value. Finally, a
standard deviation in the four sectoral specialization indicators is associated with a one to 18% increase in
the standard deviation of MES.
Table 8 reports the results for the lending specialization indicators based on the hand-collected reported
exposures and are organized in a similar fashion as Table 6. We add each of the four accounting-based
sectoral specialization indicators (as speciﬁed in Subsection 2.2) one-by-one to the regression setup reported
in the introduction to Section 3, resulting in 16 different speciﬁcations. Unlike in Table 6, we no longer
include bank ﬁxed effects due to the very limited within variation in each of the lending specialization
indicators.
Our results show no signiﬁcant relationship between lending specialization and average annualized re-
turns, which is consistent with the pairwise correlation results (see panel B of Table 5) and the results in
Table 7. None of the four lending specialization variables enters signiﬁcantly at conventional levels. Lend-
ing specialization is thus not reﬂected in stock returns. We ﬁnd a signiﬁcant relationship between three out
of four lending specialization indicators and the volatility of stock returns. Speciﬁcally, banks with a higher
dispersion and a higher concentration of sectoral portfolio shares, as measured both by the CR3 and the
HHI, show a higher volatility of stock returns. On the other hand, we do not ﬁnd any signiﬁcant relationship
between lending specialization and market-to-book values of banks. Banks with sectorally more specialized
lending portfolios are thus not traded with a discount. Hence, the results in the pairwise correlation table
(panel B of Table 5) might have been spurious and disappear after properly controlling for other factors that
affect the franchise value. Finally, accounting-based sectoral specialization is signiﬁcantly and positively
21
related to MES in both a regression and pairwise correlation framework.
In summary, both total volatility and the systemic risk exposures are higher for banks with a more
specialized (less diversiﬁed) sectoral business loan portfolio. The results for our herding measure are, as
with the market-based similarity measures, opposite (whenever signiﬁcant) to the theoretical predictions.
The more the bank’s sectoral loan portfolio differs from the average bank in the country the riskier the bank
is perceived to be. Our results are based on a very broad cross-section of countries and banks, especially
the large sample with return-based exposure data. Pooling these banks and countries and testing for a linear
relationship between banks’ sectoral specialization and their performance might mask signiﬁcant variation
across banks, across countries and over time. We will explore such variation in the next section.
4 Lending Specialization and Bank Performance: Cross-country
Heterogeneity
To the best of our knowledge, we are the ﬁrst to examine lending concentration using an international sample
of banks. Most previous studies relied on proprietary data from a central bank and were hence restricted
to one country. This feature of the sample creates the opportunity to directly examine how the impact of
lending specialization on bank performance varies with bank and country characteristics. Moreover, we
will also exploit that our return-based sample covers a long time period (of ten years) including the global
ﬁnancial crisis. In particular, we will examine whether the relationships are different in (i) the pre-crisis
versus (post-)crisis period, (ii) developed versus developing countries and (iii) countries with and without
explicit guidelines on bank asset diversiﬁcation, (iv) banks above and below the median size, (v) banks
above and below the median Lerner index of market power, (vi) banks below and above the loan-asset ratio,
and (vii) banks below and above the median of non-interest income share, where the last two are indicators
of the extent to which banks engage in traditional intermediation business.
22
More speciﬁcally, we will estimate equations of the following form:
below
P erf ormancei;t = 1 Specializationi;t 1 I (Y < Y median )+
above
1 Specializationi;t 1 I (Y Y median )+ (5)
Xi;t 1 controls + ui + vt + "i;t
In order to exploit sufﬁcient cross-country and cross-time variation, we use our larger and longer sample
to explore whether the relationship between lending specialization and performance of banks varies depend-
ing on the characteristics Y . Speciﬁcally, we interact the (factor-based) lending specialization indicators
with two dummy variables that split the sample according to characteristic Y . Results are reported in Ta-
ble 9. We focus on two speciﬁc measures of performance – the market-to-book value (panel A) and the
Marginal Expected Shortfall (panel B). As before, we only report the information of interest, i.e. the coefﬁ-
cient of each of the two sub-groups and their t-statistics, as well as the number of observations, the R2 and
below above
the p-value of a test whether 1 = 1 . To facilitate the comparison with the previous results, we
reproduce, in the ﬁrst column, the results without the sample split.
First, we split the sample into the period before and after the on-set of the Global Financial Crisis. The
relationship between lending specialization and bank performance might vary between the pre- and the cri-
sis periods for two reasons. First, different monetary regimes may affect the amount and quality of loans
outstanding. The pre-crisis era is characterized by cheap credit, massive loan growth and a riskier pool
of borrowers (Buncic and Melecky (2013)). During and after the crisis, banks have been applying stricter
lending standards. Second, after the meltdown of 2007/8, the stock market might perceive lending spe-
cialization differently, incorporating the lessons from the crisis. We therefore test whether the relationship
between the market-to-book value and the different measures of lending specialization varies across the pe-
riods 2002 to 2006 and 2007 to 2011. We ﬁnd that in most cases, the negative relationship between lending
specialization and market-to-book values is more pronounced for the crisis period than for the pre-crisis
period. Speciﬁcally, the number of Signiﬁcant Sectoral Factors and Differentiation from other banks enter
23
negatively and signiﬁcantly only for the crisis period, but not for the pre-crisis period. The Dispersion of
sectoral factor exposures enters signiﬁcantly for both periods but with a larger coefﬁcient estimate for the
crisis period. Only in the case of the Sectoral Contribution to R2 does the pre-crisis interaction enter signiﬁ-
cantly and with a larger coefﬁcient than the crisis-interaction. Overall, the negative relationship between the
risk-adjusted returns of banks and lending specialization seems to be stronger for the (post-)crisis than the
pre-crisis period.
The results in panel B of Table 9 show that the positive relationship between the MES and lending
specialization is in most cases more pronounced for the post-crisis than the pre-crisis period. Speciﬁcally,
the number of Signiﬁcant Sectoral Factors, the Sectoral Contribution to R2 and the Dispersion enter with
substantially larger coefﬁcients for the crisis than for the pre-crisis period. The number of Signiﬁcant Sec-
toral Factors enters even negatively and signiﬁcantly for the pre-crisis period. Only in the case of sectoral
Differentiation from the country-average does the pre-crisis coefﬁcient enter with a larger coefﬁcient than
the crisis-interaction. Overall, this suggests that while our results are not driven by the crisis period, they are
more pronounced for this period. Since the start of the global ﬁnancial crisis, lending specialization and the
resulting exposures to certain sectors is perceived to be more detrimental for both value as well as exposure
to systemic risk.
