Cross Country Comparison of Efficiency in Investment

Banking
.
Nemanja Radi
University of Rome Tor Vergata, Doctorate of Research in Banking and Finance, Via Columbia 2, Rome, Italy, e-mail:radic_n@yahoo.com
Franco Fiordelisi
University of Rome Federico Caff, Department of Management and Law, Rome, Italy, e-mail:fiordeli@uniroma3.it
Abstract
This paper aims to identify the framework for comparing investment banks
efficiencies across nations. In order to overcome traditional limitations two
methods are adopted: first, where separate frontiers are estimated to check for the
existence of structural differences between the countries; and second method which
accounts for the influences of environmental factors on the industry, by including
indicator of these factors in a definition of a common frontier. We use translog cost
and profit function in order to measure X-efficiency. Data set consist from more
than 900 investment banks from G7 countries (US, UK, Japan, Italy, Germany,
France and Canada) and Switzerland over the period 2000-2007.
1
1. Motivation and Introduction
Investment banking industry on the world level has gone through incredible
transformation due to cross border activities and consolidation. Today more and
more banks are crossing international borders and providing services around world.
Having in mind the business of investment banking and newest trends, we can say
that efficiency of these types of banks is important for several reasons. First,
investment bank engage in public and private market transaction for corporations,
governments and investors, and by doing so is making benefits for all the
participants. Second, efficiency of these institutions affect the financial markets and
the ability of investment banks to minimize costs or maximizes profits is important
both for them and for their clients. Third, by exercising their powers and by
improving their efficiency, these institutions improve certain industry segments
(this refers to boutique investment banks).
We can define investment banking as the intermediation between issuers and
investors through the core function of advisory, M&A, debt capital markets and
equity capital markets.
A number of environmental trends as well as the creativity and dynamism of
their professional teams have shaped todays investment banks. Most of the
researchers and analysts agree that the key drivers of the phenomenal secular
growth of the business have been: GDP growth and stock market prices;
globalization through cross border investment flows (cross border mergers and
acquisitions in the developed world as well as direct and portfolio investment in
emerging markets); the accumulation of assets managed by institutions (growing
share of GNP wealth managed by institutions such as pension funds has created a
well structured market for investment banks); securitization (has represented a
direct economical transfer from commercial to investment banks); deregulation.
1
Gardener and Molyneux (1995) have identified similar factors that affect and
influence the evolution of investment banking, such as: real per capita income and
wealth, economic forces that directly affect investment banking services through
technological advances, the regulatory framework affects, distribution of property
rights and the way that they are exercised.
Due to specific nature of this research and complexity of the investment banking
business, we provide the literature definition of the same, where: Investment
banks business can be categorized into five main areas: broking (the broking of
securities is commodity business in which firms appeal to customers mainly on
1
For further readings see Davis (2003)
2
price and integrity); trading (the trading of securities drives on market volatility);
investment banking (represents the underwriting of new issues and advisory work
also referred to as Mergers and Acquisitions); fund management (includes both
retail and wholesale fund management); interest spread (income derivatives from
borrowed funds).
2
There are two basic types of investment banks: full-service and boutique. Full
service investment banks (also known as the Wall Street bulge bracket) offer
clients a range of service including underwriting, merger and acquisition advice,
trading, merchant banking and prime brokerage. For example, Goldman Sachs
offers services in investment banking, trading and principal investments, asset
management and securities service; Merrill Lynch in capital markets, investment
banking and advisory, wealth management, investment management, insurance and
banking. Boutique investment banks specialize in particular segments of the
market. They do not offer a range of service and are not part of larger financial
institution.
3
For example, Greenhill is specialized for Advisory services in M&A,
Financial restructuring and Merchant banking, while Lazard offers Financial
advisory and Asset management services.
4
Beside these two basic types of investment banks it is important to mention
financial holding companies, which operate full-service investment banking, and
can besides that offer clients large sums of credit (for example Citigroup, HSBC,
Credit Suisse, JP Morgan Chase and Bank of America).
5
For example, business
segments for HSBC Group are personal financial services, commercial banking,
corporate, investment banking and markets and private banking.
Investment banking represents a revenue motivated business. In order for
management to maximize their share of that revenue, they have to improve some or
all sectors of investment banking business. Investment banks diversify their
business lines in order to have earnings more stable. For most investment banks
today, investment banking represents only a portion of their overall income. By
looking at the JP Morgan Chase annual report for 2006, we notice that income by
line of business was 24% for retail financial services, 24% for card services, 27%
for investment bank, 10% for asset management, 8% treasury and securities
services and 7% for commercial banking. On the other hand Morgan Stanley
income source for 2007 were 19% for asset management, 24% for global wealth
management and 57% for institutional securities.
6
2
For further readings see Gardener and Molyneux (1995)
3
For further readings see Liaw (2006)
4
Sources: Annual Reports for 2006 of the given banks.
5
For further readings see Liaw (2006)
6
Source was Morgan Stanley annual report for 2007.
3
Managing the cost base was always dominated by human capital, and influenced
by investment in global infrastructure and product platforms. However, managing
the cost bases raises also the issue of the relationship between cost and revenues,
which means that banks can match its cost base to its revenue generating potential
over time. Here the challenge is to either reduce cost base or to find other business
to grow.
7
The most valuable tangible assets are people and as such, the biggest expenses
are compensation and benefits. Other operating expenses are generally less then
compensation expenses, and are known as communication and technology,
occupancy and depreciation, brokerage, clearing and exchanges fees, marketing
and advertisements, office supplies, exc.
