Apply Vector Autoregression Model to Construct an Industrial Credit Risk Model with Macroeconomics and Industrial Factors

• Main Authors: LIU, CHENG, 劉誠
• Other Authors: LEE, MENG-FONG
• Format: Others
• Language: zh-TW
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/hk9y6n

碩士 === 國立臺北大學 === 統計學系 === 105 === The main risk in the financial sector is credit, market and operational risk. There are 60% of credit risk, 30% of operational risk, and 5% of market risk and other risks. Credit risk is the major risk of a bank. It is also an important reason caused a bank to produce overdue loan losses. In the past, banks only paid attention to the financial statements and credit information of the enterprises for the credit risk, and the risk assessment of the enterprises reduced to formality. The risk assessment was only focused on chatting and subjective description and lacked of objective or quantify information to forecast the risk. Therefore, this study considers quantitative information on macroeconomic and industrial risks to establish an industrial risk assessment model in order to reduce the credit default rate effectively. Taking the electronics industry as an example, the overall default rates of listed companies of electronic industry in Taiwan from 2002 to 2015 are adopted as of this study. Referring to past studies, 25 variables related to macroeconomic and industrial factors are collected for model building. There are 17 significant variables related to industry default rates selected by stepwise regression. Next, principal component analysis is applied to obtain two principle components called macroeconomic index and industry index. A vector autoregression model was established for default rates, macroeconomic index, and industry index to construct an industrial credit risk default prediction model.