The second sample split is based on GDP per capita. As documented by Imbs and Wacziarg (2003),
sectoral specialization varies across countries at different income levels, with important repercussions for
sectoral specialization by the ﬁnancial system.12 Economically less developed economies feature a higher
12
In addition, lending concentration may also be more prevalent in emerging market economies due to different ownership
structures compared to Anglo-Saxon countries. In the latter, ﬁrms in general have dispersed shareholders and own other ﬁrms (if
any) via subsidiaries. In emerging market economies, on the other hand, ﬁrms are often held in groups with complicated ownership
structures. La Porta et al. (1999) document that approximately 25 percent of the ﬁrms in a sample of rich countries are members
of pyramids. Lins (2003) reports statistics for 18 emerging markets and reports that 66% of the ﬁrms for which the management
group is the largest blockholder of the control rights of a ﬁrm use pyramids to increase their control rights.
24
degree of specialization, which might make sectoral specialization of banks more of a necessity and less
damaging for systemic stability.13 On the other hand, the positive effect of lending specialization on bank
performance might be more pronounced at higher levels of economic development, where banks have the
necessary expertise and institutional background to exploit such economies of scale. We therefore split our
sample into countries below and above the median GDP per capita (Slovak Republic in 2005, with a value
of 6775 for GDP per capita (constant 2000 US$)). Lending specialization is penalized more strongly in
countries with high GDP per capita, both in terms of franchise value and MES. The impact of sectoral factor
specialization on the market-to-book value (MES) is more negative (positive) in developed versus devel-
oping countries, for each of the four market-based sectoral specialization measures. In fact, in developing
countries, there is often an insigniﬁcant or even a positive (negative) and signiﬁcant effect of specialization
on market-to-book-value (MES). The negative effects of banks’ sectoral specialization on performance and
contribution to systemic risk are therefore stronger in more developed countries, outweighing any beneﬁts
that might arise from loan specialization. On the other hand, there is only limited evidence of negative ef-
fects of sectoral specialization for banks in developing countries, a ﬁnding that has important implications
for the regulatory reform debate on the global level.
In column (3), we split the sample according to whether banks in a country are subject to speciﬁc
regulatory guidelines regarding asset diversiﬁcation. On the one hand, such guidelines might mitigate the
negative impact of specialization by forcing diversiﬁcation along other dimensions, including geographical.
On the other hand, such diversiﬁcation rules might be the result of a sectorally very specialized banking
sector. The information is taken from the World Bank questionnaire on Bank Regulation and Supervision
(Cihak et al. (2012) and Barth et al. (2013)), which asks: "Are there any regulatory rules or supervisory
guidelines regarding asset diversiﬁcation?". Countries with no such guidelines are in the below group,
13
On the other hand, Brown (2013) shows that household income reductions due to banking crisis are more signiﬁcant in the
middle-income countries than in the high-income countries, and are mostly due to the labour market channel.
25
whereas countries that answer ’YES’ to the question are in the above (median) group. In our sample,
the number of countries having such rules or guidelines varies over time. In the early years 50% of the
countries had guidelines, whereas to the end of the sample period approximately 66% of the countries in
our sample have guidelines. Regarding the franchise value, we ﬁnd that lending specialization (each of the
four measures) leads to a lower market-to-book ratio in countries with no explicit rules or guidelines on
asset diversiﬁcation. In contrast, in countries that do have such guidelines, we ﬁnd a signiﬁcantly positive or
insigniﬁcant effect. Similarly, regarding the exposure to systemic risk, we also ﬁnd that the impact is much
more positive in countries without rules or guidelines. This result could indicate that stock market investors
trust lending and sectoral specialization more whenever there are regulatory rules or there is a supervisor
monitoring the exposure limits. In the absence of such rules or supervisory oversight, stock market investors
seem to put more faith in the risk diversiﬁcation gains rather than the specialization skills in screening and
monitoring.
In column (4), we split the sample according to bank size, with the median bank size being 4; 023 million
US$ (ln(TA)=8.30). On the one hand, larger banks might be better able to exploit economies of scale as well
as beneﬁts of risk management and diversiﬁcation, so that the overall effect of sectoral specialization might
be lower. On the other hand, smaller banks might be forced to specialize in niche markets and might adopt
the necessary risk management tools to do so. The results in Panel A results suggest that the effect of
sectoral specialization on banks’ franchise value is not signiﬁcantly different for large and small banks. The
coefﬁcient estimates of Sectoral Contribution to R2 , Dispersion and Differentiation enter negatively and
signiﬁcantly. While the coefﬁcient estimates are larger for smaller banks, the difference to the coefﬁcient
estimates for large banks is not signiﬁcant in any of the regressions. The results in Panel B suggest that
specialization by large banks has in three out of four cases a larger effect on MES than specialization by
small banks. With the exception of Sectoral Contribution to R2 , the coefﬁcient estimate for large banks
enters with a signiﬁcantly larger positive coefﬁcient than for small banks. This suggest that if larger banks
26
specialize into speciﬁc sectors, this has more systemic repercussions.
In column (5) we split the sample according to banks’ market power, as gauged by the Lerner index,
which is the relative markup of price over marginal cost. The median Lerner index in our sample is 24:8%.
On the one hand, banks with more market power might be better able to beneﬁt from sectoral specialization
as they can offset risks stemming from lack of diversiﬁcation with higher pricing power as well as better
screening (see e.g. Petersen and Rajan (1995) who show that more competition reduces relationship lending,
and Hauswald and Marquez (2006) who show that competition reduces lending to informationally opaque
borrowers). In addition, many papers ﬁnd that more competition (less pricing power) leads to more risk-
taking by banks (see, e.g., Beck et al. (2013)). Specialization may then be perceived as a risky gamble by
low market power banks. On the other hand, banks in more competitive environment might be able to tap
more sophisticated hedging and other risk management tools to counter the effect of sectoral specialization.
