On July 20, 2015, the Federal Reserve Board adopted a final rule (GSIB Rule) regarding capital surcharges applicable to US global systemically important bank holding companies (GSIBs) and issued a final order applying enhanced prudential standards to General Electric Capital Corporation (GECC Order).[1] The GSIB Rule was accompanied by a white paper that explains how the surcharges in the GSIB Rule were calculated (White Paper).[2] These three documents provide interesting information about the Fed’s approach to systemic risk.

The GSIB Rule

The GSIB Rule imposes substantial capital surcharges on eight US bank holding companies that are considered to represent the highest level of systemic risk (in order of putative riskiness (from highest to lowest): JPMorgan Chase, Citigroup, Bank of America, Goldman Sachs, Morgan Stanley, Wells Fargo, Bank of New York Mellon, State Street). The surcharges will be incorporated into the calculation of a GSIB’s maximum payout amount (e.g., its ability to pay dividends) along with its capital conservation buffer. They will be phased in beginning on January 1, 2016 and will be fully effective on January 1, 2019. For the companies currently treated as GSIBs, the scale of maximum surcharges requires up to an additional 5.5% of common equity tier 1 capital. Beyond that level, there can be additional increases to 6.5% plus an additional 0.5% for any 100 basis point increase in the score calculated pursuant to the GSIB Rule beyond the level at which the 6.5% surcharge applies. Based on the examples used in the preamble to the GSIB Rule and in the White Paper, the highest surcharge that would currently apply if the GSIB Rule were fully effective is 4.5%.

Although the fact that such surcharges will be imposed is important, in some ways the reasoning and the data behind the choice of a scale for the surcharges is more interesting. That reasoning and the data are presented essentially as the best the Fed can do at the moment but also as subject to re-evaluation over time as more data and better analyses become available. Looking at the Fed’s presentation can therefore provide clues about the ways in which the Fed tries to understand systemic risk; by implication, it also suggests what kinds of approaches could be potential alternatives.

The GSIB Rule is based largely on (i) a set of methods for measuring the systemic riskiness of a bank holding company; (ii) a decision that the amount of capital a bank holding company has is the best measure of its ability to avoid financial failure; and (iii) the development and calibration of a scale to measure the size of any capital surcharge.

A bank holding company is treated as a GSIB if it has a score of at least 130 on one of the measures of its systemic riskiness, the one referred to as method 1. Interestingly, under each method used by the Fed, the eight GSIBs measured substantially higher on the scale of systemic riskiness than did the next lowest bank holding company in the list. The ultimately relevant score for the purpose of calculating any capital surcharge is the higher of the scores given by two different measures. Method 1, the first measure, is a uniformly weighted set of values measuring:

  • size (based on total exposures);
  • interconnectedness (based on intra-financial system assets, intra-financial system liabilities and securities outstanding);
  • substitutability (based on payments activity, assets under custody and underwritten transactions in debt and equity markets);
  • complexity (measured by the notional amount of OTC derivatives, trading and available-for-sale securities and Level 3 assets [assets whose value must be deduced, rather than observed]); and
  • cross-jurisdictional activity (measured by cross-jurisdictional claims and liabilities).

Method 2, the second measure, is a different weighted set of values measuring size, interconnectedness, complexity and cross-jurisdictional activity as in method 1 but with a fifth component measuring and weighting short-term wholesale funding rather than substitutability. There are three calculational differences as well. Method 1 requires reliance on international data to be published periodically and uses that data together with data from the individual bank and the relevant weighting factor (20% for each of the five categories, evenly divided among the subcategories, if any) to compute a score. Method 2, on the other hand, uses some international data that has already been published and is therefore fixed for the time being and combines that data with the relevant weighting factor to compute coefficients for each (sub)category that are then multiplied by the individual bank’s data in each (sub)category to give scores that are then summed. In other words, method 2 calculates part of the result in advance using the same formula that is used by method 1 but different international data, which can be done because the particular international data used in method 2 already exists.

