Banking reforms and the evolution of cost efficiency in Indian public sector banks Sunil Kumar Received: 28 November 2010 / Accepted: 27 January 2012 / Published online: 8 March 2012 Ó Springer Science+Business Media, LLC. 2012 Abstract This paper analyses the trends of cost efficiency and its components across Indian public sector banks (PSBs) during the post-deregulation period spanning from 1992/1993 to 2007/2008. The study also examines the issue of convergence in cost, technical and allocative efficiency levels of Indian PSBs. The empirical results indicate that deregulation has had a positive impact on the cost efficiency of Indian public sector banking industry over the period of study. Further, technical efficiency of Indian PSBs followed an uptrend, while allocative efficiency followed a path of deceleration. We note that in Indian public sector banking industry, the cost inefficiency is mainly driven by technical inefficiency rather than allocative inefficiency. The convergence analysis reveals that the inefficient PSBs are not only catching-up but also moving ahead of the efficient ones, i.e., the banks with the low level of cost efficiency at the beginning of the period are growing more rapidly than the highly cost efficient banks. In sum, the study confirms a strong presence of r- and b-convergence in cost efficiency levels of Indian public sector banking industry. Keywords Data envelopment analysis Public sector banks Cost efficiency Technical efficiency Allocative efficiency Convergence JEL Classifications G21 G15 1 Introduction From the early 1970s through the late 1980s, the role of market forces in the Indian banking system was almost missing, and excess regulation in terms of high liquidity S. Kumar (&) Faculty of Economics, South Asian University, New Delhi 110067, India e-mail: [email protected]123 Econ Change Restruct (2013) 46:143–182 DOI 10.1007/s10644-012-9121-8
Transcript
Banking reforms and the evolution of cost efficiencyin Indian public sector banks
Sunil Kumar
Received: 28 November 2010 / Accepted: 27 January 2012 / Published online: 8 March 2012
� Springer Science+Business Media, LLC. 2012
Abstract This paper analyses the trends of cost efficiency and its components
across Indian public sector banks (PSBs) during the post-deregulation period
spanning from 1992/1993 to 2007/2008. The study also examines the issue of
convergence in cost, technical and allocative efficiency levels of Indian PSBs. The
empirical results indicate that deregulation has had a positive impact on the cost
efficiency of Indian public sector banking industry over the period of study. Further,
technical efficiency of Indian PSBs followed an uptrend, while allocative efficiency
followed a path of deceleration. We note that in Indian public sector banking
industry, the cost inefficiency is mainly driven by technical inefficiency rather than
allocative inefficiency. The convergence analysis reveals that the inefficient PSBs
are not only catching-up but also moving ahead of the efficient ones, i.e., the banks
with the low level of cost efficiency at the beginning of the period are growing more
rapidly than the highly cost efficient banks. In sum, the study confirms a strong
presence of r- and b-convergence in cost efficiency levels of Indian public sector
banking industry.
Keywords Data envelopment analysis � Public sector banks � Cost efficiency �Technical efficiency � Allocative efficiency � Convergence
JEL Classifications G21 � G15
1 Introduction
From the early 1970s through the late 1980s, the role of market forces in the Indian
banking system was almost missing, and excess regulation in terms of high liquidity
S. Kumar (&)
Faculty of Economics, South Asian University, New Delhi 110067, India
requirements and state interventions in allocating credit and determining the prices
of financial products resulted in serious financial repression. Realizing the presence
of the signs of financial repression and to seek an escape from any potential crisis in
the banking sector, the Government of India (GOI) embarked on a comprehensive
banking reforms plan in 1992 with the objective of creating a more diversified,
profitable, efficient and resilient banking system. The main agenda of the reforms
process was to focus on key areas: (1) restructuring of PSBs by imparting more
autonomy in decision making, and by infusing fresh capital through recapitalization
and partial privatization; (2) creating contestable markets by removing entry barriers
for de novo domestic private and foreign banks; (3) improving the regulatory and
supervisory frameworks; and (4) strengthening the banking system through
consolidation. To meet this agenda, the policy makers heralded an episode of
interest-rates deregulation, standardized minimum capital requirements as per Basel
norms, prudential norms relating to income recognition, assets classification and
provisioning for bad loans, and changes in the legal and supervisory environment.
Given the broad sketch of banking reforms portrayed above, one may ask
whether the efficiency performance of Indian PSBs since the launching of reforms
in 1992 has improved or not. In this paper, we made an attempt in this direction. In
particular, our endeavour here is to evaluate the performance of PSBs in the post-
reforms period by looking at the trends of cost efficiency (CE) and convergence in
its levels across banks. Our analysis evolves in two steps. First, using the data of 27
PSBs over a period 16 years (1992/1993–2007/2008), we calculate cost, technical
and allocative efficiency scores for individual PSBs using the technique of data
envelopment analysis (DEA). Second, we use traditional cross-sectional regression
approach for investigating the presence of r- and b-convergence in the efficiency
levels. In the contemporary literature, a similar approach has been used by Tomova
(2005), Mamatzakis et al. (2007), Weill (2008), and Brack and Jimborean (2009) to
examine the convergence in bank efficiency levels across European countries, and
by Daley and Matthews (2009) for testing the convergence in efficiency levels of
Jamaican banks.
The paper augments the extant literature related to the efficiency of Indian banks
in two major directions. The first contribution of this paper is to provide an
extensive analysis of variations in cost efficiency and its components across PSBs
during the post-reforms period. The existing studies on the efficiency of Indian
banks have reported the results for specific groups of banks (particularly defined by
ownership and size) rather than those of individual banks.1 However, we may get a
misleading picture from a group-specific analysis if one or a set of some out-
performing bank(s) supersede the dismal efficiency levels of the remaining banks of
the group. The bank-wise results reported in the present study avoid the problem of
dominance of one bank over others within the same group, and would be more
useful in designing micro-level policies in the banking industry. Our second
contribution to the literature is to add an analysis of testing of the convergence in
cost efficiency levels of Indian banks. To the best of our knowledge, no attempt has
1 See, for example, Ataullah et al. (2004), Shanmugam and Das (2004), Sensarma (2005, 2006, 2008),
Rezvanian et al. (2008), and Das and Ghosh (2006, 2009).
144 Econ Change Restruct (2013) 46:143–182
123
been made in this direction in the literature on Indian banking industry so far. The
results of the present study could serve as a performance-diagnostic test for the
regulators to see the impacts of the deregulation process launched in the early
1990 s on the cost efficiency of Indian public sector banking industry.
Our empirical investigation suggests that deregulation has had a positive impact
on the performance of Indian public sector banking industry in terms of cost and
technical efficiencies over the entire study period (1992/1993–2007/2008). How-
ever, improvement in cost efficiency has been noticed to be more pronounced in the
years belonging with the second phase of reforms (1998/1999–2007/2008) relative
to first phase (1992/1993–1997/1998). The analysis of sources of cost inefficiency
reveals that on an average, the cost inefficiency in Indian public sector banking
industry originates primarily due to technical inefficiency (i.e., managerial problems
in using the financial and physical resources) rather than allocative inefficiency (i.e.,
regulatory environment in which banks are operating). Finally, the study reports the
presence of strong r- and b-convergence in cost efficiency levels of Indian PSBs
during the deregulatory regime. Overall, Indian public sector banking industry has
not only experienced significant efficiency gains during the post-reforms period but
also witnessed convergence in cost efficiency levels among PSBs.
The rest of the paper unfolds as follows. In the next section, we present the
relevant literature review. Section 3 provides an overview of the Indian banking
system and delineates the major areas of banking reforms in India. Section 4
presents the conceptual framework for measuring cost efficiency and its components
using DEA approach. Specification of banks’ inputs and outputs, and data are
presented in Sect. 5. Section 6 discusses the empirical findings, and finally, Sect. 7
concludes the paper.
2 Relevant literature review
The literature on banking efficiency offers a myriad of studies over the past few
decades, but most of these studies are limited to developed countries such as the US
and European Union. Interested parties can refer the survey articles by Berger et al.
(1993), Berger and Humphrey (1997), Berger and Mester (1997), Ashton and
Hardwick (2000), Casu and Molyneux (2001), Mokhtar et al. (2006), and Fethi and
Pasiouras (2010). In recent times, many countries have engaged in a process of
deregulation and liberalisation of their banking systems with the avowed objective
of improving the efficiency and performance of banks. The rationale behind
introducing deregulatory and liberalisation measures like interest rates deregulation,
removal of entry barriers for private and foreign banks, etc. in the banking system is
that these measures unleash the competitive forces in the system, which in turn
compel the banks to bring the output-input combination to the optimal production
frontier and induce them to produce financial services at lower costs. This led to the
publication of a large number of research papers, which explore the efficiency
performance of banks in the wake of financial deregulation and liberalization.
Nonetheless, there are mixed findings in the literature about whether deregulation
helped the banks to improve efficiency or not (Berger and Humphrey 1997). Results
Econ Change Restruct (2013) 46:143–182 145
123
appear to vary depending on the country, bank ownership, and size (Avkiran
2000).
Notable studies which reported a positive impact of deregulation on the
efficiency and productivity of banks are Berg et al. (1992) for Norway; Zaim (1995),
Isik and Hassan (2002a, b) for Turkey; Maghyereh (2004) for Jordan; Leightner and
Lovell (1998) for Thailand; Chen et al. (2005), Jiang et al. (2009) for China; Patti
and Hardy (2005), Qayyum (2008), Burki and Niazi (2010) for Pakistan; Mukherjee
et al. (2000), Alam (2001) for US; Kumbhakar et al. (2001), Kumbhakar and
Lozano-Vivas (2005), Lozano-Vivas (1998) for Spain; Avkiran (2000), Sturm and
Williams (2004) for Australia; Rebelo and Mendes (2000), Canhoto and Dermine
(2003) for Portugal; Ali and Gstach (2000) for Austria; Lopez-Cortes (1997) for
Mexico; Ariss (2008) for Lebanon; Kondeas et al. (2008) for European Union
countries; Hermes and Nhung (2010) for ten Latin American and Asian countries;
Fethi et al. (2011) for Egypt. In contrast to aforementioned studies, there are studies
that reported a negative effect of deregulatory measures on the efficiency and
productivity of banks. Some prominent studies in this context are Humphrey (1991,
1993), Humphrey and Pulley (1997), Grabowski et al. (1994), Elyasiani and
Mehdian (1995), Wheelock and Wilson (1999) for US; Grifell-Tatje and Lovell
(1996) for Spain; Kumbhakar and Wang (2007) for China; Christopoulos and
Tsionas (2001) for Greece; Denizer et al. (2000, 2007) for Turkey; Rizvi (2001) for
Pakistan; Hao et al. (2001), Mahadevan and Kim (2001) for Korea; Cook et al.
(2001) for Tunisia.
In Indian context too, though the majority of studies portrayed a positive impact
of deregulation and liberalization policies on the efficiency and productivity of
Indian banks yet a few studies also reported an adverse or insignificant effect of
these policies on the performance of banks. The studies of Bhattacharyya et al.
(1997a, b), Ram Mohan and Ray (2004), Shanmugam and Das (2004), Ataullah
et al. (2004), Reddy (2004), Das et al. (2005), Chatterjee (2006), Mahesh and
Rajeev (2006), Sensarma (2006), Howcroft and Ataullah (2006), Zhao et al. (2007),
Jaffry et al. (2007), Rezvanian et al. (2008), Mahesh and Bhide (2008), Kumar and
Gulati (2009), Tabak and Tecles (2010), Sreeramulu et al. (2010) and Sahoo and
Mandal (2011) broadly concluded that the deregulatory policy regime has had a
positive and favourable impact in terms of efficiency improvement and productivity
surge of Indian banking industry. In addition, Gourlay et al. (2006) noted
considerable efficiency gains flowing from the post-reforms mergers among Indian
banks. The studies that reported either an adverse or insignificant effect of
deregulation on the performance of Indian banks include Kumbhakar and Sarkar
(2003), De (2004), Sensarma (2005), Galagedera and Edirisuriya (2005), Das and
Ghosh (2006), Sensarma (2008) and Kalluru and Bhat (2009).
