Content uploaded by Andrew Worthington
Author content
All content in this area was uploaded by Andrew Worthington
Content may be subject to copyright.
<additional information>
D
D
Di
i
is
s
sc
c
cu
u
us
s
ss
s
si
i
io
o
on
n
n
P
P
Pa
a
ap
p
pe
e
er
r
r
a
a
an
n
nd
d
d
W
W
Wo
o
or
r
rk
k
ki
i
in
n
ng
g
g
P
P
Pa
a
ap
p
pe
e
er
r
r
S
S
Se
e
er
r
ri
i
ie
e
es
s
s
Malmquist Indices of Pre and Post-Deregulation Productivity,
Efficiency and Technological Change in the Singaporean Banking
Sector
Boon L. Lee*
School of Economics and Finance, Queensland University of Technology, Brisbane, Queensland, Australia.
Andrew C. Worthington
Department of Accounting, Finance and Economics, Griffith University, Nathan, Queensland, Australia.
Wai Ho Leong
Economics Division, Ministry of Trade and Industry, Singapore
Working/Discussion Paper # 228
February 2008
Abstract:
By the end of the 1990s, the Singaporean government had recognised the need to open up its banking sector so as
to remain competitive in the global economy. The Monetary Authority of Singapore thus began deregulation of the
banking sector in 1999 to strengthening the competitiveness of local banks relative to their foreign competition
through mergers. This paper employs a nonparametric Malmquist productivity index to provide measure of
productivity, technological change and efficiency gains over the period 1995-2005. The findings reveal some total
factor productivity growth associated with deregulation and scale efficiency improvement largely from mergers
amongst the local banks.
JEL Classifications: G21, D24
Keywords: Efficiency, productivity; deregulation; Malmquist indices; banking
* Corresponding author: School of Economics and Finance, Queensland University of Technology, GPO Box 2434, Brisbane, QLD
4001, Australia. Tel: +61 (0)7 3138 5389; Fax. +61 (0)7 3138 1500; email: bl.lee@qut.edu.au
1. Introduction
Since the beginning of the 1980s, financial institutions in many parts of the world have
undergone changes brought about by deregulation, globalisation, privatisation and the rapid
pace of development in information technology. This phenomenon is very evident in countries
such as Australia, Belgium, France, Germany, Italy, Japan, South Korea, the Netherlands,
Norway, Spain, the United Kingdom and the United States. Similar developments are found in
the Singaporean banking sector whereby recent regulatory changes have been spurred by the
challenges of global competition. Singapore’s central bank, the Monetary Authority of
Singapore (MAS), recognized the need to deregulate its financial sector and open its domestic
banking and insurance industries to foreign competition. This was undertaken not only to
remain competitive in the global economy, but also to strengthen its banking system in terms of
the quality of banking services and to maintain or increase market share. Both the Singaporean
government and the MAS were well aware of the small stature of local banks by international
standards and lags behind international banks “…in terms of technology, expertise, range and
quality of service to customers” (MAS, 1999, p.1). Recent technological developments have led
to banking services being no longer restricting to ‘bricks and mortar’ over-the-counter services
with e-banking becoming more prevalent. This new approach to banking enables foreign banks
to extensively reach out to domestic customers, which, in a matter of time, would further reduce
and neutralise the advantages of an extensive branch network and implicit and explicit
government protection (MAS, 1999).
In 1997, MAS reviewed its regulatory policies and in 1999, launched the first phase of a five-
year programme aimed at liberalising the banking sector in Singapore. The programme,
essentially aimed at the development and upgrading of local banks, had three key features: (i)
an increase in competition from giving access to foreign banks to enter the domestic market; (ii)
strengthening the corporate governance of local banks and attracting leadership talent so as to
<additional information>
reach a level of autonomy mature enough to make professional management decisions; and (iii)
lifting the forty percent foreign shareholding limit. With the onset of deregulation, the role of
MAS changed from regulation to supervision with the aim to “…monitor and differentiate
among institutions by giving the stronger and well-managed ones more operational flexibility
while maintaining stricter controls on the weaker ones” (MAS, 1998, p. 29).
The onus was now on banks to effectively govern themselves through the setting-up of
Nominating Committees to offer appointments to key management positions. The five-year
programme, which includes a package of new banking privileges and licences for foreign
banks, opened up the domestic banking sector in terms of the issuing of full banking licenses,
known as Qualifying Full Banks (QFBs), to foreign banks. The first phase of the programme
saw four foreign banks being awarded QFB privileges in October 1999. These comprised
ABNO Amro Bank MV, Banque Nationale de Paris, Citibank NA, and Standard Chartered
Bank. In addition, an additional eight Qualifying Offshore Banks (QOBs) and eight wholesale
bank licenses were granted in the first phase of the programme.1
The second phase of the programme launched in June 2001 saw MAS freeing up the wholesale
bank market by awarding twenty wholesale bank licences over the following two years and the
upgrading of existing QOBs and offshore banks to wholesale bank status. In December 2001,
MAS awarded two QFBs and sixteen wholesale bank licences, of which eight were converted
1 Wholesale banks are permitted to engage in the same range of banking services as QFBs, except for the
acceptance of Singapore dollar fixed deposits of less than S$250,000 per deposit from non-bank customers, and the
payment of interest on Singapore dollar current accounts operated by resident individuals. Offshore banks, besides
having the same restrictions imposed on Wholesale banks, have slightly more restrictions on dealings with
residents in terms of the acceptance of interest-bearing deposits from resident non-bank customers other than
approved financial institutions. Further, the credit limit was limited to S$300 million to non-bank customers who
are Singapore residents. But with liberalisation, the QOB privileges were relaxed and allowed to have their lending
limit raised to S$1 billion, from the previous limit of S$300 million. QOBs were also be allowed to accept S$
funds from non-bank customers through swap transactions.