The other two calculational differences relate to the replacement in method 2 of the category of substitutability by the category of short-term wholesale funding. The first of these two differences concerns the maximal potential systemic riskiness score that can be assigned to substitutability and short-term funding. The score for substitutability is capped at 100. The score for short-term wholesale funding has a theoretical cap of 350, but its actual value equals that amount times a fraction defined as the individual bank’s average weighted wholesale funding amount divided by its average risk-weighted assets. Since the value of that fraction seems unlikely to ever reach 1, the highest realistic score for short-term wholesale funding will presumably be substantially less than 350. The weighting referred to in the numerator of the fraction relates to the fact that shorter remaining maturities of short-term funding are given higher weights than longer ones. The weights differ depending on which of four categories the funding is assigned to. The second of the two calculational differences relates to the fact that in calculating scores using method 1, a GSIB will compute the ratio between its own value for a category and a particular international average for that category. When calculating the method 2 score for short-term funding, however, the GSIB will compute the ratio between its own short-term funding and its own average risk-based assets, which renders the resulting score the most individualized value of all the systemic risk measurements used in computing the risk-based capital surcharge.

The Fed gives a brief explanation for its choice of short-term wholesale funding as a category to use instead of substitutability. One alternative mentioned by commenters on the original proposal was to use one of the liquidity measures already applied to bank holding companies. In the Fed’s view, short-term wholesale funding is more indicative than existing liquidity measures of the risk of contagion because of the likely consequences of (reactions to) a bank holding company’s failure to maintain its typical level of such funding.

In addition, the Fed makes some interesting comments about the scoring system it has created. For example, the Fed states that it may not be a simple matter for a bank holding company to reduce its surcharge by restructuring its short-term wholesale financing because many reductions in the measure (in the numerator of the fraction in the formula) of the company’s short-term wholesale financing may also reduce the amount of risk-weighted assets in the denominator, thus potentially leaving the value of the fraction little changed.

On the other hand, given the large value of the coefficient for Level 3 assets in method 2, it is possible (although not noted by the Fed) that reducing the amount of Level 3 assets held could to some extent reduce a company’s method 2 score and therefore the amount of its surcharge. Because the surcharge amounts are, however, associated with scores that are set out in bands of 100, the aggregate amount of any changes would in many cases need to be substantial in order to cause a significant reduction in the surcharge.

Surcharging Capital 

Although capital is important to any financial institution that takes risk, it is not immediately obvious that GSIBs should have more capital than other large bank holding companies, as opposed to more liquidity, for example, or less of a certain kind of risk. There seem to be several reasons for this choice. Section 165 of the Dodd-Frank Act requires the Fed to establish, for bank holding companies with consolidated assets of $50 billion or more, prudential standards that are more stringent than those applicable to smaller bank holding companies. Among the more stringent standards are risk-based capital requirements; contingent capital requirements are also authorized, although more stringent risk-based capital requirements need not be applied if the structure or activities of the relevant company render it inappropriate. Increased capital is also a measure used internationally as a way of reducing the systemic risk of an institution, making the use of a capital measure internationally compliant. In addition, a new capital surcharge can easily be integrated into the existing capital conservation buffer, something which would not have been so straightforward if a non-capital measure had been chosen. Nevertheless, depending on the way an institution’s capital is deployed, it may ultimately prove to be the case that one or more other aspects of its financial structure is more indicative of its ability to withstand stress.

Interestingly, on July 22, 2015, the Office of Financial Research published an issue of its OFR Brief Series[3] that casts some doubt on using capital as the primary determinant of safety, without, however, doing more than suggest how factors other than capital weakness can lead to systemic instability. In addition, on July 24, the Bank of England published its Staff Working Paper No. 536, “The impact of liquidity regulation on banks,”[4] that contains interesting information on the relationship between higher liquidity requirements, bank lending and bank balance sheets in the UK.

The White Paper

This more technical document that accompanies the GSIB Rule explains in somewhat more detail than the preamble to the GSIB Rule the derivation of the formula used to compute the capital surcharge. The derivation is based on the notion that the relevant value to be measured or derived is expected loss, which is calculated by multiplying (a) the amount of systemic loss given default by (b) the probability that the loss will occur. The requirement imposed on GSIBs is that their expected loss should not be greater than that of a bank holding company that is just below the threshold for the application of the capital surcharge (the reference institution). Since a bank holding company below the threshold will have a smaller asset base, it follows that to keep the product of its systemic loss and the probability of that loss equal to the expected loss for the reference institution, the probability of default for a GSIB must be reduced. Looking at historical data, the Fed calculated the probability that a bank holding company’s return on risk-weighted assets would be at or below a particular value in a particular year. This probability is then used to determine the probability that a bank holding company’s capital will be diminished to a level that is at or below the level at which it would be considered to be failing. This value is then plugged into an equation that sets the expected loss of the GSIB (which is assumed in the formula to have additional capital equal to the desired surcharge amount) equal to the expected loss of the reference institution. Because of some assumptions made about the probability that there will be a loss on risk-weighted assets, the amount of the capital surcharge required for a GSIB turns out to be a logarithmic function of the ratio of the loss given default for the reference institution to the loss given default for the GSIB, and not a function of the absolute amount of the GSIB’s loss given default.

