本研究共計選取我國全體銀行產業中之28家銀行，採用縱橫平滑移轉迴歸模型(Panel Smooth Transition Regression Model, PSTR)，進行各解釋變數受逾放比高低變化之非線性影響，並進一步對銀行經營績效受控制變數之影響進行評估與衡量。研究結果顯示，逾放比和銀行經營績效ROA、ROE間確實存在非線性關係。 以ROA為績效代理變數情況下，低逾放比銀行(逾放比低於門檻值)之資本適足率、負債淨值比及存放比與ROA呈顯著負相關，而備抵呆帳覆蓋率對ROA無顯著影響；高逾放比之銀行(逾放比高於門檻值)其備抵呆帳覆蓋率及負債淨值比與ROA呈顯著負相關，而資本適足率及存放比對ROA無顯著影響。同理，以ROE為績效代理變數時，低逾放比銀行(逾放比低於門檻值)其負債淨值比及存放比對ROE則有負相關，而備抵呆帳覆蓋率及資本適足率對ROE無顯著影響；高逾放比率之銀行(逾放比高於門檻值)其備抵呆帳覆蓋率及資本適足率與ROE呈顯著負相關，存放比則與ROE呈顯著正相關，而負債淨值比對ROE無顯著影響。 This research adopts twenty-eight banks in nation’s entire financial industry, utilizing Panel Smooth Transition Regression Model, PSTR to examine independent variable’s nonlinear effect on Overdue-ratio’s volatility, and further on assessing, evaluating the effect on banks’ performance. This research determines that it does exist in nonlinear relation between Overdue Ratio, bank’s performance Return on Asset (ROA) and Return on Equity (ROE). ROA as variable, low overdue-ratio bank’s bank of international settlement ratio, debt/equity ratio and loan-to-deposit ratio have negative relation on ROA; high overdue-ratio bank’s coverage ratio and debt/equity ratio appears to have negative relation on ROA. In a similar way, employing ROE as the variable, low overdue-ratio bank’s bank of international settlement ratio and its coverage ratio have no significant effect, debt/ bank of international settlement ratio and loan-to-deposit ratio have negative relation on ROE; high overdue-ratio bank’s coverage ratio and debt/equity ratio increase without positive advantage toward ROE.
