A temporary relaxation of leverage standards

Euro zone: a temporary relaxation to the rule... before it was even introduced

In September^[1], the SSM, as the single banking supervisor, announced that it had allowed the euro zone banks under its direct supervision to exclude from the calculation of their leverage ratios cash (notes and coins) and reserves with the Eurosystem (deposits made under the deposit facility and balances held in Eurosystem reserve accounts, including required reserves)^[2]. This decision was taken after the single supervisor established, after consultation with the ECB (as the monetary authority), the existence of “exceptional circumstances” that justify such an exclusion. The banks will benefit from this relaxation from 30 September 2020 until 27 June 2021, the eve of the planned introduction of the leverage constraint. The supervisor has reserved the right to extend this exclusion beyond June 2021, by means of a possible recalibration of the leverage requirement.

At present, in the European Union, the leverage ratio is not a binding requirement (Box 1). However, banks are obliged to declare and report it. The public disclosure requirement has created an implied market requirement, at the minimum level of 3% recommended by the Basel Committee since 2011.

On the basis of figures to end-March 2020, the ECB estimates that the exclusion would have increased the aggregate leverage ratio for “significant” banks (112 banks in Q1 2020) by around 30 basis points, from 5.36% to 5.66%. We calculate that in Q2 2020, the improvement in the ratio would have been 42 basis points, from 5.32% to 5.74%^[3]. Setting aside its temporary nature, the effect of this relaxation is thus likely to be less than that estimated for the major US banks (see below)^[4].

United states: a relaxation that favours bank holding companies more than their depository institution subsidiaries

A 120 basis points improvement in leverage ratios at the 8 US G-SIBs

In the USA, the translation into US law of the Basel requirement, the Supplementary Leverage Ratio (SLR)^[5], has been relaxed under two rules^[6]:

• The rule, finalised in November 2019, excluded, from the definition of the leverage exposure of custodial banks (banks predominantly engaged in custody, safekeeping and asset servicing activities), part of their excess reserves held with the central bank^[7] (equivalent to the proportion of client deposits linked to these business areas). This exclusion covers not only deposits at the Fed, but also those with central banks in other OECD countries. The rule came into force on 1 April 2020. To qualify as a ‘custodial banking organisation’ and be eligible for this measure, a bank holding company must have a ratio of assets under custody to total assets of at least 30:1 (as an average over the preceding four quarters). This relaxation allowed a significant improvement in SLR ratios for Bank of New York Mellon (BONY) and State Street in Q2 2020 (of 270bp and 260bp respectively). The other US G-SIBs have seen differing levels of benefit from these measures, depending on the make-up of their businesses. Overall, the November 2019 rule reduced the average leverage exposure of the 8 US G-SIBs by 11% and increased their average SLR ratio by 78 basis points. Without this exclusion, the average SLR ratio for these G-SIBs would have been 6.2%, against a published figure of 7% (Table 1 and Chart 1).
• The temporary rule finalised in April 2020 went on to exclude Treasury securities and reserves held with the Fed from the definition of leverage exposure for all US bank holding companies subject to the Basel leverage ratio. The rule will be in force from 1 April 2020 to 31 March 2021. Taken in isolation, we calculate^[8] that the April 2020 rule reduced the average leverage exposure of the 8 US G-SIBs by 14% and increased their average SLR ratio by 100 basis points. Overall, the benefits of excluding deposits with the Fed is similar to that from the exclusion of Treasuries. Looking at individual banks, however, there are discernable discrepancies (Table 2).

The benefits from the two rules are not cumulative, as part of the reserves that can be excluded under the April 2020 rule can also be so under the November 2019 rule. Using the information available, we have estimated the maximum benefit available to each bank from these two relaxations. Overall, the November 2019 and April 2020 rules reduced the leverage exposure of the 8 US G-SIBs by 18% and increased their average SLR ratio by 124 basis points (Table 3 and Chart 1). In Q2 2020, the margin over the minimum requirement (5% for US G-SIBs) was, on average, 202 basis points; it would have been 78 basis points without the relaxations (124 basis points lower).