8
Our focus in this article is on the efficiency frontier - how close investment
banks are to a best-practice frontier. Since engineering information on the
technology of financial institutions is not available, studies of frontier efficiency
rely on accounting measures of costs, outputs, inputs, revenues, profits, etc. to
impute efficiency relative to the best practice within the available sample.
For example, Berger and Humphrey (1997) stated, Frontier analysis provides
an overall, objectively determined, numerical efficiency value and ranking of firms
that is not otherwise available (Berger and Humphrey 1997, p. 2). Same authors
concluded that in terms of applications role of efficiency analysis is: to inform
government policy (e.g., by assessing the effects of deregulation, mergers, and
market structure on industry efficiency); to address research issues (e.g., by
determining how efficiency varies with different frontier approaches, output
definitions, and time periods); and to improve managerial performance (e.g., by
identifying best-practice and worst-practice branches within a single
firm) (Berger and Humphrey 1997, p. 46).
Today more than ever investment banking business has changed due to
globalization, deregulation and the accumulation of assets, innovations, aggressive
expansion and rivalry amongst industry leaders. In the new conditions to stay
competitive and successful investment bank needs to have a strong product line, the
ability to provide clients with an integrated solution, a strong global presence,
financial strength, integrity and teamwork.
Identification, analyses and measurement some of these factors could tell us
whether current efficiency measurements are sufficient to meet new trends in
investment banking and whether investment bank efficiency is determined by
structural characteristics (such as environmental factors) or technological progress.
7
For further readings see Davis (2003)
8
For further readings see Liaw (2006)
4
This paper contributes to the existing literature for the following reasons: First,
with providing a focus on the investment banking, which is surprisingly
inadequately explored (Berger and Humphrey 1997 cites no studies on investment
banks, and the only paper analyzing this sector was from Beccalli 2004) we
contribute to existing literature. Second, we perform cost and profit efficiency
comparison of the investment banking industries in G7 countries (US, UK, Japan,
Italy, Germany, France and Canada), through introduction of the appropriate
environmental variables in the cost and profit frontier estimations. Our goal is to
obtain a proper comparison of banking efficiency across countries by using a global
best practice econometric frontier whereby the banks in each country can be
compared against the same standard. Third, we conduct completely separate
frontiers analysis to check for the existence of structural differences between the
countries.
We choose to consider cost and profit efficiency comparison of the investment
banking industries because markets are becoming more competitive and current
differences in productive efficiency and costs among them, will determine each
country banking structure and future competitiveness. Reason for consideration of
alternative profit efficiency is in the case where assumptions underlying cost and
profit efficiency are not met. Dietsch and Lozano-Vivas (2000) said To predict
the effects of an expected increase in cross-border competition, knowledge of the
differences or similarities in the current banking costs and productive efficiencies
between countries is important (Dietsch and Lozano-Vivas 2000, p. 987).
Research regarding international comparisons of banking efficiency can be
summarized into three groups (Berger 2007). First, comparisons of the efficiencies
of banks in different nations, with all banks measured against a common frontier.
Second, comparisons of the efficiencies of banks in different nations, with banks
from each nation measured against their own nation-specific frontier. Third,
comparisons of the efficiencies of foreign-owned versus domestically owned banks
within the same nation, with both types of banks measured against the same nation-
specific frontier. As it can be seen from same study all three types of comparisons
have limitations, but only the third category addresses the key issues why cross
border consolidation among developed nations is so low and why foreign banks
presence is much higher in developing nations.
As Berger and Humphrey (1997) said, cross-country comparisons are difficult to
interpret because the regulatory and economic environments faced by financial
institutions are likely to differ importantly across nations and because the level and
quality of service associated with deposits and loans in different countries may
differ in ways that are difficult to measure. On the other side, they can provide
valuable information regarding the competitiveness of banks in different countries,
5
a concern of particular importance in the increasingly harmonized European market
for banking services and the perhaps more globalized financial markets of the
future.
6
2. Literature Review
This section reviews the existing literature on the influence of the environmental
variables on the banking efficiency studies. It is organized as follows: section 2.1.
examines the studies on the efficiency of investment banks, whereas section 2.2.
analyses the issue of the influence of environmental variables on banking
performance.
Section 2.1. The efficiency of the investment banks
Literature review by Berger and Humphrey (1997) quotes no studies on
efficiency of investment banks. This is due to the difficulties of modeling
successfully the peculiar nature of their production process (variables
identification) and partially to the lack of good quality data. For example, the same
authors mentioned only five studies that compare efficiency levels across countries
where three of these studies took Nordic countries for comparison, and other two
cross-country studies were applied for 11 OECD and 8 developed countries. As
well, most financial institution efficiency studies have been applied to the U.S.
banking industry.
Further motivation for our study comes from recent interest in comparison of
banking efficiency.
International comparisons of bank efficiency, literature review from Berger
(2007), investigate 100 studies that compare bank efficiencies across nations. These
comparisons differ in terms of how efficiency is measured. Studies that have
compared efficiency of different nations by using common frontier have mainly
focused on several European nations, and U.S (they have examined mainly
developed nations). Efficiency comparisons of different nations by using nation
specific frontiers have been applied for depository financial institutions and
insurance companies covering mainly U.S. and individual European nations (most
of these single-nation efficiency studies do not focus on international comparisons).
A number of recent studies have expanded the bank efficiency literature by
comparing the efficiencies of foreign-owned versus domestically owned banks
within the same nation using the same nation-specific frontier and they have been
dealing with developed and developing nations. Generally, problems with these
studies were that their results arent distinguished by the nation of origin of the
foreign owned banks, where only the most comprehensive developed nation studies
have identified the nation of origin of the foreign owned banks.
