再探銀行產業的市場競爭度——東亞八國的實證分析

Re-visit the Market Competition of the Banking Industries in Eight East Asian Countries

本文使用非結構法之PR檢定法，探討東亞八國銀行業於1994年到2008年的亞洲金融風暴前後市場競爭程度。與過去相關文獻主要差異，在於應用切齊隨機邊界模型，檢定各樣本國家銀行產業是否處於長期均衡狀態。此模型的優點為不必依照傳統方式，將所有樣本的應變數——資產報酬率——加1調整成正值，以利再取自然對數後進行迴歸分析，可避免扭曲樣本蘊含之訊息並獲得具備一致性的係數估計式。實證結果發現各樣本國銀行業在亞洲金融風暴前後，皆為獨占性競爭或完全競爭，業者間競爭程度頗高。運用切齊隨機邊界模型，發現六個國家的銀行業達到長期均衡狀態；唯若使用傳統方法，卻出現六個國家銀行業未達到長期均衡狀態，顯示傳統方法容易導致迴歸係數估計式缺乏一致性。

This paper investigates the degree of market competition in the banking industries of eight East Asian countries over the period of 1994-2008, using the ‘H-statistic’ proposed by Panzar and Rosse (1987). Differing from previous works, this paper applies the newly developed model, i.e., the censored stochastic frontier (CSF) model, to test whether these markets under consideration have achieved the long-run equilibrium. The model is superior to the conventional one that requires researchers adding unity to dependent variable ROA for all observations to force the transformed dependent variable having a positive value and being able to take the natural logarithm. Evidence is found that all of the banking industries in our sample are characterized as either monopolistic, or perfect, competition, intead of monoply, or oligopoly. The CSF model confirms that six out of the eight banking markets are already in the long-run equilibrium, which validates the use of Panzar-Rosse model. Conversely, the employment of the conventional approach leads to a rejection of the long-run equilibrium in six out of the eight markets.