https://scholars.lib.ntu.edu.tw/handle/123456789/49704
標題: | 臺灣銀行業失敗預警模型之研究 Analyzing the Early Warning of Bank Failure Models in Taiwan |
作者: | 林威助 Lin, Wei-Chu |
關鍵字: | 截斷;規模參數;無參數模型;半參數模型;參數模型;Cox比例危機模型;參數比例危機模型;加速失敗時間模型;censor;scale parameter;non-parametric model;semi-parametric model;parametric model;Cox proportional hazards model;parametric proportional hazards models;accelerated failure-time models | 公開日期: | 2008 | 摘要: | 本研究目的在利用銀行財務指標分析銀行預警模型,探討2002年至2008年間臺灣地區正常與危機銀行的危機函數模型,實證研究以2001年為基準。利用因素分析中主成份分析法選取各種財務指標,將因素透過直交轉軸之變異數最大法,發現預警銀行失敗函數的六項因素,包含營運、規模、資產品質、資本適足、流動性、成長性等因素。實證結果發現存活分析之模型以參數模型配適程度較半參數模型與無參數模型為佳,由加速失敗時間模型迴歸結果發現係數具顯著之解釋能力,且與預期方向一致。最後顯示在Log-logistic模型下,財務指標因素可以正確預測銀行失敗的論證。 The purpose of this research is to analysis the early warning of bank failure models using bank financial indicators and discuss the hazard function of healthy and default banks in Taiwan from 2002 to 2007, the year 2001 is a benchmark year. Factor analysis using Principal components method, and rotating the factor-loading matrix by the varimax method, this research shows that a bank will fail is a function of variables related to six financial indicators, including banking operations , assets quality, bank scale, capital adequacy, liquidity and growing. Empirical results show that the parametric model is superior to the semi-parametric model and the non-parametric model, as the majority of the regressors in the accelerated failure-time model are significant with the correct sign. Overall, the Log-logistic Model is advanced to indicate how the financial indicators can lead to the failure of the banks, and demonstrate good accuracy. |
URI: | http://ntur.lib.ntu.edu.tw//handle/246246/179311 |
顯示於: | 經濟學系 |
檔案 | 描述 | 大小 | 格式 | |
---|---|---|---|---|
ntu-97-P95323008-1.pdf | 23.53 kB | Adobe PDF | 檢視/開啟 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。