Capital requirements may reduce the chance of a financial institution’s failure—in theory, the more capital that an institution has, the better it may withstand a run. While capital requirements do serve a necessary function vis-à-vis certain exogenous correlation risks affecting the industry generally, contagious runs are likely to overwhelm any plausible capital requirement, due to the staggering losses that inevitably follow from asset fire sales. Indeed capital requirements at any plausible level will be insufficient to prevent contagion, as it is unlikely that short-term debt-holders even take an institution’s solvency into account (which in any event will be difficult for them to determine) during a run—better safe than sorry. Further, and crucially, capital requirements only apply to banking organizations and a few specific nonbanks (e.g., the three nonbank SIFIs, for which the requirements have not yet been determined). So they cannot stop contagion in the nonbanking sector, an important feature of the 2008 crisis. Capital requirements may even increase contagion risk in the nonbank sector as they force activities with high capital requirements out of banks into the nonbank sector.
Government-imposed capital requirements are generally thought necessary to mitigate the ex ante effects of regulatory safety nets, particularly deposit insurance, on banks’ risk taking and leverage. A significant unintended effect of deposit insurance is to transform deposits from “de jure overnight debt financing,” which imposes a high degree of market credit discipline, into “de facto patient debt financing,” which is much less sensitive to the bank’s riskiness.1 As a result the cost of short-term bank debt from deposits is relatively insensitive to the bank’s leverage even as this leverage increases. While banks clearly have an incentive to pile on this relatively “cheap” debt, the cost of the bank’s increased risk of default is borne by the deposit insurance fund, effectively allowing banks to extract rent from the insurer. The imposition of capital requirements stems this moral hazard based on the proposition that “capital reduces incentives for incurring risks.”2 Another justification for government-imposed capital requirements is the belief that “capital serves as a buffer against unexpected losses.”3 Unexpected losses and consequent deleveraging can result in fire sales, causing negative knock-on effects on otherwise healthy banks that hold the same or similar assets.4 Capital requirements potentially mitigate these negative externalities, which can arise regardless of deposit insurance or even the presumption of public bailouts. A third justification for capital requirements is that “a capital requirement provides the supervisor with room for intervention before the bank becomes insolvent.”5 Even if capital requirements could achieve these objectives, they cannot eliminate, or even significantly reduce, the risk of contagion. They are no substitute for strong weapons to fight contagion.
Following the financial crisis, the Basel Committee on Banking Supervision, the international body entrusted since 1988 with bank capital requirements, issued a reform proposal for capital regulation, entitled “Basel III,” as part of a series of initiatives sponsored by the Group of 20 (“G20”) nations.6 US regulators issued a notice of proposed rulemaking in June 2012 implementing key provisions of Basel III, which was adopted in a final rulemaking in July 2, 2013.7 The centerpiece of Basel III is a series of amendments to the capital adequacy standards embodied in the worldwide framework for capital regulation created by Basel I and extensively revised and expanded under Basel II.8 These amendments specify three broad revisions to the Basel I and II architecture: (1) increases in minimum mandatory bank capital requirements, (2) new measures to control countercyclicality in capital regulation, and (3) new restrictions on what instruments qualify as capital and adjustments to risk-weightings. Besides proposing the Basel III amendments, the Basel Committee has expanded the stress testing requirements established under Basel II.
Basel III capital requirements provisional phase-in schedule
Note: Press Release, Minimum Capital Standards, supra note 1160, at 7. a. CET1 is common equity tier I capital. http://www.bis.org/publ/bcbs198.pdf at 5.
Basel III establishes minimum capital ratios for different definitions of capital, set forth in tabular form in table 14.1. A new requirement for a minimum common equity (almost equivalent to tangible equity) provides for a capital ratio expressed as a percentage of risk-weighted assets (RWA). It mandates an increase from 2 percent to a base level of 4.5 percent by 2015.9 Basel III further provides for a cumulative increase of 2 percent in minimum tier I capital (the old Basel I definition of capital), raising the minimum tier I ratio from 4 percent currently to 6 percent by the start of 2015. Minimum tier II capital (again an old Basel I measure containing weaker elements of capital) will be reduced from a minimum of 4 to 2 percent of RWA, thus providing that total tier I plus tier II capital ratio stay at its present 8 percent level—this is expressed as minimum total capital in table 14.1. In addition to imposing higher “basic” equity and tier I capital ratios, Basel III requires financial institutions to institute a supplementary common equity capital “conservation buffer” equivalent to an additional 2.5 percent of RWA, to both measures of capital, to be fully implemented by the start of 2019.10 Thus, in total, the minimum common equity capital requirement imposed under Basel III at the conclusion of the phase-in period will amount to 7 percent of RWA, inclusive of the capital conservation buffer (4.5 percent plus 2.5 percent).11 Once phase-in of the regime is completed by the end of 2018, minimum tier I capital and total capital under Basel III will amount to 8.5 percent (6 percent plus 2.5 percent) and total capital, tier I = tier II to 10.5 percent (8 percent plus 2.5 percent) of RWA.