<additional information>
from QOBs. By May 2003, eight other wholesale bank licences were awarded. The second
phase of the programme focused on the replacement of the restricted bank licenses with
wholesale bank licenses. thus allowing a wider range of banking activities to be conducted
(hence the renaming of restricted banks to wholesale banks). This move restructured MASs’
three-tiered banking license of Full, Restricted and Offshore banks, towards a more streamlined
two-tiered licensing regime of Full and Wholesale banks. The upgrade of Qualifying Offshore
Banks and Offshore Banks to Wholesale Banking status for the period 1998-2005 is evident in
Table 1 with the rising number of Wholesale banks and the falling number of Offshore banks.
Revisions to the QFB licenses were also carried out in the second phase with an increase in the
number of locations from ten to fifteen. Prior to revision of the QFB licenses, foreign banks
were allowed up to ten locations, of which five could be branches. The new privileges attached
to the QFB license now increased the limits of foreign banks to fifteen locations, of which ten
could be branches and the remainder as off-site automated teller machines (ATMs). In addition,
QFBs could also provide debit services through Electronic Funds Transfer at Point of Sale
(EFTPOS) networks, thus enhancing competition in retail banking through the permitted issue
of debit cards to consumers.
As shown in Table 1, as of March 1999 there were 142 commercial banks, comprising 9 local
banks, 22 full banks, 13 wholesale banks (previously termed restricted banks) and 98 offshore
banks. By March 2006, there were 108 commercial banks, of which 5 were local banks, 24
were full banks, 35 were wholesale banks and 34 were offshore banks. The drop in the number
of local banks was the result of mergers and acquisitions: namely, DBS acquired the Post Office
Savings Bank (POSB) in 1998; Keppel and Tat Lee merged to become Keppel-Tatlee in 1998;
the United Overseas Bank (UOB) and Overseas Union Bank (OUB) merged to become UOB in
2002, the Overseas Chinese Banking Corporation (OCBC) and Keppel-Tatlee Bank merged in
<additional information>
2002; and the Industrial & Commercial Bank (ICB) Limited and UOB merged in August 20022.
The driving force behind the government’s desire for the consolidation of the local banks was
the issue of size. Then Deputy Prime-Minister Lee expressed the rationale as follows: “…the
logic of Singapore’s position is inescapable. If we want strong banks, then they have to be big
banks and if they are big banks, then we must have fewer banks. This is the reality in many
small countries”3.
INSERT Table 1.
Strangely enough in the midst of deregulation, Singapore’s largest bank, the Development Bank
of Singapore (DBS), was not completely privatised, and though publicly listed was still partly
government-owned in terms of shareholdings4. Further, despite more foreign banks being
granted QFB licenses, the privileges were still rather limited, even after the revisions in 2001.
For instance, foreign QFBs are limited to sharing ATMs amongst themselves and are not
permitted to access the local banks' ATM networks. Arguably, the most important reason for
this quasi-market is national interest. MAS still has an important role in the form of supervision
over the smaller banks. The strengthening of corporate governance to maintain a high
prudential standard is vital to the survival of local banks in order to compete with foreign
banks.
However, failure to effectively supervise can have dire consequences, as realised in the 1991-
1993 banking crises in Norway and Sweden. Many banks suffered severely from substantial
credit losses as a result of poor management and failure to appropriately evaluate the risk-
levels. In addition, the financial system problems are associated with the deflation of real estate
values (Bartholomew 1994; Ball 1994). The eventual outcome was government intervention
2 Prior to merger, the Industrial & Commercial Bank (ICB) was a subsidiary of UOB.
3 Quote from article by Angela Tan, BG Lee: Singapore to Stay Open to Global Players, Business Times, June 3,
2001.
4 Temasek Holdings (Pte) Ltd, a company wholly owned by the Ministry for Finance Incorporated, is the
investment arm of the Singapore government. It effectively owns 420,170,835 shares (28%) of DBS holdings, of
which 15.7% is owned by Maju Holdings Pte Ltd, a wholly-owned subsidiary of Temasek Holdings (Pte) Ltd.
<additional information>
through the issue of general guarantees and bailing out banks (Lindblom, 1994). In Norway, the
two largest banks in Norway; Denorske Bank (DnB) and Christiania Bank og Kreditkasse
(CBK), were nationalised. The government’s long-term goal was to retain a substantial minority
position (20-33.33%) over these two banks. In Sweden, the measures were less drastic with
government bailouts (Skandinaviska Enskilda Banken in Sweden).
As recently as 2007, some of the major US banks, such as the Bank of America and Citibank,
are now facing a similar crisis due to falling housing prices and problems in mortgage loans.
The problems faced in economies with liberalised financial services is that the risk level rises as
a result of gaining greater market share from increasing loans due to competition.
Consequently, the failure to recognise the mortgage crisis would suggest that there was a lack
of appropriate governing body in monitoring due to complete deregulation. Hence, while
Singapore may be moving towards a more liberalised banking service, it still has some form of
monitoring embedded in its financial system in the form of MAS monitoring the smaller banks
as well as the government having some share of assets in Singapore’s largest bank, DBS. In
terms of the level of government involvement, it still plays a substantial role as indicated in
Singapore’s financial freedom index in the Index of Economic Freedom 2008 produced by the
Heritage Foundation. Singapore obtained an index of 50 whereas the top ten economies had
financial freedom index ranging between 70 and 90 thus showing the level of government
involvement in Singapore financial services.
In this paper, the main objective is to determine whether the merger of local banks as a result of
deregulation improved productivity over the period 1995 to 2005. The estimates of productivity
growth in Singapore’s banking over the period 1995-2005 are derived using the Malmquist
productivity index. Sources of any productivity change are established by decomposing the
Malmquist productivity index into changes in productive efficiency (catching-up up to the best-
practice frontier) and changes in the production frontier (technological change). While a myriad
of factors may have contributed to changes in bank productivity over this period, deregulation
<additional information>
is arguably the most significant event within the banking sector.
