EVALUATING COST PERFORMANCE OF BANKS IN THE ASIA PACIFIC

Economic Papers Special Edition, December 2006 pp. 61-70

EVALUATING COST PERFORMANCE OF BANKS IN THE ASIA PACIFIC

KYM BROWN* and MICHAEL SKULLY*

This paper aims to evaluate the cost efficiency of banks in the Asia-Paclfic region and test whether the operating performance of banks in poorer economies improves with the inclusion of environmental proxies. Our basic cost eficiency model finds that Australia and Singapore are the most cost eficient banking economies. With the inclusion of environmental factors into our model, the relative cost performance of banks from poorer economies does not improve, perhaps suggesting that their weaker operating environment is not the cause of the banks’ lower results.

JEL codes: G2 1 , L25

Keywords: Banks, Cost Efficiency, DEA

1 Introduction

Banking and financial market integration is increasing in the ASEAN and APEC regions, leading to increased competition between local and foreign banks. Hence the measurement of bank cost efficiency in the Asia-Pacific is important for policy makers and bankers to enable them to understand how these potential changes impact on the incumbent local banks. Therefore the major research question posed is: How efficient are Asia-Pacific commercial banks? And more specifically, does the efficiency cost performance of banks from economies with poorer operating environments improve significantly when allowance is made for their differences such as competition, financial development, bank regulation and population. Efficiency is conventionally defined relative to other banks within a particular sample, and cost efficiency is based on estimating the minimal inputs possible, for a given set of outputs, based on the characteristics of those banks. Consequently, efficiency estimates based on different samples (involving different firms) are not directly comparable. A bank with an efficiency score of 100% in one sample, for example, may be relatively inefficient when compared with a different group of banks. This illustrates the problems encountered in designing across-country efficiency studies.

Traditionally, cost efficiency was measured via the bank’s cost to income ratio. Unfortunately, this did not properly reflect the intermediation process or allow for

Department of Accounting and Finance, Monash University. The authors greatly acknowledge Kevin Davis’ most useful editorial suggestions when revising our initial paper for this publication.

61

KYM BROWN & MICHAEL SKULLY

factors beyond the bank management’s control. By using the non-parametric Data Envelopment Analysis (DEA) technique, multiple inputs and outputs can produce a single cost efficiency measure, and therefore overcome these limitations.

However, a further problem is that country characteristics may create differing environments within which banks operate, and so may bias efficiency comparisons. So a direct application of DEA to a regional sample would ignore the impact of these environmental factors. In contrast, an economy-by-economy application would mean non-comparability between the different countries. Therefore our approach assumes a common efficiency frontier, after environmental effects have been considered. This seems only fair, given that the competition within an economy, its financial development, bank regulation and demographics are all factors outside a management’s control but may still affect cost efficiency.

This paper therefore examines the cost efficiency performance, with environmental considerations, of commercial banks across the Asia-Pacific region. Section 2 surveys the relevant literature and the methodology is detailed in Section 3. Sources and details of data are provided in Section 4, whilst the results are portrayed in Section 5 . The conclusions are contained in Section 6.

2 Literature Review

2.1 Bank Efficiency Literature

The first generation of bank efficiency research used ratio analysis within (initially) a single-country setting (e.g. Alhadeff, 1954; Benston, 1965; Bell and Murphy, 1968). The second generation came with new linear programming techniques and was based on cost efficiency measurement, often using the structure, conduct, and performance (SCP) research paradigm (e.g. Ruthenberg and Elias, 1996; Altunbas and Molyneux, 1996). Further methodological refinements led to a range of third-generation bank efficiency measures. These have included (i) the measurement and comparison of cost and profit efficiencies (e.g., Berger and Mester, 1997); (ii) the comparisons of differing efficiency models on the one data set (e.g., Allen and Rai, 1996; Weill, 2002); and finally the inclusion of environmental variables especially in cross-country literature (e.g., Dietsch and Lozano-Vivas, 2000).

