The Bank of England have released their Staff Working Paper No. 635 which suggests that both before and after the crisis, a 1 percentage point increase in capital requirements lowered annual loan (risk weighted asset) growth by 8 (12) basis points.
This paper has applied a well-established approach to examine the effects of capital requirements on bank capital ratios, lending activity and balance sheet growth. We explicitly consider the impact these requirements have on banks’ targeted, long-run capital ratios and how banks alter lending supply, balance sheet size, asset portfolio risk and regulatory capital levels in attempt to meet such targets. We also examine these effects – and the driving factors – both before and after the crisis as a step toward determining whether bank behaviour changed. Understanding whether and how bank behaviour changed will be important for developing tools aimed at quantifying the effects of capital requirements on bank balance sheet adjustments in general and lending supply in particular.
Using confidential regulatory returns data from 1989 to 2013, we find a positive association between banks’ capital ratios and capital requirements. Our findings suggest that banks reacted to higher requirements by raising capital ratios even when those requirements did not necessarily appear to bind. But banks did not raise capital ratios by the full amount of an increase in requirements. Rather, our results indicate that they tended to raise capital ratios by only about 30% of the increase in requirements within the first six months, and by around 90% of the change in the long-run.
We also find that to achieve these higher ratios, banks, in addition to increasing regulatory capital levels overall, reduced lending, assets and risk-weighted assets. The results, however, suggest that the effects on lending, risk-weighted asset and total asset growth were relatively modest, though we document some difference between pre- and post-crisis adjustments that are worth highlighting. Before the crisis, banks reduced loans, assets, and risk-weighted assets in addition to raising regulatory capital levels in response to an increase in capital requirements. During the crisis, however, they placed greater emphasis on reducing lending. After the crisis, this behaviour changed, with banks placing more emphasis on reducing balance sheet size and raising capital and, in particular, better-quality, tier 1 capital in response to heightened regulatory pressures.
These empirical findings can provide the basis for quantifying the effects of policies that change capital requirements. We illustrate this idea with a simple policy experiment involving the introduction of a series of higher countercyclical capital charges similar in spirit to what has been included in the final Basel III package of regulatory reforms. Simulation results confirm the modest effects on lending growth overall and more clearly delineate the possible structural change in bank behaviour from the pre- to post-crisis periods.
Although the findings depend on a partial equilibrium view of banks’ balance sheet adjustment behaviour and do not take into account feedback effects from the real economy, they are useful for highlighting shortcomings with the previous capital regime and providing evidence on the intended effects of more recent efforts to address these shortcomings. In particular, our results show that banks, in an effort to alter capital ratios before the crisis, tended to focus on raising the cheapest – and least loss absorbent – capital elements to the extent permitted under regulation. That finding confirms concerns about the efficacy of capital requirements for mitigating losses and strengthening the resilience of individual banks and financial stability more broadly. Evidence on post-crisis behaviour, however, suggests that policies designed to modify such behaviour have been somewhat effective, with banks placing greater emphasis on adjusting capital ratios by raising better-quality, tier 1 capital together with less emphasis on altering balance sheet risk. Still, there remains substantial evidence that banks have a preference for first raising lower-quality capital to meet higher capital requirements, which supports ongoing policies aimed at strengthening the quality of capital.
Another implication of our results is that the relatively slow speed (c. 30% in the short-run) at which banks adjust capital ratios suggests that capital requirements may not be the most effective way to strengthen bank resilience quickly. In particular, the long-run adjustment to an increase in capital requirements of around 90% suggests that if a macroprudential policymaker targets say a 4 percentage point increase in capital ratios, it may need to recommend a higher percentage point increase to achieve that goal. This result supports the use of other regulatory tools, such as loan-to-income (LTI) flow limits for residential mortgages, to counteract emerging risks expeditiously.
Finally, while this study’s results provide insights into how banks responded in the past to changes in capital requirements set at the bank level, policymakers will need to use considerable judgment in how to use them in impact assessment. This is because bank behaviour could be different under Basel III, which includes a macroprudential regime. Our results pointing to a structural change in bank behaviour after the end of the crisis support this idea and highlight just how important judgment will play in making inferences about bank balance sheet behaviour going forward.