The impact of board structure on bank loan herding via mediation of underperformance

ABSTRACT

When weak governance practices cause poor bank lending performance, banks are more likely to engage in ‘loan herding’ to avoid a sustained performance deterioration. Using the three main types of Chinese banks as our sample, this study first confirms the existence of loan herding and the positive effects of a weak board structure on poor lending performance (i.e. underperformance). Then, after finding positive impacts of this underperformance on loan herding, we examine whether board structure variables negatively affect loan herding. We find that higher director compensation, a moderate average age of directors, and higher director education levels significantly reduce bank loan herding, whereas board duality and the presence of very old directors significantly increase bank loan herding. Finally, we study the mediation effect of loan performance on the relationship between board structure and loan herding. Our results show that board governance quality affects loan performance, which, in turn, affects loan herding.

KEYWORDS:

• Board structure
• Chinese banks
• corporate governance
• herding behavior
• lending

Notes

1. The five SOCBs are the Agricultural Bank of China, Bank of Communications, Industrial and Commercial Bank of China, China Construction Bank and Bank of China. The twelve JSCBs are Ping An Bank, Shanghai Pudong Development Bank, Hua Xia Bank, China Minsheng Bank, China Merchants Bank, Industrial Bank, China Everbright Bank, China Citic Bank, China Guangfa Bank, China Zheshang Bank, Hengfeng Bank and China Bohai Bank. The seven CCBs are the Bank of Ningbo, Bank of Nanjing, Bank of Beijing, Bank of Shanghai, Bank of Jiangsu, Bank of Hangzhou and Bank of Changsha.

2. We subtract this term to reduce the problem, with more variation in the increasing loan ratios for industries with only a few loans.

3. j is the number of industries in $\varphi$, and k is the number of regressors. ${\chi }^2\left(j-k\right)$ is the chi-squared distribution with j − k degrees of freedom.

4. Because many studies suggest that the effects of governance variables and control variables on performance are set during the same period, we use governance variables for the same period in our analysis (Andres and Vallelado 2008; Grove et al. 2011; Liang, Xu, and Jiraporn 2013).

5. To conserve space, we do not show the results of the correlation analysis between board structure variables and loan herding.