The results in Panel A suggest that the negative effect of sectoral specialization on the franchise value is
signiﬁcantly stronger for banks in more competitive environments. In all four row blocks does the coefﬁcient
for banks below the median Lerner index enter negatively and signiﬁcantly larger than for banks with a
Lerner index above the median. The results in Panel B provide mixed evidence on the relative importance
of market power for the relationship between sectoral specialization and MES. In the case of Signiﬁcant
Sectoral Factors and Sectoral Contribution to R2 , the coefﬁcient for banks with market power below the
median is signiﬁcantly more positive than for banks with market power above the median. There is no
signiﬁcant difference in the case of Dispersion, where the coefﬁcients for both banks with high and low
market power enter positively and signiﬁcantly, while in the case of Dispersion from the average sectoral
specialization, the positive and signiﬁcant relationship between sectoral specialization and MES is stronger
for banks with high market power. This suggests that it is banks with high market power and specializing
away from the country average that add to systemic risk more than other banks.
Finally, we also inspect whether a bank’s business focus in lending and traditional intermediation income
27
affects the specialization-performance relationship. In column (6), we split the sample according to the loan-
asset ratio (median value = 65%). Banks with a higher loan-asset ratio and thus more traditional business
model focusing on ﬁnancial intermediation might suffer more from sectoral specialization as they have less
opportunities to diversify along other dimensions, as for example non-lending activities. The results in
column (6) conﬁrm this conjecture, as for three of the four sectoral specialization measures, the coefﬁcient
estimates enters with a higher coefﬁcient in the regressions of Franchise Value and MES. The only exception
is Differentiation from the country-average, where there is no signiﬁcant difference between banks below
and above the median of the loan-asset ratio. Finally, in column (7) we split the sample according to the
median non-interest income share. The intuition behind this sample split is similar to the split according to
the loan-asset ratio; banks with a higher non-interest income share and thus less of a focus on traditional
intermediation business might be less affected by sectoral specialization. The results in column (7), however,
do not provide any evidence to this effect, with most coefﬁcients not being signiﬁcantly different for banks
below and above the sample median of the non-interest income share.
5 Conclusions
This paper gauges the relationship between sectoral specialization of banks and their performance and con-
tribution to systemic risk. While theory and previous country-speciﬁc evidence provides ambiguous ﬁnd-
ings, our results, based on a broad cross-country and cross-bank sample show that banks with higher sectoral
specialization have higher volatility, lower market-book values and a higher contribution to systemic risk,
while there is no signiﬁcant relationship with proﬁtability. Critically, we ﬁnd signiﬁcant variation in these
relationships over time, across countries and across banks. The relationship between sectoral specialization,
volatility and risk holds across our sample period 2001 to 2011, but is stronger for the post-crisis period
starting in 2007. The relationship is stronger for more developed economies and for countries without
regulatory rules on diversiﬁcation. We also ﬁnd that the relationship between sectoral contribution and sys-
28
temic risk contribution is larger for larger banks and banks with more market power, while the relationship
between sectoral specialization and the market-book values is stronger for banks in more competitive envi-
ronments. Finally, we ﬁnd some evidence that the relationship is stronger for banks with more traditional
intermediation models.
Our ﬁndings are a ﬁrst attempt at empirically gauging the relationship between sectoral specialization
and bank performance and risk, using novel data and approaches. While a ﬁrst exploration, they provide
already some important policy messages. First, while there seem to be no signiﬁcant beneﬁts in terms of
proﬁtability from diversifying or specializing, sectoral specialization brings with it signiﬁcant risks. How-
ever, these risks vary signiﬁcantly across countries, across market structures and banks’ business models.
Any regulatory response to sectoral specialization has to take this variation into account.
29
References
Acharya, Viral, Iftekhar Hasan, and Anthony Saunders, 2006, Should banks be diversiﬁed? evidence from
individual bank loan portfolios, The Journal of Business 79, 1355–1412.
Acharya, Viral V., and Sascha Steffen, 2013, The greatest carry trade ever? understanding eurozone bank
risks, National Bureau of Economic Research Working Paper Series No. 19039.
Acharya, Viral V., and Tanju Yorulmazer, 2007, Too many to fail–an analysis of time-inconsistency in bank
closure policies, Journal of Financial Intermediation 16, 1–31.
Acharya, V.V., L. H. Pedersen, T. Philippon, and M. Richardson, 2012, Measuring systemic risk, CEPR
Discussion Papers 8824 .
Barth, James R., Jr. Caprio, Gerard, and Ross Levine, 2013, Bank regulation and supervision in 180 coun-
tries from 1999 to 2011, National Bureau of Economic Research Working Paper Series No. 18733.
Beck, Thorsten, Olivier De Jonghe, and Glenn Schepens, 2013, Bank competition and stability: Cross-
country heterogeneity, Journal of Financial Intermediation 22, 218–244.
Boeve, Rolf, Klaus Duellmann, and Andreas Pﬁngsten, 2010, Do specialization beneﬁts outweigh concen-
tration risks in credit portfolios of german banks?, Bundesbank Discussion Paper No 10/2010 .
˘ S2011
Brown, Martin, 2013, The transmission of banking crises to households: Lessons from the 2008âA ¸
crises in the eca region, World Bank Policy Research Working Paper 6528 .
Brown, Stephen J., and William N. Goetzmann, 1997, Mutual fund styles, Journal of Financial Economics
43, 373–399.
Buncic, Daniel, and Martin Melecky, 2013, Equilibrium credit : the reference point for macroprudential
supervisors, World Bank Policy Research Working Paper Series 6358 .
30
Cai, Jian, Anthony Saunders, and Sascha Steffen, 2013, Syndication, interconnectedness and systemic risk,
unpublished manuscript .
Cihak, Martin, Asli Demirguc-Kunt, Maria Soledad Martinez Peria, and Amin Mohseni-Cheraghlou, 2012,
Bank regulation and supervision around the world: A crisis update, World Bank Policy Research Working
Paper 6286 .
De Jonghe, O., 2010, Back to the basics in banking? a micro-analysis of banking system stability, Journal
of Financial Intermediation 19, 387–417.
Demirguc-Kunt, Asli, and Harry Huizinga, 2010, Bank activity and funding strategies: The impact on risk
and returns, Journal of Financial Economics 98, 626–650.