Studies involving a consideration of environmental variables have tried to proxy and allow for country-specific environmental factors (e.g., Chaffai, Dietsch and Lozano- Vivas, 200 1 ; Dietsch and Lozano-Vivas, 2000; Lozano-Vivas, Pastor and Hasan, 2001 ; Saythe, 2006; Lozano-Vivas, Pastor and Pastor, 2002). These studies (primarily focused on developed countries and the European Union) suggest that environmental factors can have a strong influence on the end results. This paper therefore aims to apply this methodology to banks in the Asia-Pacific. It is hypothesised that banks operating in poorer economic environments would perform better if they operated in healthier economies.

Another relevant factor is size. It may be expected that as a bank increases in size, it may become more cost efficient. While studies of bank scale-efficiency give mixed results (e.g., Shaffer, 1985; Hunter and Timme, 1986; Gilligan et al., 1984; Cavallo and Rossi, 2001), effects of scale still need to be considered in cost-efficiency studies. We expect to find a U-shaped cost curve, but note that in a country such as China, the state

62

EVALUATING COST PERFORMANCE

ownership of large banks may distort the results. We also hypothesise that stock exchange listed banks are likely to be more efficient due to the discipline imposed by external shareholders.

2.2 Asian Bank Literature

In a study of the pre-Asian crisis period, Laeven (1999) linked efficiency with risk factors. Better performing banks had not only high efficiency levels but also lower risks. Banks with higher cost inefficiencies were found in countries with greater government involvement in lending, smaller bank size and more restrictive regulations. The riskiest banks were family-owned, followed by company-owned banks, whilst the least risky banks were foreign owned. An individual 'efficiency frontier' was derived for each country, given the differing environments in each country. Therefore efficiency levels across countries were not comparable.

In his technical efficiency analysis of the largest banks in seventeen Asian-Pacific countries, Sathye (2006) also experimented with environmental factors and developed frontiers at both the national and regional levels. He found that banks in high-income countries generally reported higher efficiency change. Williams and Nguyen (2005) examined the five crisis economies of Indonesia, Korea, Malaysia, the Philippines and Thailand over 1990-2003 using Stochastic Frontier Analysis (SFA). Within the region, state-owned banks were less efficient than private banks, and privatised banks generally performed better. Foreign banks generally acquired profitable banks, whose efficiency often fell on initially on acquisition, but then improved. In the pre-crisis period, Karim (2001) found that, of four Asian countries, the Thai banks were the most cost efficient, followed by Malaysia, Indonesia, and finally the Philippines. State banks were usually less efficient than privately owned banks. Karim (2001) found increasing returns to scale existed up to US$3 billion (there were diseconomies of scale thereafter) but continued cost efficiency gains. The cost inefficient countries had more restrictive regulatory banking systems and a smaller banking system size as well as greater government intervention. Cost inefficiencies increased prior to the 1997 crisis.

This study builds on this prior work by focusing on the post-Asian crisis period.

3 Methodology

The cost efficiency of commercial banks is measured here using the non-parametric Data Envelopment Analysis (DEA) technique. DEA has an advantage of not requiring a specific functional form (and can deal with inputs and outputs measured in quite different ways). It does not allow, however, for 'statistical noise' in the data and assumes that the banks within the sample are similar. The base model is the Banker, Charnes and Cooper (1984) (BCC) input model, which may be expressed as follows:

Min 0 subject to Yr L YO 8. r

xr 5 ex, err= I t10"

63


where Y is the matrix of output vectors; X is the matrix of input vectors; (&,Yo) is the unit being rated; eT denotes a row-vector of 1's; t is the vector of intensity variables; and t9 is the efficiency score-a quantity between 0 and 1. The approach can be applied assuming either a variable return to scale (VRS) or a constant return to scale (CRS).