Only a few studies have been made on efficiency of investment firms like Anolli
and Resti 1996, and Beccalli 2004. Study from Beccalli (2004) has introduced two
7
new methods for cross country comparisons of the cost efficiency of UK and Italian
investment firms over the period 1995-1998. The first method shows differences
between the efficiency of the two countries by incorporating environmental
variables into the cross country common frontier. The second method shows
differences in the efficiency of the domestic versus foreign investment firms in the
two countries, by testing the ability to monitor and control on a cross-border basis.
Methodology used is based on parametric stochastic frontier approach (SFA) in
order to model cost efficiency. Data in the study are taken from financial
statements from both countries. The author found important to control for
environmental variables since they had significant influence on cost efficiency and
profitability in her research. In terms of cross country operations, it was found that
more efficient firms go abroad, exporting a more efficient model while less
efficient firms attract foreign investment firms with higher efficiency.
Over the past decade, substantial research has been done for measuring the
efficiency of financial institutions, mainly commercial banks. Different efficiency
concepts (cost, profit and alternative profit), different efficiency measurement
methods (parametric and non parametric) have been employed to improve current
methodology. Next, we give overview some of these researches.
Looking at the study from Berger and Mester (1997) we can realize that there is
still little information and no consensus on the sources of the substantial variation
in measured efficiency, although there has been significant research regarding the
efficiency of financial institutions.
There is a consensus in the literature that differences in frontier efficiency among
financial institutions exceed inefficiencies attributable to incorrect scale or scope of
output. However, there is really no consensus on the preferred method for
determining the best-practice frontier against which relative efficiencies are
measured (Berger, Hunter & Timme 1993).
Recent studies such as those from Hughes and Mester (1993), McAllister and
McManus (1993), Mester (1996), Berger and DeYoung (1997), Altunbas, et. al.
(2001), suggest that risk characteristics need to be incorporated in the underlying
industry cost or profit functions because, `unless quality and risk are controlled for,
one might easily miscalculate a bank's level of inefficiency'. What these studies
have in common is their focus only on one country.
Earlier consideration leads us to investigate efficiency and synergies on both cost
and revenue side. We can find many studies dealing with banking changes in cost
and profit efficiency, but they are mainly limited to US and Europe, while no study
has treated the investment banking cross-country cost and profit efficiency. Some
of the studies that have analyzed universal banking (which includes investment
8
banking in their business) together with traditional banking are analyzed in next
paragraphs.
Allen and Rai (1996) use distribution-free approach (DFA) and stochastic
frontier approach (SFA) for a systematic comparison of X-inefficiency measures
across 15 developed countries under different regulatory environments. The authors
estimate a global cost function for international banks to test for both input and
output inefficiencies. Results for 1988-1992 data (in the form of balance sheet and
income statement) suggest the prevalence of input X-inefficiencies far outweighs
that of output inefficiencies, and that the distribution-free model overestimates the
magnitude of X-inefficiencies relative to the stochastic cost frontier approach.
Vander Vennet (2002) used a parametric methodology in order to measure cost
and profit efficiency of European financial conglomerates and universal banks in
1995-1996. The sample consists of 2.375 EU banks from seventeen countries for
which all the variables were available from their published annual statements.
Results show that financial conglomerates are more revenue efficient than
specialized banks and that universal banks are more efficient on both cost and
revenue side. The author suggests, Further research should examine the sources
of the efficiency differences between various types of banks (Vander Vennet 2002,
p. 280).
Section 2.2. The importance of the environmental variables in the studies of
banking efficiency and performance
Berger et al. (1993) and Berger and Humphrey (1997) confirm that efficiency
scores differ markedly across studies. According to Mester (1993, 1997) and
Berger and Mester (1997), the failure to account for heterogeneity is a likely
candidate to cause this instability of efficiency results. Consequently, controlling
for heterogeneity results in efficiency scores that more accurately reflect
managements ability to minimize costs and maximize profits was also recognized
by Bos et al. (2008).
Cross-border comparison of efficiency was somewhat of a paradox, since banks
were compared to a common efficient frontier while assuming that different
countries have access to the same technology. Some research papers were working
on country specific environmental factors in order to avoid this technology problem
(Lozano-Vivas et al. 2002, Dietsch and Lozano-Vivas 2000).
According to Dietsch and Lozano-Vivas (2000), considering environmental
conditions while measuring banking efficiency differences across countries is
important because these differences should take into account the way in which
banking services are produced. In the research from Beccalli (2004), author also
9
proves the importance of environmental variables, for the cross-country
comparisons of the cost efficiency of UK and Italian investment firms.
Looking at the cross-country differences in banking efficiency Valverde, et al.
(2007) showed using data on large banks across 10 European countries for the
period 1996-2002, that they are roughly equally efficient after controlling for
differences in business environment, banking costs, and bank productivity.
Parametric approach for measuring cost efficiency used in this study was the
distribution free approach (DFA). Results suggest that the large banks in each of
the 10 countries had almost identical average efficiency values and since no
country have a strong efficiency advantage, it seems likely that state efforts to
promote national champions through favorable mergers may determine the
outcome.
In existing studies that estimate the efficiency of banks in a cross-national
scenario, the standard approach is to construct a common efficient frontier for all
firms, regardless of their home country. However, this standard approach is unable
to compare the different banking systems on an equal footing, because it does not
account for cross-country differences in regulation, economic and demographic
conditions, which are beyond the control of bank managers.
