Does bank efficiency matter? Market value relevance of bank efficiency in Australia

Abstract

The stochastic frontier analysis is employed to investigate efficiency of publicly listed Australian banks over the period 1985 to 2008. The results suggest that technical, cost and profit efficiency of Australian banks have improved over time. Large banks have attained a higher level of cost efficiency but a lower level of technical efficiency compared to small banks. No substantial difference between the two groups is found in terms of profit efficiency. A panel regression of bank stock return on bank efficiency suggests that an improvement in technical, cost or profit efficiency contributes to the market value of a bank. Thus, the shareholder wealth maximization goal is aligned with the goal of maximizing bank efficiency in the Australian context.

Keywords:

• bank efficiency
• SFA analysis
• distance function
• market model

JEL Classification::

• D61
• G21
• D21
• G28

Notes

¹As Berger and Mester (1997) suggest that SFA has an advantage over DEA in terms of estimation of economic efficiency. Kempkes and Pohl (2010) also argue that the DEA technique may not be applicable to organizations with heterogeneous structure.

²Sturm and Williams (2004, 2008) find that major Australian banks use their market power as a barrier to entry to new foreign entrants.

³Some international studies (e.g. Chu and Lim, 1998; Beccalli et al., 2006; Guzman and Reverte, 2008) also observe that bank stock prices reflect DEA-based TE. In contrast to these studies, we derive efficiency scores from the SFA approach and examine the relation between different measures of efficiency (technical, cost and profit efficiency) and bank stock prices using a multi-factor model of stock returns.

⁴These are the Australian and New Zealand Banking Group Limited (ANZ), Commonwealth Bank of Australia (CBA), National Australia Bank (NAB) and Westpac Banking Corporation (WBC).

⁵For empirical analysis, noninterest income is used as a proxy for underlying noninterest bearing assets and off-balance sheet activities (see, e.g. Kirkwood and Nahm, 2006).

⁶The same software is also used to estimate the cost and profit frontier models.

⁷The detailed specification test results for the alternative assumptions regarding the stochastic frontier TE model are available from the corresponding author on request.

⁸The sample includes the ANZ, CBA, NAB, WBC, the Bendigo Bank (BEN), Adelaide Bank (ADB), Suncorp-Metway Group (SUN), St. George Bank (SGB), BOQ and Macquarie Group (MQG). Due to data unavailability, the number of observations is not the same in all models, which is reported in the corresponding result tables.

⁹Since ADB merged with the BEN at the end of 2007, descriptive statistics for 2007 rather than 2008 are presented.

¹⁰The detailed estimation results are available from the corresponding author on request.

¹¹The effect of time trend on an efficiency frontier is assessed by differentiating the distance production function, cost function or profit function with respect to time trend t.

¹²Note that the data are not available before 1988 for the big-four banks and before 1993 for the regional banks.

¹³Similar practice is also observed by Choi and Elyasiani (2011) in the insurance industry, where many customers are inclined to buy insurance products from large insurers even at a higher price.

¹⁴Given our small sample, we have ruled out the possibility of allowing βs to vary over time or across banks.