逾放比是觀察銀行的經營管理以及金融環境的穩定狀況的關鍵指標,逾放比過高會影響金融穩定與安全。本研究主要在探討國內銀行的銀行特性與總體經濟變數對逾放比的影響。就銀行特性方面,挑選出負債比率、存放比率、存放款利差、每股盈餘、董監事持股比率、經理人持股比率等變數;總體經濟方面則挑選失業率與經濟成長率等變數。由於過往文獻甚少將銀行資產規模設為門檻變數,以探討在不同的資產規模情況下,是否會對逾放比有不同的影響,所以本研究利用Gonza’lez, Terasvirta and Dijk(2004,2005)修正後的縱橫門檻迴歸模型,稱為縱橫平滑移轉迴歸模型(Panel Smooth Transition Regression Model),並加入銀行資產規模做為門檻變數,檢定分析銀行特性與總體經濟變數對企金逾放比、消金逾放比及住宅逾放比的影響。
Non-performing loans ratios are the key indicator of observing banks’ management and financial stability. It will affect financial stability and security when NPLs ratio is too high. This study aimed to investigate the impact of bank-specific and macroeconomic determinants on NPL ratios in Taiwan. For bank-specific determinants, we chose the variables of debt ratio, ratio of loans-to-deposits, deposit and loan spreads, earnings per share, directors and supervisors’ shareholding ratio, managers’ shareholding ratio etc. As for macroeconomic determinants, we chose the variables of unemployment rate and growth rate of GDP. Due to the past literatures seldom set banks’ assets sizes as the threshold variables in order to investigate whether they would have difference impacts on NPLs ratios in the different assets sizes. So this paper used Gonza’lez, Terasvirta and Dijk(2004,2005) modified panel threshold regression model, called the Panel Smooth Transition Regression model and added banks’ assets sizes as the threshold variables to analyze the effects of bank-specific and macroeconomic determinants on business, mortgage and consumer NPLs ratios.