Duellmann, Klaus, and Nancy Masschelein, 2007, A tractable model to measure sector concentration risk
in credit portfolios, Journal of Financial Services Research 32, 55–79.
Han, Rui, and Martin Melecky, 2013, Financial inclusion for ﬁnancial stability : access to bank deposits and
the growth of deposits in the global ﬁnancial crisis, World Bank Policy Research Working Paper 6577 .
Hauswald, Robert, and Robert Marquez, 2006, Competition and strategic information acquisition in credit
markets, Review of Financial Studies 19, 967–1000.
Hayden, Evelyn, Daniel Porath, and Nataljav Westernhagen, 2007, Does diversiﬁcation improve the perfor-
mance of german banks? evidence from individual bank loan portfolios, Journal of Financial Services
Research 32, 123–140.
Imbs, Jean, and Romain Wacziarg, 2003, Stages of diversiﬁcation, American Economic Review 93, 63–86.
La Porta, Rafael, Florencio Lopez-De-Silanes, and Andrei Shleifer, 1999, Corporate ownership around the
world, The Journal of Finance 54, 471–517.
31
Laeven, Luc, and Ross Levine, 2009, Bank governance, regulation and risk taking, Journal of Financial
Economics 93, 259–275.
Lakonishok, Josef, Andrei Shleifer, and Robert W. Vishny, 1992, The impact of institutional trading on stock
prices, Journal of Financial Economics 32, 23–43.
Lins, Karl V., 2003, Equity ownership and ﬁrm value in emerging markets, Journal of Financial and Quan-
titative Analysis 38, 159–184.
Liu, Chen, 2011, Herding behavior in bank lending: Evidence from us commercial banks, Unpublished
manuscript .
Petersen, Mitchell A., and Raghuram G. Rajan, 1995, The effect of credit market competition on lending
relationships, The Quarterly Journal of Economics 110, 407–443.
Sharpe, William, 1992, Asset allocation: Management style and performance measurement, Journal of
Portfolio Management 18, 7–19.
Smirlock, Michael, Thomas Gilligan, and William Marshall, 1984, Tobin’s q and the structure-performance
relationship, The American Economic Review 74, 1051–1060.
Stevens, Jerry L., 1990, Tobin’s q and the structure-performance relationship: Comment, The American
Economic Review 80, 618–623.
Stiglitz, Joseph E., and Andrew Weiss, 1981, Credit rationing in markets with imperfect information, The
American Economic Review 71, 393–410.
Stiroh, K.J., and A. Rumble, 2006, The dark side of diversiﬁcation: The case of u.s. ﬁnancial holding
companies, Journal of Banking and ﬁnance 30, 2131–2161.
32
Tabak, Benjamin M., Dimas M. Fazio, and Daniel O. Cajueiro, 2011, The effects of loan portfolio concen-
tration on brazilian banks’ return and risk, Journal of Banking and Finance 35, 3065–3076.
ter Horst, Jenke R., Theo E. Nijman, and Frans A. de Roon, 2004, Evaluating style analysis, Journal of
Empirical Finance 11, 29–53.
Wagner, Wolf, 2010, Diversiﬁcation at ﬁnancial institutions and systemic crsises, Journal of Financial In-
termediation 19, 373–386.
Winton, Andrew, 2000, Don’t put all your eggs in one basket?: Diversiﬁcation and specialization in lending,
Wharton Financial Institutions Center Paper 00-16 .
33
Table 1: Sectoral factor loadings and return-based sectoral specialization measures
This table contains information on sectoral factor exposures, their signiﬁcance as well as sectoral specialization mea-
sures based on these factor exposures. The sectoral exposures are obtained from a regression of a bank’s stock return on
the returns on a broad market index as well as on the return to 10 different sectoral indices. We estimate such a regres-
sion for each bank and for each year using daily returns, yielding a panel database on sectoral exposures that varies at
the bank-year frequency. The panel dataset of estimated exposures consists of 12,689 bank-year observations, covering
2,005 banks from 77 countries over a ten year period starting in 2002. Panel A reports for each estimated factor loading
the mean and standard deviation across 12,689 observations, as well as the ﬁfth, ﬁftieth and ninety-ﬁfth percentile of
the panel of estimated factor loadings. In panel B, we report for each factor the relative frequency of observing t-stats
in ﬁve groups. Based on the estimated sectoral exposures, we compute four time-varying bank-speciﬁc measures of the
intensity of sectoral specialization of which summary statistics are reported in panel C. A detailed description of the
construction of these four measures is provided in the text as well as in Table 4.
Panel A: Summary Statistics on Sectoral Factor Loadings
variable mean sd p5 p50 p95
1= Oil & gas (OILGS) 0.002 1.010 -1.442 -0.008 1.495
2= Basic materials (BMATR) -0.006 0.854 -1.226 0.012 1.087
3= Industrials (INDUS) 0.008 0.586 -0.787 0.000 0.840
4= Consumer goods (CNSMG) 0.004 0.646 -0.903 0.002 0.885
5= Healthcare (HLTHC) -0.007 0.739 -1.131 -0.001 1.105
6= Consumer services (CNSMS) 0.006 0.611 -0.873 0.005 0.873
7= Telecommunications (TELCM) -0.009 0.468 -0.756 0.000 0.693
8= Utilities (UTILS) 0.021 0.442 -0.627 0.024 0.651
9= Technology (TECNO) 0.009 0.955 -1.354 0.000 1.381
10= Financials(FINAN) 0.094 0.308 -0.234 0.037 0.604
Panel B: Frequency table of t-stats of sectoral factor loadings
t<-1.65 -1.65<=t<-1 -1<=t<1 11
Sectoral Contribution to R2 Contribution to R2 of the sectoral factors
Dispersion (factors) Dispersion (st.dev) of sectoral betas
Differentiation (factors) Euclidean distance between bank’s return-based
exposures and country’s average exposures (excluding bank in average)
37
Panel B: Sectoral Specialization Indicators based on Accounting data
Dispersion (accounting) Dispersion (st.dev) of sectoral acc. Exposures
Sectoral CR3 Cumulative exposure of largest three acc. Exposures
Sectoral HHI Concentration (HHI) of sectoral acc. Exposures
Differentiation (accounting) Euclidean distance between a bank’s exposures and
the country’s average exposures (excluding bank in average)
Panel C: Performance Measures
Return Annualized Average Daily Stock Return
Volatility Annualized Volatility of Daily Stock Return
Franchise Value Market-to-Book value of Equity
MES Marginal Expected Shortfall (5%, wrt LOCAL banking sector)
Table 5: Correlation matrices
This table contains three panels of correlation matrices. In each panel, the pairwise correlation coefﬁcients as well
as the p-value are reported. In panel A, we report the correlation coefﬁcients among and between the performance
measures and the sectoral specialization measures based on factor loadings (corresponding with the larger sample of
12634 observations). In panel B, we report the correlation coefﬁcients among and between the performance measures
and the sectoral specialization measures based on accounting exposures (smaller sample of 1392 observations). In panel
C, we report the correlation coefﬁcients among and between all sectoral specialization measures. The number-letter
combination in the column headers (which differ by panel) correspond with a variable in the rows.