The paper is divided into five sections. Section 2 describes the Malmquist productivity index
and its decomposition. Section 3 describes the inputs and outputs employed and the limitations
faced. Section 4 presents the results in terms of productivity change, technological change and
efficiency change and assess their significance in relation to deregulation. The paper concludes
with some brief remarks.
2. Malmquist Productivity Index
The current study employs the nonparametric input-oriented Malmquist productivity index that
decomposes productivity change into technical change and technical efficiency change. This
approach has been adopted by many studies analysing productivity at the industry level,
including Färe, Grosskopf, Lindgren, & Roos (1992) in the pharmaceutical industry,
Hjalmarsson and Veiderpass (1992) in electricity retail distribution and Price and Weyman-
Jones (1996) in the gas industry, among others. In terms of banking and finance services,
related studies include Berg, Forsund and Jansen (1992), Fukuyama (1995), Gilbert and Wilson
(1998), Worthington (1999), Rebelo and Mendes (2000), Alam (2001), Mukherjee, Ray and
Miller (2001), Isik and Hassan (2003), Casu, Girardone and Molyneux (2004), Sturm and
Williams (2004) and Rezitis (2006).
INSERT FIGURE 1
The framework can be illustrated by Figure 1, following Coelli, Rao & Battese (1998). In this
diagram, a production frontier representing the efficient level of output (y) that can be produced
from a given level of input (x) is constructed, and the assumption made that this frontier can
shift over time. The frontiers (F) thus obtained in the current (t) and future (t+1) time periods
are labelled accordingly. When inefficiency is assumed to exist, the relative movement of any
given financial institution over time will therefore depend on both its position relative to the
<additional information>
corresponding frontier (technical efficiency) and the position of the frontier itself (technical
change). If inefficiency is ignored, then productivity growth over time will be unable to
distinguish between improvements that derive from a financial institution ‘catching up’ to the
frontier, or those that result from the frontier itself shifting up over time.
Now, for any given financial institution in period t, say, represented by the output/input bundle
zt, an input-based measure of efficiency can be deduced by the horizontal distance ratio 0N/0S.
That is, inputs can be reduced in order to make production technically efficient in period t (ie.
movement onto the efficient frontier). By comparison, in period t+1 inputs should be multiplied
by the horizontal distance 0R/0Q in order to achieve comparable technical efficiency to that
found in period t. Since the frontier has shifted, 0R/0Q exceeds unity, even though it is
technically inefficient when compared to the period t+1 frontier.
It is possible using the input-orientated Malmquist productivity index to decompose this total
productivity change between the two periods into technical change and technical efficiency
change. Input-orientation refers to the emphasis on the equiproportionate reduction of inputs,
within the context of a given level of output. Studies such as Berg, Forsund and Jansen (1992),
Fare, Grosskopf and Lovell (1994), Fukuyama (1995), Gilbert and Wilson (1998), and Rebelo
and Mendes (2000) employed this approach. Following Fare, Grosskopf and Lovell (1994), the
input-oriented Malmquist productivity change index is expressed as:
2
1
1
11111
111 ),(
),(
),(
),(
),,,( ⎥
⎦
⎤
⎢
⎣
⎡×=
+
+++++
+++
tt
I
t
tt
I
t
tt
I
t
tt
I
t
tttt
I
txyD
xyD
xyD
xyD
xyxyM (1)
where the superscript I indicates an input-orientation, M is the productivity of the most recent
production point (xt+1, yt+1) (using period t + 1 technology) relative to the earlier production
point (xt, yt) (using period t technology), D are input distance functions, and all other variables
<additional information>
are as previously defined. Values greater than unity indicate positive total factor productivity
(TFP) growth between the two periods. An equivalent way of writing this index is:
2
1
1111
11111
111 ),(
),(
),(
),(
),(
),(
),,,( ⎥
⎦
⎤
⎢
⎣
⎡×=
++++
+++++
+++
tt
I
t
tt
I
t
tt
I
t
tt
I
t
tt
I
t
tt
I
t
tttt
I
txyD
xyD
xyD
xyD
xyD
xyD
xyxyM (2)
or
P
E
M
⋅
=
(3)
where
),(
),( 111
tt
I
t
tt
I
txyD
xyD
E+++
= (4)
2
1
1111
11 ),(
),(
),(
),( ⎥
⎦
⎤
⎢
⎣
⎡×=
++++
++
tt
I
t
tt
I
t
tt
I
t
tt
I
txyD
xyD
xyD
xyD
P (5)
and M (Malmquist TFP) is the product of a measure of technical progress P as measured by
shifts in the frontier measured at period t + 1 and period t (the geometric mean of the two ratios
in the square bracket) and a change in efficiency E over the same period (the term outside the
square bracket). Using this approach, four efficiency/productivity indices are provided for each
financial institution along with a measure of technical progress over time. These are: (i)
technical efficiency change (i.e. relative to a constant returns-to-scale technology); (ii)
technological change; (iii) pure technical efficiency change (i.e. relative to a variable returns-to-
scale technology); (iv) scale efficiency change; and (v) TFP change. Coelli, Rao & Battese
(1998) discuss the linear programs necessary to calculate these indices and the DEAP Version
2.1 software used in this analysis.
<additional information>
3. Data and Input/Output Specification
The data consist of annual observations of twenty-six commercial banks over the period 1995 to
1999 and ten commercial banks over the period 2000 to 2005. The data are drawn from the
audited financial statements of the banks, purchased from the Accounting and Corporate
Regulatory Authority (ACRA) (previously known as the Registry of Companies and
Businesses) in Singapore. Seven other commercial banks were excluded through the technical
requirement for a balanced panel of data: the Bank Nationale De Paris, Paribas Merchant
Banking Asia, Bank of Tokyo, Union Bank of Switzerland, Mitsubishi Bank, Tat Lee Bank,
and HSBC Investment Bank.