Basel III also prescribes a new discretionary “countercyclical buffer” ranging from 0 percent to a maximum 2.5 percent of common equity (or “fully loss absorbing” equivalents; i.e., tier I capital) for banks located in countries where “excess credit growth … is resulting in a systemwide buildup of risk.”12 It ties the countercyclical buffer to supervisory discretion, prescribing it for use in overheated credit markets that promote rising asset values with an accompanying procyclical increase in bank leverage. The decision to implement this buffer and the assessment of its magnitude would be made by a designated national authority identified by the member jurisdiction.13 Both the countercyclical and conservation buffers are analogous to forms of loan loss reserves intended to be drawn down to absorb unanticipated credit losses.14
These basic capital requirements will apply to all banks in the United States, except those banks with less than $500 million in consolidated assets and certain savings and loan companies.15 In July 2011, the Basel Committee released a consultative document outlining a proposed methodology for determining which banks were systemically important, called “G-SIBs.”16 In November 2013, the FSB released an updated list of twenty-nine banks designated as G-SIBs, which included eight US banks and replaced four of the original G-SIBs with four new banks.17 In November 2014, the FSB released an updated list, which contained the same institutions as the 2013 list and added one more bank, bringing the total number of G-SIBs to 30.18 Based primarily on cross-jurisdictional activity, size, connectedness, substitutability, and complexity, G-SIBs are thus required to hold an additional common equity buffer ranging from 1 percent to 2.5 percent of RWA.19
The Federal Reserve proposed in December 2014 to implement surcharges for its G-SIBs by proposing a somewhat different methodology than adopted by the FSB.20 The proposed rule differs from the Basel requirements, as it includes reliance on wholesale funding in its calculation of the risk-based capital surcharge applied to G-SIBs. According to the proposed rule, the increased capital charge for reliance on wholesale funding would “help the resiliency of the firm against runs on its short-term wholesale funding,” thereby reducing the risk of the firm’s failure, and help internalize the cost of using wholesale funding.”21 The proposed rule specifically cites the systemic risk arising from reliance on wholesale funding, noting that under difficult market conditions, institutions may be forced to conduct fire sales of assets to meet the withdrawals of short-term creditors. The resulting contagion from these fire sales is the source of systemic risk.22 Thus it is clear that the Fed sees capital as a key bulwark against contagion.
The proposed rule would double a G-SIB’s capital surcharge as compared to the Basel Committee’s approach.23 Table 14.2 displays the Federal Reserve estimates, with effective capital surcharges projected between 1 and 4.5 percent for each G-SIB.24 Method 1 is the Basel method and method II is the method proposed by the Fed. The total G-SIB capital surcharge for all eight banks will be $209.3 billion. This represents a capital surcharge increase of $94.6 billion as compared to the aggregate surcharge under the Basel standard.25 A fundamental flaw with this rule, and with a recent Federal Reserve calibration of the G-SIB surcharge, is that each bases the surcharge on an interconnectedness analysis, increasing the surcharge for more interconnected entities.26 As shown in this book, there is no empirical evidence that interconnectedness in its own right is a significant problem or that decreasing interconnectedness would reduce the risk of contagion.
Estimated scores and surcharges
Note: Estimates of Buckingham Research Group.
The Basel approach is in large part based on requiring capital in relationship to the riskiness of assets, the RWA approach. This is a highly problematic enterprise given the difficulties of actually weighting assets for risks. Basel I was built around the so-called standardized approach where the regulators specified risk-weights, an inherently impossible task. Attempts to set prices for goods and services, a much easier task than pricing risk, have generally failed in the past. Basel I required more capital for unsecured loans to private borrowers (100 percent RWA) than it did for any residential mortgages, including subprime (50 percent), and assigned the same zero risk-weight to all OECD sovereign bonds, which included Greece as well as the United States. Underweighting of some risks while overrating of others distorted capital allocation.27
Recognizing that standardized risk-weights did not work, Basel II gave large sophisticated banking institutions the right to use models, with certain parameters, to estimate different risks, market, credit, and operational risks. But this grants banks a large amount of discretion to use models that underestimate risk, thereby artificially bolstering reported capital ratios. Basel has already substituted more standardized calculations for models with respect to resecuritization exposures including CDOs of ABS.28 And the Basel Committee is currently reexamining risk-weighting models as part of the Basel III process and may seek to use standardized methodology to set RWA floors—models could only increase but not decrease risk estimates.29 But this just brings back the problems of the standardized approach. The general point here is that we should not have a lot of faith in the ability of capital requirements to limit risk, let alone avoid contagion.