Correlation between performance measures and Sectoral Specialization Indicators based on Factor Loadings (large sample: 12634 obs.)
Variables A1 A2 A3 A4 A5 A6 A7 A8
Return (=A1) 1.000
Volatility (=A2) -0.398 1.000
(0.000)
Franchise Value (=A3) 0.071 -0.212 1.000
(0.000) (0.000)
MES (=A4) -0.269 0.406 -0.146 1.000
(0.000) (0.000) (0.000)
Signiﬁcant Sectoral Factors (=A5) 0.010 0.026 -0.001 0.129 1.000
(0.247) (0.004) (0.918) (0.000)
Sectoral Contribution to R2 (=A6) -0.036 -0.011 -0.034 0.378 0.289 1.000
(0.000) (0.220) (0.000) (0.000) (0.000)
Dispersion (factors) (=A7) -0.065 0.437 -0.168 0.229 0.507 0.205 1.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Differentiation (factors) (=A8) -0.093 0.330 -0.114 0.049 0.413 -0.017 0.777 1.000
(0.000) (0.000) (0.000) (0.000) (0.000) (0.053) (0.000)
Correlation between performance measures and Sectoral Specialization Indicators based on Accounting Data (small sample: 1392 obs.)
Variables B1 B2 B3 B4 B5 B6 B7 B8
Return (=B1) 1.000
Volatility (=B2) -0.327 1.000
(0.000)
Franchise Value (=B3) -0.082 0.066 1.000
(0.002) (0.013)
MES (=B4) -0.406 0.746 0.109 1.000
(0.000) (0.000) (0.000)
Dispersion (accounting) (=B5) -0.008 0.060 0.073 0.094 1.000
(0.752) (0.025) (0.007) (0.000)
Sectoral CR3 (=B6) -0.003 0.068 0.080 0.080 0.923 1.000
(0.896) (0.011) (0.003) (0.003) (0.000)
Sectoral HHI (=B7) -0.012 0.054 0.069 0.097 0.983 0.875 1.000
(0.646) (0.044) (0.010) (0.000) (0.000) (0.000)
Differentiation (accounting) (=B8) -0.051 0.099 0.202 0.109 0.586 0.569 0.606 1.000
(0.056) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000)
Correlation between Sectoral Specialization Indicators based on Factor Loadings and Accounting Data (small sample: 1363 obs.)
Variables C1 C2 C3 C4 C5 C6 C7 C8
Signiﬁcant Sectoral Factors (=C1) 1.000
Sectoral Contribution to R2 (=C2) 0.206 1.000
(0.000)
Dispersion (factors) (=C3) 0.531 0.227 1.000
(0.000) (0.000)
Differentiation (factors) (=C4) 0.365 -0.058 0.595 1.000
(0.000) (0.031) (0.000)
Dispersion (accounting) (=C5) -0.010 -0.048 0.016 0.054 1.000
(0.720) (0.078) (0.545) (0.046)
Sectoral CR3 (=C6) -0.009 -0.067 0.024 0.081 0.923 1.000
(0.741) (0.013) (0.377) (0.003) (0.000)
Sectoral HHI (=C7) -0.009 -0.026 0.014 0.036 0.983 0.875 1.000
(0.743) (0.341) (0.593) (0.189) (0.000) (0.000)
Differentiation (accounting) (=C8) -0.021 -0.143 0.001 0.059 0.586 0.569 0.606 1.000
(0.433) (0.000) (0.976) (0.028) (0.000) (0.000) (0.000)
38
Table 6: Drivers of Bank Performance
This table shows the results of regressing a bank performance metric (return, volatility, franchise value or MES) on the set of control variables. In
the left hand side, we report the regression results of the full sample. In the right hand side, we report the regression results for the smaller sample.
The control variables capture various dimensions of a bank’s business model that might inﬂuence performance and stability. Size is computed as the
natural logarithm of total assets expressed in 2007 US dollars. We measure a bank’s share of non-interest income to total operating income, by dividing
other operating income (which comprises trading income, commissions and fees as well as all other non-interest income) by the sum of interest income
and other operating income. The other bank-speciﬁc variables are proxies for leverage (capital-to-asset ratio), the funding structure (share of deposits
in sum of deposits and money market funding), asset mix (loans to assets ratio), proﬁtability (return-on-equity), annual growth in total assets as well
as expected credit risk (Loan Loss Provision to Total Assets). The independent variables are lagged one year to mitigate concerns regarding reverse
causality. We winsorize all variables at the 1 and 99 percentile level to mitigate the impact of outliers. Next to the set of control variables, we also
include year ﬁxed effects as well as bank ﬁxed effects. The standard errors are clustered at the bank level. Numbers in parentheses are t-statistics.