Overall, as illustrated by the empirical evidences provided above, the effect of
deregulation on the efficiency and productivity of the banking sector seems highly
dependent on the specific economic environment of each country. The reported
adverse effect in a few studies may be due to the short-term effects of liberalization
such as credit rationing, high spreads and weakening loan quality (Musonda 2008).
These problems tend to be exacerbated under an unstable macroeconomic
environment which is often associated with the early years of reforms. This
146 Econ Change Restruct (2013) 46:143–182
123
suggests that the hypothesis stating that deregulation always improves efficiency
and productivity may be rejected.
3 Developments in Indian banking sector: an overview
3.1 Banking sector in India
The Reserve Bank of India (RBI) is the central bank of the country that regulates the
operations of banks, manages the money supply and discharges other myriad
responsibilities that are usually associated with a central bank. The banking system
in India comprises commercial and cooperative banks, of which the former accounts
for more than 90% of the assets of the banking system. Within the category of
commercial banks, there are two types of the banks: (1) schedule commercial banks
(i.e., which are listed in Schedule II of the Reserve Bank of India Act, 1934); and (2)
non-scheduled commercial banks. Depending upon the pattern of ownership,
scheduled commercial banks can be classified into three types: (1) Public sector
banks which include (a) State Bank of India (SBI) and its associate banks,
(b) Nationalized banks, and (c) other public sector banks; (2) Private Sector Banks
consist of private domestic banks (which can further be classified as old private
banks that are in business prior to 1992, and de novo private banks that are
established after 1992), and foreign banks; and (3) Others comprising Regional
Rural Banks (RRBs) and Local Area Banks.
Of these, PSBs have a countrywide network of branches and account for over
70% of total banking business. The contribution of PSBs in India’s economic and
social development is enormous and well documented. They have a strong presence
in rural and semi-urban areas, and employ a large number of staff. On the other
hand, de novo private domestic banks are less labour-intensive, have limited number
of branches, have adopted modern technology, and are more profitable. Though
foreign banks are more techno-savvy and have carved a niche in the market, but
they confine their operations in major urban centres. PSBs sponsor the RRBs and
their activities are localized. Further, RRBs serve the needs for rural credit and have
a diminutive share (about 3%) in the commercial banking industry of India. It has
been observed that the market share of PSBs in terms of investments, advances,
deposits, and total assets is more than 70%. About 87% of branches of the
commercial banks in India belong to PSBs. Further, their share in the total
employment provided by commercial banking industry is about 81%. In brief, PSBs
command a lion’s share of the Indian banking industry.
3.2 Banking reforms in India
Up until the launching of banking reforms in 1992, the Government of India (GOI)
used the banking system as an instrument of public finance (Hanson and Kathuria
1999). Substantial and increasing volumes of credit were channeled to the
government at below-market rates through high and increasing cash reserve
requirements (CRR) and statutory liquidity requirements (SLR) in order to fund a
Econ Change Restruct (2013) 46:143–182 147
123
large and increasing government fiscal deficit at relatively low cost (Sen and Vaidya
1997).2 The commercial banks, especially PSBs, were obliged to allocate a
substantial part of their total loan portfolio to ‘priority’ sectors (such as agriculture
and small-scale industries) at a rate that was below the market rate. Furthermore,
interest rates on both deposits and advances were completely administered by the
RBI. There was virtually no autonomy to the banks even in taking the decision to
open new bank branches. The government also tightly regulated the licensing of
market entry of new domestic and foreign banks, which resulted into domination of
the banking industry by PSBs. PSBs stumbled downhill throughout the period
between 1980–1992 as the non-performing assets continued to pile up whilst
standard assets were doing little to generate any significant profits for the banks.
Besides these, there were several weaknesses in the organizational structure of
banks like lack of delegation, weak internal controls, and non-transparent
accounting standards (Mohan and Prasad 2005). In sum, all the signs of financial
repression such as excessively high-reserve requirements, credit controls, interest
rate controls, strict entry barriers, operational restrictions, pre-dominance of state-
owned banks, etc., were present in the Indian banking system.
To get rid of the distressed banking situation, the GOI embarked on a strategy of
reform measures in the financial sector, in general and banking sector, in particular.
Note that the banking reforms in India had two distinct phases. The first phase of
reforms introduced consequent to the release of the Report of the Committee on the
Financial System (Chairperson: M. Narasimham 1991). The focus of this phase of
the reforms was on economic deregulation targeting at relaxing credit and interest
rates controls, and removing restrictions on the market entry and diversification. The
second phase of reforms, introduced subsequent to the recommendations of the
Committee on Banking Sector Reforms (Chairperson: M. Narasimham 1998), was
targeted to enhance prudential regulations, and improve the standards of disclosure
and levels of transparency so as to minimize the risks that banks assume and to
ensure the safety and soundness of both individual banks and the Indian banking
system as a whole. On the whole, the key objective of the banking reforms was to
transform the operating environment of the banking industry from a highly
regulated system to a more market-oriented one, with a view to increase
competitiveness and efficiency (Sarkar 2004).
Although the broad contours of reform measures in the banking sector have been
provided by the aforementioned committees but a large number of committees and
working groups have been constituted for addressing the specific issues in the
banking sector. For example, Janakiraman Committee (1992) investigated irregu-
larities in fund management in commercial banks and financial institutions.
Padmanabhan Committee (1996) focused on the on-site supervision of banks, and
recommended the implementation of CAMELS rating methodology for on-site
supervision of the banks. Khan Committee (1997) suggested specific measures for
bringing about harmonization in the lending and working capital finance by banks
and Development Financial Institutions (DFIs). Verma Committee (1999)
2 By 1991, the pre-emption under the cash reserve ratio and the statutory liquidity ratio, on an
incremental basis, had reached 63.5% of net demand and time liabilities.
148 Econ Change Restruct (2013) 46:143–182
123
concentrated on restructuring of weak PSBs. The committee identified three weak
banks, viz. Indian Bank, United Commercial Bank and Union Bank of India, and
suggested introducing Voluntary Retirement Fund enabling the bank to reduce
excess manpower. Vasudevan Committee (1999) recommended the strategy on
upgradation of the existing technology in the banking sector. Mittal Committee
(2000) made vital recommendations on the regulatory and supervisory frameworks
for internet banking in India. Mohan Committee (2009) which is popularly known
as the Committee on Financial Sector Assessment has suggested significant
measures to improve the stability and resilience of the Indian financial system.
In the post-1992 period, the reform measures have been taken in six directions for
improving the efficiency and profitability of Indian banks (see, Reddy 1998, 2002;
Ahluwalia 2002; Shirai 2002, Rangarajan 2007; Roland 2008, for details). First, for
making available a greater quantum of resources for commercial purposes, the
statutory pre-emption has gradually been lowered.3 Second, the structure of
administered interest rates has been almost totally dismantled in a phased manner.4
Third, the burden of directed sector lending has been gradually reduced by
(a) expanding the definition of priority sector lending, and (b) liberalizing lending
rates on advances in excess of Rupee 0.2 million. Fourth, entry regulations for
domestic and foreign banks have been relaxed to infuse competition in the banking
sector.5 Fifth, the policy makers introduced improved prudential norms related to
capital adequacy,6 asset classification7 and income recognition in line with
international norms, as well as increased disclosure level.8 Sixth, towards
strengthening PSBs, the GOI recapitalized public sector banks to avert any
financial crisis and to build up their capital base for meeting minimum capital
adequacy norms.9
To sum up, we note that during the last 19 years, the policy makers adopted a
cautious approach for introducing reform measures in the Indian banking sector.
3 The combined pre-emption under CRR and SLR, amounting to 63.5% of net demand and time
liabilities in 1991 (of which CRR was 25%) has since been reduced and presently, the combined ratio
stands below 35% (of which, the SLR is at its statutory minimum at 25%).4 Prior to 25th October, 2011, except saving deposit account, non-resident Indian (NRI) deposits, small
loans up to Rupee 0.2 million and export credit, all the interest rates were fully deregulated. Recently,
RBI also deregulated the savings bank deposit interest rate.5 In 1993, the RBI issued guidelines concerning the establishment of new private sector banks. Nine new
private banks have entered the market since then. In addition, over twenty foreign banks have started their
operations since 1994.6 India adopted the Basel Accord Capital Standards in April 1992. An 8% capital adequacy ratio was
introduced in phases between 1993 and 1996, according to banks ownership and scope of their operations.
Following the recommendations of Narasimham Committee II, the regulatory minimum capital adequacy
ratio was later raised to 10% in the phased manner.7 The time for classification of assets as non-performing has been tightened over the years, with a view to
move towards the international best practice norm of 90 days by end 2004.8 From 2000 to 2001, the PSBs are required to attach the balance sheet of their subsidiaries to their
balance sheets.9 The GOI has injected about 0.1% of GDP annually into weak public sector banks (Hanson 2005;
Rangarajan 2007). During the period 1992/1993–2001/2002, GOI contributed some Rupee 177 billion,
about 1.9% of the 1995/96 GDP, to nationalized banks (Mohan and Prasad 2005).
Econ Change Restruct (2013) 46:143–182 149
123
The foremost objective of the banking reforms process was to improve the
performance of PSBs in their operations and to inculcate a competitive spirit in
them. Against this backdrop, we confine our analysis to PSBs which constitute the
most significant segment of the Indian banking sector.
4 Methodological framework for computing efficiency scores
4.1 Concept of cost efficiency and measurement techniques
An analytical framework to measure cost efficiency10 of a firm dates back to the
seminal work of Farrell (1957). Measuring cost efficiency requires the specification
of an objective function and information on market prices of inputs. If the objective
of the production unit is that of cost minimization, then a measure of cost efficiency
is provided by the ratio of minimum cost to observed cost (Lovell 1993). A cost
efficiency measure provides how close a bank’s cost is to what a best-practice
bank’s cost would be for producing the same bundle of outputs (Weill 2004). In
Farrell’s framework, the cost efficiency (CE) is composed of two distinct and
separable components: technical efficiency (TE)—the ability of a firm to produce
existing level of output with the minimum inputs (input-oriented), or to produce
maximal output from a given set of inputs (output-oriented); and allocative
efficiency (AE)—the ability of a firm to use the inputs in optimal proportions, given
their respective prices. Allocative efficiency relates to prices, while technical
efficiency relates to quantities (Barros and Mascarenhas 2005). The level of
technical efficiency is related to managerial decisions while allocative efficiency is
related to the regulatory environment or macroeconomic conditions (Lovell 1993).
Therefore, technical inefficiency is caused and correctable by management, and
allocative inefficiency is caused by regulation and may not be controlled by the
management (Hassan 2005). Matthews et al. (2008) interpreted allocative ineffi-
ciency (i.e., sub-optimal factor mix) as bureaucratic rent-seeking inefficiency. In
fact, the relationship between CE, TE, and AE is expressed as: CE = TE 9 AE.
Most empirical studies pertaining to the measurement of cost efficiency in
banking industry applied either parametric or non-parametric methods. These
approaches use different techniques to envelop the observed data and make different
accommodations for random noise and for the flexibility in the structure of the
production technology (Lovell 1993). In parametric approaches, a specific
functional form of the production function like Cobb-Douglas and transcendental
logarithmic (translog), etc. is required to specify a priori. The efficiency is then
assessed in relation to this function with constant parameters and will be different
depending on the chosen functional form. The most commonly used parametric
methods are the Stochastic Frontier Approach (SFA), the Thick Frontier Approach
(TFA), and the Distribution Free Approach (DFA). In contrast, non-parametric
approaches do not specify a functional form, and involve solving linear program, in
10 In banking efficiency literature, the term cost efficiency is being used interchangeably with economic
efficiency, X-efficiency and overall efficiency.
