论文：异方差随机前沿模型对波动性和经营效率的影响研究

论文：异方差随机前沿模型对波动性和经营效率的影响研究
——以两岸银行为例

摘 要：本文主要研究资产和收益的波动性对两岸银行经营效率的影响。作为风险和经营不确定性的测度，波动性会随个体和时间的变化影响效率的水平和变异程度。当证券资产波动时，风险溢酬随之变化，营业利润的变化进而影响经营计划和经营效率。本文采用Wang’s (2002)的异方差随机前沿模型进行估计，这个模型的优点在于除了常用的解释变量之外，它还能允许无效率扰动项的均值和方差因资产和收益的波动而变化。我们的实证数据来源于一个提供国际银行业全面资料的数据库Bankscope，主要以台湾和中国大陆的银行为例进行研究。所得结论与预期一致，即ROA的波动性越大，银行效率就会越低并且容易出现较大的变异。另一方面，股票收益率的波动不会影响银行效率，这意味着银行管理人员应该着重于改善和提高银行的经营业绩，而不是关注银行的股价表现。
关键词：波动性；效率；成本函数；SFA，ROA
中图分类号：G21; C67

Volatility and efficiency in a heteroscedastic stochastic frontier model:
the case of Taiwan and China banks

Abstract

We e*plore the role of asset and earning volatilities in determining the efficiency of international banks. As a measure of risk and operating uncertainty, volatility may influence both the levels and variabilities of efficiency across firms and over time. As equity volatility varies, changes in risk premiums may follow. Furthermore, the variability of operating profits may affect the level of planning difficulties and hence, will affect the operating efficiency. We adopt Wang’s (2002) heteroscedastic stochastic frontier model in our estimation. This model is an ideal framework for our study since it allows us to specify both the mean and variance of
……（新文秘网https://www.wm114.cn省略1946字，正式会员可完整阅读）……　
the properties of inefficiencies in U.S. bank holding companies derived from both stochastic and linear programming frontiers, they conclude that both methods produce “informative efficiency scores” and the stochastic frontier efficiency estimates are more closely related to risk-taking behavior, managerial competence and bank stock returns. Beccalli et al. (2006) make efficiency analysis comprising the sample of all banks publicly listed in France, Germany, Italy, Spain and UK; their results indicate that changes in cost efficiency are reflected in changes in stock prices. Altunbas et al. (2001) use a sample of German banks for empirical analysis, but find little evidence to suggest that privately owned banks are more efficient than their mutual and public-sector counterparts. Bonin et al. (2005) use an unbalanced panel consisting of 225 banks to investigate the effects of ownership, they find that foreign-owned banks are more cost-efficient and provide better service, in particular if they have a strategic foreign owner, while the remaining government-owned banks are less efficient in providing services. Berger et al. (2009) analyze the efficiency of Chinese banks over 1994-2003, their findings suggest that Big four banks are least efficient; foreign banks are most efficient; and minority foreign ownership is associated with significantly improved efficiency. While Lensink et al. (2008) draw a different conclusion; they use the SFA model for 2095 commercial banks in 105 countries and find that foreign ownership negatively affects bank efficiency.
The e*isting literature has very little empirical study on the relationship between stock return’s volatility and bank efficiency, not to mention some financial ratio’s volatility. In our paper, the empirical application is quite different. We adopt Wang’s (2002) heteroscedastic stochastic frontier model in the estimation, which allows us to specify both the mean and variance of the inefficiency turbulence and investigate the non-monotonic effects on efficiency.
The remaining of the paper is organized as follows. Section 2 introduces the theoretical Wang’s (2002) heteroscedastic stochastic frontier model. Section 3 specifies our empirical model and describes the data in details. Section 4 provides an empirical analysis and e*plains the relationship between volatility and bank efficiency. Section 5 concludes our study.

2.Method
The stochastic frontier approach, independently proposed by Aigner et al. (1977) and Meeusen and van den Broeck (1977), modifies the traditional assumption of deterministic production frontier. Those two works specify a composed error component which equal to the sum of two parts, i.e. a one-sided error that measures the non-negative inefficiency effects and random factors not controlled by the decision-making unit (DMU).
Furthermore, a number of studies e*tend SFA to accommodate panel data. Early e*tended panel data models consider the inefficiency effects as either time-invariant (e.g., Pitt and Lee, 1981; Schmidt and Sickles, 1984) or time-varying (e.g., Battese and Coelli, 1992; Lee and Schmidt, 1993). Moreover, Battese and Coelli (1995) apply single-stage ML procedure to investigate the determinants of inefficiency among DMUs and assume that inefficiency effects are a function of some DMU-specific factors. Recent efforts in modeling heteroscedasticity in inefficiency effects (uit) consider more fle*ible specification in two ways: Kumbhakar et al. (1991), Huang and Liu (1994), and Battese and Coelli (1995) assume the mean of uit (i.e., μit) would differ among DMUs. Caudill et al. (1995) assume μit is constant but allow the variance of uit () to be observation-specific.
More recently, Wang (2002) combines the feature of traditional models and those e*tended models above, and allow both μit and to be observation-specific. Suppose that total costs for the ith bank in year t represents as TCit. Yit and Pit are the input vector and the price of input vector, respectively. The heteroscedasticity stochastic frontier model specification for cost function can be presented as bellow:

where νit is the stochastic error term with i.i.d. normal distribution. Thi ……（未完，全文共30387字，当前仅显示5465字，请阅读下面提示信息。收藏《论文：异方差随机前沿模型对波动性和经营效率的影响研究》）