Without a common benchmark it is difficult to compare efficiency levels and
rankings (Coelli et al., 2005; Bos and Schmiedel, 2007). Most recent studies
therefore estimate a common benchmark, but seek to control for systematic
differences across banks that are not due to inefficiency.
9
In this paper, we therefore have to account for potential differences arising from
certain country-specific aspects of the banking technology on the one hand and
from the environmental and regulatory conditions on the other. In particular, the
economic environments are likely to differ significantly across countries. Three
categories of environmental variables are taken into account: (1) those that describe
the main macroeconomic conditions, which determine the banking product demand
characteristics, (2) variables that describe the structure of the banking industry, and
(3) those that characterize the accessibility of banking services. More explanation
about given variable is provided in the data and sample section of the paper.
9
Deprins and Simar (1989), Kumbhakar and Lovell (2000) observe that it can be difficult to determine if an
exogenous variable is a characteristic of production technology or a determinant of productive efficiency.
10
3. Methodology
Primary objective of our empirical analysis is to identify the framework for
comparing investment banks efficiencies across nations.
10
Two methods are
adopted: first, where separate frontiers are estimated to check for the existence of
structural differences between the countries; and second method which accounts for
the influences of environmental factors on the industry, by including indicator of
these factors in a definition of a common frontier.
11
We use stochastic frontier approach to model cost and profit efficiency.
Efficiency is estimated by using the parametric Stochastic Frontier analysis
(originally independently proposed by Aigner, Lovell and Schmidt (1977) and
Meeusen and Van den Broeck (1977)). This model can be expressed in the
following form:
Y
i
= x
i
+
i

i
=
i
+ u
i
i=1,...,N (1)
Where:
- Y
i
is the (logarithm of the) cost of production of the i-th firm;
- x
i
is a k1 vector of (transformations of the) input prices and output of the i-th
firm;
- is a vector of unknown parameters;
-
i
is disentangled in two main components: The first is the random error term (
i
),
accounting for measurement errors, bad luck and other factors unspecified in the
cost function. The
i
are assumed to be iid normal random variables with mean zero
and constant variance V
2
, |N(0,
V
2
)| and independent of the u
i
; The second term is
a non-negative cost inefficiency term (u
i
), added to the cost frontier representing
minimum cost. It is generally assumed to have a half normal or truncated normal
distribution, with variance equal to
U
2
, |N(0,
U
2
)|
12
.
Firm-specific estimates of technical inefficiency, u
i
, can be calculated by using
the distribution of the inefficiency term conditional on the estimate of the
10
According to Berger and Hannan (1994), efficiency measurements problems partially come from the fact that the
measured efficiencies of the different industries may not be comparable to each other at all.
11
Introduction of two methods is performed in order to overcome traditional limitations.
12
Assuming a half-normal distribution with mean zero implies that most banks are closely located to the frontier and
with small level of inefficiency. Another possibility is to relax this a priori assumption and estimate the mean of the
truncated normal distribution from the data.
11
composed error term,
i
(Jondrow et al., 1982). The mean of this conditional
distribution for the half normal model is shown as:
13
1
1
]
1

,
_

+
+

i
i
i
i i
F
E
) / ( 1
) / (
1
) / (
2

(2)
Where F(.) and (.) are respectively the standard normal distribution and the
standard normal density function. E(u
i
/
i
) is an unbiased but inconsistent estimator
of u
i
. The ratio of variability (standard deviation,) for u and can be used to
measure the relative inefficiency of a firm, where =
u
/
v
is a measure of the
amount of variation stemming from inefficiency relative to noise for the sample.
Estimates of bank specific cost efficiency are obtained by calculating:
CE
i
= [exp (-u
i
)]
-1
(3)
This measure takes on a value between 0 and 1. Cost efficiency equals one for a
fully efficient bank that operates on the efficient stochastic frontier.
14
The method of maximum likelihood is proposed for simultaneous estimation of
the parameters of the stochastic frontier and the model for the technical inefficiency
effects. We utilize the parameterization of Battese and Corra (1977) who replace

V
2
and
U
2
with
2
=
V
2
+
U
2
and =
U
2
/(
V
2
+
U
2
). The parameter must lie
between 0 and 1, where a value of zero means that all the deviations from the
frontier are due to random error and a value of one indicates that all deviations are
due to inefficiency. The technical efficiency of production for the i-th firm is
defined by equation:
TE
i
, = exp (-u
i
,) (4)
The prediction of the technical efficiencies is based on its conditional
expectation, given the model assumptions.
15
We choose to consider also alternative profit efficiency. Reason for
consideration of alternative profit efficiency is in the case where assumptions
underlying cost and profit efficiency are not met and are violated by the data.
According to Berger and Mester (1999, pg.3), profit maximization is superior to
cost minimization for most purposes because it is the more accepted economic goal
13
For further readings see Beccalli (2004, pg. 1368)
14
For further readings see Bos and Schiemdel (2007, pg.2086)
15
For further readings see Battese and Coelli (1995, pg.328)
12
of firms owners, who takes revenues as well as costs into account when making
decisions.
16
The frontier definition is the same as in the cost case, except for the dependent
variable: we replace total cost with total profit and the inefficiency term (u
i
) is
subtracted as in the production case, given that the frontier represents maximum
profit. Efficiency is given by the ratio of observed profit to frontier maximum profit
(the ideal best practice for which u
i
=0), equal to:
17
PE
i
= [exp (-u
i
)] (5)
In order to successfully perform first method of our research we specify both
stochastic cost and profit function for each country to verify whether or not
structural variables are the same in each country. Then we specify common
stochastic frontier with two cases (case I with only endogenous structural variables,
and case II with exogenous environmental variables).