150 Econ Change Restruct (2013) 46:143–182
123
which an objective function envelops the observed data; then efficiency scores are
derived by measuring how far an observation is positioned from the envelope or
frontier (Delis et al. 2009). The most widely used non-parametric approaches are
Data Envelopment Analysis (DEA) and Free Disposal Hull (FDH). However, no
consensus has been reached in the literature about the appropriate and preferred
estimation methodology (Iqbal and Molyneux 2005; Staikouras et al. 2008). On the
basis of review of methodological cross-checking studies, Musonda (2008) noted
that there is really no loss of generality in using either methodology to analyze the
efficiency in banking, and the choice of the approach adopted is a matter of
convenience and is largely dictated by the data used in the analysis and ease of
application.
4.2 Data envelopment analysis models
For getting a convenient decomposition of cost efficiency, this paper uses data
envelopment analysis (DEA) to estimate empirically the cost, technical and
allocative efficiency scores for individual PSBs. We made use of DEA models with
reference to a technology with constant returns-to-scale.11 The computational
procedure used to implement the DEA approach to the measurement of cost
efficiency and its components is of three steps. The first step is to obtain a measure
of TE as introduced by Charnes et al. (1978). Consider N banks each producing Mdifferent outputs using K different inputs. The K 9 N input matrix, X, and the
M 9 N output matrix, Y, represent the data of all N banks, while for the individual
bank these are represented by the vectors xi and yi. The input-oriented measure of
technical efficiency for a particular bank is calculated as:
minh;k
h
subject to
Yk� yi;Xk� hxi;k� 0
ð1Þ
where h is the scalar and k is a M 9 1 vector of constants. If h = 1 the bank is
technically efficient as it lies on the frontier, whereas if h\ 1 the bank is inefficient.
11 Even though the true technology could be different from constant returns-to-scale (CRS), but we adopt
the CRS specification of technology on account of the following reasons. First, given the small sample
size like ours (27 banks in each years), one may get a distribution with many observations having
efficiency score equal to 1 using variable returns-to-scale (VRS) specification. This implies that one may
not get better discrimination of sampled units under VRS specification of technology in case of small
sample size. Second, regarding the use of VRS specification of technology, Noulas (1997) stated that the
assumption of CRS allows the comparison between small and large banks. In a sample where a few large
banks are present, the use of VRS framework raises the possibility that these large banks will appear as
being efficient for the simple reason that there are no truly efficient banks. Avkiran (1999) also mentions
that under VRS each unit is compared only against other units of similar size, instead of against all units.
Pasiouras et al. (2007) point out that the assumption of VRS is more suitable for large samples. The
prominent studies that made use of CRS assumptions for measuring cost and technical efficiencies in
banking system include Aly et al. (1990), Ariss et al. (2007), Hassan and Sancez (2007), Pasiouras et al.
(2007), Rezvanian et al. (2008) among others.
Econ Change Restruct (2013) 46:143–182 151
123
Note that the linear programming problem (1) must be solved N times, once for each
bank in sample. A value of h for bank i represents its TE score i.e., TEi.
The second step is to calculate cost efficiency by solving the following linear
program (see, Fare and Grosskopf 1985; Ferrier et al. 1993, for details).
minx�i ;k
w0ix�i
subject to
Yk� yi;Xk� x�i ;k� 0
ð2Þ
where wi denotes the vector of input prices for bank i. The solution of linear
program (2) yields a cost-minimizing input vector x�i , and we get minimum costs as
w0ix�i . Comparing minimum costs to observed costs w
0ixi of bank i gives cost effi-
ciency as:
CEi ¼w0
ix�i
w0ixi
The third step involves the calculation of allocative efficiency component
residually as the ratio of the measure of cost efficiency to the Farrell input-oriented
measure of technical efficiency. Thus, a measure of (input-mix) allocative efficiency
for bank i is obtained as:
AEi ¼CEi
TEið3Þ
This relationship facilitates the decomposition of cost efficiency as CEi ¼TEi � AEi. Note that the measures of cost, technical and allocative efficiencies
range between 0 and 1. Corresponding to these efficiency measures, the measures of
inefficiency can be obtained as CE�1i � 1
� �; TE�1
i � 1� �
; and AE�1i � 1
� �, respec-
tively (Isik and Hassan 2002b; Welzel and Lang 1997).
5 Data and measurement of input and output variables
There is no consensus on what constitute the inputs and outputs of a bank (Casu and
Girardone 2002; Sathye 2003). In the literature on banking efficiency, there are
mainly two approaches for selecting the inputs and outputs for a bank: (1) the
production approach, also called the service provision or value added approach;
and (2) the intermediation approach, also called the asset approach.12 Consistent
with most of the recent literature on banking efficiency, we follow the interme-
diation approach since it is ‘concerned with the overall costs of banking and is
appropriate for addressing questions concerning the economic viability of banks’
(Ferrier and Lovell 1990). Table 1 provides the description of the variables used in
the measurement of cost efficiency and its components.
12 The interested readers can consult, for instance, Berger and Humphrey (1992), Mlima and
Hjalmarsson (2002), Tortosa-Ausian (2002) and Hafferman (2005), for details on these approaches.
152 Econ Change Restruct (2013) 46:143–182
123
The output vector contains two output variables: (1) net-interest income, and (2)
non-interest income. The variable ‘net-interest income’ connotes net income
received by the banks from their traditional activities like advancing of loans and
investments in government and other approved securities. The output variable ‘non-
interest income’ accounts for income from fee generating off-balance sheet items
such as commission, exchange and brokerage, etc. The inclusion of ‘non-interest
income’ enables us to capture the recent changes in the production of services as
Indian banks are increasingly engaging in non-traditional banking activities. As
pointed out by Siems and Clark (1997), the failure to incorporate these types of
activities may seriously understate bank’s output, and this is likely to have statistical
and economic effects on estimated efficiency.
Some notable banking efficiency analyses that include ‘non-interest income’ as
an output variable are Isik and Hassan (2002a, b), Drake and Hall (2003), Sufian
(2006), Sufian and Majid (2007), Hahn (2007) among others. Further, a majority of
the studies on efficiency of Indian banks has also included ‘non-interest income’ in
the chosen output vector (see, for example, Das 1997, 2000; Saha and Ravisankar
2000; Mukherjee et al. 2002; Sathye 2003; Ram Mohan and Ray 2004; Chakrabarti
and Chawla 2005; Ray 2007; Kumar and Gulati 2009). It is worth noting here that
our choice of output variables is consistent with the managerial objectives that are
being pursued by the Indian banks. In the post-reforms years, intense competition in
the Indian banking sector has forced the banks to reduce all the input costs to the
minimum and to earn maximum revenue with least inputs. In this context, Ram
Mohan and Ray (2004) rightly remarked that in the post-liberalization period,
Table 1 Definition of variables used in efficiency measurement
(1) Price of physical capital (w1) (Rent, taxes and lighting ? printing and
stationary ? depreciation on bank’s property ?
repairs and maintenance ? insurance)/Physical
capital
(2) Price of labour (w2) (Payment to and provisions for employees)/labour
(3) Price of loanable funds (w3) (Interest paid on deposits ? interest paid on
borrowings from RBI and other agencies)/
Loanable funds
10 lacs = 1 million
Source: Author’s elaboration
Econ Change Restruct (2013) 46:143–182 153
123
Indian banks are putting all their efforts in the business of maximizing incomes from
all possible sources.
The input variables used for computing cost efficiency are (1) physical capital,
(2) labour, and (3) loanable funds, which are proxied by fixed assets, staff, and
deposits plus borrowings, respectively. Correspondingly, the prices of these inputs
are worked out as per unit price of physical capital, per employee wage bill, and cost
of loanable funds. The details on the definitions of these variables are given in the
Table 1. The required data on the variables used for computing various efficiency
measures have been culled out from the various issues of ‘Statistical Tables Relating
to Banks in India’, an annual publication of Reserve Bank of India (RBI) and
‘Performance Highlights of Public Sector Banks’, an annual publication of Indian
Banks’ Association (IBA). In the terminal year of the study, 28 PSBs were operating
in India and data on the IDBI Ltd. (a new public sector bank) were available only
after 2004/05. Therefore, we excluded IDBI Ltd. bank from our sample and
confined the study to 27 PSBs that were operating in India during the period
spanning from 1992/1993 to 2007/2008.13 Following Barman (2007) and Roland
(2008), we bifurcated the entire study period into distinct sub-periods: (1) first phase
of banking reforms (1992/1993–1998/1999), and (2) second phase of banking
reforms (1999/2000–2007/08). To compute CE, TE and AE scores, the analysis has
been carried out with real values of the variables (except labour) which have been
obtained by deflating the nominal values by the implicit price deflator of gross
domestic product (GDP) at factor cost (base 1993/94 = 100). Following Denizer
et al. (2007), we normalized all the input and output variables by dividing them by
the number of branches of individual banks for the given year. The main purpose of
using this normalization procedure is that it reduces the effects of random noise due
to measurement error in the inputs and outputs.
6 Empirical results
This section delineates the trends of cost efficiency and its sources, namely technical
and allocative efficiencies, in Indian public sector banking industry at industry and
bank levels during the post-deregulation period. Also, the results concerning
convergence in efficiency levels across PSBs are presented here. Instead of
constructing a ‘grand or inter-temporal frontier’14 as suggested by Tulkens and van
den Eeckaut (1995), and implemented by Bhattacharyya et al. (1997a) for
estimating the efficiency scores of individual banks, we followed Isik and Hassan
(2002b), Pasiouras et al. (2007) and Kumar and Gulati (2009), and estimated
separate annual efficient frontiers for obtaining year-by-year efficiency estimates.
Isik and Hassan (2002b) pointed out the following two advantages of this approach.
First, it is more flexible and, thus, more appropriate than estimating a single multi-
year frontier for the banks in the sample. Second, it alleviates, at least to some
13 No public sector bank failed or exited the market during the study period. The exit of banks from the
market has taken place only in the private and foreign banking segments of the Indian banking sector.14 The ‘grand frontier’ envelops the pooled input–output data of all banks in all years.
154 Econ Change Restruct (2013) 46:143–182
123
extent, the problems related to the lack of random error in DEA by allowing an
efficient bank in one year to be inefficient in another under the assumption that the
errors owing to luck or data problems are not consistent over time. In addition, the
efficiency estimates obtained from grand frontier are generally over-stated because
they are affected by technological progress in the industry. Thus, we believe that our
efficiency estimates are more reliable and accurate than what can be obtained from
the grand frontier which envelops the pooled input–output data of all PSBs in all the
years.
Since DEA results are influenced by the size of the sample, some discussion on
the adequacy of sample size is provided here. The size of the sample utilized in the
present study for estimating separate annual efficient frontiers is consistent with the
various rules of thumb available in the DEA literature. Cooper et al. (2007) provide
two such rules that can be expressed as: N �max K �M; 3ðK þMÞf gwhere
N = number of DMUs, K = number of inputs, and M = number of outputs. The
first rule of thumb states that sample size should be greater than equal to product of
inputs and outputs. While the second rule states that number of observations in the
data set should be at least three times the sum of the number of input and output
variables. Given K = 3 and M = 2 in the present study, the sample size (N = 27)
used for estimating separate annual efficient frontiers exceeds the desirable size as
suggested by the above-mentioned rules of thumb to obtain sufficient discriminatory
power. The sample size in this study is feasible and larger than that used in some of
the studies in the DEA literature (see, for example, Avkiran 1999).
6.1 Trends in cost (in)efficiency at industry level
Panel A of Table 2 provides year-wise mean cost, technical and allocative
efficiency scores for Indian public sector banking industry and its distinct sub-
groups. The results show that there are noticeable variations across years in cost
efficiency levels, and there seems to be an uptrend in the cost efficiency of Indian
public sector banking industry. The cost efficiency increased consistently from 71%
in 1992/1993 to 80.6% in 1997/98, and then declined gently and reached to the level
of 76.3% in 2001/02. Subsequently, a precipitous uplift in cost efficiency has been
noticed which ceased at the level of 86.7% in 2006/07. However, cost efficiency
turned down and attained a level of 81.6% in the terminal year (2007/08). We
further note that the average level of cost efficiency (inefficiency) in Indian public
sector banking industry is 79.6% (25.6%). This figure of cost efficiency implies that
the average bank in the sample could have produced the same level of outputs using
only 79.6% of the cost actually incurred, if it was producing on the cost frontier
rather than at its current location. On the other hand, the figure of cost inefficiency
implies that during each year of the study period, an average bank needed 25.6%
more resources and, thus, incurred more cost to produce the same output as
compared to the efficient bank. This divulges that Indian public sector banks, in
general, have not been successful in employing best-practice production methods
and achieving the maximum outputs from the minimum cost of inputs. Apparently,
there exists substantial room for significant cost savings if Indian PSBs use and
allocate their productive inputs more efficiently.
