台灣的銀行經營效率之研究

The Study on Performance Efficiency of Taiwan’s Bank

彭家瑞(Chia-Jui Peng)

指導教授 ： 李正文

中原大學/商學院/商學博士學位學程/博士(2015年)

https://doi.org/10.6840/CYCU.2015.00179

摘要

銀行業有別一般的製造業性質，銀行業經營業務範圍及經營條件是受限制於政府法令規範，營運獲利的主要來源深受執政當局的利率及財稅政策的影響。近年來世界各國金融業互動愈趨連動，由一國的不動產不良債權引發了全球金融危機，已凸顯了這種不愉快的事實。雖然台灣金融體系未受到全球金融風暴太大影響，但展望未來金融業競爭力發展趨勢，政府部門應掌握銀行業經營效率的變化，確使金融市場的中介角色運作正常。
本研究採用超級效率來分析，以差額變數為基礎的衡量模型、縱橫資料模型、非線性縱橫平滑轉換模型等三種主要研究方法，建構不同理論模型以三篇文章方式來分析台灣銀行業整體經營效率值及財務面持續性變化的情形。其中，縱橫資料模型及非線性縱橫平滑轉換模型，可以確實掌握銀行業經營績效動態持續性的程度，當經營績效的持續性愈高時，代表整體銀行業的經營模式具有高度的穩定性，並證明波動取決於政策面代理變數的門檻變量，另外，當門檻值以平滑移轉的方式表現，可以解決以往研究門檻值跳躍的不合理性現象。因此，其結果更能真實反映政策執行上的真實情況，對執政當局及銀行管理階層在擬定政策方向時，可就影響層面上提供精準的參考依據。本研究發現如下:
1.台灣銀行業整體效率值具有跨期或多期遞延的動態持續性效果，各效率值持續性平均值達89.39%，證明台灣銀行業在經營上具有相當穩定的成長性，而且各銀行在以平緩的速度調整經營規模及方式時，經營效率值變化將相對於若干滯後期表現出來，亦即其經營效率即具有一定的動態持續性。若以其它外生及環境解釋變數來說明各效率值的變化原因，其最高的解釋能力極低(10.61%)，因此不適用此研究。
2.以縱橫平滑轉換模型分析台灣銀行業經營效率具有動態非線性的持續性，導入了代表貨幣政策中的銀行存放款利差為轉換變數，探討在銀行利差下的門檻效果。這評估基礎是在PSTR模型下進行，這模型可以解釋影響銀行經營績效在政府貨幣政策下的改變情形，另一方面能夠解決潛在的解釋時間中的異質性及對持續性影響程度。實證結果顯示台灣銀行業在經營效率值CRS和VRS存在持續性的門檻效果變化，且銀行業在經營效率值持續性存在非線性的變動，在不同的區間隨時間而變動。
3.在探討台灣不動產景氣對銀行財務績效之影響，以抑制房價的財政政策替代變數房價綜合趨勢分數及貨幣政策的替代變數貸款負擔率為轉換變數，評估基礎在PSTR模型下進行。在房價位於高檔區間時，會創造銀行財務績效有較高的持續性，反之，在房價呈現向下趨勢時，銀行的財務績效則會受其它外生因素的干擾，財務績效持續性則不明顯，此結果說明調整稅賦的財政政策，確實會影響銀行獲利。其次，當轉換變數為貸款負擔率時，銀行財務績效幾乎完全受外生因素影響，本身財務績效無遞延效果，意謂當政府降低房貸成數、調降利率及大量減少購屋意願時，銀行獲利持續性呈現高度不穩定現象。在本研究實證中，門檻值可以做為政策執行面的警示指標。
綜而言之，以往文獻對企業經營的效率或財務績效，較多從系統風險(匯率、利率、物價、景氣循環等)進行分析，忽略或無法完整捕捉到企業本身的非系統風險(匯兌損益、經營模式、研發進程等)，且無法清楚說明政府政策面的影響性。因此，藉由本研究的整體估計概念，將有效評估出銀行業持續性變化情形，並可掌握到政府政策面的對銀行業的效率及財務績效的影響程度。