To define a common frontier we use the following translog
18
specification
19
:


+ + + + +
+ + + + +
j j
j Ej j Ej
i
EE E
j
j i ij
i j i
j i
i j
ij
j
j i ij j j i i
p E m y E k E E h E f p y g
p p d y y c p b y a a TC
ln ln ln ln ln ln
2
1
ln ln ln
) ln ln ln ln (
2
1
ln ln ln
1 0
(6)
where
TC
i
is total cost for the i-th firm;
y
i
are the output quantities;
p
i
are the input prices.
E is the firm level of equity capital.
20
As usual, symmetry and linear homogeneity restrictions are imposed standardising
total cost TC and input prices p
i
by the last input price.
16
Some of the studies employ an alternative profit function in which the firm maximizes profits given output
quantities, rather than taking output prices as exogenous (Berger, Cummins and Weiss, 1996; Humphrey and Pulley,
1997; Akhavein, Berger, and Humphrey, 1997; Berger and Mester, 1997).
17
Fiordelisi and Ricci (2006, pg.11)
18
Berger and Mester (1997) used the distribution free approach and stochastic frontier approach for both translog and
the Fourier specification of the cost and profit function, and have concluded that difference between two methods are
not relevant. Same was observed and stated also by Vander Vennet (2002).
19
For further readings see Vander Vennet (2002, pg.264)
20
Equity is included into equation (as suggested by some authors such as Altunbas et al. 2000, Vander Vennet 2002,
Beccalli 2004, Bos and Schmiedel 2007) as a measure of financial capital. It is treated as a netput, specifying
interaction terms with other output and input prices
13
In this model we dont account for possible strong heterogeneity in the sample (so
efficiency estimates could be biased). Many authors have stressed the importance
of accounting for heterogeneity in the frontier definition.
According to Dietsch and Lozano-Vivas (2000) the environmental conditions faced
by financial institutions are likely to differ substantially and the specific
environmental conditions of each country play an important role in the definition
and specification of the common frontier of different countries.
In order to account for heterogeneity we follow Coelli et al. (1999) approach where
there are two different ways for including environmental conditions or firms
specific factors, that the authors specify as Case 1 and Case 2 Model.
21
Case 1 - Environmental factors have a direct influence on the production structure
One possibility is to consider that environmental conditions/firm specific factors
have a direct influence on the production structure. In this case we have to include
some control variables in the deterministic portion of the stochastic frontier: it
implies assuming that every firms face a different production function (Coelli et
al. 1999, p. 254).
So well have:


+ + +
K
k
M
j
i i ji j ki k i
u v z x y
1 1
0
ln ln ln

(7)
where we account for M environmental/firm specific factors z
j
assuming different
values for each i-th firm.
This specification can be straightforwardly adjusted for the cost case by assuming
the natural log of total cost as dependent variable and changing the sign of the
inefficiency component (u
i
). Using the translog specification, the deterministic
portion of the cost frontier is the following:
22

+ + + + + +
+ + + + +
j
M
j
j j
j
j Ej j Ej
i
EE E
j
j i ij
i j i
j i
i j
ij
j
j i ij j j i i
z p E m y E k E E h E f p y g
p p d y y c p b y a a TC
1
1 0
ln ln ln ln ln ln ln
2
1
ln ln ln
) ln ln ln ln (
2
1
ln ln ln

(8)
Case 2 - Environmental factors influence the inefficiency distribution
21
Same approach was used by Fiordelisi and Ricci (2006) for Banc assurance in Europe.
22
For further readings see Fiordelisi and Ricci (2006, pg.13)
14
A second possibility is to include the environmental/firm specific variables not
directly in the production frontier, but to use them for modelling the inefficiency
distribution.
As noted by Battese and Coelli (1995), the stochastic frontier production function is
estimated in the first stage under the assumption that the inefficiency effects (error
term) are identically distributed, while in the second stage the predicted technical
efficiencies are regressed upon a number of factors, hence suggesting the
inefficiency effects are not identically distributed. A more appropriate approach
involves the specification of a model in which both relations are estimated in a
single stage. This accounts for a stochastic frontier production function in which
the technical inefficiency effects are a function of firm characteristics.
23
The inefficiency components u
i
are assumed to be distributed independently, but
not identically. For each i-th firm the technical inefficiency effect is obtained as
truncation at zero of a normal distribution N(
i
,
2
) where the mean
i
is a function
of M factors representing the firm specific environment (Fiordelisi and Ricci 2006):

+
M
j
ji j i
z
1
0
(9)
The deterministic portion of the frontier remains the same as in equation 6 .
In this case we are supposing that all firms share the same technology, and
environmental/firm specific factors have an influence only on the distance between
each firm and the best practice.
The resulting efficiency estimates are incorporating the effect of environmental
factors and can be viewed as gross measure of efficiency.
24
23
For further readings see Coelli, et al.(1999, pg.255)
24
For further readings see Bos, et al.(2005, pg.11)
15
Data and variables
This study comprises banks balance sheet and annual reports data of G7
countries (US, UK, Japan, Italy, Germany, France and Canada) and Switzerland
over the 2000-2007 period. The data were compiled from the International bank
Credit Analysis Bankscope Database.
In order to estimate separate regional and common frontiers, the sample selection
requires us to consider only those countries, for which a sufficient large number of
observation is available. Number of observation is 992.