Econ Change Restruct (2013) 46:143–182 155
123
In order to analyze the group-specific behaviour of the mean TE over the entire
study period and distinct sub-periods, we followed the prevalent grouping criterion
in Indian public sector banking industry and bifurcated the PSBs into two groups
Table 2 Mean cost, technical, and allocative efficiency scores in Indian public sector banking industry:
(1) CE, TE and AE stands for cost, technical and allocative efficiencies, respectively; and (2) The arrows : and ;indicate that mean CE, TE and AE of the bank has increased and decreased, respectively in the second phase of
reforms relative to what has been observed during the first phase of reforms
Source: Author’s calculations
156 Econ Change Restruct (2013) 46:143–182
123
namely, State Bank of India group (SBI group) and group of nationalized banks (NB
group). In the Indian banking system, this grouping of public sector banks is vital
and matters to the policy makers and analysts because of the following differences
in institutional characteristics of these groups in terms of ownership, functions and
organizational structure (see, Arun and Turner 2002; Maheshwari 2006, for more
details). First, the SBI, India’s largest commercial bank in terms of branches and
assets, was established under the State Bank of India Act, 1955 and its 7 subsidiary
banks which were established under the State Bank of India Act, 1959. While the 19
nationalized banks were established under the two Acts, i.e., Banking Companies
(Acquisition and Transfer of Undertakings) Act, 1970 and the Banking Companies
(Acquisition and Transfer of Undertakings) Act, 1980. Thus, the banks in SBI and
NB groups are governed by the different statutes. Second, the Reserve Bank of India
(RBI) owns the majority share of SBI, while the shares of subsidiary banks are
owned by the SBI. On the other hand, nationalized banks are wholly owned by the
Government of India. Third, SBI besides carrying out its normal banking functions
also acts as an agent of the Reserve Bank of India. SBI undertakes most of the
government business transactions (including major borrowing programmes),
thereby earning more non-interest income than nationalized banks. However, this
privilege has not been bestowed upon the nationalized banks. Fourth, the SBI has a
well-defined system of decentralization of authority, while in case of nationalized
banks the organizational structure differs from the bank to bank. The inter-group
analysis reveals that, over the years under reference, the average cost efficiency
levels ranged between 82.1 and 98.1% for SBI group, while the same ranged
between 61.7 and 85.8% for NB group. Further, the average levels of cost efficiency
(inefficiency) for SBI and NB groups are about 91.8% (8.9%) and 74.5% (34.2%),
respectively. Looking at these figures of average levels of cost efficiency, we can
safely infer that SBI and its associate banks score over nationalized banks in terms
of minimizing cost of producing financial outputs in the production process.
The comparative analysis for distinct sub-periods highlights that the average cost
efficiency of Indian public sector banking industry has increased by about 4.3%
(81.5% vis-a-vis 77.2%). The straightforward implication of this finding is that the
average cost inefficiency in Indian public sector banking industry has decreased
during the second phase relative to the first phase (29.5% vis-a-vis 22.7%). This
should not be surprising because at the time of introduction of second phase of
banking reforms, the PSBs had almost fully adjusted to liberalization, enhanced
competition, and new prudential regulations of the banking sector. Further, it has
been identified that the observed increase in the average cost efficiency during the
second phase was entirely contributed by the nationalized banks. The average cost
efficiency of NB group has been found to be 78.4% for the second phase compared to
69.4% for the first phase, indicating a 9% increase in input cost-saving potentials. On
the other hand, the average cost efficiency of SBI group declined by 6.8% between
these two phases. This is evident from the fact that the average cost efficiency of SBI
group for the second phase has been observed to be 95.7% against 88.9% for the first
phase. The results clearly show the increase in average cost efficiency of the NB
group was responsible for the observed upturn in the average cost efficiency of the
Indian public sector banking industry during the second phase of reforms.
Econ Change Restruct (2013) 46:143–182 157
123
To test whether the differences in average cost efficiency between the sub-
periods are statistically significant or not, we applied non-parametric Kruskal–
Wallis test (see Panel B of Table 2). The observed values of H-statistics for public
sector banking industry as a whole has been noted to be 3.248, which is greater than
the critical value of v2 = 2.706 at 10% level of significance. Hence, we reject the
null hypothesis of no differences in average cost efficiency levels between the sub-
periods. This suggests that cost efficiency in Indian public sector banking industry
as a whole has improved significantly during the second phase of reforms relative to
first one. For SBI and NB groups, we also rejected the null hypothesis of no
differences in average cost efficiency levels between the sub-periods. On the whole,
we note that (1) average cost efficiency of SBI group has declined significantly
during the second phase relative to the first phase; (2) average cost efficiency of NB
group has increased significantly during the second phase of reforms in comparison
of the first phase; and (3) the effect of substantial cost efficiency gains in NB group
is neutralized to a large extent by the significant losses of efficiency in SBI group,
and thus has resulted in no considerable efficiency gains in Indian public sector
banking industry with the progress of the deregulation process. Further, the presence
of the phenomenon of changes in technical (in)efficiency are offset by changes in
allocative (in)efficiency suggests that there is a trade-off between one type of
inefficiency against another.
6.2 Sources of cost (in)efficiency
Recall that technical and allocative efficiencies are two mutually exclusive
components of cost efficiency. Thus, cost inefficiency incorporates both allocative
inefficiency from failing to react optimally to relative prices of inputs, and technical
inefficiency from employing too much of the inputs to produce a certain output
bundle (Gjirja 2004). Table 2 also gives the year-wise mean technical and allocative
efficiency scores for Indian public sector banking industry and its distinct segments.
From Panel B of the table, we note that over the years under evaluation, the average
technical efficiency is 84.4%, indicating that an average bank wasted about 18.5%
of factor inputs in the production process by operating off the efficient production
frontier. The observed level of average allocative efficiency is 94.4%, pointing that
average bank incurred about 5.9% more production cost by choosing the incorrect
input combination given input prices. For determining the dominant source of cost
inefficiency, we make a comparison of the relative sizes of technical and allocative
inefficiency levels. We note that, for all the sample years, allocative efficiency is
consistently higher than technical efficiency, which signals that technical ineffi-
ciency (i.e., underutilization or wasting of inputs) has greater significance than
allocative inefficiency (i.e., choosing the incorrect input combination given input
prices) as a source of cost inefficiency within all inefficient PSBs. This result
suggests that the observed cost inefficiency in Indian public sector banking industry
originates primarily due to managerial problems in using the resources rather than
the regulatory environment in which PSBs are operating. Apparently, the managers
of PSBs operate relatively efficient with respect to the optimal combinations of
inputs given their prices and technology, yet they are not efficient in transforming
158 Econ Change Restruct (2013) 46:143–182
123
bank inputs into outputs and avoiding waste in the production process. Turning to
the segment-wise analysis, we note that average cost inefficiency in NB group is
primarily driven by technical inefficiency rather than allocative inefficiency.
However, in SBI group, both technical and allocative inefficiencies are roughly the
equally important source of cost inefficiency.
Turning to the impact of deepening of the process of banking reforms, it has been
observed that average technical efficiency of Indian public sector banking industry
has increased by 5.3% in the second phase of reforms than what has been observed
in the first phase (86.7% vis-a-vis 81.4%). Further, this gain in the average technical
efficiency has been observed to be statistically significant as indicated by the
rejection of the null hypothesis in Kruskal–Wallis test. Regarding average allocative
efficiency, we note that an ascent in the intensity of reforms did not bring any
significant change in its level. The acceptance of the null hypothesis in Kruskal–
Wallis test confirms this. The segment-wise analysis reveals that in the second phase
of reforms, a statistically significant gain in average technical efficiency in tune to
9.5% has noted for NB group, while a statistically significant decline in average
technical efficiency in order of (-)4.4% has been observed for SBI group. The
analysis further reveals that a statistically significant decline in average allocative
efficiency by (-)2.6% has taken place in SBI group. Nevertheless, average
allocative efficiency shown a negligible increase during the second phase of
reforms, which is further observed to be statistically insignificant. Peeping deep into
the results, we note that what so ever increase in cost and technical efficiencies in
public sector banking industry has taken place during the study period that was
contributed by the significant improvement in technical efficiency of nationalized
banks. In fact, the drag in allocative efficiency of the banks belonging to SBI groups
is not only responsible for a decline in allocative efficiency of the public sector
banking industry as whole, but also offset, to a great extent, the effect of gains in
technical efficiency of nationalized banks on the cost efficiency of the public sector
banking industry as a whole.
6.3 Growth rates analysis
To ascertain a more concrete picture about the trends of efficiency measures, we
relied on the trend growth rates of efficiency measures for the entire study period and
distinct sub-periods. For computing the average annual growth rate of efficiency
scores for the entire study period, we estimated a log-linear trend equation:
ln Et ¼ aþ bt þ et, where Et is mean efficiency score in the year t (t = 1,2,…,T) and
et denotes the stochastic error term. Following Boyce (1986), a kinked exponential
model has been used for estimating the growth rates for the sub-periods. The
regression equation in kinked exponential model for estimating the growth rates for
sub-periods takes the form: ln Et ¼ aþ b1ðDt þ ð1� DÞkÞ þ b2ð1� DÞðt � kÞ þ et,
where D is a dummy variable (D = 1 for first sub-period and 0 for second sub-
period), k is the midpoint of the two discontinuous series (k = 7.5 in the present
study). The OLS estimates of b1 and b2 (i.e., b1 and b2) gives the growth rates for the
first and second sub-periods, respectively.
Econ Change Restruct (2013) 46:143–182 159
123
Panel D of Table 2 provides the growth rate estimates of cost efficiency and its
components. We note that cost efficiency of Indian public sector banking industry
grew at a modest rate of 0.868% per annum over the entire study period. Further, it
has declined at the rate of (-)0.845% per annum for SBI group and recorded a
decent growth rate of 1.761% per annum for NB group. The analysis of growth rates
for the distinct sub-periods reveals that (a) in SBI group, the declining trend of cost
efficiency was more pronounced in the second phase, (b) in NB group, cost
efficiency grew at the rate of 1.967% during the second phase which is about half a
percent more than what has been noticed during the first phase, and (c) the effect of
decent growth in cost efficiency in NB group during the second phase of reforms
was offset to a great extent by a pronounced decline in the same in SBI group. This
led to a very slight improvement in growth of cost efficiency of Indian public sector
banking industry during the second phase of reforms relative to the first one
(0.894% vis-a-vis 0.829%).
Turning to the growth rates of disaggregate components of the cost efficiency, we
note that, over the entire study period, technical efficiency of Indian public sector
banking industry followed an uptrend, while allocative efficiency followed a path of
deceleration. This is evident from the figures of growth rates at 0.962% per annum
and (-)0.064% per annum for technical efficiency and allocative efficiency,
respectively. In Indian public sector banking industry, the components of cost
efficiency moved in opposite directions, and they are counterbalancing in nature.
The followings may be the prominent reasons for a decline in the allocative
efficiency during the study period. First, the observed increase in allocative
inefficiencies may be due to fluctuations and instability of factor prices which arise
due to the difficulties of implementing an inflation targeting framework for
stabilizing expectations under high exchange rate volatility. Second, another
possible reason for an increase in allocative inefficiency may be regulatory in
nature. Deterioration in allocative efficiency may be occurred due to an introduction
of stringent regulatory restrictions primarily in the area of maintaining the capital
adequacy ratio as per Basel norms during the post-reforms years. From no norm of
capital adequacy in the pre-reforms period, Indian banking system is implementing
Basel I and II norms in a phased manner during the post-reforms years. Although the
government has injected capital into the PSBs that had their capital adequacy ratio
(CAR) below Basel norms, but still a few banks faced problems in the complacency
of the capital adequacy ratio during the period of analysis. Third, in the post-reforms
years, PSBs have increasingly used equity market to raise funds. This exposed
banks to the consequences of the imperfections inherent in this market. This may
have led to distortions in the process of allocating resources in Indian public sector
banking industry. Fourth, another reason of decline in allocative efficiency of PSBs
may be that due to the dismantling of administered interest rate regime and entry of
new private banks since 1994, PSBs started competing for loans and deposits.