In addition to RWA capital requirements, Basel III provides for a non–risk-weighted tier I capital leverage ratio, provisionally set at 3 percent of total assets and subject to adjustment during the phase-in.30 The United States currently employs two leverage ratio concepts. The “generally applicable” leverage ratio is the ratio of a firm’s tier I capital to its total consolidated balance sheet assets (but not off-balance sheet items).31 Under the current US regime, all BHCs and insured depository institutions (“IDIs”) must maintain a minimum generally applicable leverage ratio of 4 percent.32 Furthermore IDIs that are subject to the advanced approaches risk-based capital rules (“advanced approaches” IDIs) must comply with a “supplementary” leverage ratio requirement under which they must maintain a generally applicable leverage ratio of 5 percent to be considered “well-capitalized” for purposes of prompt corrective action and financial services holding company status. The “supplementary” leverage ratio is the ratio of tier I capital (the old Basel definition) to total leverage exposure, comprising total consolidated balance sheet assets plus many off-balance sheet exposures, including derivatives positions.33 US regulators have proposed that this supplementary leverage ratio for the largest US BHCs and their IDIs be 5 percent and 6 percent, respectively.34
The case for the use of a leverage ratio to set capital requirements is bolstered by the deficiencies of the RWA approach. Andrew G. Haldane, Executive Director of Financial Stability at the Bank of England, has stated that in a financial environment filled with uncertainty, Basel III’s complex risk-weighting system for capital adequacy may be less optimal than a simpler one. He finds that in the period leading up to the recent financial crisis, simple leverage ratios had greater power in predicting the failure of large global banks than the more complex risk-weighted measures of the Basel approach. Haldane determines that a simple market-based leverage ratio outperforms the more complex Basel III tier I capital ratio by a factor of 10 to 1 as an indicator of bank solvency. In addition Basel III’s increased reliance on internal risk models has led in Haldane’s judgment to overly complex measures with thousands, if not millions, of parameters.35 Calibration of these parameters will likely require decades of data collection. In terms of capital regulation, and in particular the Basel approach, less may be more.
Though intended “as a backstop to the risk-based measures,”36 a leverage requirement that does not rely on RWA seems to run counter to the basic premise of the Basel initiative from its start in 1988 that capital adequacy should not be judged without considering the riskiness of the assets.37 A leverage ratio requires precisely the same amount of capital for high- and low-risk assets. Effectively, the risk-weight assigned to all assets under a leverage ratio is 100 percent, regardless of the actual riskiness of the asset. As a result, if the regulatory cost of capital for a bank is the same for high- and low-risk assets, this may give bank management an incentive to increase return on equity by investing in high-risk assets with higher returns. Such incentives are inconsistent with prudent risk management and sound banking practice. In theory, this could be avoided to some extent by employing both RWA and non–RWA-based requirements together. Another consequence of a binding leverage ratio is that when the Federal Reserve begins to reduce the size of its balance sheet then banks will be less able to buy and hold Treasuries and mortgage-backed securities to offset the decreased Fed role in these markets.38
While Basel III increases minimum capital requirements through the direct measures described above, it also effectively increases the amount of capital required by restricting the range of instruments eligible for inclusion in the calculation of tier I capital.39 In particular, Basel III calls for eliminating certain elements from qualifying common equity tier I capital (“CET1”), to be phased in gradually through 2018.40 These eliminations include certain deferred tax assets, mortgage servicing assets, and significant investments in the common shares of unconsolidated financial institutions.41
Finally, the Basel Committee addressed the stress-testing requirements established under Basel II through expansion of the internal capital adequacy assessment process (“ICAAP”).42 In general, stress tests are an analysis of a bank’s capital adequacy under varying adverse economic scenarios. Stress tests augment the Basel capital requirements by requiring financial institutions to plan for highly adverse events. Basel requires banks to have a comprehensive stress-testing program that aims to address the possibility of severe shocks and changes in market conditions.43 US regulators issued a final rule in November 2011 requiring all US-domiciled BHCs with consolidated assets of $50 billion or more to submit to an annual Comprehensive Capital Analysis and Review (“CCAR”).44 Each year, covered BHCs must develop and submit a three-year capital plan to the Federal Reserve. The Federal Reserve will then carry out a supervisory stress test based on a predetermined stress scenario that projects earnings, and losses. The Federal Reserve uses its own nondisclosed model to calculate losses. Based on the resulting determination of the adequacy of each firm’s capital plan, the BHC may be required to forego capital distributions to shareholders in the form of dividends or share buybacks, or raise additional capital.45 Some believe that US stress tests rather than the Basel RWA or leverage requirements are the binding constraint on the amount of capital banking organizations must hold, given the amount of capital that must make up for losses driven by the adverse scenarios.
