Three Essays on Banking Concentration

Abstract

Banks warrant special attention because of the key role they play in providing liquidity to the market, transforming assets, managing risks, and monitoring borrowers. Over the past few decades, the US banking system consolidated considerably which resulted in a more concentrated system where the majority of assets are controlled by a few excessively large institutions. This dissertation examines concentration of the US banking sector and its relationship with the real economy, idiosyncratic bank stability, and financial market volatility. Chapter 1 investigates the association between banking concentration and the real economy through the bank failures channel. To this end, we build a system of equations that estimates the association between banking concentration and the real economy by employing quarterly U.S. data from 1984 through 2013. The first equation tests the association between bank concentration and the rate of bank failure using an autoregressive Poisson model which allows for more accurate estimates than linear models. The remaining three equations model respectively, real GDP growth, unemployment, and inflation as functions of the rate of bank failure. Three interesting results are obtained. First, there is a threshold below which increasing concentration causes a reduction in bank failures and above which an increase in failures. Second, as bank failures increase, economic growth slows while unemployment and inflation both increase. Third, our results imply that the U.S. banking system is more than twice as concentrated as the optimal level as determined by the minimum rate of bank failure and is having a detrimental effect on the real economy. Our results suggest that while the Dodd-Frank Act of 2010 introduced legislation aimed in part at restricting the level of banking concentration, additional reductions in concentration may be necessary to strengthen the economy. Chapter 2 investigates the association between banking concentration and idiosyncratic bank stability after the passage of the Riegle-Neal Interstate Banking and Branching Efficiency Act of 1994 into legislation. First, we model individual bank stability as a non-linear, as opposed to a linear, function of banking concentration allowing us to determine if rising concentration increases (decreases) bank stability up to a certain point and decreases (increases) it thereafter. Second, we differentiate between large and small banks by introducing an interaction term between concentration and bank size allowing us to determine if size-based differences alter the concentration-stability relationship. Third, we employ a fixed effects instrumental variable model and correct for reverse causality between bank stability and bank concentration. Our findings indicate that large and small banks react very differently to changes in concentration. As concentration exceeds a certain threshold, small banks become less stable, hold less capital, are less profitable, and hold more volatile portfolios. The results are the reverse for large banks. We also find that as concentration increases, large banks increasingly contribute to systemic risk, despite the fact that their idiosyncratic risk is reduced. Chapter 3 investigates the association between financial market volatility and banking concentration. Research on this relationship has been sparse and remains ambiguous. A main difficulty with achieving this task is the low frequency (quarterly) nature of the concentration data relative to the high frequency (daily) volatility data. To overcome this problem, we employ a GARCH-MIDAS volatility model which allows us to test the relationship between data with dissimilar frequencies. We consider the sample period 1986:1 to 2013:4. Our results indicate that higher levels of banking concentration are positively associated with higher volatility in the US stock, options, and corporate bond markets and negatively associated with the US government bond volatility. These finding fill a major void in the literature and have implications for regulators and policy makers.