Deposits being a component of loanable funds (a critical input in the production
process) have been acquired by PSBs at the prices quite higher than those
correspond to cost-minimization level.
The segment-wise analysis reveals that in SBI group, both components followed
a declining trend over the entire study period. However, in NB group, these
160 Econ Change Restruct (2013) 46:143–182
123
components showed a positive trend, and growth in technical efficiency was more
impressive than that of allocative efficiency. Further, the analysis of growth rates for
distinct sub-periods reports (1) a negative trend in both components of cost
efficiency in SBI group during the second phase relative to a positive trend during
the first phase, (2) a noticeable improvement in the growth of technical efficiency of
Indian public sector banking industry as a whole and its segment of nationalized
banks during the second phase relative to first one, and (3) the allocative efficiency
of Indian public sector banking industry has shown a decelerating trend during the
latter phase relative to the former. Overall, the analysis manifests that in Indian
public sector banking industry, the growth in technical efficiency contributed
positively to the growth of cost efficiency, and the deceleration in allocative
efficiency actually drags it.
6.4 Inter-bank analysis
Table 3 provides the average cost, technical and allocative efficiency scores for
individual PSBs over the entire study period and distinct sub-periods. The perusal of
table gives that there is heterogeneity in the level of average cost efficiency across
PSBs. United Bank of India presents the lowest level of cost efficiency (58.6%), and
State Bank of Hyderabad (95.3%) displays the highest ones. Further, in 6 PSBs, the
magnitude of average cost inefficiency is found to be less than 10%. These banks
are State Bank of Hyderabad (95.3%), State Bank of India (93.9%), State Bank of
Indore (94.6%), State Bank of Mysore (93.8%), State Bank of Patiala (93.2%), and
Corporation Bank (91.3%). We can rightly designate these banks as ‘marginally
cost inefficient’ banks. It is significant to note here that (1) out of 6 marginally cost
inefficient banks, 5 banks belong to SBI group, (2) all the observed marginally
inefficient banks have both high levels of technical and allocative efficiencies, and
(3) average technical efficiency is more than average allocative efficiency in all
these banks.
In the remaining 21 PSBs, the average cost efficiency ranged between 58.6 and
89.6%, indicating that the extent of cost inefficiency lies in the range between 11.6
and 70.6%. These banks can be categorized as ‘distinctively cost inefficient banks’.
Two points are noteworthy here that (a) in 20 distinctively cost inefficient banks,
cost inefficiency emanates primarily due to technical inefficiency rather than
allocative inefficiency, (b) the three banks viz., Indian Bank (67.4%), UCO Bank
(58.9%) and United Bank of India (58.6%) which were identified as weak banks by
the Committee on the Banking Sector Reforms (1998) and Working Group on
Restructuring of Weak Public Sector Banks (1999) are the least cost efficient banks
in the sample.
The comparative analysis of average cost efficiency between the sub-periods
provides the following points: (1) the average cost efficiency has improved in 17
PSBs during the latter phase of reforms relative to first one; (2) of 8 PSBs which
belong to SBI group, the average cost efficiency in 6 banks recorded a downturn in
the latter phase compared to the earlier phase; (3) the three weak banks (Indian
Bank, UCO Bank and United Bank of India) followed an upturn in the average cost
efficiency over the second phase of reforms compared to first phase; and (4) out of
Econ Change Restruct (2013) 46:143–182 161
123
Ta
ble
3M
ean
cost
,te
chnic
alan
dal
loca
tiv
eef
fici
ency
sco
res
inIn
dia
np
ub
lic
sect
or
ban
kin
gin
du
stry
:a
ban
k-w
ise
anal
ysi
s
Ban
kE
ffici
ency
mea
sure
s
Co
stef
fici
ency
(CE
)T
ech
nic
alef
fici
ency
(TE
)A
llo
cati
ve
effi
cien
cy(A
E)
Per
iod
of
stud
y
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stud
y
per
iod
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Sta
teB
ank
of
Bik
aner
and
Jaip
ur
0.9
60
0.8
45;
0.8
96
0.9
91
0.8
67;
0.9
21
0.9
69
0.9
74:
0.9
72
Sta
teB
ank
of
Hy
der
abad
0.9
45
0.9
60:
0.9
53
0.9
84
0.9
94:
0.9
90
0.9
60
0.9
65:
0.9
63
Sta
teB
ank
of
Ind
ia0
.997
0.8
93;
0.9
39
0.9
98
0.9
47;
0.9
70
0.9
98
0.9
43;
0.9
67
Sta
teB
ank
of
Ind
ore
0.9
78
0.9
22;
0.9
46
0.9
79
0.9
74;
0.9
76
0.9
99
0.9
46;
0.9
69
Sta
teB
ank
of
My
sore
0.9
31
0.9
44:
0.9
38
0.9
82
0.9
93:
0.9
89
0.9
48
0.9
50:
0.9
49
Sta
teB
ank
of
Pat
iala
0.9
69
0.9
03;
0.9
32
0.9
91
0.9
92:
0.9
92
0.9
78
0.9
10;
0.9
40
Sta
teB
ank
of
Sau
rash
tra
0.9
84
0.8
06;
0.8
84
0.9
98
0.8
59;
0.9
20
0.9
85
0.9
42;
0.9
61
Sta
teB
ank
of
Tra
van
core
0.8
89
0.8
36;
0.8
59
0.9
53
0.9
02;
0.9
24
0.9
33
0.9
29;
0.9
31
All
ahab
adB
ank
0.6
24
0.7
52:
0.6
96
0.6
77
0.7
96:
0.7
44
0.9
25
0.9
46:
0.9
37
An
dh
raB
ank
0.6
92
0.8
67:
0.7
91
0.7
35
0.9
42:
0.8
51
0.9
46
0.9
19;
0.9
31
Ban
ko
fB
aro
da
0.8
68
0.7
94;
0.8
27
0.9
47
0.8
58;
0.8
97
0.9
18
0.9
28:
0.9
24
Ban
ko
fIn
dia
0.6
83
0.7
51:
0.7
21
0.7
41
0.7
92:
0.7
70
0.9
24
0.9
49:
0.9
38
Ban
ko
fM
ahar
ash
tra
0.6
97
0.7
73:
0.7
40
0.7
29
0.7
99:
0.7
68
0.9
49
0.9
68:
0.9
60
Can
ara
Ban
k0
.798
0.7
65;
0.7
80
0.8
57
0.8
16;
0.8
34
0.9
33
0.9
40:
0.9
37
Cen
tral
Ban
ko
fIn
dia
0.6
19
0.7
33:
0.6
83
0.6
43
0.7
71:
0.7
15
0.9
61
0.9
52;
0.9
56
Corp
ora
tio
nB
ank
0.9
20
0.9
07;
0.9
13
0.9
68
0.9
97:
0.9
84
0.9
52
0.9
10;
0.9
28
Den
aB
ank
0.7
93
0.8
31:
0.8
14
0.8
34
0.8
79:
0.8
59
0.9
50
0.9
44;
0.9
47
Ind
ian
Ban
k0
.570
0.7
54:
0.6
74
0.6
65
0.7
81:
0.7
30
0.8
69
0.9
63:
0.9
22
Ind
ian
Ov
erse
asB
ank
0.6
62
0.8
22:
0.7
52
0.7
04
0.8
52:
0.7
87
0.9
42
0.9
65:
0.9
55
162 Econ Change Restruct (2013) 46:143–182
123
Ta
ble
3co
nti
nu
ed
Ban
kE
ffici
ency
mea
sure
s
Co
stef
fici
ency
(CE
)T
ech
nic
alef
fici
ency
(TE
)A
llo
cati
ve
effi
cien
cy(A
E)
Per
iod
of
stud
y
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stud
y
per
iod
Fir
st
ph
ase
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Ori
enta
lB
ank
of
Co
mm
erce
0.8
97
0.8
33;
0.8
61
0.9
65
0.9
86:
0.9
77
0.9
27
0.8
45;
0.8
81
Pu
nja
ban
dS
ind
Ban
k0
.579
0.7
83:
0.6
94
0.6
13
0.8
30:
0.7
35
0.9
46
0.9
47:
0.9
47
Pu
nja
bN
atio
nal
Ban
k0
.743
0.8
57:
0.8
07
0.7
72
0.8
89:
0.8
38
0.9
61
0.9
64:
0.9
62
Sy
nd
icat
eB
ank
0.6
39
0.7
71:
0.7
13
0.6
83
0.8
23:
0.7
62
0.9
32
0.9
37:
0.9
35
UC
OB
ank
0.5
20
0.6
44:
0.5
89
0.5
51
0.6
76:
0.6
21
0.9
46
0.9
53:
0.9
50
Un
ion
Ban
ko
fIn
dia
0.7
54
0.7
72:
0.7
64
0.8
32
0.8
50:
0.8
42
0.9
12
0.9
09;
0.9
10
Un
ited
Ban
ko
fIn
dia
0.4
44
0.6
96:
0.5
86
0.4
61
0.7
22:
0.6
08
0.9
59
0.9
66:
0.9
63
Vij
aya
Ban
k0
.686
0.7
87:
0.7
43
0.7
11
0.8
33:
0.7
79
0.9
64
0.9
45;
0.9
53
The
arro
ws:
and;
ind
icat
eth
atm
ean
CE
,T
Ean
dA
Eo
fth
eb
ank
has
incr
ease
dan
dd
ecre
ased
,re
spec
tiv
ely
inth
ese
con
dp
has
eo
fre
form
sre
lati
ve
tow
hat
has
bee
n
ob
serv
edd
uri
ng
the
firs
tp
has
eo
fre
form
s
So
urc
e:A
uth
or’
sca
lcu
lati
on
s
Econ Change Restruct (2013) 46:143–182 163
123
18 PSBs that belong to NB group, only in 4 banks, namely Bank of Baroda,
Corporation Bank, Oriental Bank of Commerce and Canara Bank, a decline in
average cost efficiency has been observed during the second phase of reforms. The
above result indicates that though at the industry level, no significant change in
average cost efficiency has been observed, but at the level of individual banks
noticeable improvement in average cost efficiency has been observed with the
deepening of the process of banking reforms since 1998/1999. The main reason for
insignificant improvement in the cost efficiency at the industry level is that the
downturn in the average cost efficiency among most of the banks in SBI group
offsets the effect of an ascent in average cost efficiency in the majority of banks
belonging to NB group.
As far as the components of cost efficiency are concerned, we observe that
average technical efficiency has increased in the 20 PSBs during the second phase of
reforms relative to first one. This indicates that operating efficiency of a majority of
PSBs has improved with the increase in the intensity of reforms. The connotation of
this finding is that PSBs have learnt to avoid the waste of inputs in transforming
outputs with the deepening of reforms. Further, we note that in 15 PSBs, the average
allocative efficiency has increased relatively in the latter phase of reforms compared
to first one. Thus, the majority of PSBs have also learnt to organize the inputs in the
cost-minimizing way given their prices. On the whole, we observed that a majority
of PSBs exhibited a decline in both technical and allocative inefficiencies with the
ascent of deregulation in Indian banking industry.
