Role of ownership and governance in bank risk and performance: an econometric study
The banking sector is central to the economy, but has recurrent dysfunctions. Following the Global Financial Crisis of 2007-2009, regulators have attempted to reform governance in banks. However, previous empirical studies on the effects of governance structures have important gaps. Using an econometric framework with novel simultaneous equations models and new dependent variables, I investigate whether corporate governance and ownership have significant effects on bank risk and performance. I employ a novel data set combining financial data from the Bankscope database with governance and ownership data collected painstakingly by hand from annual reports and Basel Pillar 3 disclosures of UK banks over the period 2003-2012. My findings are supported by interpretation of relevant literature and are summarised as follows (stated along with policy implications in parentheses for which features of banking should be encouraged, based on normative assumptions stated in section 9.3). My work shows that the effects of a particular ownership or governance structure can be attributed to the ways in which categories of decision-maker within the bank are empowered by that structure, and that factors relating to information processing capability have important effects. Mutual and foreign ownership each have negative effects on risk and return because of managerial incentives and information asymmetries, respectively, without either affecting provision of investment to the wider economy. A foreign parent also increases the probability of bank failure (implying mutuality is socially beneficial while foreign ownership is not). A higher NED ratio reduces the probability of bank failure, as does having a remuneration committee, because of greater accounting for risk in decisions (implying they are desirable). The presence of an independent Chairman increases risk because it weakens CEO accountability and confuses decision-making (implying it is undesirable). An independent CRO (as a full Board member) may have similar effects. A higher proportion of Directors with no previous financial services experience increases both returns and the probability of failure because of weaker use of information (implying it is undesirable). Permission to use IRB models lowers risk and return because it provides information to empower risk-averse agents, again without affecting credit provision to the wider economy (implying it is desirable). I report other novel findings on effects of ownership, governance, remuneration and size. These results can guide bank reform.