Large sample: 2005 banks from 77 countries, 2002-2011 Small sample: 317 banks from 57 countries, 2007-2011
VARIABLES Return Volatility Franchise Value MES Return Volatility Franchise Value MES
Bank Size (orthogonalized) -27.106*** -5.951*** -0.361*** 0.369*** -3.732 -7.765* -0.394** -0.744
(2.345) (1.410) (0.065) (0.108) (11.415) (4.703) (0.194) (0.484)
Non-Interest Income Share -53.127*** -12.370*** -0.401 0.095 -46.249 -24.336* -1.084** -1.882
(8.744) (4.378) (0.256) (0.333) (35.211) (13.977) (0.445) (1.442)
39
Equity to Total Assets 347.149*** 9.705 2.594*** -4.355*** 228.280 -53.338 6.207 -3.813
(40.019) (20.894) (0.897) (1.628) (177.442) (69.831) (3.770) (7.752)
Share of Deposits in Total Funding 104.780*** 18.038*** 1.186*** -1.606*** -16.731 1.381 0.890 1.997
(10.310) (6.032) (0.317) (0.501) (53.604) (19.886) (0.643) (2.064)
Loans to Total Assets 95.833*** 23.674*** 2.046*** -2.358*** -15.038 58.200** 0.509 5.356*
(13.524) (8.153) (0.381) (0.616) (56.507) (25.054) (1.113) (2.775)
Return on Equity -88.137*** -50.317*** -1.060*** 1.131** -21.656 -48.328*** -1.013 -2.265
(10.284) (6.046) (0.301) (0.461) (44.945) (17.685) (0.994) (1.974)
Growth in Total Assets 16.432*** 2.301 0.498*** 0.039 0.303 9.959** 0.162 0.963*
(3.328) (1.813) (0.083) (0.154) (12.494) (4.765) (0.164) (0.538)
Loan Loss Provisions to Total Assets -14.579*** 0.751 -0.281*** 0.308*** 5.087 -3.075 -0.223** -0.252
(1.400) (0.772) (0.039) (0.058) (5.900) (2.270) (0.097) (0.227)
Constant -126.296*** 5.803 -0.959** 3.797*** 37.860 32.410* 0.442 0.606
(16.156) (9.854) (0.478) (0.773) (44.691) (17.494) (0.595) (2.067)
Observations 12689 12689 12684 12689 1426 1426 1426 1426
Adjusted R-squared 0.340 0.616 0.768 0.607 0.398 0.736 0.796 0.747
Bank Fixed Effects YES YES YES YES YES YES YES YES
Year Fixed Effects YES YES YES YES YES YES YES YES
Clustered SE Bank Bank Bank Bank Bank Bank Bank Bank
Robust standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1
Table 7: Lending Specialization and Bank Performance: Evidence from Stock Return-based Sectoral Specialization Measures
To gauge the relationship between lending specialization and bank performance and fragility, we add each of the four return-based sectoral specialization
indicators one-by-one to a regression of each of the four dependent variables on the control variables (as reported in the left hand side part of Table
6). This results in 16 different speciﬁcations. We summarize the results by reporting only the information of interest. More speciﬁcally, for each
regression, we only report the coefﬁcient and t-statistic on the return-based sectoral specialization indicator as well as the number of observations and
the adjusted R-squared of the regression. The table is constructed as follows. Each of the four sectoral specialization indicators is reported in a different
column. The dependent variable varies by block of rows. The sample consists of 12689 observations, on 2005 banks from 77 countries, 2002-2011.
Return
Signiﬁcant Sectoral Factors Sectoral Contribution to R2 Dispersion (factors) Differentiation (factors)
Coeff 0.093 -6.287 -1.245 -0.052
t-statistic 0.789 -0.601 -0.497 -0.139
Observations 12689 12639 12689 12689
Adjusted R-squared 0.340 0.341 0.340 0.340
Volatility
40
Signiﬁcant Sectoral Factors Sectoral Contribution to R2 Dispersion (factors) Differentiation (factors)
Coeff 0.043 11.583*** 17.284*** 1.788***
t-statistic 0.818 2.616 14.266 9.228
Observations 12689 12639 12689 12689
Adjusted R-squared 0.616 0.619 0.631 0.622
Franchise Value
Signiﬁcant Sectoral Factors Sectoral Contribution to R2 Dispersion (factors) Differentiation (factors)
Coeff -0.004* -0.988*** -0.391*** -0.039***
t-statistic -1.950 -5.949 -9.532 -6.391
Observations 12684 12634 12684 12684
Adjusted R-squared 0.768 0.771 0.771 0.769
MES
Signiﬁcant Sectoral Factors Sectoral Contribution to R2 Dispersion (factors) Differentiation (factors)
Coeff 0.011** 6.981*** 1.259*** 0.080***
t-statistic 2.169 14.190 12.560 5.388
Observations 12689 12639 12689 12689
Adjusted R-squared 0.607 0.620 0.616 0.608
*** p<0.01, ** p<0.05, * p<0.1; t-stats based on robust standard errors clustered at bank level
Each regression includes control variables, bank and year ﬁxed effects.
Table 8: Lending Specialization and Bank Performance: Evidence from Accounting-based Sectoral Specialization Measures
To gauge the relationship between lending specialization and bank performance and fragility, we add each of the four accounting-based sectoral specialization indicators
one-by-one to a regression of each of the four dependent variables on the control variables (as reported in the right hand side part of Table 6). The only difference with
respect to Table 6 is that we do not include bank ﬁxed effects due to the lack of within bank variation in the sectoral specialization measures. This results in 16 different
speciﬁcations. We summarize the results by reporting only the information of interest. More speciﬁcally, for each regression, we only report the coefﬁcient and t-statistic on
the accounting-based sectoral specialization indicator as well as the number of observations and the adjusted R-squared of the regression. The table is constructed as follows.
Each of the four sectoral specialization indicators is reported in a different column. The dependent variable varies by block of rows. The regressions are based on a smaller
sample (compared to table 7) of 1426 observations, covering 317 banks from 57 countries over the period 2007-2011.