The current study is an extension of Leong and Dollery (2004) which focused solely on the
commercial banks. The sample size of 26 banks is in some way representative of the banking
industry. Bank sizes ranging from SG$1.9 billion to SG$106 billion in 2000 allows the study to
analyse productivity growth based on the utilisation of inputs and not driven by the institution
size which is not a focus of the current study.
The issue of determining outputs and inputs is highly dependent on the development process on
what banks actually produce. This has been an on-going contentious issue in the banking
literature (see Berger and Humphrey, 1992). In general, there are two main approaches to
classifying outputs and inputs; the production approach and the intermediation approach. The
production approach employed in studies like Sherman and Gold (1985), Berg, Forsund and
Jansen (1992), Berg and Humphrey (1992), and Drake (2001) consider deposit-taking
institutions as the producers of services associated with the loans and deposit accounts. Hence,
loans and deposits are ‘produced’ with inputs like capital and labour. In contrast, the
intermediation approach consider financial institutions as intermediaries and that the sole
<additional information>
purpose of banks is to raise funds through deposits and/or borrowed wholesale funds (managed
liabilities) and transform these into loans and other earning assets. This approach thus identifies
loans and other earning assets as outputs while deposits and borrowed funds together with
capital and labour as inputs. Studies that adopted the intermediation approach include Millar
and Noulas (1996), Gilbert and Wilson (1998), Rebelo and Mendes (2000), and Drake (2001).
In the context of Singapore’s commercial banks, Leong and Dollery (2004) noted that the
quantum of high value-added deposits compared to time and savings deposits is relatively
small. Further, given the fact that foreign banks are legally restricted in their ability to accept
Singapore dollar deposits, this would imply that their revenue share of interest-bearing assets
would be larger than deposits (Leong and Dollery, 2004). It is based on these rationales that the
current study employs the intermediation approach and identifies one output: loans to non-bank
customers (y1) and three inputs: customer deposits (x1), fixed assets (x2), and personnel/staff
costs (x3)5. All monetary values are converted into 2000 prices using the GDP deflator of
financial and business services drawn from various issues of the Yearbook of Statistics
published by the Department of Statistics, Singapore.
It is important to note that some banks do not provide the personnel/staff costs (x3) input for the
years 1995 and 1996. Since the focus of this paper is the efficiency performance before and
after deregulation, a two-stage approach of the Malmquist productivity index is adopted. First,
for the years 1995 to 1999, which represents the period before deregulation, only two inputs,
customer deposits (x1) and fixed assets (x2) are considered. Second, from 1999 to 2005 which
represents the period of deregulation, all three inputs are used as these data are available from
the financial statements of each bank.
5 Other studies used number of employees while the current study used staff costs. Conventionally, the former
would be used but as some banks’ financial statements did not provide this information, we used staff costs as a
proxy to labour input.
<additional information>
For the period 2000-2005, the sample size was reduced from 26 to 10 due to the following
reasons. First, the reduced sample was due to a significant portion of Japanese banks (Singapore
branches) having shutdown from 1997 onwards due to bankruptcies faced by major financial
institutions in Japan. This is further worsened by the recession in Japan from 1997 to 1998.
Second, some other banks, including the Deutsche Bank Aktiengesellschaft, Calyon Merchant
Bank Asia Ltd and Credit Suisse (Singapore) Ltd, were excluded due to missing data in their
annual reports/financial statements for certain years. Barclays Bank PLC was excluded as its
data provided unusual figures in loans.
4. Empirical results
Table 2 reports the sample means of inputs and outputs by year for the period 1995-1999 while
Table 3 reports the same information for the period 1999-2005. Before deregulation, the most
interesting indication is the low average annual growth rate of loans (output) compared to its
inputs. Largely accounting for the poor growth was the onset of the Asian financial crisis in
1997. Although Singapore weathered the Asian financial crisis better than many Asian
economies, it was still affected by it due to its close economic integration with other regional
economies. The effects flowed-on to the wholesale and retail trade, hotels and restaurants, as
well as its financial services sector, with a slowdown in growth in these sectors. In turn the
effects reduced the level of loans as well as the accumulated level of fixed assets.
INSERT TABLE 2
INSERT TABLE 3
Since 2000, with the gradual implementation of deregulation and recovery from the financial
crisis, growth rates for loans have improved at an average rate of 6.60 percent per annum.
However, when compared to the growth rates of inputs, this would suggest that there was little
<additional information>
productivity growth. The above comparisons of means, while interesting, do not provide any
productivity change analysis. Such an analysis is based on the Malmquist indices of
productivity as detailed in Section 2 based on the assumption that banks operate under constant
returns-to-scale. These results are presented and analysed below.
Three primary results are derived from the Malmquist indices of productivity growth over the
sample period. First, the measurement of productivity change. Second, the decomposition of
productivity change into efficiency change (i.e. a ‘catch-up’ effect) and technological change
(i.e. a ‘frontier-shift’ or ‘best-practice frontier’ effect). Third, the ‘catch-up’ effect is further
decomposed into technical efficiency and scale efficiency: this helps explain the main sources
of improvement.
