In a recent paper, Natasha Sarin and Lawrence Summers ask the provocative question “Have big banks gotten safer?” The authors point out that by several capital markets-based measures, including stock price volatility and CDS spreads, banks appear to be riskier now than they were before the crisis, even as bank capital and liquidity standards have been substantially raised over that same period of time. Sarin and Summers conclude that the most likely explanation is that banks’ franchise values have declined. Overall, we think the paper makes a compelling case and raises some important questions.
Highlighting the Decline in Market-to-book Ratios in the Post-crisis Period
Sarin and Summers use several financial market indicators for the top 6 U.S. banks as well as the largest foreign banks to show that measures of volatility and expected returns appear to be higher in the post-crisis period relative to pre-crisis levels.1 To explain their empirical findings the authors provide 3 possible explanations: (i) Markets underestimated the risks banks were taking in the pre-crisis period; (ii) measures of regulatory capital have severe shortcomings; or (iii) the franchise value of banks has substantially declined in the post-crisis period. The paper concludes that its empirical findings can be explained with an increase in market leverage — the ratio of the market value of equity to assets — of banks in the post-crisis period due to a decrease in the franchise value of banks.
Consistent with the decline in banks’ franchise values, market leverage for the largest U.S. banks has declined in the post-crisis period as shown by the red line in Exhibit 1. Given that bank equity risk increases as the ratio of market equity to assets declines, this result is consistent with the paper’s finding that bank equity risk has risen. Meanwhile, as shown by the blue line in Exhibit 1, the higher capital requirements introduced with Basel III and Dodd-Frank have themselves led to a sizable increase in the book leverage for the largest U.S. banks.2
One important question the paper raises, and that has been overlooked in its media coverage, is why market-to-book ratios declined in the post-crisis period. Market value depends on the ability to generate earnings and increase those earnings over time. The tightening of regulations that has occurred since the crisis, while increasing loss absorbency, has also reduced the profitability of banks. As shown in the left panel of Exhibit 2, return-on-equity at large banks remains well below levels observed prior to the start of the past financial crisis.3 The low level of interest rates and flat yield curve are also holding down bank earnings and the prospect for future earnings. Indeed, as shown in the right panel of Exhibit 2, large banks’ net interest margins are hovering around their lowest point over the past two decades. Moreover, the low interest rate environment is also damping the earning prospects of mid-sized banks, perhaps even more than those of large banks, and the increased liquidity requirements will likely lead to higher deposit rates going forward, which negatively impacts mid-sized banks deposit franchise values. These factors likely explain why Sarin and Summers’ empirical results are the strongest for the smallest institutions in their sample.
Market value also depends on the discount factor applied to those earnings. The period prior to the recent financial crisis was the end of a decade-long episode referred to the “Great Moderation.” All investors have increased the odds they place on periods of heightened volatility, recession, and defaults. In addition, as recently emphasized by Mark Carney, the Governor of the Bank of England, uncertainty about future regulations can add substantially to the riskiness of bank earnings. For both reasons, risk premiums have risen accordingly, not only reducing bank equity values further but also contributing to the widening of spreads on bank debt reported by the authors.
Finally, predictability in earnings is valued by investors. With respect to large banks, however, investors currently cannot know what future capital requirements will be, and have good reason to believe they will be higher. Not only are new capital standards still being introduced but also for those banks subject to the U.S. stress tests, capital requirements effectively change every year. The resulting uncertainty and pessimism likely dampens the outlook for further earnings growth.
In sum, we agree with the authors that
…it appears plausible that a large part of the reason for declines in franchise value is regulatory activity and the prospect of future regulation. There is a possibility that by further eroding bank franchise value, further regulatory actions could actually increase systemic risk. p. 34
Indeed, we think it is critical that the costs and risks from additional regulations owing to reduced franchise value be taken into account when assessing whether a new regulation should be adopted.
In summary, banks’ strength and social value both derive from banks’ ability to take appropriate risks. Banks can be made completely safe by requiring them to fund themselves with long-term debt and equity and hold nothing but Treasury bills, but such a strategy would neither allow the banks to perform the socially valuable maturity and liquidity transformation associated with taking deposits and making loans, nor enable the banks to make sufficient earnings to survive over time.
Safer for Whom?
Another point that some discussions of the paper have missed is that its finding that risk is steady or increasing is not inconsistent with taxpayers now being at reduced risk from the prospect of bank failure. Given its emphasis on market measures of risk, the paper, in asking “Have big banks gotten safer?” necessarily appears to be asking whether big banks have gotten safer for investors in their equity and long-term debt.4 It seems clear that the allocation of bank risk has changed significantly through post-crisis regulatory reform, which has been focused on ensuring that the costs of any defaults will be borne solely by investors in the equity and long-term debt of banks. Thus, holding risk constant, the waterfall of losses has changed materially for U.S. (and many foreign) banks.
Most notably, holding company shareholders and long-term creditors are now in a clear first-loss position, and legal reforms in how resolution is conducted make it more likely that they will suffer loss rather than being bailed out; conversely, short-term creditors at the operating company level (including depositors) are now clearly senior to holding company creditors, and thus less likely to suffer loss. Thus, not only have credit ratings agencies downgraded holding company debt to reflect an absence of government support but they have also upgraded operating company debt to reflect its improved position in the waterfall. Furthermore, significant increases in bank liquidity to some extent reduce the chance of default but also make an orderly failure more likely, with greater ability to inflict losses on creditors. Thus, the increased volatility of, and greater risk premiums applied to, holding company equity and debt quite possibly indicate that reforms are working as intended. As noted, the paper does not address this issue, but we highlight it here, as some of the reporting on the paper appears to erroneously assume that its finding extend to the question of whether banks have gotten more or less safe to taxpayers.
Predicting Bank Failure
Lastly, while the finding is not central to the paper, we also note that the paper incorrectly assumes that simple measures of leverage are better predictors of bank failure than risk-adjusted regulatory capital ratios. However, a recent TCH research note shows that risk-based measures of capital adequacy, such as the tier 1 capital risk-based ratio, were better predictors of bank failure than the tier 1 leverage ratio during the past financial crisis. In particular, banks that survived the crisis reported a tier 1 risk-based ratio that was about 30 percent higher than banks that failed. In contrast, the leverage ratio of banks that survived was only slightly higher than the leverage ratio of banks that failed. In fact, approximately one-third of the banks that failed during the recent financial crisis had leverage ratios above 10 percent just prior to the crisis.5
1Sarin and Summers also expand their analysis to the top fifty financial institutions in the United States.
2 Book leverage is defined as common equity tier 1 capital divided by risk-weighted assets. The increase in book leverage is similar if average tangible assets are used instead of risk-weighted assets. For comparability with Sarin and Summers, the largest banks include the top 6 U.S. banks.
3 Large banks are defined as the top 33 bank holding companies each quarter and it is meant to be a proxy for bank holdings companies above $50 billion and subject to the U.S. stress tests.
4 Specifically, Sarin and Summers evaluate both bank equity and publicly traded debt, but for U.S. banks, the latter consists almost exclusively consisting of long-term debt issued by the holding company.
5 Moreover, using market measures of risk to assess capital adequacy at banks would lead to unnecessary swings in lending and real economic activity since market leverage is an extremely procyclical capital requirement and it would force banks to excessively deleverage in response to negative shocks to the equity market.