Return
Dispersion (accounting) Sectoral CR3 Sectoral HHI Differentiation (accounting)
Coeff 12.288 2.317 4.880 -6.599
t-statistic 0.450 0.271 0.425 -0.813
Observations 1426 1426 1426 1392
Adjusted R-squared 0.413 0.413 0.413 0.410
Volatility
41
Dispersion (accounting) Sectoral CR3 Sectoral HHI Differentiation (accounting)
Coeff 32.380* 12.301** 11.772* 7.272
t-statistic 1.938 2.311 1.722 1.383
Observations 1426 1426 1426 1392
Adjusted R-squared 0.416 0.419 0.416 0.417
Franchise Value
Dispersion (accounting) Sectoral CR3 Sectoral HHI Differentiation (accounting)
Coeff -0.790 -0.435 -0.265 0.490
t-statistic -0.664 -1.235 -0.523 1.319
Observations 1426 1426 1426 1392
Adjusted R-squared 0.273 0.275 0.273 0.274
MES
Dispersion (accounting) Sectoral CR3 Sectoral HHI Differentiation (accounting)
Coeff 6.221*** 1.759** 2.551*** 0.821
t-statistic 2.689 2.410 2.703 1.144
Observations 1426 1426 1426 1392
Adjusted R-squared 0.360 0.359 0.360 0.365
*** p<0.01, ** p<0.05, * p<0.1; t-stats based on robust standard errors clustered at bank level
Each regression includes control variables and year ﬁxed effects.
Table 9: Sample splits: panel A (Franchsie Value) and panel B (MES)
This table contains information on how the impact of lending specialization on banks’ franchise value and marginal expected shortfall varies with
country- and bank characteristics. The dependent variable in Panel A is the market-to-book value of equity (franchise value). It is the Marginal
Expected Shortfall in Panel B. Each panel is constructed in a similar fashion. Each panel consists of four subpanels corresponding with a return-based
sectoral specialization measure. The regression setup differs in only one dimension from the one in table 7 and 8. We now interact the (factor-based)
lending specialization indicators with two dummy variables that split the sample according to the characteristic mentioned in the column title. For
example, in the ﬁrst subpanel, the independent variable of interest is the number of signiﬁcant sectoral factor exposures. The baseline result, which
is also in Table 6, is repeated in the ﬁrst column. This is the ’unconditional’ effect. In subsequent columns, we examine whether the effect of the
number of signiﬁcant factors on the MES varies over time (the period 2002-2006 versus 2007-2012) or with country characteristics: developed vs.
developing countries (sample split according to median value of GDP per capita), and countries with no guidelines on asset diversiﬁcation (below) or
with guidelines (above). We also examine the scope for differential effects of lending specialization on franchise value and MES for (i) large versus
small banks, banks with market power (the Lerner index), banks with a focus on lending (loan-to-asset ratio), and banks with a focus on non-tradition
banking (non-interest income share) We only report the coefﬁcients of interest and the associated t-statistic. Signiﬁcance at the conventional levels is
indicated with *** (1 per cent), ** (5 per cent) or * (10 per cent). We also report the number of observations and R-square. Furthermore, the last row
in each block reports the p-value of a t-test for the equality of the coefﬁcient below or above the sample mean. In each regression, we include a set of
control variables (as in table 6) as well as bank and time ﬁxed effects. Standard errors are clustered at the bank level.
42
Panel A: Franchise Value (Market to Book value of Equity)
pre/post 2007 GDP per capita Asset Diversiﬁcation Size Market Power Loan/Asset NII-share
Signiﬁcant Sectoral Factors
Signiﬁcant Sectoral Factors (=X1) -0.004*
t-stat -1.950
X1 * [column title below median] 0.003 0.004 -0.017*** -0.004 -0.014*** -0.000 -0.007***
t-stat 0.971 0.811 -6.028 -1.252 -5.303 -0.154 -2.580
X1 * [column title above median] -0.011*** -0.006** 0.012*** -0.005 0.005** -0.008*** -0.001
t-stat -3.707 -2.542 3.557 -1.551 2.041 -2.971 -0.373
Observations 12684 12684 12684 12684 12684 12684 12684 12684
Adjusted R-squared 0.768 0.769 0.768 0.770 0.768 0.769 0.768 0.768
p-value(below=above) 0.003*** 0.082* 0.000*** 0.800 0.000*** 0.042** 0.163
Sectoral Contribution to R2
Sectoral Contribution to R2 (=X2) -0.988***
t-stat -5.949
X2 * [column title below median] -1.510*** 1.801*** -1.630*** -1.232*** -1.404*** -0.682*** -1.081***
t-stat -6.943 4.144 -9.164 -5.292 -7.198 -3.283 -5.534
X2 * [column title above median] -0.759*** -1.328*** 0.989*** -0.789*** -0.616*** -1.199*** -0.766***
t-stat. -3.676 -7.440 3.485 -3.666 -3.407 -6.658 -4.064
Observations 12634 12634 12634 12634 12634 12634 12634 12634
Adjusted R-squared 0.771 0.771 0.772 0.774 0.771 0.771 0.771 0.771
43
p-value(below=above) 0.007*** 0.000*** 0.000*** 0.137 0.000*** 0.010*** 0.151
Dispersion (factors)
Dispersions (factors) (=X3) -0.391***
t-stat -9.532
X3 * [column title below median] -0.247*** -0.111* -0.546*** -0.418*** -0.501*** -0.340*** -0.381***
t-stat -3.884 -1.662 -10.922 -7.448 -10.489 -6.968 -7.530
X3 * [column title above median] -0.470*** -0.445*** -0.045 -0.361*** -0.239*** -0.437*** -0.348***
t-stat -8.263 -9.487 -0.782 -6.882 -5.077 -9.056 -7.368
Observations 12684 12684 12684 12684 12684 12684 12684 12684
Adjusted R-squared 0.771 0.772 0.772 0.774 0.771 0.772 0.772 0.771
p-value(below=above) 0.013** 0.000*** 0.000*** 0.431 0.000*** 0.063* 0.580
Differentiation (factors)
Differentiation (factors) (=X4) -0.039***
t-stat -6.391
X4 * [column title below median] -0.012 -0.009 -0.060*** -0.044*** -0.059*** -0.036*** -0.037***
t-stat -1.249 -0.856 -7.482 -5.028 -8.178 -4.587 -4.706
X4 * [column title above median] -0.052*** -0.046*** 0.007 -0.033*** -0.010 -0.042*** -0.037***
t-stat -6.668 -6.571 0.766 -3.996 -1.262 -5.424 -4.507
Observations 12684 12684 12684 12684 12684 12684 12684 12684
Adjusted R-squared 0.769 0.770 0.770 0.771 0.769 0.770 0.769 0.769
p-value(below=above) 0.001*** 0.002*** 0.000*** 0.368 0.000*** 0.529 0.964
*** p<0.01, ** p<0.05, * p<0.1; t-stats based on robust standard errors clustered at bank level
Each regression includes control variables, bank and year ﬁxed effects.
