CEO overconfidence in the U.S. banking industry

The financial crisis that started in 2008 has had a significant impact around the global. In the US alone, it is estimated that approximately $14 trillion of wealth has been lost (Luhby 2009). A considerable body of literature has focused on the role of excessive risk-taking by banks in the years preceding the crisis (Laeven & Levine 2009; Fouque & Langsam 2013; Weiß et al. 2014). We still have limited knowledge, however, about why some financial institutions take more excessive risks than others, despite the fact that these financial institutions are subject to the same financial regulations and macroeconomic environment (Lo 2011; Barone-Adesi et al. 2013; Xiong 2013). This concern about excessive risk taking by banks has encouraged scholars to study the possible causes. A common argument is that the subprime and financial crisis was not due to regulatory failure, but psychological failure (Shefrin 2009; Barberis 2011). If this is the case, it would suggest that any explanation of the causes of the recent banking crisis that focuses only on bank characteristics, failure of financial regulations and market discipline may be incomplete. First of all I find that managerial overconfidence could influence the language style and tone, tested by DICTION, of the Management Discussion and Analysis (MD&A) section of annual reports, as well as the whole 10-K report. The results tentatively suggest that the MD&A contains vocabulary that can be used to capture CEO overconfidence. Secondly, using a sample of 189 U.S bank holding companies involving 261 CEOs from 1994 to 2014, I examine whether CEO overconfidence affects BHCs’ risk-taking choices, and its subsequent systemic risk effects. The results suggest that managerial overconfidence is positively associated with systemic risk. I also find that mortgage-backed securities and provisioning for loan-losses can serve as the drivers of systemic instability. Finally, I examine the impact of CEO overconfidence on bank diversification. The results suggest that CEO overconfidence does appear significantly to influence the degree of non-traditional banking revenue and asset diversification. The results also indicate that overconfident CEOs tend to be detrimental for bank performance when the degree of business complexity increase. BHCs with overconfident CEOs, and with a higher degree of asset diversification, are associated with higher systemic risk as measured by MES. In contrast, BHCs with overconfident CEOs can curtail tail risk exposures if these BHCs are geared towards a higher degree of revenue diversification