How Do Capital Requirements Impact Banking Sector Risk-Taking and Market Shares of Big and Small Banks?

During the financial crisis of 2008, numerous banks failed, and many received government support. In response to this crisis, Congress passed the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010 to strengthen the banking sector and thwart future financial crises. This policy change tightened capital and liquidity requirements in order to deter risky bank lending, help banks respond to future funding shocks, and support banks in maintaining a sufficiently high charter value.¹

In their paper "Capital Buffers in a Quantitative Model of Banking Industry Dynamics," Philadelphia Fed economist Pablo D'Erasmo and Philadelphia Fed visiting scholar Dean Corbae of the University of Wisconsin-Madison and the National Bureau of Economic Research (NBER) study how much the Dodd-Frank Act has impacted bank exit rates, patterns of lending, loan interest rates levels, and the market shares of big and small banks. They do this by developing a model with imperfect competition-that is, where a set of large banks that have market power interact with a wide set of competitive banks. More generally, they examine (i) how regulatory policies affect banking sector stability and market structure, and (ii) the feedback effects of market structure on the efficacy of policies.

The authors analyze differences in large- and small-bank behavior. Large banks, they document, generally have smaller capital buffers than do small banks. (That is, large banks have less capital in excess of what is required by regulation.) This is primarily because large banks have more diverse funding sources, which leave them less susceptible to disruptions in funding (that is, to liquidity shocks). Furthermore, Corbae and D'Erasmo report, virtually all of the entry and exit from the U.S. banking sector has occurred among small banks. These "fringe" banks tend to enter the sector during expansions of the gross domestic product and exit during economic downturns.

Over time, the banking sector has become increasingly concentrated. The number of U.S. banks fell from 11,000 in 1984 to under 5,000 in 2016. Meanwhile, the market share (as measured by assets) of the top 10 banks rose from 27 percent in 1984 to 58 percent in 2016. This translates into an approximate 50/50 asset split between the 10 largest banks and the rest of the sector. The rising concentration motivated the authors to construct a model of the U.S. banking sector with imperfect competition, whereby some banks are dominant and have market power.

Corbae and D'Erasmo use their model of banking industry dynamics to estimate the impact of regulatory policies on major characteristics of the banking sector. They employ U.S. banking system data obtained from Call Reports. (Insured banks must submit these data quarterly to the Federal Reserve.) Their model incorporates frictions in the banking sector, including equity issuance costs, bankruptcy costs, and the existence of government deposit insurance and limited liability.

The authors' model predictions are consistent with existing features of the U.S. business cycle, bank lending concentration, and market stability. Moreover, they use their model to conduct policy counterfactuals based on the Dodd-Frank Act, considering imperfectly competitive loan markets. Under this act, bank capital requirements rose from 4 percent (following the Basel II rules) to 8.5 percent (following Basel III).² The authors show that a rise in capital requirements had the intended policy effect of decreasing the bank exit rate over the long run. At the same time, a higher capital requirement reduced bank entry, leading to a more concentrated banking sector. Further, the higher capital requirement incentivized big banks to strategically work over the short run to gain loan market share (which they financed by cutting dividends, retaining more earnings, and issuing more equity). The authors find that the net effect of more large-bank lending and less small-bank lending was a 7 percent decline in overall lending in the short run and a 9 percent decline in the long run, while the rise in the capital requirement led to only a modest rise in the interest rate on loans in the long run.³

The authors also examine how market structure affects the predictions of policy experiments. They compared a model with perfect competition to a baseline with imperfect competition in the loan market. In response to an increase in the capital requirement, they find, bank entry and exit rates fall much more in the long run in an imperfectly competitive market than in a perfectly competitive market. Thus, under imperfect competition, there is a lower likelihood of a banking sector crisis. The increase in capital requirements also translates into modest welfare gains in the long run in the model with imperfect competition. In contrast, under perfect competition, they find, the change in capital requirements leads to an unintentional rise in the likelihood of a banking crisis and modest welfare losses in the long run. The authors explain the different results by pointing out that under imperfect competition, large banks can increase lending and deter weaker small banks from entering the sector, suggesting that, in this case, "more competition leads to more fragility."

In sum, using their dynamic model of the U.S. banking industry, Corbae and D'Erasmo find that regulatory policies significantly impact the banking sector's market structure, including the market power of large banks. In turn, these regulatory policies improve market efficiency and stability in the banking sector. Furthermore, their research suggests that market structure affects the efficacy of policy, indicating a potential tradeoff between the level of banking sector competition and bank stability.⁴