Table 1. reports the number of banks, by distinguishing for countries.
Table 1. Overview of the selected sample
Country/Year 2001 2002 2003 2004 2005 2006 2007 Total
Canada 4 4 4 5 4 4 4 29
France 5 3 4 6 12 12 9 51
Germany 12 12 13 13 11 15 16 92
Italy 3 3 2 5 8 11 9 41
Japan 21 19 21 24 25 24 24 158
UK 17 18 18 31 40 42 35 201
USA 9 11 11 12 10 11 11 75
Switzerland 55 47 48 50 49 48 48 345
Total 126 117 121 146 159 167 156 992
NOTE: We select banks with available balance sheets statements in Bankscope
for the years 2000-2007.
In the literature, the definition of the bank inputs and outputs varies across
studies and mainly depends on what a researcher pictures a bank to be.
Following Hughes and Mester (2008), outputs are typically measured by the
dollar volume of the banks assets in various categories. Inputs are typically
specified as labor, physical capital, deposits and other borrowed funds, and
sometimes equity capital. However, there is reasonable agreement about the
specification of most of the important inputs and outputs for financial institutions.
All agree that loans and other major assets of financial institutions should count as
outputs (Berger & Humphrey 1997).
Accordingly, investment banks inputs are defined as price of labor (P
1
), price of
physical capital (P
2
) and price of funds (P
3
). More precisely, the price of labor
16
equals the personnel expenses over total assets. The price of physical capital is
measured as operating expenses less labor and interest expenses over total fixed
assets. Price of funds equals total interest expenses over total funds.
Having in mind specific characteristics of investment banking business and
outputs that they produce, we consider added value approach where outputs are
defined based on their share of value added. Outputs are defined as loans (Y
1
) and
other earning assets (Y
2
). The variable equity controls for the differences in equity
capital risks across banks. In order to estimate cost and profit efficiency scores, we
use the total cost (TC) as the sum of interest, commission, fee, trading, and total
operating expenses and total profit (TP) as pre-tax profit, as our depended
variables.
In the table 2. we gives overview of the variables and their estimation:
INSERT TABLE 2.
Table 3. displays the description, mean, standard deviation, as well as minimum
and maximum values of all the input prices, outputs and depended variable.
INSERT TABLE 3.
In order to account for heterogeneity we follow Coelli et al. (1999) approach
where there are two different ways for including environmental conditions or firms
specific factors: Case I, environmental factors have a direct influence on the
production structure; Case II, environmental factors influence the inefficiency
distribution.
To decide on which firm specific factors to account for heterogeneity in our
research we choose on the basis of empirical literature evidence in this area.
Therefore we have to account for potential differences arising from certain country-
specific aspects of the banking technology on the one hand and from the
environmental and regulatory conditions on the other. In particular, the economic
environments are likely to differ significantly across countries. Three categories of
environmental variables are taken into account: (1) those that describe the main
macroeconomic conditions, which determine the banking product demand
characteristics, (2) variables that describe the structure of the banking industry
25
,
and (3) those that account for profitability.
25
According to Liaw (2006) major types of risk investment banks face include market risk, credit risk, operating
risk, reputation risk, legal risk and founding risk. The ability to properly and effectively identify, asses, monitor and
manage each type of risk is critical to an investment banks financial soundness and profitability.
Davis (2003) describes investment banking risk similar to Liaw, where according to him investment banks face
credit, market and operational as main risks, and unexpected and product risk as additional risks.
17
The Table 4. And 5. explains environmental variables (for Case I and Case II)
selected for our research, together with the studies that have used them previously.
INSERT TABLE 4.
INSERT TABLE 5.
18
Preliminary results
In our beginning we have been looking to make both separate frontiers and
common frontier for all countries. Due to a small number of bunks for some
countries we have excluded separate frontiers from our analysis.
For common frontier we have:
In the base model with only structural variables, where we have obtained next
results:
Base Model Cost Efficiency Results for Common Frontier
COST EFFICIENCY
Country Mean Std. Dev. Min. Max.
G7 countries & Switzerland (n=992)
0.70377 0.10965 0.17337 0.94976
Canada (n=29)
0.73448 0.04812 0.64020 0.81969
France (n=51)
0.62904 0.12983 0.33409 0.88071
Germany (n=92)
0.67939 0.12122 0.37095 0.91237
Italy (n=41)
0.62869 0.13480 0.19082 0.85540
Japan (n=158)
0.69358 0.09933 0.42994 0.92084
UK (n=201)
0.69940 0.14349 0.17337 0.94976
USA (n=75)
0.73091 0.04780 0.59030 0.83738
Switzerland (n=345)
0.72897 0.08022 0.26725 0.86095
Base Model Profit Efficiency Results for Common Frontier
PROFIT EFFICIENCY
Country Mean Std. Dev. Min. Max.
G7 countries & Switzerland (n=992)
0.86930 0.08043 0.00000 0.99992
Canada (n=29)
0.86708 0.05786 0.75589 0.95603
France (n=51)
0.88475 0.05572 0.72805 0.97988
Germany (n=92)
0.86041 0.08344 0.58211 0.99233
Italy (n=41)
0.90522 0.05161 0.77428 0.99159
Japan (n=158)
0.86249 0.04985 0.54403 0.99982
UK (n=201)
0.86295 0.07341 0.62283 0.99992
USA (n=75)
0.81241 0.15018 0.00000 0.99887
Switzerland (n=345)
0.88450 0.07380 0.51010 0.99685
19
The Case I, where environmental factors have a direct influence on the production
structure, we find not suitable for our data set (nevertheless we have run the model
and found evidence that support previous claim).