The inter-bank analysis of trend growth rates of cost efficiency and its
disaggregate components is provided in the Table 4. The results show that (1) the
cost efficiency in the majority of banks that belong to SBI group followed a
declining trend. This is evident from the fact that, of 8 PSBs in SBI group, 6 banks
registered a negative growth rate over the entire study period; (2) barring 4 PSBs,
the remaining 15 banks belonging to NB group experienced an increasing trend in
cost efficiency. The highest growth in cost efficiency has been observed in United
Bank of India (5.89%), followed by Bank of Maharashtra (4.65%) and Punjab and
Sind Bank (4.10%); and (3) in 20 PSBs, cost efficiency and its disaggregate
components evolved with the same tendency. That is, an increasing (decreasing)
trend in cost efficiency is followed by the increasing (decreasing) trend in technical
and allocative efficiencies. This underscores the presence of a phenomenon of co-
movement in the growth of cost, technical and allocative efficiencies in Indian
public sector banking industry.
Turning to the analysis for the distinct sub-periods, it has been observed that the
number of banks having a positive trend in cost efficiency (technical efficiency,
allocative efficiency) during the second phase of reforms was 18(18, 12), while this
number stood at 17(18, 19) during the first phase. This highlights that the number of
banks showing downtrend in allocative efficiency has increased considerably during
the latter phase of reforms. In a great majority of banks in SBI group, a declining
trend in cost efficiency and its components has been noticed in the second phase.
Further, only 9(8, 6) PSBs experienced an improvement in the growth rate of cost
efficiency (technical efficiency, allocative efficiency) in the second phase relative to
the first one. This conveys that no considerable improvement in the growth of cost
164 Econ Change Restruct (2013) 46:143–182
123
Ta
ble
4G
row
thra
tes
of
cost
,te
chnic
alan
dal
loca
tiv
eef
fici
enci
es:
anin
ter-
ban
kan
alysi
s
Ban
kE
ffici
ency
mea
sure
s
Co
stef
fici
ency
(CE
)T
ech
nic
alef
fici
ency
(TE
)A
llo
cati
ve
effi
cien
cy(A
E)
Per
iod
of
stud
y
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stud
y
per
iod
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Sta
teB
ank
of
Ind
ia-
3.5
78
1.4
37
-0
.53
3-
2.6
89
1.3
26
-0
.231
-0
.56
90
.37
2-
0.0
11
Sta
teB
ank
of
Bik
aner
and
Jaip
ur
1.0
75
-0
.39
20
.20
50
.487
-0
.137
0.1
16
0.5
89
-0
.25
90
.086
Sta
teB
ank
of
Hy
der
abad
-0
.45
9-
1.9
58
-1
.34
8-
0.6
37
-0
.452
-0
.528
0.1
78
-1
.51
2-
0.8
25
Sta
teB
ank
of
Ind
ore
2.1
68
-3
.25
9-
0.8
35
1.3
93
-1
.043
-0
.052
0.7
61
-2
.72
2-
1.4
78
Sta
teB
ank
of
My
sore
1.5
06
-0
.52
40
.30
10
.953
-0
.272
0.2
26
0.5
54
-0
.25
10
.077
Sta
teB
ank
of
Pat
iala
1.0
11
-3
.10
7-
1.6
45
0.0
21
-0
.107
-0
.055
0.9
91
-3
.00
3-
1.6
13
Sta
teB
ank
of
Sau
rash
tra
-0
.13
9-
4.3
11
-2
.81
90
.690
-4
.386
-2
.476
-0
.82
40
.06
8-
0.2
94
Sta
teB
ank
of
Tra
van
core
-2
.55
41
.28
4-
0.1
82
-1
.062
0.4
36
-0
.173
-1
.49
40
.84
5-
0.1
06
All
ahab
adB
ank
3.9
86
1.1
97
2.3
31
3.0
90
1.6
11
2.2
13
0.8
82
-0
.40
30
.120
An
dh
raB
ank
4.8
22
2.7
22
3.5
76
7.4
28
0.9
99
3.6
13
-2
.59
81
.72
00
.052
Ban
ko
fB
aro
da
-2
.64
30
.80
7-
0.5
96
-3
.496
1.0
08
-0
.506
0.8
53
-0
.20
10
.228
Ban
ko
fIn
dia
1.7
73
1.1
55
1.4
06
1.4
17
0.9
50
1.1
40
0.3
57
0.2
07
0.2
68
Ban
ko
fM
ahar
ash
tra
8.7
96
1.3
70
4.6
57
6.8
85
1.4
80
3.8
63
1.2
32
-0
.22
10
.370
Can
ara
Ban
k-
2.0
11
0.7
20
-0
.39
0-
2.3
31
1.2
09
-0
.230
0.3
27
-0
.48
8-
0.1
57
Cen
tral
Ban
ko
fIn
dia
7.4
16
-0
.94
02
.03
47
.282
-0
.764
1.9
80
0.1
35
-0
.17
5-
0.0
49
Corp
ora
tio
nB
ank
0.1
40
0.3
86
0.2
86
1.9
86
-0
.440
0.5
46
-1
.84
40
.82
7-
0.2
59
Den
aB
ank
2.4
74
0.7
65
1.4
60
2.9
43
0.4
54
1.4
66
-0
.47
40
.30
6-
0.0
11
Ind
ian
Ban
k-
2.1
21
7.7
74
3.7
95
-5
.242
7.5
40
1.9
76
3.0
92
0.2
40
1.4
00
Econ Change Restruct (2013) 46:143–182 165
123
Ta
ble
4co
nti
nu
ed
Ban
kE
ffici
ency
mea
sure
s
Co
stef
fici
ency
(CE
)T
ech
nic
alef
fici
ency
(TE
)A
llo
cati
ve
effi
cien
cy(A
E)
Per
iod
of
stud
y
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stud
y
per
iod
Fir
stp
has
e
of
refo
rms
Sec
ond
ph
ase
of
refo
rms
En
tire
stu
dy
per
iod
Ind
ian
Ov
erse
asB
ank
-5
.42
18
.07
81
.59
6-
6.0
05
7.9
82
0.8
64
0.5
90
0.0
92
0.2
95
Ori
enta
lB
ank
of
Com
mer
ce0
.34
0-
1.0
97
-0
.29
21
.379
-0
.561
0.2
28
-0
.24
0-
0.8
66
-0
.609
Pu
nja
ban
dS
ind
Ban
k4
.38
53
.90
94
.10
34
.346
4.0
31
4.1
59
0.0
38
-0
.12
5-
0.0
59
Pu
nja
bN
atio
nal
Ban
k1
.71
12
.66
62
.27
71
.381
2.6
25
2.1
19
0.3
27
0.0
50
0.1
63
Sy
nd
icat
eB
ank
4.9
11
0.2
33
2.1
35
3.6
71
0.7
15
1.9
17
1.2
44
-0
.48
50
.218
UC
OB
ank
2.6
14
1.6
80
2.0
60
2.2
57
1.6
40
1.8
91
0.3
72
0.0
29
0.1
68
Un
ion
Ban
ko
fIn
dia
-1
.55
72
.23
61
.17
7-
1.0
58
1.7
54
0.6
11
-0
.49
50
.48
60
.087
Un
ited
Ban
ko
fIn
dia
10
.68
82
.60
75
.89
39
.848
2.9
38
5.7
48
0.8
22
-0
.32
60
.238
Vij
aya
Ban
k-
0.8
91
-0
.19
7-
0.4
98
-0
.341
2.8
60
0.6
97
0.0
39
-1
.05
6-
0.6
35
So
urc
e:A
uth
or’
sca
lcula
tions
166 Econ Change Restruct (2013) 46:143–182
123
efficiency and its components has been noticed in Indian public sector banking
industry with the ascent in the intensity of reforms since 1998/1999. Considering
both the sub-periods separately, we noticed the appearance of co-movement in trend
growth rates of cost, technical and allocative efficiencies in the majority of PSBs.
By and large, the results of growth rates of cost efficiency are in consonance with
the changes in average cost efficiency levels between the first and second phases of
banking reforms. The inter-bank analysis indicates that to a large extent, the India’s
experience with banking reforms offers a success story to be emulated by other
developing economies, since the majority of the PSBs experienced a positive trend
in cost efficiency during the reforms period.
The aforementioned empirical findings vividly indicate a positive trend in the
cost efficiency levels of Indian public sector banking industry during the post-
reforms years, but some discussion on what derived this improvement is warranted
here. In this context, the most significant factor is the heightened competition in the
Indian banking sector during the post-reforms period due to relaxed entry norms for
de novo private domestic and foreign banks. To keep their survival intact in the
highly competitive environment, the PSBs, especially the weak ones, started
allocating resources efficiently, and changed their behavioural attitude and business
strategies. Further, in their drive to achieve higher levels of operating efficiency,
Indian PSBs during the post-reforms years, primarily concentrated on the
rationalization of the labour force and branching, and reduction in the cost of
financial transactions. For making optimal use of labour force, these banks evolved
policies aimed at ‘rightsizing’ and ‘redeployment’ of the surplus staff either by way
of retraining them and giving them appropriate alternate employment or by
introducing a ‘voluntary retirement scheme (VRS)’ with appropriate incentives.
Consequently, the labour cost per unit of earning assets fell from 2.44% in 1992/
1993 to 0.95% in 2007/2008. With the objectives of cutting the cost of day-to-day
banking operations in the long run, and retaining their existing customers and
attracting new ones by providing new technology-based delivery channels (like
internet banking, mobile banking and card based funds transactions), PSBs made a
heavy investment in technology during the post-reforms years. Between September
1999 and March 2008, PSBs incurred an expenditure of Rupee 15015 crore
(1 crore = 10 million) on computerization and development of communication
networks (Reserve Bank of India 2006). The computerization of branches and
installation of ATMs are two major areas in which the use of technology is clearly
visible. By end-March 2008, about 93.7% branches of PSBs were fully comput-
erized, of which 67.7% branches of nationalized banks and 95% of SBI and its
associates were under core banking solutions. The number of both on-site and off-
site ATMs by PSBs increased from 3,473 at the end of March 2003 to 34,789 at the
end of March 2008. On the whole, the post-reforms period witnessed an enhanced
level of IT usage by PSBs which might have contributed to efficiency improvement.
Another major influential factor that contributed to cost efficiency gains is that
due to profound changes in the regulatory and legal frameworks, there has been a
better recovery of non-performing loans which led to an improvement in the assets
quality of the PSBs. This is evident from the fact that in public sector banking
segment, the quantum of net NPAs as a percentage of net advances declined from
Econ Change Restruct (2013) 46:143–182 167
123
10.7% in 1994/1995 to 0.99% in 2005/06. Among the various channels of recovery
available to banks for dealing with bad loans, SARFAESI Act15 and the debt
recovery tribunals (DRTs) have been the most effective in terms of the amount
recovered (Reserve Bank of India 2008). Due to better recovery of NPAs, the share
of net-interest income in total income of PSBs has increased significantly. Further,
in the Indian banking industry, the off-balance sheet activities business has soared
during the post-reforms years. This has led to increase in ‘other income’ of the
PSBs. The improvement in efficiency could also be attributable to the fact that there
has been a change in the orientation of PSBs from social objectives towards an
ascent of profitability, particularly given that with the dilution of the government
equity in most of these banks, a stake of private investors is involved. The capital
market discipline imposed on PSBs since 1992/1993 when these banks were
allowed to raise capital from the stock market has also led to significant efficiency
gains. From the above discussion, we may infer that cost efficiency gains in Indian
public sector banking during the post-reforms years stemmed not only due to cost-
curtailing measures adopted by PSBs, but also occurred due to measures aiming at
augmenting income-generating capacity of banks.
6.5 Convergence in efficiency levels
6.5.1 Testing of r-convergence
The concept of convergence as used in the present study refers to the tendency for
two or more banks to become similar in terms of efficiency levels. Therefore, if the
banks with low levels of efficiency at the beginning of the period grow more rapidly
than those with the high initial level of efficiency, convergence occurs, implying that
the less efficient banks are catching-up. The literature spells out two different
concepts of convergence: (1) r-convergence; and (2) b-convergence (see Barro et al.