Environmental variables are then added as non-discretionary components to the inputs and outputs (known as a single-stage model) if their individual influence meets the statistical criterion proposed by Lozano-Vivas, Pastor and Pastor (2002).' The adjusted environmental model is as follows:

Min B subject to Yz> Yo (2) 8. f

zr > X,

err= 1 t > 0 "

xr < ex,

where Z represents the matrix of selected environmental outputs, and Zo is the corresponding vector of the unit being rated. Environmental factors are included as outputs in the model.

4 Data

The Bankscope database provided data for 322 domestic commercial banks from the twelve largest economies in the Asia Pacific for the year 2004 (see later in Table 5) . The data was based on unconsolidated results (thus reflecting only banking activities) and converted to US dollars using the exchange rate at balance date. The environmental factors used to serve as proxies for the different operating conditions in each economy were obtained as shown in Table 1.

TABLE 1 ENVIRONMENTAL VARIABLE SOURCES

Variable Source Market structure Bankscope database sample Financial Development IMF IFS statistics

Asian Development Bank UNCTAD Bank for International Settlements World Federation of Exchanges Heritage Foundation, 2004 Index of Economic Freedom, SBS World Guide (2005) CIA Factbook (2004)

Bank Regulation Population Density Income per capita

' The methodology for selecting variables will be made available by the authors on request. 64


5 Results

5.1 Input and Output Variables

We adopt the ‘intermediation approach’ to the bank production process and assume the inputs are deposits, employee costs, and physical capital while the outputs are loans and other earning assets. To allow for cross-country differences in employee costs, we have used a proxy for the number of employees by dividing personnel costs by total assets. Physical capital is measured using other operating expenses (i.e., excluding employee or interest costs) divided by total fixed assets plus other non-earning assets (as done previously by Weill, 2004).

5.2 Environmental Variables

To allow for cross-country differences in operating conditions, thereby providing a more equitable measure of cost efficiency, proxies for competition, financial development factors, prudential regulation, density of population, and income per capita are all considered. These are included as discretionary variables within the model. Competition, measured by a bank concentration ratio based on the assets of the five largest banks as a percentage of all bank assets, minus 100 is shown in Column A, Table 2. Financial development is an index based on macroeconomic health, the depth of financial markets, and turnover of financial products. These three factors have then been averaged to compute an index score (out of ten-as shown in Column B, Table 2) . The more financially developed the country, the greater would seem the potential for efficient banks.

TABLE 2 ENVIRONMENTAL VARIABLES

A B C D E Economy Competition Financial Bank Density of Income

( I Concentration Development Regulation Population per capita - 1001) Index (5-index) (p ersonskm’) US$

Australia 13.4 3.3 4.0 2.7 26,900 China (PRC) 15.3 2.7 1 .o 129.9 1,290 Hong Kong (PRC) 20.4 5.8 4.0 6,834.3 26,8 10 Indonesia 26.8 0.9 1 .o 115.5 1,140 Japan 51.6 3.5 1 .o 338.8 37,180 Malaysia 46.5 3.3 1 .o 68.2 4,650 New Zealand 5.0 2.6 4.0 14.8 20,3 10 Philippines 38.6 1.5 2.0 273.3 1,170 Singapore 0.4 5.0 3.0 6,290.3 24,220 Republic of Korea 36.7 2.7 2.0 487.3 13,980 Taiwan 61.2 3.3 2.0 663.3 25,300 Thailand 30.7 1.9 2.0 122.5 2,540

Note: Columns A and C are converted so that all variables measured with a larger value are likely to have a positive impact on bank-cost efficiency.

