A better way to measure bank risk

| Article

In response to the global banking crisis, regulators and policy makers worldwide have united behind efforts to increase financial institutions’ minimum capital requirements and to limit leverage, hoping to reduce the likelihood of future bank distress.1 As of this writing, the debate over proper capital requirements continues, with major implications for the industry and the economy—yet there have been few specifics on which ratios should be targeted or at what levels.

To shed some light on the discussions, we analyzed the global banking crisis of 2007 through 20092 to identify relationships that different types of capital and capital ratios have to bank distress.3 Our analysis is observational, based on historical data, and not a real-world experiment, which would have required randomly selected financial institutions to hold different capital levels to gauge their effects. As a result, the findings do not definitively establish how institutions might perform in the future if minimum capital ratios were changed, but we believe that the evidence we provide is a valuable input for current policy discussions.

We found that one capital ratio—the ratio of tangible common equity (TCE)4 to risk-weighted assets—outperforms all others as a predictor of future bank distress. We also found that requiring a minimum leverage ratio would not have offered any insights that couldn’t have been found by studying the right capital ratio. And, not surprising, we found that a higher bar on capital requirements, while reducing the likelihood of bank distress, comes at an increasing cost.

One capital ratio outperforms the rest

Among the various ratios, the one that offers the greatest clarity into the likelihood of bank distress actually measures TCE (the portion of equity that is neither preferred equity nor intangible assets) against risk-weighted assets, or RWA (Exhibit 1). TCE, like Tier 15 capital, can absorb losses because it offers banks the contractual flexibility either to eliminate repayments entirely or to defer them for extended periods of time. It can also absorb losses whether or not a bank remains a going concern. Moreover, our analysis found that the measures most commonly regulated currently—those based on the combined Tier 1 plus Tier 26 capital levels—are the least useful, in part because banks can seldom use Tier 2 capital to absorb a loss if they are to continue operating. For example, unrealized gains on securities may be unavailable in times of severe economic stress, and subordinated debt may trigger default if payments are deferred.

Exhibit 1
From the analysis

In addition, banks have successfully arbitraged capital ratios traditionally watched by regulators through the banks’ increasing use of non-common-equity instruments, such as cumulative preferred stock and trust-preferred securities, that qualify for treatment as Tier 1 capital but could be issued at lower cost than common equity. This practice weakens the ability of an institution to absorb losses and the ability of regulations to limit its riskiness.

Leverage ratios add little benefit

Our analysis also found that an additional leverage ratio would not have offered any insight into the likelihood of bank distress beyond that provided by the TCE/RWA ratio. The same number of banks are affected (and the same amount of distress avoided) whether or not limits are placed on leverage.

This finding does not prove that regulating leverage ratios is a bad idea. It does suggest, however, that the rationale must be based on other considerations. For example, leverage ratios might protect the liability side of the balance sheet against greater-than-expected haircuts on repurchase (or repo) financing, which could precipitate a systemic crisis. They also might help prevent future errors in risk weighting and regulatory arbitrage of risk weightings. But the use of leverage ratios has also arguably created an incentive for the growth of off-balance-sheet activities, which remove certain assets from the leverage ratio calculation and increase risk while circumventing additional capital requirements.

Lowering risk has a cost

While it is possible to lower a bank’s level of risk by increasing its TCE/RWA ratio, the trade-off is higher costs. Reducing the number of banks at risk through a higher capital base decreases the returns on equity (ROE) for the industry (Exhibit 2). For instance, a TCE/RWA ratio of 10 percent would have affected all of the banks that became distressed during the recent crisis but would have required an incremental $1.45 trillion in capital7 and reduced industry-wide average ROEs by an extraordinarily high 560 basis points. In addition to the impact on ROEs, increasing the required capital levels would likely have macroeconomic costs, including the effects of a short-term contraction in the availability of credit and the potential long-term effects of reduced lending levels, which result in lower GDP growth.8

Exhibit 2
Costly security

One test for regulators is wisely balancing the incremental benefits of higher capital requirements against the costs that they impose on financial institutions, borrowers, and society more broadly. For example, our analysis indicates that requiring banks to hold a TCE/RWA ratio in the range of 6.5 to 7.5 percent would have affected 83 percent of banks that became distressed while requiring $540 billion in incremental capital and a decrease in ROE of 260 basis points.

In the effort to prevent future banking crises, regulators would do well to set minimum capital requirements by balancing the benefits of reduced distress with the costs that come from higher capital requirements.

Explore a career with us