Stress tests have their own methodological problems. First, the extreme adverse scenarios are very extreme. In the 2014 CCAR, projecting economic conditions over nine quarters, from Q4 2013 to Q4 2015, the scenario made the following assumptions (compared to what really happened in parentheses): real GDP decline of 4.75 percent (increased 2.4 percent), unemployment rate 11.5 percent (5.6 percent), deflation 1 percent (inflation 0.8 percent), equity prices fall 50 percent (11.4 percent increase, all time high), house prices fall 25 percent (4.6 percent gain).46 Second, the Fed’s model to analyze bank losses is a black box—we do not know what it is, so we cannot judge if it is any good. The Fed has argued that it if revealed its model, it would be gamed by the banks. Even if this is true, the fact remains that we have no way of judging the quality of the model.
In the context of contagion concerns, the stress test design really falls short. CCAR does not consider the possibility of runs by short-term creditors in the adverse scenarios.47 The Office of Financial Research has pointed to two primary limitations of the current stress test designs: (1) a lack of fire-sale or run considerations and (2) a failure to capture potential feedback effects of external shocks, such as the potential reduction of credit availability to the firm resulting from its losses.48 If capital requirements were actually stressed by runs, they would be sorely wanting, due to fire sales, and we would realize that capital requirements are a false palliative for anti-contagion weapons.
The Basel III framework has been criticized by some for still not providing for enough capital.49 In the United States, the Federal Reserve has already increased the capital requirement for US banks beyond the Basel requirement.50 But these new capital requirements cannot prevent or provide an adequate buffer against contagion. During the crisis of 2008, most of the largest banks had a regulatory capital ratio that exceeds the new Basel III minimum ratio. Leading up to the financial crisis, in 2007, the average regulatory capital ratio for the top 20 US banks was 11.7 percent, which exceeded the then existing regulatory minimums by 50 percent.51 And as table 14.3 indicates, their ratio of tier I common equity to risk-weighted assets was higher than the new Basel III requirement of 4.5 percent and many had higher than the 7 percent requirement that includes the 2.5 percent buffer.52
Large US bank capital ratios in 2007
By this time the major US investment banks had also implemented Basel II, pursuant to regulation by the SEC.53 Yet, despite being effectively compliant under the Basel II framework before the financial crisis, these institutions still did not hold enough capital to survive the crisis without public support.
In fact the 2009 IMF Global Financial Stability Report found that risk-weighted capital adequacy ratios were unable to identify which institutions would require government assistance, even finding that capital ratios were higher on average for commercial banks needing intervention.54 Furthermore bank default risk as measured by CDS spreads did not correlate meaningfully with regulatory capital ratios during the crisis, feeding “doubts … in relation to the efficacy of the capital index tier I ratio as a safeguard against the risk of future default.”55 The fact that new minimum Basel III capital requirements, which were met by large banks during the crisis, did not capitalize US financial institutions sufficiently to avoid public support in the crisis, undercuts expectations for the framework’s future performance.
A bill proposed in the US Senate in 2013 by Senators Sherrod Brown and David Vitter would substantially increase capital requirements for large banks and discard risk-weights in favor of an exclusive leverage requirement.56 Under this proposal, regional banks would be required to hold 8 percent equity capital to total assets, while the largest banks would have a 15 percent requirement.57 Regulators would be given the authority to increase these capital requirements further on a case-by-case basis as institutions grow larger.58 However, even these increased levels of bank capital are not sufficient to have a meaningful effect in reducing the risk of contagion. The Brown–Vitter proposal also, incidentally, would prohibit the Federal Reserve and other banking regulators from providing discount window lending, short-term credit insurance, and other federal support programs to nonbank financial institutions.59 Another bill passed by the House, the FORM Act, as previously discussed, would further restrict but not prohibit the Federal Reserve’s ability to lend to nonbanks.60 If either bill were implemented, it would dramatically increase the risk of contagion because short-term creditors would be much more inclined to run in the absence of any lender-of-last-resort support.61
One paper authored by economists at the Bank of England in 2011 called for bank equity capital requirements of between 16 percent and 20 percent of RWA.62 A government-appointed commission of noted policy makers in Switzerland proposed heightened capital requirements for UBS AG and Credit Suisse as an additional prophylactic measure for the Swiss economy, under which both banks would be required to hold total capital of 19 percent and be subject to a 10 percent common equity minimum due to their designation as “too big to fail” (TBTF).63 This “too-big-to-fail” legislation incorporating the so-called Swiss finish received final approval from both houses of the Swiss parliament in September 201164 and became effective on March 1, 2012, although implementation will be phased in through January 1, 2019.65 Again, these higher capital requirements are still too low to have a meaningful impact on the probability of avoiding contagion.