INSERT TABLE 4
Table 4 shows the mean annual figures for total factor productivity (henceforth TFP), efficiency
change and technological change, as well as the number of banks on or above the frontier for
the periods 1995-1999 and 2000-2005. On examining the changes in productivity, efficiency
and technology for the period 1995 to 1999, there was a mean increase in TFP of 1.2 percent
due to improvements in efficiency (48.8 percent), but dampened by a decrease in technological
change by 32 percent. Table 4 clearly shows efficiency change being the main driver of TFP. It
is interesting to note that TFP in 1998 was below unity due to a decline in efficiency change,
rather than technological change because of the effects of the Asian financial crisis. The
implication from this is that many of the banks (19 of them) must have improved through best-
practice measures in reaction to the contagion from the Asian financial crisis. For the period
2000 to 2005, mean TFP fell by 3.6 percent due to technological regress although there was
evidence of ‘catch-up’ of around 23.4 percent. A finding that is similar to studies on banks is
the efficiency change score. The relatively high efficiency change scores, before and after
deregulation, are in line with other studies on banking efficiency such as Elyasiani and Mehdian
(1995) for US banks, Favero and Papi (1995) for Italian banks, and Christopoulos and Tsionas
(2001) for Greeks banks.
<additional information>
In regards to technological change, its mean score in post-deregulated period compared to the
previous period showed signs of improvement (from -32 percent to -21.9 percent), whereas for
technical change, this fell from 48.8 percent to 23.4 percent. What this suggests is that in
general, banks have begun to adopt best-practice with the adoption of new forms of innovation
to improve banking services such as e- banking which improves efficiency and enhances
competition and convenience to customers. This is evident from Table 4, which shows
technological change of over 1.00 in 2004 and 2005. Prior to 2004, technological change of less
than 1.00 reflects most banks still in the process of introducing e-banking as part of their
service. Wu, Hsia and Heng (2006) identified that e-banking was a disruptive innovation for the
incumbent banks and required massive changes in the areas of both technological knowledge
and business model. Such change require significant amounts of time and thus from 2000 to
2003, the technological change was less than unity.
An interesting issue to note is that with the onset of deregulation of Singapore’s banking
services, the period 2000-05 exhibit lower TFP than before liberalisation. Economic theory
dictates that with deregulation, the level of competition increases and in turn improves
efficiency and productivity. Whilst this may not seem to hold true from the findings of Table 4,
it is important to note that the sample size differs between the two periods and that the TFP
score that is being examined is only an average score which may be exaggerated as a result of
poor performance by just a few banks (ie. outliers). A more concise analysis on TFP would be
at the firm level which is examined in Section 4.2. Nevertheless, to ascertain the contributions
to the fall in mean TFP between these two periods, we further examine efficiency change as this
indicator showed a deprovement between the two periods. Efficiency change is decomposed
into pure technical efficiency and scale efficiency and their scores are presented in Table 5.
INSERT TABLE 5
As mentioned earlier the main driver of TFP change for both periods was efficiency change. A
decomposition of this indicator into pure technical efficiency (PTE) and scale efficiency (SE)
<additional information>
would provide more evidence for TFP growth. For the period 1995-1999, most banks were
operating efficiently except for the year 1998, which exhibits the effects of the Asian financial
crisis in 1997. In 1998, 17 banks were operating inefficiently, indicating that these banks could
have saved, on average, 25.3 percent of (that is, 1 – E) in input quantities if they had adopted
best practice technology. The productivity losses for this year are attributed to the decrease in
scale efficiency of around 33.4 percent (1 – SE) which indicates failure to adopt best-practice
management. This is expected when investments in banks fall during a financial turmoil
resulting in surplus resources and thus poor allocation of available resources. For the period
2000-2005, most banks were operating efficiently with both technical efficiency and scale
efficiency contributing towards the change in efficiency with some meaningful “catch-up”.
Tables 6 and 7 present the mean productivity scores for each bank for the periods 1995-99 and
2000-05, respectively. The main aspect of this part of the discussion is to determine whether the
local banks have shown any productivity improvement before and after deregulation of the
Singapore financial services. The local banks are OUB holdings, KTB Ltd, UOB holdings,
DBS Bank Ltd, and OCBC Holdings. For the period 1995-99, of the 26 banks, 12 had a TFP
score above 1.00 which indicates productivity growth. Amongst these 12 banks, 9 of which are
mainly investment or merchant banking operations and have TFP scores above the retail banks.
These are Credit Suisse (Singapore) Ltd, Societe Generale Asia (Singapore) Ltd, Bank of
America (Singapore) Ltd, Morgan Guaranty Trust Company of New York, JP Morgan Chase
Bank, N.A., Deutsche Bank Aktiengesellschaft, Royal Bank of Canada, and The Industrial
Bank of Japan Ltd. One postulate is that these banks are more nimble and globally focused
business with very diverse portfolios. This meant that their production functions were more
geographically diversified and would thus emerged less battered by the Asian financial crisis
compared to their domestically oriented retail peers.
INSERT TABLE 6
<additional information>
The local banks performed modestly in that 2 of the 5 local banks had TFP over 1.00, namely
OUB Holdings and KTB Ltd. TFP growth for UOB holdings and DBS Bank Ltd fell by 5.7
percent and 8.3 percent, respectively. OCBC Holdings was the worst performer amongst the
local banks with TFP growth falling by 11.5 percent. All five local banks however experienced
some form of ‘catch-up’ attributed to improvements in pure technical efficiency - OUB
Holdings (3.239), UOB holdings (3.03), DBS Bank Ltd (3.744) and OCBC Holdings (3.394).
INSERT TABLE 7
In the deregulated period, there was some improvement in TFP amongst the local banks.
