Risk-return in the banking industry using quantile regression: Evidence from cross-straits banking industry

• Main Authors: Chin-Yuan Lin, 林金源
• Other Authors: Hsiou-Jen Kuo
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/56202602730054609301

博士 === 國立中山大學 === 財務管理學系研究所 === 101 === The enigma of risk-return relationships has long posed problems in the field of banking research. This study employed data related to cross-strait banking to investigate the risk-return relationship between 2005 and 2011.Traditional OLS optimization techniques capture only central behaviors, and misidentify the relationship between bank risk and profitability, including the amount, significance, and even sign; therefore, this study departs from conventional research in the modeling of parameters related to risk-return regression and proposes a novel, conditional quantile regression method (hereafter QR), to survey the dynamics of the relationship between risk and return among banks in Taiwan, Hong Kong, and China. This study employed ROE as a proxy variable for bank returns, using loan/total assets (LO) as a proxy variable for bank risk. Risk-return relationships for banks were analyzed using OLS regression and QR. The study period covered the period of the subprime lending crisis; therefore, data was categorized into two groups: a pre-subprime crisis group and a post-subprime crisis group. Data was also classified into three groups according to LO level: low LO group, middle LO group and high LO group. This enabled the effects of the subprime crisis and the impact of risk exposure to be clearly differentiated. Analysis of OLS regression demonstrated that risk and return among banks in Taiwan were negatively related over the entire study period, the pre-subprime crisis group, the low and the middle LO group. This means that increasing the risk assumed by banks would result in reduced profits for these banks. In addition, our empirical findings demonstrate that the risk-return relationship varied across the quantiles of bank profitability in the three LO ranges, both before and after the subprime crisis. Furthermore, variations in profitability were often the result of the business strategies employed. This indicates that grouping banks with different business strategies to facilitate analysis disregards the impact of business strategy on returns and may be one of the reasons for previous inconsistencies in empirical results. While OLS regression results showed a positive risk-return relationship associated with banks in China and Hong Kong, QR results indicate a positive risk-return relationship in all quantile groups, with the exception of banks of Hong Kong in the upper-quantile of the middle LO group and in the lower-quantile of the high LO group. These results support the theory of a positive risk-return relationship; however, it deviates from the negative risk-return relationship observed in Taiwanese banks. In a comparison of loan quality between banks in Taiwan and those in Hong Kong, based on BDTI-LO relationships we discovered that the negative risk-return relationship in Taiwan could be attributed to poor loan quality. Thus, despite efforts of the banking industry in Taiwan to increase the loan ratio for higher ROE, the widespread issue of poor loan quality remains. If loan quality cannot be improved, the blind pursuit of loan expansion will leave the banking industry in Taiwan susceptible to higher operating risk without improving ROE.