Panel B: Marginal Expected Shortfall
pre/post 2007 GDP per capita Asset Diversiﬁcation Size Market Power Loan/Asset NII-share
Signiﬁcant Sectoral Factors
Signiﬁcant Sectoral Factors (=X1) 0.011**
t-stat 2.169
X1 * [column title below median] -0.028*** -0.013 0.026*** 0.005 0.018*** -0.002 0.011*
t-stat -4.400 -1.068 3.978 0.754 2.624 -0.231 1.732
X1 * [column title above median] 0.046*** 0.017*** -0.009 0.017** 0.003 0.023*** 0.009
t-stat 5.771 3.400 -1.359 2.127 0.515 3.523 1.477
Observations 12689 12689 12689 12689 12689 12689 12689 12689
Adjusted R-squared 0.607 0.609 0.607 0.608 0.607 0.607 0.607 0.607
p-value(below=above) 0.000*** 0.022** 0.000*** 0.234 0.080* 0.006*** 0.821
Sectoral Contribution to R2
Sectoral Contribution to R2 (=X2) 6.981***
t-stat 14.190
X2 * [column title below median] 1.168** -3.072*** 8.607*** 8.289*** 8.607*** 4.531*** 6.994***
t-stat 2.110 -3.294 17.023 13.514 14.488 7.971 11.817
X2 * [column title above median] 9.537*** 8.188*** 1.970*** 5.899*** 5.524*** 8.720*** 5.975***
t-stat. 16.552 15.998 3.789 8.641 10.859 15.579 10.702
Observations 12639 12639 12639 12639 12639 12639 12639 12639
Adjusted R-squared 0.620 0.631 0.624 0.627 0.621 0.623 0.623 0.619
44
p-value(below=above) 0.000*** 0.000*** 0.000*** 0.007*** 0.000*** 0.000*** 0.085*
Dispersion (factors)
Dispersions (factors) (=X3) 1.259***
t-stat 12.560
X3 * [column title below median] 0.746*** 0.338* 1.477*** 0.883*** 1.274*** 1.032*** 1.077***
t-stat 5.329 1.782 12.570 6.785 10.546 8.952 8.324
X3 * [column title above median] 1.541*** 1.472*** 0.770*** 1.657*** 1.238*** 1.466*** 1.238***
t-stat 11.571 13.268 5.858 11.112 10.524 11.056 10.229
Observations 12689 12689 12689 12689 12689 12689 12689 12689
Adjusted R-squared 0.616 0.617 0.617 0.618 0.617 0.616 0.617 0.615
p-value(below=above) 0.000*** 0.000*** 0.000*** 0.000*** 0.785 0.003*** 0.295
Differentiation (factors)
Differentiation (factors) (=X4) 0.080***
t-stat 5.388
X4 * [column title below median] 0.110*** 0.014 0.106*** 0.048** 0.063*** 0.063*** 0.054***
t-stat 4.577 0.513 5.867 2.456 3.506 3.487 2.829
X4 * [column title above median] 0.065*** 0.099*** 0.024 0.121*** 0.104*** 0.096*** 0.095***
t-stat 3.501 5.948 1.093 5.247 5.539 4.809 4.877
Observations 12689 12689 12689 12689 12689 12689 12689 12689
Adjusted R-squared 0.608 0.608 0.608 0.609 0.608 0.608 0.608 0.608
p-value(below=above) 0.145 0.008*** 0.002*** 0.015** 0.066* 0.169 0.110
*** p<0.01, ** p<0.05, * p<0.1; t-stats based on robust standard errors clustered at bank level
Each regression includes control variables, bank and year ﬁxed effects.
45
A List of countries and number of bank-year observations by country
Country Large Sample Small sample Country Large Sample Small sample
ARGENTINA 62 15 MACEDONIA (FYROM) 8 0
AUSTRALIA 59 25 MALAYSIA 108 5
AUSTRIA 70 16 MEXICO 65 10
BAHRAIN 42 10 MOROCCO 37 3
BANGLADESH 128 0 NIGER 2 0
BELGIUM 26 10 NIGERIA 6 2
BERMUDA 20 4 NORWAY 146 25
BOSNIA AND HERZEGOVINA 6 0 OMAN 28 5
BOTSWANA 9 0 PAKISTAN 122 0
BRAZIL 145 10 PERU 59 10
BULGARIA 21 0 PHILIPPINES 93 14
CANADA 103 0 POLAND 120 51
CHILE 57 25 PORTUGAL 51 15
CHINA 62 58 QATAR 35 15
COLOMBIA 60 17 ROMANIA 35 5
CROATIA 96 10 RUSSIAN FEDERATION 65 40
DENMARK 331 10 SAUDI ARABIA 74 33
ECUADOR 35 0 SERBIA 14 0
46
EGYPT 66 1 SINGAPORE 37 15
FINLAND 18 9 SLOVAKIA 23 4
FRANCE 247 24 SLOVENIA 24 0
GERMANY 109 15 SOUTH AFRICA 64 20
GREECE 82 20 SPAIN 93 15
HONG KONG SAR, CHINA 87 35 SRI LANKA 56 0
HUNGARY 14 0 SWEDEN 40 20
ICELAND 4 0 SWITZERLAND 114 10
INDIA 272 70 TAIWAN, CHINA 222 45
INDONESIA 122 17 THAILAND 119 34
IRELAND 10 5 TUNISIA 47 0
ISRAEL 80 30 TURKEY 100 45
ITALY 205 30 UGANDA 6 0
JAPAN 915 290 UKRAINE 26 0
JORDAN 53 9 UNITED ARAB EMIRATES 93 34
KAZAKHSTAN 26 10 UNITED KINGDOM 77 20
KENYA 71 0 UNITED STATES OF AMERICA 6426 113
KOREA REPUBLIC OF 74 0 VENEZUELA 106 8
KUWAIT 41 13 VIETNAM 0 3
LEBANON 44 14 ZIMBABWE 7 0
LITHUANIA 48 0
LUXEMBOURG 21 5
Large sample: 12689 observations, on 2005 banks from 77 countries, 2002-2011
Small sample: 1426 observations, on 317 banks from 57 countries, 2007-2011