OCBCs’ improvement in TFP from -11.5 percent to 0.7 percent would suggest that the merger
with KTB was the driving force. This outcome is supported by observing the scale efficiency,
whereby OCBC improved from -59.5 percent to 27.5 percent which would imply improvements
in operating size and management practices. OCBC’s improvement in technological change
(from 0.644 to 0.713) also suggest the benefits gained from best-practices as a result of
acquiring the Bank of Singapore Limited (BOS) in 2000, which during the dot-com era in 2000,
was Singapore's first pure internet bank. The merger of UOB and OUB had similar results to
the OCBC merger. UOBs’ TFP of 0.2 percent in the deregulated period was about the average
of OUB (1.4 percent) and UOB (-5.7 percent) in the pre-deregulated period. With the merger,
the improvements are clearly shown in the scale-efficiency scores from 0.432 (OUB) and 0.436
(UOB) to 1.195 (OUB merged). DBS experienced TFP growth higher than pre-deregulated
period with its merger with POSB in 1998. The above analysis would suggest that deregulation
which leads to mergers improves efficiency as mergers remove the redundancies and raises the
level of efficiency.
The performance of the Singaporean branches of foreign banks also showed some unusual
results. First, Bank of America improved in TFP from 33.3 percent to 42.4 percent largely due
to improvements in technological change from -35 percent in the period 1995-99 to a growth
rate of 12.6 percent in 2000-05. Decomposing the efficiency change into technical efficiency
<additional information>
and scale efficiency shows the latter falling from 2.052 to 1.265. Second, Standard Chartered
Bank improved in TFP from 4.1 percent in 1995-99 to 12.9 percent in 2000-05 with
improvements made in technological change indicating a move towards best-practice decision-
making. Third, Dresdner Bank Aktiengesellschaft showed significant improvement with a TFP
growth from -43.2 percent in 1995-99 to 8.7 percent in 2000-05. The main improvement was
largely in both efficiency change (-11.2 percent to 46.3 percent) and technological change (-
36.1 percent to-25.7 percent). The significant improvement in efficiency change is attributed to
pure technical efficiency from -15.7 percent to 39.7 percent for the same periods. Fourth,
Citibank N.A. experienced significant TFP growth from -7.4 percent in 1995-99 to 1.6 percent
in 2000-05 largely driven by technological change: there was no ‘catch-up’. Finally JP Morgan
Chase Bank, Mizuho Corporate Bank Ltd and Royal Bank of Canada fared poorly in the
deregulated period with falls in TFP from 29.9 to -8.1 percent, -16 to -31.1, and 17.1 to -34.5
percent, respectively. Falling TFP for JP Morgan Chase Bank and Royal Bank of Canada was
attributed to deterioration in efficiency change, primarily scale efficiency while for Mizuho
Corporate Bank’s poor performance this was attributed to falling pure technical efficiency from
36.7 to 0 percent.
5. Concluding Remarks
This paper analysed productivity growth in Singapore’s banking sector before and following
deregulation. Using a two-stage approach, the Malmquist productivity index allowed a
comparison of the changes in productivity in terms of efficiency change and technological
change between the pre-deregulated period and post-deregulated period. Two outcomes were
revealed in our findings. First, the results from our study follow a similar pattern to Gilbert and
Wilson (1998) for Korean banks, Mukherjee, Ray and Millar’s (2001) for US banks, Casu,
Girardone and Molyneux (2004) on European banks, and Rezitis (2006) on Greek banks that
<additional information>
deregulation improves productivity growth. In the deregulated period, 7 of the 10 banks
experienced some productivity growth, mainly driven by improving best practices
(technological change). Second, although no significant ‘catch-up’ was evident, deregulation
improved operational size (i.e. scale efficiency) with the several bank mergers. This was one of
the main findings in the current study which aimed at determining the outcome from the
mergers of local banks.
Whilst the study has provided some promising results, it should be noted that one of the main
limitations of the current study was the use of a small sample size for the second period. A large
sample size would have provided more robust results, especially when using the Malmquist
productivity index model. Nonetheless, this is a first step towards examining the level of
efficiency of Singapore banks since deregulation. Future studies on this would aim at not only
increasing the sample size, but to improve on the data outputs where available, such as non-
lending activities (securities), risk-adjusted off-balance sheet items, and other earning assets.
So has Singapore benefited from deregulating its banking sector? In the years since 2000, there
has been some improvement, although it is relatively insignificant. This was the immediate
response to the growing foreign competition which resulted in the mergers of many local banks
into just a few conglomerates. However, as shown in Lindblom (1994) and the recent crisis of
mortgage defaults experienced by some of the major US banks in 2007, complete liberalised
financial services can still falter largely due to failure in risk-management and the lack of
appropriate counter-measures (ie. like a prudential authority or governing body overseeing the
operations). In the case of Singapore, the process of deregulation is ongoing, and as such it is
still in its infancy in terms of deregulation. However, with a governing body like the MAS
whose role is to supervise and monitor the operations of banks, and Singapore’s sound
economic management which has weathered the effects of the Asian Financial Crisis,
Singapore is no-doubt in a position prepared for such a crisis.
<additional information>
References
Alam, I.M.S., 2001, A Nonparametric Approach for Assessing Productivity Dynamics of Large U.S. Banks,
Journal of Money, Credit and Banking 33 (1), pp. 121-139.
Tan, A., 2001, BG Lee: Singapore to Stay Open to Global Players, Business Times, June 3.
Ball, M., 1994, The 1980s property boom, Environment and Planning A 26, pp. 671-695.
Bartholomew, P.F., 1994, Comparing Depository Institution Difficulties in Canada, the United States and the
Nordic Countries, The Journal of Housing Research 5 (2), pp. 303-309.
Berg, S.A., F.R. Forsund and E.S. Jansen, 1992, Malmquist indices of productivity growth during the deregulation
of Norwegian banking, 1980-89, Scandinavian Journal of Economics 94, pp. S211-S228.
Berger, A., and D. Humphrey, 1992, Measurement and Efficiency Issues in Commercial Banking, in Z. Griliches,
ed.: Output Measurement in the Services Sectors (University of Chicago Press, Chicago Ill).
Casu, B., C. Girardone and P. Molyneux, 2004, Productivity change in European banking: A comparison of
parametric and non-parametric approaches, Journal of Banking & Finance 28, pp. 2521–2540.