關鍵字

動態效率持續性 ； 縱橫平滑轉換模型 ； DEA超級差額變數效率模型 ； 縱橫資料模型

並列摘要

Financial industries are different from traditional manufacturing industries. The operative businesses of the former are limited by government regulation, and the source of profitability is strongly affected by the fiscal policy and interest rate set by the authorities. Recently, global financial industries have interacted closely; bad debts from the real estate industry worsened the financial crisis worldwide and clearly highlighted the unfortunate situation. Although Taiwan’s financial system was not significantly affected by this crisis, the government should control changes in the operative performance of banks to establish their roles as financial intermediates.
This study adopts super efficiency based on the slack measurement model and the smooth transition regression model for panel and nonlinear panel data. Three essays are constructed with regard to the different models to analyze the persistent financial changes in the overall operative efficiency value of banks. The smooth transition regression models for panel and nonlinear panel data can capture the operative performance levels of banks in terms of dynamic persistence. A high persistence in operative performance implies that the operative model of the bank is highly stable. In addition, the use of the threshold of policy as a proxy variable to determine volatility is confirmed. The smooth transition behavior of the threshold can explain the arbitrary situation of threshold jumping in regimes, as detailed in previous literature. Hence, our results can reflect the actual situation for policy execution, as well as serve as a reference and provide suggestions for the policy-making process. This paper contributes to the followings.
1. The empirical result indicates that the overall efficiency value exerts cross-period or multi-lagged effects of dynamic persistence on banks. Furthermore, the average efficiency value of persistence reaches 89.39%; this finding proves that banks in Taiwan display stable growth rate in operation. If the operative scale and method of individual banks are smoothly adjusted to match, then the operative efficiency value can change following the lagged period. Therefore, operative efficiency exists in a specific persistence. If we explain the efficiency change through exogenous variables, then the explanation power reaches only 10.61%, which is insufficient for reference or analysis.
2. We have proven that the efficiency value of banks is determined under either cross-period or multi-lagged dynamic effects; thus, we observe the changes in the efficiency persistence of banks in Taiwan. Upon employing interest rate differential as the proxy variable and monetary policy as the transition variable, we investigate the threshold effects under interest rate differential. An estimation is made based on the PSTR model to explain the change in the operative performance of banks under monetary policy; the model can resolve heterogeneity and understand the level of influence of performance persistence. Empirical results suggest that the operative efficiencies of the CRS and the VRS of banks are detected in the threshold effects on persistence; these efficiencies change in nonlinearity, and vary with time in different regimes.
3. We examine the business cycle of the real estate industry, which affects the financial persistence of banks in Taiwan, to analyze the effects of anti-housing-price policy on the financial performance persistence of banks. We adopt housing price trend score as the proxy variable for fiscal policy and housing affordability as monetary policy (the transition variable) to estimate the effect under the PSTR model.
First, empirical results indicate that when housing price trend score acts as a transition variable, the threshold value is 124.4871 and financial performance persistence fluctuates between 0.2363 and 0.5972. When housing price is at a high threshold level, financial performance enhances persistence. On the contrary, if housing price exhibits a downward trend, then the performance of banks can be affected by exogenous factors. The persistence of banks becomes insignificant. This result illustrates that the fiscal policy of tax adjustment affects bank profit. Moreover, strict anti-housing-price policies implemented by the government can influence the persistence of profit gain in banks.
Second, the threshold is 28.5110 and persistence varies between 0.0369 and 0.4415 when housing affordability is adopted as a transition variable. Therefore, financial persistence is almost unaffected by exogenous factors when housing affordability value is lower than 28.5110. Financial performance does not depend on the persistence effect; thus, interest and willingness to purchase a house decrease when the government reduces the ratio of bank loans. The profitability of banks can be highly destabilized. Hence, anti-housing-price policy must be implemented smoothly; a sudden application can lower persistence in the financial performance of banks. As a result, the financial market is destabilized and indirectly stunts economic growth. This conclusion can serve as a reference for the government, and the researched threshold can act as an alarm indicator in policymaking.
Finally, traditional investigations into operative efficiency or financial performance mainly analyze system risk factors (e.g., exchange rate, interest rate, price, and business cycle). Such examinations either ignore or cannot capture the non-system risks (e.g., gains or losses in exchange rate, operative model, and R&D process) of an enterprise. Moreover, these works cannot determine the effects of policies. Therefore, the current study develops the concept of top-down estimation to assess the persistence change in the performance of banks and determine the effect of government policies on efficiency and financial performance.

並列關鍵字

panel data ； DEA super slack variable efficiency model ； panel smooth transition regression ； dynamic efficiency persistence

參考文獻

Altunbas, Y., Liu, M. H., Molyneux, P. and Seth, R. (2000), “Efficiency andrisk in Japanese banking”, Journal of Banking and Finance, 24, pp.1605–1628.

Andersen, P. and Petersen, N. C. (1993), “A Procedure for ranking efficient units in data Envelopment Analysis”, Management Science, 39(10), , pp.1261-1264.

Arshadi, N. and Lawrence, E. C. (1987), “An empirical investigation of new bank performance” , Journal of Banking and Finance, 11, pp.33-48.

Barr, R.S., Seiford, L.M. and Siems, T.F. (1993), “An envelopment-analysis approach to measuring the managerial efficiency of hanks”, Annals of Operations Research, 45, pp.1-19.

Barth, J.R., Lin, C., Ma, Y., Seade, J. and Song, F. M. (2013) ,“Do bank regulation, supervision and monitoring enhance or impede bank efficiency? ”, Journal of Banking and Finance, 37, pp.2879-2892.