1991; Barro and Sala-i-Martin 1992, 1995; Sala-i-Martin 1996a, b). Convergence of
r-type considers whether gaps between inefficient and efficient banks decline over
time. The concept of r-convergence is said to exist if the distribution of efficiency
levels across banks gets tighter over time, thus reducing some measure of dispersion
over time. It focuses on the evolution of cross-sectional distribution of efficiency
over time. The existence of r-convergence implies a tendency of efficiency levels to
be equal across banks over time. The r-convergence can be tested empirically by
regressing the standard deviations (or coefficient of variations) of the cross-sections
over time on a trend variable. Symbolically, it implies that
ln SDt or CVtð Þ ¼ aþ rt þ et ð4Þ
where SDt and CVt denote the standard deviation and coefficient of variation,
respectively of efficiency measure across all banks, ‘a’ is a constant and ‘t’ is a trend
variable. A negative and significant slope coefficient sigma (r) is taken as evidence
15 The Securitisation and Reconstruction of Financial Assets and Enforcement of Security Interest Act,
2002 (SARFAESI) empowers banks to recover their non-performing assets without the intervention of the
court.
168 Econ Change Restruct (2013) 46:143–182
123
for r-convergence, i.e., a decline in SD (or CV) of efficiency measure over time
implies a narrowing of the dispersion of efficiency levels.
Table 5 presents the regression results pertaining to r-convergence. In all the nine
regression equations given in Column 1 of Panel A, B and C, the natural logarithm of
standard deviations of cost, technical and allocative efficiency scores, respectively is
taken as a dependent variable which is regressed on trend variable t. Further, the
regression equations given in Column 2 involve the natural logarithm of coefficient ofvariations of these efficiency measures as a dependent variable and time trend t as the
explanatory variable. The results reveal that in the regression equations given in
Panels A and B, the estimated parameter r (which is the coefficient of trend variable
t) bears a negative sign and is statistically significant at 1% level of significance for
the first sub-period and entire study period; whereas it is negative but insignificant in
the regression equations for the second sub-period. Further, all the regression
equations for the first sub-period and entire study period show a reasonable goodness
of fit with the values of R2 greater than 70%. From the regression equations
pertaining to allocative efficiency, as given in Panel C, we note that the estimated
parameter r is positive and insignificant for the entire study period. However, for the
distinct sub-periods, the sign of r has been observed to be negative and insignificant.
The aforementioned empirical findings vividly highlight that dispersion in the
distribution of cost and technical efficiency scores have decreased for the first sub-
period and entire study period. This implies that the gap between both cost and
technically inefficient and efficient PSBs has declined significantly during the entire
study period, and this phenomenon of narrowing the gap was more pronounced in
the first phase of reforms relative to second one. Further, some insignificant signs of
r-convergence in allocative efficiency levels appeared in the sub-periods but on the
whole no significant convergence in allocative efficiency levels has been noted in
Indian public sector banking industry during the entire period under investigation. In
a nutshell, the results confirm the presence of strong r-convergence in both the cost
and technical efficiency levels in Indian public sector banking industry throughout
the entire study period.
6.6 Testing of b-convergence
6.6.1 Absolute b-convergence
The concept of b-convergence relates to the catch-up phenomenon. Convergence of
b-type considers whether the improvement in efficiency exhibits a negative
correlation with the initial level of efficiency. There exists b-convergence in a cross-
section of banks if the inefficient banks tend to improve in efficiency faster than
efficient ones. The existence of b-convergence can be examined empirically by
estimating a cross-sectional regression of annual average growth rates of efficiency
on the initial levels of efficiency. Thus, the testing for b-convergence involves the
estimation of the following regression equation:
E�
i;t;t�s¼ ½lnðEi;tÞ � lnðEi;t�sÞ
�s ¼ aþ b ln Ei;t�s
� �þ ei;t ð5Þ
Econ Change Restruct (2013) 46:143–182 169
123
Ta
ble
5T
esti
ng
for
r-co
nv
erg
ence
Co
lum
n1
Colu
mn
2
Pa
nel
A:
Cos
tef
fici
ency
(CE
)
En
tire
per
iod
(19
92/1
99
3–2
00
7/2
00
8)
lnðS
Dt޼�
1:5
4���
ð�1
9:4
1Þ�
0:0
54
6��� t
ð�6:6
5Þ
(R2
=7
6.0
%)
lnðC
Vt޼�
1:2
4���
ð�1
4:9
5Þ�
0:0
63
3��� t
ð�7:4
0ÞÞ
(R2
=7
9.6
%)
Fir
stp
has
e(1
99
2/1
99
3–1
99
8/1
99
9)
lnðS
Dt޼�
1:4
9���
ð�2
4:7
3Þ�
0:0
58
9��� t
ð�4:3
8Þ
(R2
=7
9.3
%)
lnðC
Vt޼�
1:1
6���
ð�1
8:6
4Þ�
0:0
74
7��� t
ð�5:3
6Þ
(R2
=8
5.2
%)
Sec
ond
ph
ase
(19
99/2
00
0–2
00
7/2
00
8)
lnðS
Dt޼�
2:1
3���
ð�1
7:4
6Þ�
0:0
18
1t
ð�0:8
4ÞÞ
(R2
=9
.1%
)lnðC
Vt޼�
1:8
6���
ð�1
3:6
1Þ�
0:0
31
7t
ð�1:3
0Þ
(R2
=1
9.5
%)
Pa
nel
B:
Tec
hni
cal
effi
cien
cy(T
E)
En
tire
per
iod
(19
92/1
99
3–2
00
7/2
00
8)
lnðS
Dt޼�
1:5
4���
ð�1
9:1
9Þ�
0:0
52
7��� t
ð�6:3
7Þ
(R2
=7
4.3
%)
lnðC
Vt޼�
1:2
8���
ð�1
5:4
2Þ�
0:0
62
3��� t
ð�7:2
4Þ
(R2
=7
8.9
%)
Fir
stp
has
e(1
99
2/1
99
3–1
99
8/1
99
9)
lnðS
Dt޼�
1:4
8���
ð�2
0:1
5Þ�
0:0
60
8��� t
ð�3:7
0Þ
(R2
=7
3.2
%)
lnðC
Vt޼�
1:2
2���
ð�1
4:7
2Þ�
0:0
74
4��� t
ð�4:0
1Þ
(R2
=7
6.3
%)
Sec
ond
ph
ase
(19
99/2
00
0–2
00
7/2
00
8)
lnðS
Dt޼�
2:0
4���
ð�1
5:0
3Þ�
0:0
28
5t
ð�1:1
8ÞÞ
(R2
=1
6.6
%)
lnðC
Vt޼�
1:8
3���
ð�1
2:6
6Þ�
0:0
43
2t
ð�1:6
9Þ
(R2
=2
8.9
%)
Pa
nel
C:
All
oca
tive
effi
cien
cy(A
E)
En
tire
per
iod
(19
92/1
99
3–2
00
7/2
00
8)
lnðS
Dt޼�
3:1
1���
ð�1
7:6
2Þþ
0:0
03
0t
ð0:1
6Þ
(R2
=0
.2%
)lnðC
Vt޼�
3:0
6���
ð�1
6:6
1Þþ
0:0
03
6t
ð0:1
9Þ
(R2
=0
.3%
)
Fir
stp
has
e(1
99
2/1
99
3–1
99
8/1
99
9)
lnðS
Dt޼�
2:9
1���
ð�8:5
1Þ�
0:0
57
1t
ð�0:7
5Þ
(R2
=1
0.0
%)
lnðC
Vt޼�
2:8
4���
ð�7:9
5Þ�
0:0
59
6t
ð�0:7
5Þ
(R2
=1
0.0
%)
Sec
ond
ph
ase
(19
99/2
00
0–2
00
7/2
00
8)
lnðS
Dt޼�
3:0
2���
ð�1
3:6
8Þ�
0:0
04
8t
ð�0:1
2Þ
(R2
=0
.2%
)lnðC
Vt޼�
2:9
7���
ð�1
2:9
5Þ�
0:0
03
4t
ð�0:0
8Þ
(R2
=0
.1%
)
(1)
**
*,
**
,an
d*
rep
rese
nt
stat
isti
cal
sig
nifi
can
ceat
the
1,
5,
and
10
%le
vel
,re
spec
tiv
ely
;an
d(2
)F
or
all
reg
ress
ion
s,th
et-
stat
isti
csv
alu
esar
ep
rese
nte
din
par
enth
eses
So
urc
e:A
uth
or’
sca
lcula
tions
170 Econ Change Restruct (2013) 46:143–182
123
where E�
i;t;t�s¼ ½lnðEi;tÞ � lnðEi;t�sÞ
�s is the i-th bank’s average growth rate of
efficiency between the periods t and t - s, respectively. s is the length of the time
period. If the regression coefficient on the initial level of efficiency bears a statis-
tically significant negative sign, i.e., if b\ 0, then we can say that there exists
absolute b-convergence. The negative coefficient of the variable ‘initial level of
efficiency’ signifies that relatively inefficient banks having higher growth rates of
efficiency that enable them to catch-up with the efficient banks. It should be
observed that the equation (5) gives absolute, also denoted unconditional, b-con-
vergence under the assumption that all PSBs face homogenous economic and
regulatory environments.
For testing the hypothesis of absolute b-convergence, we estimated the
regression model (5) and hypothesized that the average annual growth rates of
cost, technical and allocative efficiencies have a negative relationship with their
initial levels. Table 6 shows the regression results for absolute b-convergence. In
all the nine regression equations reported in Panels A, B and C, we noticed a
reasonable goodness of fit of the model. Further, the results reveal that the
estimated b coefficients in all regression equations are both negative and
statistically significant, and thus indicating a negative relationship between the
initial level of efficiency measures and growth in these measures. We noticed a
Table 6 Testing for absolute b-convergence
Panel A: Cost efficiency (CE)
Entire period (1992/
1993–2007/2008)CE�
1992=1993�2007=2008 ¼ �0:0123���ð�3:49Þ
�0:0613���ð�8:58Þ
ln CE1992=1993(R2 = 74.7%)
First phase (1992/1993–
1998/1999)CE�
1992=1993�1998=1999 ¼ �0:0165�ð�2:01Þ
�0:0948���ð�5:72Þ
ln CE1992=1993(R2 = 56.6%)
Second phase (1999/
2000–2007/2008)CE�
1999=2000�2007=2008 ¼ �0:0237���ð�3:42Þ
�0:107���ð�5:02Þ
ln CE1999=2000(R2 = 50.2%)
Panel B: Technical efficiency (TE)
Entire period (1992/
1993–2007/2008)TE�
1992=1993�2007=2008 ¼ �0:00318ð�1:06Þ
�0:0501���ð�6:99Þ
ln TE1992=1993(R2 = 66.2%)
First phase (1992/1993–
1998/1999)TE�
1992=1993�1998=1999 ¼ �0:0153�ð�1:88Þ
�0:106���ð�5:41Þ
ln TE1992=1993(R2 = 53.9%)
Second phase (1999/
2000–2007/2008)TE�
1999=2000�2007=2008 ¼ �0:00625ð�1:25Þ
�0:0729���ð�4:04Þ
ln TE1999=2000(R2 = 39.5%)
Panel C: Allocative efficiency (AE)
Entire period (1992/
1993–2007/2008)AE�
1992=1993�2007=2008 ¼ �0:00699���ð�5:02Þ
�0:0739���ð�7:11Þ
ln AE1992=1993(R2 = 66.9%)
First phase (1992/1993–
1998/1999)AE�
1992=1993�1998=1999 ¼ �0:00965���ð�4:03Þ
�0:152���ð�8:53Þ
ln AE1992=1993(R2 = 74.4%)
Second phase (1999/
2000–2007/2008)AE�
1999=2000�2007=2008 ¼ �0:00702��ð�2:69Þ
�0:0553��ð�2:07Þ
ln AE1999=2000(R2 = 14.6%)
(1) ***, **, and * represent statistical significance at the 1, 5, and 10% level, respectively; and (2) For all
regressions, the t-statistics values are presented in parentheses
Source: Author’s calculations
Econ Change Restruct (2013) 46:143–182 171
123
reasonable goodness of fit of the model in most of the estimated regression
equations reported in Table 6. Regarding the speed of absolute b-convergence in
cost efficiency levels, we found that (1) it was about 6.1% per annum during the
entire study period; and (2) it was little more in the second phase relative to first
phase (10.7% vis-a-vis 9.5%). Further, the speed of absolute b-convergence in
technical efficiency (allocative efficiency) levels was 5.0% (7.4%) per annum
during the study period under consideration. Also, the speed of convergence in
technical efficiency (allocative efficiency) levels is relatively less in the second
sub-period in comparison to first one. On the whole, the empirical findings confirm
the occurrence of absolute b-convergence in cost, technical and allocative
efficiency levels of Indian public sector banking industry during the deregulatory
regime, but the rate at which the convergence occurred has declined for technical
and allocative efficiency levels in the latter sub-period relative to former.