65

TABL

E 3

SAM

PLE

DW

RIP

7W

E S

TATI

SiYc

li 12

[104

)

MiIl

ions

USD

A

vary

r R

mca

tw

suc

MQ

Ua

lit Y

S

truc

tun

Pcr

latI

UIl

CC

rs

quflt

r Er

nnom

y lo

o. b

unks

in rr

rmpl

c)

Aw

rayc

To

hl

Lorn

Lor

r Eq

uity

! N

d

Rdu

mon

R

iium

on

Co

stto

N

ct

NdL

oum

i Li

quid

A

&se

tspc

rBun

k K

amrl

Gra

u J

otu

IAx

ki

Int~

mt

Avc

rqgc

A+

crd@

Im

um

I.o

Mo

t C

urto

mr8

1 lc

urcW

L

OU

M

argi

n A

sssk

E

quity

R

ulio

a

lAs

rts

S

TFun

diag

W

&S

T

[%D

AA

) (B

0.U

) F-

AU

strp

lin (14)

57.334.4

0.62

12.35

2.09

1.15

14.50

55.62

79.86

109.88

8.19

Hoo

pKon

y( 13)

6.781.4

3.90

19.75

2.72

1.40

10.71

56.32

41.30

58.09

74.61

Ioda

nesia

(32)

2,358.4

3.76

12.43

5.03

1.43

10.79

55.04

51.07

69.38

26.49

Japtnill9)

30.221.4

2 .m

49

5 18

0 0.15

2.90

64 93

65

D1

70.02

18.14

Mala

lrja (221

g547.2

6.14

10.50

2.71

1.24

1365

43 21

5139

62.64

31.9

3

Philippincr

(231

2,087.3

8.33

12.46

4.93

0.81

3.82

84.68

47.30

58.91

39.63

Sin

wo

rc (6

) 34,258.4

3.74

21D4

1 3

9 1.10

7.73

45 mr

51

69

66.07

17.56

rnin

n (2

31

14.321.4

1.43

6.3 3

2.57

0.48

7.46

67.32

66.37

74.91

15.56

Thai

land

112

1 14,275.9

8.54

8.58

3.43

1.31

15.5

7 54

35

65

01

17.37

15.93

Chi

na, P

copl

cr' R

epub

lic cd

381

44,4%.3

2.53

99 7

2 30

0.42

93 7

$4.78

5260

66.78

34.56

Kor

et~S

outh

(13)

58,770.4

1.78

5.22

2.81

0.84

16.62

45.88

66.00

85.70

8.80

Nm

iknl

and (6)

22.829.2

0.68

4.4 6

2.04

0.89

44.13

57.59

78.60

88.61

10.28

v1

7;:


Bank regulation is adjusted so that the higher the degree of freedom, the less the constraints on achieving efficient production. So the higher the score in Column C, the more likely is it that a country’s banks will be efficient. With the next column (D), a higher density of population should also have a positive influence on possible bank efficiency, given the efficiency savings of less geographic spread. Finally, an economy with greater income per capita as shown in Column E, is expected to have more efficient banks. Population density and financial development proved to be highly correlated in the sample, but both were included in the initial analysis, given that the result was likely dnven by the financial centres of Singapore and Hong Kong.

5.3 Descriptive Statistics

Table 3 presents the sample descriptive statistics for all the countries countries, including a number of variables sometimes used as efficiency measures. Average bank size varies markedly across the countries, suggesting the need to allow for potential scale economies. The average loan loss reserve also varies markedly, as does the apparent riskiness of banks (‘proxied’ by the equitykotal assets ratio). On a costs-to- income basis (where lowest is ‘best’), Malaysia has the most efficient banks, followed by Singapore and Korea.

5.4 Cost Efficiency Results

TABLE 4 EFFICIENCY SCORES COMPARED

Country Basic Cost Initial Environmental Cost Environmental

Australia 0.836939 1 0.904385 5 Efficiency Ranking Efficiency Model

China Hong Kong Indonesia Japan Korea South Malaysia New Zealand Philippines Singapore Taiwan Thailand

0.5 13470 0.687173 0.5 15650 0.526361 0.714508 0.48 1247 0.565510 0.387 126 0.754303 0.512779 0.435942

8 4 7 6 3 10 5 12 2 9 1 1

0.669070 0.981522 0.534644 0.907777 0.720345 0.495050 0.987568 0.404182 0.998073 0.556249 0.447235