The Case II, where environmental factors influence the inefficiency distribution,
we find suitable for our data set. For the case II for profit efficiency results we have
some problems for the negative values, but we are working on solving those
problems.
In the table below, we give overview of the exogenous firm specific factors
determining the inefficiency distribution in the cost case.
Cost Efficiency
Variable Coefficient Significance
Z1 (PD) 0.72941 0.00007
Z2 (GDP) -1.33577 0.00063
Z3 (FDII) 0.05758 0.41440
Z4 (FDIO) 0.62856 0.00001
Z5 (BAS) -0.31863 0.00006
Z6 (CAR) -1.09319 0.00000
Z7 (CONC) -0.17582 0.05987
Z8 (NII) 0.48573 0.00000
Z9 (NNII) -0.17349 0.00328
Z10 (IR) 0.37217 0.00050
Z11 (LIQ) 0.76525 0.00042
Z12 (OBSE) 0.50467 0.00000
Z13 (LB) 0.27581 0.07367
Z14 (SR) 0.26006 0.00583
Z15 (ROA) 13.70692 0.00177
Z16 (ROE) -0.55655 0.15688
NOTES: A coefficient >0 means a positive effect on the inefficiency
component u
i
, and then a negative relationship with efficiency;
the opposite for a coefficient <0.
For the p value<0.01***, p value<0.05**, p value<0.1* and p
value>0.1 not significative.
20
References
A. Lozano-Vivas, J. T. Pastor and J. M. Pastor 2002. An Efficiency Comparison
of European Banking Systems Operating Under Different Environmental
Conditions. Journal of Productivity Analysis, 18, 59-77.
A. N. Berger, A. Demirg-Kunt, R. Levine and J. G. Haubrich 2003. Bank
Concentration and Competition: An Evolution in the Making. Working paper,
The Wharton School, University of Pennsylvania.
A. N. Berger and D. B. Humphrey 1994. Bank Scale Economies, Mergers,
Concentration and Efficiency: The US Experience. Working paper, The Wharton
School, University of Pennsylvania.
A. N. Berger and D. B. Humphrey 1997. Efficiency of Financial Institutions:
International Survey and Direction for Future Research. Working paper, The
Wharton School, University of Pennsylvania.
A. N. Berger and R. DeYoung 2001. The Effects of Geographic Expansion on
Bank Efficiency. Journal of Financial Service Research 19:2/3 163-184.
A. N. Berger and T. H. Hannan 1994. The Efficiency Cost of Market Power in the
Banking Industry: A test of the Quiet Life and Related Hypotheses. Working
paper, The Wharton School, University of Pennsylvania.
A. Tuch 2006. Investment Banking: Immediate Challenges and Future
Directions.Legal Studies Research Paper, The University of Sydney.
B. N. Anand and A. Galetovic 2006. Relathionships, Competition and Structure
of Investment Banking Markets. Journal of Industrial Economics, Volume LIV,
No. 2, 0022-1821.
E. Beccalli 2004. Cross Country Comparisons of Efficiency: Evidence from the
UK and Italian investment firms. Journal of Banking and Finance, 28 1363
1383.
E. Luciano and L. Regis 2007. Bank Efficiency and Banking Sector
Development: The Case of Italy. Working Paper, International Centre for
Economic Research.
E. J. Kane, H. Unal, and A. Demirg-Kunt 1991. Capital Positions of Japanese
Banks. Working Paper, International Eonomics Department, The World Bank.
21
F. Allen and A. M. Santomero 1999. What Do Financial Intermediaries Do?
Working paper, The Wharton School, University of Pennsylvania.
J. P. Bonin, I. Hasan and P. Wachtel 2005. Bank Performance, Efficiency and
Ownership in Transition Countries. Journal of Banking and Finance, 29, 3153.
J. P. Hughes and L. J. Mester 2008. Efficiency in Banking: Theory, Practice and
Evidence. Working paper, The Wharton School, University of Pennsylvania.
J.W.B. Bos and H. Schmiedel 2007. Is There a Single Frontier in a Single
European Banking Market? Journal of Banking and Finance, 31, 20812102.
J.W.B. Bos, M. Koetter, J.W. Kolari and C.J.M. Kool 2008. Effects of
heterogeneity on bank efficiency scores. European Journal of Operational
Research.
Katrina Ellis, Roni Michaely and Maureen OHara 2005. Competition in
Investment Banking: Proactive, Reactive, or Retaliatory? Unpublished Paper,
University of California, Davis.
M. Dietsch and A. Lozano-Vivas 2000. How the Environment Determines
Banking Efficiency: A comparison between French and Spanish industries.
Journal of Banking and Finance, 24, 985-1004.
R. Vander Vennet 2002. Cost and Profit Efficiency of Financial Conglomerates
and Universal Banks in Europe. Journal of Money, Credit and Banking, Vol. 34,
No. 1, pp. 254-282.
S. C. Valverde, D. B. Humphrey and R. Lpez del Paso 2007. Do Cross Country
Differences in Bank Efficiency support a policy of national champions? Journal
of Banking and Finance, 31, 21732188.
T. Coelli, S. Perelman and E. Romano 1999. Accounting for Environmental
Influences in Stochastic Frontier Models: With Application to International
Airlines. Journal of Productivity Analysis, 11, 251-273.
Yener Altunbas, Ming-Hau Liu, P. Molyneux and R. Seth 2000. Efficiency and
risk in Japanese banking. Journal of Banking and Finance, 24, 1605-1628.