Christopoulos, D., and E.G. Tsionas, 2001, Banking Economic Efficiency in the deregulation period: Results from
heteroscedastic stochastic frontier models, The Manchester School 69, pp. 656-676.
Coelli, T., D. S. Prasada Rao, and G. E. Battese, 1998, An Introduction to Efficiency and Productivity Analysis
(Boston, MA, Kluwer).
Drake, L., 2001, Efficiency and productivity change in UK banking, Applied Financial Economics 11, pp. 557-
571.
Elyasiani, E., and S.M. Mehdian, 1995, The Comparative Efficiency performance of small and large commercial
banks in the pre and post deregulation eras, Applied Economics 27, pp. 1069-1079.
Fare, R., S. Grosskopf, and C.A.K. Lovell, 1994, Production Frontiers (Cambridge University Press, Cambridge).
Färe, R., S. Grosskopf, B. Lindgren, and P. Roos, 1992, Productivity Developments in Swedish pharmacies 1980-
1989: A non-parametric approach, Journal of Productivity Analysis 3 (1-2) pp. 85-101.
Favero, C.A., and L. Papi, 1995, Technical Efficiency and scale efficiency in the Italian banking sector: A non-
parametric approach, Applied Economics 27, pp. 385-395.
Fukuyama, H., 1995, Measuring efficiency and productivity growth in Japanese banking: a nonparametric frontier
approach, Applied Financial Economics 5, pp. 95-107.
<additional information>
Gilbert, R.A., and P.W. Wilson, 1998, Effects of Deregulation on the Productivity of Korean Banks, Journal of
Economics and Business 50, pp. 133-155.
Hjalmarsson, L. and A. Veiderpass (1992), Efficiency and ownership in Swedish electricity retail distribution,
Journal of Productivity Analysis, 3, 7-23.
Isik, I., and M.K. Hassan, 2003, Financial deregulation and total factor productivity change: An empirical study of
Turkish commercial banks, Journal of Banking and Finance 27, pp. 1455–1485.
Leong, W.H., and B. Dollery, 2004, The Productive Efficiency of Singapore Banks: An Application and extension
of the Barr et al. (1999) Approach, The Singapore Economic Review 49 (2), pp. 273-290.
Lindblom, Ted, 1994, Credit Losses in Nordic Banks, in Jack Revell, ed.: The Changing Face of European Banks
and Securities Markets (Macmillan, UK/St Martin’s).
Monetary Authority of Singapore, 1997/98 Annual Report, Singapore.
Monetary Authority of Singapore, 1999, Liberalising Commercial Banking and Upgrading Local Banks, Statement
by the Monetary Authority of Singapore, May, Singapore.
Millar, S.M., and A.G. Noulas, 1996, The technical efficiency of large bank production, Journal of Banking and
Finance 20, pp. 495-509.
Mukherjee, K., S.C. Ray, and S.M. Miller, 2001, Productivity growth in large US banks: The initial
postderegulation experience, Journal of Banking and Finance 25 (5), pp. 913–939.
Price, C.W., and T. Weyman-Jones, 1996, Malmquist Indices of Productivity Change in the U.K. Gas Industry
Before and After Privatization, Applied Economics 28, pp. 29-39.
Rebelo, J., and V. Mendes, 2000, Malmquist Indices of Productivity Change in Portuguese Banking: The
Deregulation Period, International Advances in Economic Research 6(3), pp. 531-543.
Rezitis, A.N., 2006, Productivity Growth in the Greek Banking Industry: A Non-Parametric Approach. Journal of
Applied Economics 9 (1), pp. 119-138.
Sherman, H.D., and F. Gold, 1985, Bank branch operation efficiency: evaluation with data envelopment analysis,
Journal of Finance 9, pp. 297-316.
Sturm, J.E., and B. Williams, 2004, Foreign Bank Entry, Deregulation and Bank Efficiency: Lessons from the
Australian Experience, Journal of Banking and Finance 28 (7), pp. 1775-99.
Worthington, A.C., 1999, Malmquist Indices of Productivity Change in Australian Financial Services, Journal of
International Financial Market, Institutions and Money 9 (3), pp. 303-320.
<additional information>
Wu, J-H., T-L. Hsia, and Michael, S.H. Heng, 2006, Core Capabilities for Exploiting Electronic Banking, Journal
of Electronic Commerce Research 7 (2), pp. 111-122.
<additional information>
Table 1. Number of commercial banks in Singapore by license type, 1998-2006
1998 1999 2000 2001 2002 2003 2004 2005 2006
Locala 12 9 8 8 6 5 5 5 5
Foreign Full
Bk
22 22 23 23 22 22 23 24 24
Wholesale
Bk
13 13 16 20 33 31 37 35 34
Offshore
Bk
107 98 93 82 59 59 50 47 45
Total 154 142 140 133 120 117 115 111 108
N
otes: All local banks are full banks. Fi
g
ures at March end.
Source: MAS Annual Report, 2005/06.
<additional information>
yt+1
0 x
y
yt
xt+1 xt
zt+1
zt
Ft+1
Ft
N P L S Q R
Figure 1. Malmquist index and productivity change over time
<additional information>
Table 2: Means of inputs and outputs, 1995 - 1999 (in millions of SG$ at 2000 prices),
1995 1996 1997 1998 1999 Average
Annual growth
rate,
1995-1999
y1: loans to non-bank customers 8,060 10,103 10,852 9,921 8,840 2.31
x1: customer deposits 5,677 7,202 8,108 9,470 10,815 16.11
x2: fixed assets 116 150 111 185 204 14.11
Note: Mean is the average of 26 financial institutions.
<additional information>
Table 3: Means of inputs and outputs, 1999 - 2005 (in millions of SG$ at 2000 prices),
1999 2000 2001 2002 2003 2004 2005 Ave.