I 3 9 4 6 10 2 12 1 8 1 1

The basic model with a frontier across all economies was run with the two output and three input variables, with and without environmental variables. The results are listed in Table 4. Where environmental variables are excluded, Australian banks were the most cost efficient, followed by the Singapore and South Korea banks. The least efficient banks on average were then in the Philippines, and then Thailand and Malaysia. The environmental factors were added to the basic model individually. The most influential variable was population density and it was therefore selected first. Once all the environmental factors had been selected, the efficiency was recalculated. Both the number of filly efficient banks, and overall efficiency scores increased, as expected.

67


It had been hypothesised that banks operating in weaker economies, such as Indonesia, Thailand, and the Philippines, would have improved their ranking once their environmental differences were considered. However, our results did not reflect this. Indeed, economies such as Hong Kong, Japan, New Zealand, and Singapore actually experienced significantly improved efficiency scores. Given that population density was the first variable selected from the environmental variables, it was expected that geographically larger economies would be compensated for this.

Scale efficiencies are provided in Table 5. The results illustrate a classic U-shaped cost curve. As expected, the environmental factors improved the efficiency score for all banks.

TABLE 5 SCALE EFFICIENCY RESULTS

Size Category Basic Cost Model Environmental (USD Billions) Model > $100 billion 0.933035 0.96601 7 > US $50 billion > US $20 billion > US $10 billion > US $5 billion > US $2 billion > US $1 billion < US $1 billion

0.748714 0.90801 7 0.6 5 0444 0.856602 0.602141 0.775797 0.412980 0.677612 0.385302 0.606096 0.412603 0.540293 0.589566 0.652729

It was expected that stock-exchange listed banks would be more efficient than non- listed banks given the role of market discipline. However, the average efficiency score for the two groups was similar at around 58%. The listed Chinese banks were more efficient, which was as expected. For Taiwan, the Philippines, and Hong Kong, however, unlisted banks scored higher, which may reflect the importance of foreign banks shareholdings in particularly the last two countries’ domestic banks.

Finally, both the basic model and the environmental model efficiency scores were compared to the traditional measure of cost efficiency, the cost to income ratio. The cost-to-income ratio was found to be significantly correlated to the basic model but not to the environmental model measure, suggesting that a strict focus on cost to income ratios ignores the environmental differences.

6 Conclusions

This paper examined the cost efficiency of banks in the Asia-Pacific and whether the including environmental variables in our model would improve the efficiency scores of banks operating in the weaker economies. Contrary to our expectations, banks in the weaker economies improved only minimally as a result, although this may reflect exclusion of other relevant environmental factors. The improved efficiency scores were instead for Hong Kong, Japan, Malaysia, and New Zealand, possibly reflecting the influential environmental factor of population density for three of these countries.

68


The environmental model ranked Singapore, New Zealand, Hong Kong, Japan, and Australia respectively as having the most efficient banks. Given the development of their financial markets, and the importance of some as financial centres, it is perhaps not surprising that they might have the most efficient banks in the region. What is suggestive from these results is that such higher efficiency, even after allowing for environmental variables, may enable banks from these countries to expand competitively into other Asian countries due to their lower cost operations.

While the results from this model are interesting, further research is necessary to understand fully the causes of relative bank efficiency in the Asia Pacific region. A limitation of this paper is that only cost efficiency was analysed. A bank with higher profit efficiency may in fact incur higher costs in order to service its clients. So future research should also measure profit efficiencies, perhaps again taking account of environmental factors in the model. Further work is also being undertaken by the authors to examine the results from a multi-year study of efficiency, which may shed light on trends over time, rather than the single-year results presented here.