22
Table 2. Variable definition
Type of Variable Symbol Variable Name Calculation
Dependent variable TC Total Cost
TC is obtained as the sum of interest expense,
commission expense, fee expense, trading expense
and total operating expenses
Dependent variable TP Total Profit TP is obtained as pre-tax profit
Input Price P1 Price of Labor
P1 is calculated as personnel expenses over total assets
Input Price P2
Price of Physical
Capital
P2 is calculated as other administrative expenses and
other operating expenses over total fixed assets
Input Price P3 Price of Funds
P3 is calculated as total interest expenses over total
funds
Output Y1 Loans Y1 represents loans
Output Y2
Other Earning
Assets
Y1 represents other earning assets
Control variable E Total Equity E represents equity capital
23
Table 3. Descriptive statistics of outputs, inputs and control variables
Variable Description Mean Median Std. Dev. Min. Max.
TC Total Cost 1418986.326 202828.997 5356962.056 3090.235 75506000.000
TP Pre-Tax Profits 156199.286 25009.733 919083.984 -12831000.000 10426000.000
Y1 Loans 5328236.750 543002.175 17842688.739 88.129 248222594.000
Y2 Other Earning Assets 23635462.258 1089135.883 72202139.572 104.866 796332000.000
P1 Price of Labour 0.045 0.018 0.068 0.000 0.558
P2 Price of Physical Capital 683809.602 121872.506 2279613.974 1961.157 24233012.268
P3 Price of Funds 0.056 0.024 0.264 0.000 4.571
E Total Equity 1405882.276 304844.799 3922136.233 1089.681 39038000.000
24
Table 4. Environmental variables definition for Case I
Variable type Variable name Symbol Studies
Investment banking
risk exposure
Capital risk exposure CAR
Dietsch and Lozano-Vivas 2000, Lozano-Vivas et al.
2002 as control variable, Altunbas et al. 2000,
Athanasoglou et al. 2006, Brissimis et al. 2008,
Lapetite et al. 2008
Insolvency risk exposure IR Lapetite et al. 2008
Liquidity risk exposure LIQ
Altunbas et al. 2000 as proxy, Demirguc-Kunt and
Huizinga 2004, Brissimis et al. 2008, Fiordelisi and
Molyneux 2009
Securities risk exposure SR ????
CAR=equity/assets
IR=(100+average ROE)/SDROE
LIQ=Liquid assets/assets
SR=Total securities/total assets
25
Table 5. Environmental variables definition for Case II
Variable type Variable name Symbol Studies
Social and
macroeconomic
conditions
Population density PD
Dietsch and Lozano-Vivas 2000, Lozano-Vivas et al.
2002, Bos et al. 2005, Valverde et al. 2007, Fiordelisi
and Molyneux 2009
GDP per capita GDP
Salas and Saurina 2003, Athanasoglou et al. 2006,
Valverde et al. 2007, Fitzpatrick and McQuinn 2007,
Brissimis et al. 2008, Fiordelisi and Molyneux 2009
FDI Inflows FDII Beccalli 2004
FDI Outflow FDIO Beccalli 2004
Banking structure
(industry specific,
determinants and risk)
Bank asset size
a
BAS
Altunbas et al. 2000 as proxy, Lozano-Vivas, et al.
2002 asset quality as control variable, Valverde, et al.
2007 use total asset per bank, Athanasoglou et al. 2006,
Lapetite et al. 2008, Fiordelisi and Molyneux 2009
Capital risk exposure
b
CAR
Dietsch and Lozano-Vivas 2000, Lozano-Vivas et al.
2002 as control variable, Altunbas et al. 2000,
Athanasoglou et al. 2006, Brissimis et al. 2008,
Lapetite et al. 2008
Herfindhal index of concentration
c
CONC
Dietsch and Lozano-Vivas 2000, Vander Vennet 2002,
Athanasoglou et al. 2006, Fiordelisi and Molyneux
2009
Income diversification
d
NII Lapetite et al. 2008, Fiordelisi and Molyneux 2009
Income diversification
e
NNII Lapetite et al. 2008, Fiordelisi and Molyneux 2009
Insolvency risk exposure
f
IR Lapetite et al. 2008
Liquidity risk exposure
g
LIQ
Altunbas et al. 2000, Demirguc-Kunt and Huizinga
2004, Brissimis et al. 2008, Fiordelisi and Molyneux
2009
Off-Balance Sheet Exposure
h
OBSE Casu and Girardone 2007
Publicly listed bank
i
LB Beccalli 2004, Fiordelisi and Molyneux 2009
Securities risk exposure
j
SR ????
ROA
k
ROA Athanasoglou et al. 2006, Lapetite et al. 2008
Profitability
ROE
l
ROE
Berger et al. 1993, Allen and Rai 1996, Lozano-Vivas
et al. 2002, Vander Vennet 2002, Beccalli 2004,
Athanasoglou et al. 2006, Lapetite et al. 2008
a
BAS= total assets
b
CAR is calculated as equity over assets
c
Obtained as the sum of the squares of market shares for all banks operating in the industry
d
NII = net interest income/net profits
e
NNII = net non-interest income/net profits
f
IR=(1+average ROE)/SDROE
g
LIQ is calculated as liquid assets over assets
h
OBSE is measured as off-balance sheet items over total assets
i The bank is publicly listed or otherwise, where 1 = listed; 0= non-listed
j
SR is calculated as total securities over total assets
k
ROA=Net profits/total assets
l
ROE=Net profits/Shareholders equity (total assets - total liabilities)
26