The White Paper then considers some of the uncertainties that result from using existing data on returns on risk-weighted assets but decides to use the formula it developed because it is simple and conservative. Its conservativeness derives from the fact that regulatory interventions during the financial crisis in all likelihood prevented some losses from occurring. In addition, it considers two alternatives to using expected loss values. One would require an economy-wide cost-benefit analysis. This was rejected by the Fed for two reasons: (i) the determinations made on this basis by the Basel Committee on Banking Supervision are not specifically related to risk in the US, a requirement that the Fed sees as being implied by the wording of Section 165 of the Dodd-Frank Act; and (ii) there is no presently available way to calculate such a cost-benefit analysis in a practical manner. The other approach would calculate the amount of additional capital required to financially offset any too-big-to-fail subsidy. The Fed declined to follow this approach because it would require data that is not currently available.

The GECC Order

As a systemically important financial institution (SIFI), GECC is subject in principle to certain more stringent (“enhanced”) prudential standards than non-SIFIs. In reaction to this, GE has decided to sell much of GECC’s commercial lending and leasing and all of its consumer lending businesses and to focus more on industrial activities and financial activities directly related thereto. GECC’s issuances of commercial paper will also be reduced to $5 billion, and GE has now guaranteed all tradable debt securities and commercial paper issued or guaranteed by GECC. The Fed, in the GECC Order, has shown that it is willing to postpone the application of at least some (but not all) of the standards that might otherwise have applied, but it has basically reserved the right to apply them if the disposition of GECC’s financial activities does not proceed as indicated by GE. GE has also indicated that it will apply to have its SIFI designation removed in 2016.

This willingness of the Fed to compromise is interesting in and of itself. It is also interesting as a matter of the regulatory management of systemic risk to see which standards are being imposed now and which will not be imposed until later, if at all. For the time being, GECC will be required to comply with risk-based capital, leverage ratio and liquidity coverage ratio (LCR) standards. In connection with those requirements, it must file certain additional reports. Its risk-based capital requirements will be calculated using the standardized approach as if GECC were a bank holding company. GECC will calculate its LCR monthly until 2018, and will in 2016 be required to reach only 90% of the otherwise required level.

Unless the Fed is satisfied with the progress of GECC’s divestitures, in 2018 or 2019, as the case may be, GECC will be required to comply with additional standards relating to:

  • risk management and risk committees as applicable to bank holding companies with $50 billion or more in consolidated assets [2018];
  • board membership (majority of independent directors; chairman must be independent; all members of risk committee must be independent)[2018];
  • capital (as applicable to advanced approaches, rather than standardized approach, institutions, but without being required to use the internal ratings-based and advanced measurement approaches or to calculate an advanced measure for market risk; in other words, it will be required to recognize most elements of accumulated other comprehensive income)[2018];
  • an enhanced supplementary leverage ratio of 4% (rather than 5%)[2018];
  • capital planning (with a requirement that its tier 1 common capital ratio be at least 5% and concomitant restrictions on distributions)[2018];
  • stress testing (supervisory and company stress tests as required for SIFIs and bank holding companies with $50 billion or more in consolidated assets)[2019];
  • liquidity risk management (management, planning and testing pursuant to the standards applied to bank holding companies with consolidated assets of $50 billion or more)[2018]; and
  • affiliate transaction standards under Section 23B of the Federal Reserve Act, treating GECC as if it were a member bank, and with some special transition provisions)[2018].

On December 31, 2017 and in 2018, GECC must begin to file the additional reports that accompany the application of the heightened standards that begin to apply in 2018. In the meantime, GECC also remains subject to examination by the Fed, must continue to engage in resolution planning (i.e., prepare “living wills”), must comply with any single-counterparty credit limits that the Fed may adopt, must give the Fed prior written notice of any acquisition of a financial company with consolidated assets of $10 billion or more, may not merge with or acquire another company if the resulting company’s liabilities would exceed 10% of the aggregate liabilities of all US financial companies and must continue to satisfy the supervisory expectations for SIFIs set out by the Fed.