REFERENCES

Alhadeff, D. (1954) “Monopoly and Competition in Banking” (New York: Amo Press). Allen, L. and A. Rai (1996) “Operational Efficiency in Banking: An International

Comparison”, Journal of Banking and Finance, 20, pp. 655-672. Altunbas, Y. and P. Molyneux (1996) “Cost Economies in EU Banking Systems,

Journal of Economics and Business, 48, pp. 2 17-230. Banker, R. D., A. Charnes and W. W. Cooper (1984) “Some Models for Estimating

Technical and Scale Inefficiencies in Data Envelopment Analysis”, Management Science, 3, pp. 1078-1092.

Bell, F. W. and N. B. Murphy (1968) “Costs in Commercial Banking: A Quantitative Analysis of Bank Behaviour and its Relation to Bank Regulation”, Federal Reserve Bank of Boston, Research Report No. 4 1.

Benston, G . J. (1965) “Branch Banking and Economies of Scale”, Journal of Finance,

Berger, A. N. and L. J. Mester (1997) “Inside the Black Box: What Explains Differences in the Efficiencies of Financial Institutions?”, Journal of Banking and Finance, 21, pp. 895-947.

Cavallo, L. and S. P. S Rossi (2001) “Scale and Scope Economies in the European Banking Systems”, Journal of Multinational Financial Management, 1 1, pp. 5 15- 531.

Chaffai, M. E., M. Dietsch and A. Lozano-Vivas (2001) “Technological and Environmental Differences in the European Banking Industries”, Journal of Financial Services Research, 19, pp. 147-1 62.

Dietsch, M. and A. Lozano-Vivas (2000) “How the Environment Determines Banking Efficiency? A Comparison between French and Spanish Industries”, Journal of Banking & Finance, 24, pp. 985-1004.

Gilligan, T. W., M. L. Smirlock and W. Marshall (1984) “Scale and Scope Economies in the Multiproduct Banking Firm”, Journal of Monetaly Economics, 13, pp. 393- 405.

20, pp. 3 12-332.

69


Hunter, W. C. and S. G. Timme (1986) “Technical Change, Organizational Form, and the Structure of Bank Production”, Journal of Money, Credit and Banking, 18, pp.

Karim, M. Z. A. (2001), “Comparative Bank Efficiency across Select ASEAN Countries”, ASEAN Economic Bulletin, 18, pp. 289-304.

Laeven, L. (1999) “Risk and Efficiency in East Asian Banks”, World Bank Working Paper, No. 2255, December.

Lozano-Vivas, A., J. T. Pastor and I. Hasan (2001) “European Bank Performance Beyond Country Borders: What Really Matters?”, European Finance Review, 5, pp.

Lozano-Vivas, A., J. T. Pastor and J. M. Pastor (2002) “An Efficiency Comparison of European Banking Systems Operating Under Different Eenvironmental Conditions, Journal of Productivity Analysis, 18, pp. 59-77.

Ruthenberg, D. and R. Elias ( 1 996) “Cost Economies and Interest Rate Margins in a Unified European Banking Market”, Journal of Economics and Business, 48, pp.

Sathye, M. (2006) “Technical Efficiency of Large Bank Production in Asia and the Pacific”, Multinational Finance Journal (forthcoming).

Shaffer, S . (1985) “Competition, Economies of Scale, and Diversity of Firm Size, Applied Economics, 1 I , pp. 467476.

Weill, L. (2004) “Measuring Cost Efficiency in European Banking: A Comparison of Frontier Techniques”, Journal of Productivity Analysis, 21, pp. 133-1 52.

Weill, L. (2002) “Measuring Cost Efficiency in European Banking: A Comparison of Frontier Techniques”, paper presented at the Econometric Society ESEM Conference, Venice, Italy, 25-28 August.

Williams, J. and N. Nguyen (2005) “Financial Liberalisation, Crisis, and Restructuring: A Comparative Study of Bank Performance and Bank Governance in South East Asia”, Journal of Banking h Finance, 29, pp. 2 1 19-2 154.

152-166.

14 1-1 65.

231-249.

70

EVALUATING COST PERFORMANCE OF BANKS IN THE ASIA PACIFIC

Documents