Utilizing Survival Analysis to Predict Duration of Mail Loan Products

• Main Authors: Sheng-Chung Chen, 陳勝忠
• Other Authors: Hongming Huang
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/22622217623210190028

碩士 === 國立中央大學 === 財務金融學系碩士在職專班 === 97 === In the aftermath of the sub-prime crisis, the liquidity risk to financial institutions have become more and more important, Fund transfer pricing is also a major key factor in the current practices of asset and liabilities management. To determine a fair and reasonable pricing for the transfer of funds, each product’s duration must be accurately estimated. This study applies the Proportional Hazard Model to analyze the various factors which may effect the duration of an individual product, and to employ Survival Analysis to estimate the length of the duration. Using 93,929 mail loan transactions from T-bank during 2001 to 2008, the finding reveal that under the Proportional Hazard Model female have a longer duration than male, higher education means a shorter duration, older individuals have a longer duration than their younger peers, higher loan balance means shorter duration, higher loan rate means shorter duration. We also discover that the Extreme distribution resulted in a more accurate duration forecast from the survival analysis.