6.6.2 Conditional b-convergence
Alongside the absolute b-convergence, we also tested the presence of conditional b-
convergence using the following equation:
E�
i;t;t�s
¼ ½lnðEi;tÞ � lnðEi;t�sÞ�s ¼ aþ b ln Ei;t�s
� �þXk
j¼1
dj lnðX ji;t�sÞ þ ei;t ð6Þ
The Eq. 6 allows us to control for the variables, which might influence the
steady-state level of efficiency measure. The choice of the control variables (or
conditioning variables) Xj depends upon economic theory, a priori beliefs about
growth process, and availability of data (Ghosh 2006). Conditional b-convergence
implies a negative correlation between the growth and initial level of efficiency
measure, after controlling for factors impacting steady-state position. Thus,
conditional b-convergence holds if b\ 0. The difference between these two
concepts of b-convergence is that absolute convergence means that each bank
moves toward the same steady-state efficiency, whereas conditional convergence
suggests that each bank possesses its own steady-state efficiency to which it is
converging. The conditional convergence and absolute convergence hypotheses
coincide only if all banks have the same steady-state (Fung 2006).
In order to test the hypothesis of conditional b-convergence, we estimated the
regression model (6) with three conditioning variables and hypothesized that the
steady-state efficiency growth rate of a bank is positively related to the bank’s
profitability (PROF),16 size (SIZE)17 and negatively related to intermediation cost(IC).18 The results for testing conditional b-convergence appear in Table 7. The
regression results reveal that the variables natural logarithm of the initial level of
cost, technical and allocative efficiencies bear a negative and statistically significant
coefficient in all the regression equations except one. The results, thus, indicate the
16 The ‘profitability’ is measured in terms of return on assets (ROA).17 The variable ‘size’ is measured in terms of value of total assets.18 The ‘intermediation cost’ is measured as the ratio of operating expenses as a percentage of total assets.
172 Econ Change Restruct (2013) 46:143–182
123
Ta
ble
7T
esti
ng
for
con
dit
ion
alb
-con
ver
gen
ce
Panel
A:
Cost
effici
ency
(CE
)
Enti
reper
iod
(1992/
1993–2007/2
008)
CE�
19
92=1
99
3�
20
07=2
00
8¼�
0:0
103
ð�0:1
8Þ�
0:0
611���
ð�8:1
1Þ
lnC
E1
99
2=1
99
3þ
0:0
0082
ð0:2
6Þ
lnSIZ
E1
99
2=1
99
3�
0:0
0097
ð�0:5
0Þ
lnP
RO
F1
99
2=1
99
3�
0:0
113
ð�0:6
5Þ
lnIC
19
92=9
3(R
2=
76.0
%)
Fir
stphas
e(1
992/1
993–
1998/1
999)
CE�
19
92=1
99
3�
19
98=1
99
9¼�
0:0
55
ð�0:5
4Þ�
0:0
99���
ð�7:2
6Þ
lnC
E1
99
2=1
99
3�
0:0
0292
lnSIZ
E1
99
2=1
99
3
ð�0:5
1Þ
þ0:0
113���
ð3:2
4Þ
lnP
RO
F1
99
2=1
99
3þ
0:0
586�
ð1:8
6Þ
lnIC
19
92=1
99
3(R
2=
75.2
%)
Sec
ond
phas
e(1
999/
2000–2007/2
008)
CE�
19
99=2
00
0�
20
07=2
00
8¼�
0:1
28
ð�1:4
3Þ�
0:1
02���
ð�2:7
0Þ
lnC
E1
99
9=2
00
0þ
0:0
0433
ð0:8
2Þ
lnSIZ
E1
99
9=2
00
0þ
0:0
009
ð0:0
3Þ
lnP
RO
F1
99
9=2
00
0þ
0:0
415
ð1:6
9Þ
lnIC
19
99=2
00
0(R
2=
57.2
%)
Panel
B:
Tec
hnic
al
effici
ency
(TE
)
Enti
reper
iod
(1992/
1993–2007/2
008)
TE�
19
92=1
99
3�
20
07=2
00
8¼
0:0
288
ð0:5
3Þ�
0:0
519���
ð�6:5
8Þ
lnT
E1
99
2=1
99
3�
0:0
0133
ð�0:4
3Þ
lnSIZ
E1
99
2=1
99
3�
0:0
0016
ð�0:0
8Þ
lnP
RO
FIT
19
92=1
99
3�
0:0
138
ð�0:8
0Þ
lnIC
(R2
=67.1
%)
Fir
stphas
e(1
992/1
993–
1998/1
999)
TE�
19
92=1
99
3�
19
98=1
99
9¼�
0:0
33
ð�0:2
8Þ�
0:1
01���
ð�6:0
5Þ
lnT
E1
99
2=1
99
3�
0:0
0454
ð�0:6
9Þ
lnSIZ
E1
99
2=1
99
3þ
0:0
118���
ð3:0
0Þ
lnP
RO
FIT
19
92=1
99
3þ
0:0
610
ð1:6
6Þ
lnIC
19
92=1
99
3(R
2=
72.7
%)
Sec
ond
phas
e(1
999/
2000–2007/2
008)
TE�
19
99=2
00
0�
20
07=2
00
8¼�
0:0
828
ð�1:0
5Þ�
0:0
538
ð�1:5
5Þ
lnT
E1
99
9=2
00
0þ
0:0
0353
ð0:7
5Þ
lnSIZ
E1
99
9=2
00
0�
0:0
102
ð�0:3
6Þ
lnP
RO
F1
99
9=2
00
0þ
0:0
330
ð1:5
9Þ
lnIC
19
99=2
00
0(R
2=
47.2
%)
Panel
C:
All
oca
tive
effici
ency
(AE
)
Enti
reper
iod
(1992/
1993–2007/2
008)
AE�
19
92=1
99
3�
20
07=2
00
8¼�
0:0
513�
ð�2:1
2Þ�
0:0
76���
ð�6:4
5Þ
lnA
E1
99
2=1
99
3�
0:0
0238�
ð1:7
7Þ
lnSIZ
E1
99
2=
19
93�
0:0
00732
�0:8
9ð
Þln
PR
OF
19
92=1
99
3þ
0:0
124
ð1:5
2Þ
lnIC
19
92=
19
93
(R2
=73.9
%)
Fir
stphas
e(1
992/1
993–
1998/1
999)
AE�
19
92=1
99
3�
19
98=1
99
9¼�
0:0
431
ð�0:9
5Þ�
0:1
63���
ð�7:3
1Þ
lnA
E1
99
2=1
99
3þ
0:0
0113
ð0:4
5Þ
lnSIZ
E1
99
2=
19
93þ
0:0
0043
ð0:2
8Þ
lnP
RO
F1
99
2=1
99
3þ
0:0
159
ð1:0
3Þ
lnIC
19
92=1
99
3(R
2=
75.6
%)
Sec
ond
phas
e(1
999/
2000–2007/2
008)
AE�
19
99=2
00
0�
20
07=2
00
8¼�
0:0
329
ð�0:8
2Þ�
0:0
689��
ð�2:6
4Þ
lnA
E1
99
9=2
00
0�
0:0
0230
ð0:0
8Þ
lnSIZ
E1
99
9=2
00
0�
0:0
207
ð1:0
0Þ
lnP
RO
F1
99
9=2
00
0þ
0:0
009��
ð�2:5
6Þ
lnIC
(R2
=40.7
%)
(1)
***,
**,
and
*re
pre
sent
stat
isti
cal
signifi
cance
atth
e1,
5,
and
10%
level
,re
spec
tivel
y;
and
(2).
For
all
regre
ssio
ns,
the
t-st
atis
tics
val
ues
are
pre
sente
din
par
enth
eses
Sourc
e:A
uth
or’
sca
lcula
tions
Econ Change Restruct (2013) 46:143–182 173
123
presence of strong conditional b-convergence in efficiency levels in Indian public
sector banking industry during the post-deregulation period. Further, the estimated
coefficients of conditioning variable lnSIZE are statistically insignificant in eight
regression equations. Thus, we observe no definite relationship between size and
growth of different efficiency measures in Indian public sector banking industry.
Similarly, we also failed to get a crystal-clear relationship between the variable lnICand growth of efficiency measures. The coefficients of control variable lnPROF are
positive in four equations but statistically significant only in two regression
equations. Overall, in seven cases, the coefficients of lnPROF are statistically
insignificant. The results, thus, suggest that in Indian public sector banking industry,
the relationship between the profitability and growth of cost and technical
efficiencies is very moderate in nature. Regarding the speed of conditional
b-convergence, we note that it was 6.11% per annum during the study period under
consideration. Also, the speed of convergence in CE levels is little more in the
second sub-period (10.2% vis-a-vis 9.9%). The implication of this finding is that
there was a smooth diffusion process of new banking technology in Indian public
sector banking industry during the post-reforms years, especially during the second
phase of reforms; and this process led to decrease in inter-bank disparities in the cost
efficiency levels over time. We further note that speed of convergence in AE levels
is greater than the TE levels (7.6% vis-a-vis 5.19%).
On the whole, the empirical findings provide evidence in favour of both r-
convergence and b-convergence in CE levels across PSBs. Following Koski and
Majumdar’s (2000) terminology, we can, thus, infer that Indian public sector
banking industry witnessed the presence of both catching-up as well as leapfrogging
phenomena during the post-reforms period. This implies that the initially cost
inefficient banks in Indian public sector banking industry are not only catching-up
with the initially efficient ones (i.e., the banks with a low level of efficiency at the
beginning of the period are growing more rapidly than highly efficient banks), but
their performance is improving at such a rate which enabled them to overtake the
well-performing banks. The most plausible reason for catching-up and leapfrogging
phenomena in Indian public sector banking industry is not only the improved
performance of initially lagging banks due to rationalization of the labour force,
better recovery of non-performing loans, increased application of technology, more
optimal allocation of resources, etc., but also the deterioration in the performance of
initially well-performing banks, especially the banks belonging to SBI group.
7 Conclusions, policy implications and directions for future research
The main purpose of this paper is to analyze the evolution of the cost efficiency in
Indian public sector banking industry during the post-deregulation period. In
particular, we intend to study the trends of cost efficiency and its components across
27 PSBs during the post-deregulation period spanning from 1992/1993 to
2007/2008. Further, we aim to investigate whether the phenomenon of convergence
in cost efficiency levels has taken place in Indian public sector banking industry
during the post-deregulation years or not. To accomplish the task of measuring the
174 Econ Change Restruct (2013) 46:143–182
123
levels of cost, technical and allocative efficiencies for individual PSBs, we have
used the increasing popular methodology of data envelopment analysis (DEA). The
presence of convergence in efficiency levels has been tested by using traditional
cross-sectional regression approach.
The results indicate that on the average, PSBs operating in India were found to
exhibit substantial cost inefficiencies to the tune of 25.6%, indicating that there is
significant room for improving their competitiveness and profitability. In addition,
technical inefficiency (relative to allocative inefficiency) accounts for a major share
of the deviations from optimal costs. Nevertheless, over the years, there has been a
modest growth in the relative cost efficiency of PSBs, with nationalized banks
displaying large cost efficiency gains, and banks in SBI group experiencing a
downturn in cost efficiency. Further, we note that (1) the growth in technical
efficiency contributed positively to the growth of cost efficiency, and (2) the growth
of cost efficiency in Indian public sector banking industry was circumscribed by