The Basel III requirements also come with high and uncertain costs to the real economy. One study estimates that Dodd–Frank and Basel III regulations would collectively lead to 20 percent fewer loans and result in 600,000 fewer home sales, ultimately costing one million housing starts and 3.9 million fewer jobs.66 Although the precise economic impact on GDP is uncertain, several organizations have commissioned studies to make estimates. Studies have been published by the Macroeconomic Assessment Group (“MAG”), established by the Basel Committee, along with the FSB; the Institute of International Finance; the IMF; the Organization for Economic Co-operation and Development; the Government Accountability Office (“GAO”); and a panel including staff from the FRBNY, Bank of Italy, BIS, ECB, European Commission, and IMF.67 All of the studies predict that Basel III will have a negative impact on gross domestic product. Moreover the Institute of International Finance (IIF), an industry group, estimates that the various Basel III requirements (excluding the leverage ratio) will reduce the real GDP of the United States, the euro area, the United Kingdom, Switzerland, and Japan by about 3.2 percent, leading to 7.5 million fewer jobs being created by 2015.68 In addition the FSB, in collaboration with the IMF and World Bank, has published a study that identified unintended consequences of regulatory reforms to emerging market and developing economies, particularly as it pertains to trade finance and restrained international capital flows.69 Although mostly qualitative in nature, the analysis highlights the widespread economic impact of current regulatory reform measures.
But these studies fail to incorporate the dynamic impact that bank capital requirements will no doubt have in shifting risky activity, financed by short-term funding, to the nonbanking (i.e., “shadow banking”) sector, which is subject to no or very low capital requirements. Thus the true economic cost of these new capital requirements, which could mean more risk to the economy from nonbank failures and contagion, is not taken into account.
One solution to improving the determination of capital adequacy is to assign a larger role to the judgment to markets. The CCMR has conducted a study on capital regulation that examines a balanced approach to enhancing private market discipline, through better disclosure and requirements for capital instruments that cannot be bailed out, while strengthening the role of regulators.70 A regulatory approach recognizing the dual roles of government and the private market in determining adequate capital might result in a stronger and safer financial system than can be achieved through public regulation alone. In general, the private market is much more effective and efficient at pricing risk than regulators, so policymakers should pursue a market-based approach that enhances the market discipline of banks, relying on the market to dictate optimal levels of capital. Market signals, such as debt yields or CDS prices, may also better inform a regulator of capital deficiencies than strict adherence to government-imposed requirements.
The heaviest consideration weighing against reliance on capital requirements to control contagion, however, is that while capital cushions short-term creditors against having to absorb losses, perhaps deterring the impulse to run, it does not foreclose the risk of suffering impairment altogether. As long as a financial institution is reliant on short-term funds, in any significant amount, to support long-term investment, short-term creditors who supply those funds are exposed to potential losses incurred through fire sales. In a crisis, the rational option will be to run. When that happens, capital requirements can certainly lower public costs by ensuring that deeper reserves of private funding and capital are available to the distressed institution. What they cannot do is prevent the run in the first place, or stop it from becoming generalized to the financial system.
If the problems posed by contagion are solved through more effective means, then individual banks can be allowed to fail as a result of bank mismanagement without concerns that their failure will spread to other banks. Having effective anti-contagion weapons is just as important in dealing with the too-big-to fail problem as effective resolution procedures.
Capital will not stop contagion but it is still vitally important. A correlated negative shock causes the failure of many large financial institutions at the same time. This problem cannot be addressed by a lender of last resort because insolvent banks should not be eligible for such lending absent adequate collateral. Despite the limitations of capital requirements in addressing contagion, increased capital may provide a level of protection against correlation risk, since stronger banks can better withstand common external shocks, and ultimately limit the need for public injection of capital.