Annual
growth
rate,
1999-
2005
y1: loans to non-
bank customers 15,500 15,296 20,573 18,504 21,562 23,100 23,032 6.60
x1: customer
deposits 21,485 21,822 28,302 26,102 30,756 32,200 31,728 6.49
x2: fixed assets 426 431 605 540 567 531 526 3.51
x3: staff costs 156 187 229 223 247 264 295 10.62
Note: Mean is the average of 10 financial institutions.
<additional information>
Table 4: TFP, efficiency and technological change scores in Singaporean banks
(annual mean)
TFP change Efficiency Change Technological Change
Number of
Banks
Score Number
Efficient
Score Number
Efficient
Score Number
Efficient
1996 26 1.297 20 1.463 24 0.887 0
1997 26 1.055 11 2.057 22 0.513 0
1998 26 0.756 9 0.747 9 1.013 19
1999 26 1.015 10 2.185 24 0.465 1
Mean 1.012 1.488 0.680
2000 10 1.006 4 1.672 10 0.602 0
2001 10 1.085 4 1.263 7 0.859 2
2002 10 0.822 2 1.322 9 0.622 1
2003 10 0.806 5 1.210 8 0.666 2
2004 10 1.302 7 1.291 7 1.009 9
2005 10 0.853 2 0.810 1 1.053 8
Mean 0.964 1.234 0.781
<additional information>
Table 5: Efficiency scores in Singaporean banks, 1996-2005 (annual mean)
Efficiency Change (E) Pure Technical Efficiency
(PTE)
Scale Efficiency (SE)
Number of
Banks
Score Number
Efficient
Score Number
Efficient
Score Number
Efficient
1996 26 1.463 24 1.000 14 1.463 24
1997 26 2.057 22 2.444 25 0.841 7
1998 26 0.747 9 1.120 15 0.666 14
1999 26 2.185 24 1.812 22 1.206 21
Mean 1.488 1.492 0.997
2000 10 1.672 10 2.127 10 0.786 6
2001 10 1.263 7 1.367 10 0.923 6
2002 10 1.322 9 1.006 9 1.315 8
2003 10 1.210 8 1.122 10 1.079 6
2004 10 1.291 7 1.127 10 1.146 6
2005 10 0.810 1 0.821 6 0.986 3
Mean 1.234 1.203 1.026
<additional information>
Table 6: Ranked TFP scores by individual banks, 1995-1999 (annual mean)
Efficiency
Change
Technological
Change
Pure
Technical
Efficiency
Scale
Efficiency
TFP
Credit Suisse (Singapore) Ltd 2.803 0.792 2.177 1.287 2.219
Societe Generale Asia (Singapore) Ltd 2.377 0.714 1.325 1.794 1.697
Bank of America (Singapore) Ltd 2.052 0.650 1.000 2.052 1.333
Morgan Guaranty Trust Company of New York 1.794 0.726 1.443 1.243 1.303
JP Morgan Chase Bank, N.A. 2.061 0.630 1.303 1.582 1.299
Deutsche Bank Aktiengesellschaft 1.921 0.660 1.624 1.183 1.267
Royal Bank of Canada 1.811 0.646 1.393 1.300 1.171
OCBC Trustee Ltd 1.578 0.731 1.743 0.905 1.154
The Industrial Bank of Japan Ltd 1.536 0.693 1.319 1.164 1.065
Standard Chartered Bank 1.644 0.633 1.516 1.084 1.041
OUB Holdings 1.399 0.725 3.239 0.432 1.014
KTB Ltd 1.488 0.677 1.458 1.020 1.007
Credit Agricole (Suisse) SA 1.381 0.712 1.323 1.044 0.983
The Tokai Bank Ltd 1.329 0.713 1.270 1.046 0.948
UOB Holdings 1.321 0.714 3.030 0.436 0.943
Citibank N.A. 1.423 0.651 1.351 1.053 0.926
DBS Bank Ltd 1.284 0.714 3.744 0.343 0.917
The Sakura Bank 1.429 0.634 1.326 1.078 0.906
OCBC Holdings 1.375 0.644 3.394 0.405 0.885
ABN AMRO Asia Merchant Bank (Singapore) Ltd 1.300 0.665 1.000 1.300 0.865
Mizuho Corporate Bank Ltd 1.330 0.632 1.367 0.973 0.840
Calyon Merchant Bank Asia Ltd 1.082 0.763 0.998 1.084 0.826
Barclays Bank PLC 1.272 0.622 1.156 1.100 0.791
Sumitomo Mitsui Banking Corporation 1.155 0.672 1.061 1.089 0.777
The Asahi Bank Ltd 1.000 0.670 1.000 1.000 0.670
Dresdner Bank Aktiengesellschaft 0.888 0.639 0.847 1.048 0.568
<additional information>
Table 7: Ranked TFP scores by individual banks, 2000-2005 (annual mean)
Efficiency
Change
Technological
Change
Pure
Technical
Efficiency
Scale
Efficiency
TFP
Bank of America (Singapore) Ltd 1.265 1.126 1.000 1.265 1.424
Standard Chartered Bank 1.465 0.771 1.551 0.945 1.129
Dresdner Bank Aktiengesellschaft 1.463 0.743 1.397 1.047 1.087
Citibank N.A. 1.355 0.750 1.496 0.906 1.016
UOB Holdings 1.399 0.716 1.170 1.195 1.002
OCBC Holdings 1.372 0.713 1.076 1.275 0.978
DBS Bank Ltd 1.329 0.724 1.000 1.329 0.962
JP Morgan Chase Bank, N.A. 1.351 0.680 1.327 1.018 0.919
Mizuho Corporate Bank Ltd 0.957 0.719 1.000 0.957 0.689
Royal Bank of Canada 0.676 0.968 1.176 0.575 0.655