With a few exceptions, depository institutions have not excluded their reserves and Treasuries from the calculation of their ratios

Whilst the temporary exclusion of reserves and Treasuries from leverage exposure (April 2020 rule) is automatic for bank holding companies, it is optional for depository institutions. In May, US regulators effectively extended the new SLR calculation method to all depository institutions with balance sheets of more than USD250 billion (the Category II and III banks) and subsidiaries of US G-SIBs – on the condition, however, that they submitted their dividend payment programmes (including intra-group dividends) to their supervisors.

Although the new rule has clear attractions with regard to the prudential leverage requirement (depository institution subsidiaries of G-SIBs generally carry on their balance sheets the bulk of central bank reserves for the consolidated group, whilst their minimum leverage requirement is more severe, at 6%), it also limits the payment of dividends from G-SIB subsidiaries to their holding companies (and thus from the holding companies to their shareholders^[9]).

As a result, amongst depository institution subsidiaries of G-SIBs, only Goldman Sachs Bank has opted for a change in the definition of its leverage exposure (improving its SLR ratio in the process by 220 basis points in Q2 2020, Chart 2). The depository institution subsidiaries of BONY and State Street are also exceptions in that they took fairly full advantage of the November 2019 rule change (depository institution subsidiaries of bank holding companies designated as custodial banks can, like their parent companies, exclude part of their excess central bank reserves).

As a consequence of these three exceptions (GS Bank, The Bank of New York Mellon et State Street Bank), the SLR ratio of the main depository institution subsidiaries of the 8 G-SIBs declined by only 20 basis points on average between Q4 2019 and Q2 2020. In Q2 2020, the average ratio was 78 basis points above the minimum level of 6%.

Limited room for manoeuvre

Although the relaxations have automatically improved leverage ratios, there are still significant constraints on balance sheets.

Bank balance sheets have expanded considerably since mid-March. Drawings against confirmed credit lines and issuance of guaranteed loans under the schemes introduced by the authorities have increased balance sheet assets. Although like central bank reserves, loans covered by a government guarantee come with a zero risk weighting (in the calculation of risk-weighted capital ratios) they are included in full in the calculation of leverage exposure.

Banks have also seen their central bank reserves increase substantially following the amplification of monetary policy measures. However, the changes made only allow for a temporary exclusion of reserves from the calculation of leverage exposure (other than for the US rule specific to custodial banks). Given the lasting nature of the reserves created, unless the central banks reduce the size of their balance sheets (which looks unlikely in the short term) there would be justification for excluding them from the denominators of leverage ratios for an extended period.

Lastly, the assessment of systemic importance scores will remain a function of total leverage exposure (i.e. not corrected for reserves). Growth in bank balance sheets as a result of the exceptional measures introduced to support business could thus result, towards the end of the year, in an increase in G-SIB scores and hence in higher CET1 capital requirements. This would then require efforts to rationalise balance sheets.

Towards an increase in G-SIB surcharges?

In the USA, the April 2020 rule explicitly seeks to neutralise the effect of exclusions applied to total exposure in the calculation of the G-SIB surcharge. In the euro zone, the SSM has not discussed the issue of the G-SIB surcharge, suggesting alignment with the US position. In other words, unless there are changes to the US SLR rule and to European regulations before the end of this year (or a recommendation in this direction from the Basel Committee), the assessment of the systemic importance of banks and the determination of their capital surcharges will continue to be based on total exposures, that is to say including central bank reserves (and Treasuries in the case of US banks).

Higher systemic importance scores

In the USA, the calculation method for G-SIB scores allows us to follow their quarter-on-quarter movements (Box 2). In Q1 2020, the systemic importance scores for US G-SIBs all rose, with the notable exceptions of Wells Fargo (where balance sheet growth is capped^[10]) and Morgan Stanley. Most noticeably, the total score for JP Morgan increased by 100 basis points over the quarter, against 56 basis points in the same period last year, due to a sharp rise in its complexity score^[11] and its international lending and commitments.

On the basis of Q2 2020 data, published on 22 September, the scores at JP Morgan, Bank of America, Citigroup and Goldman Sachs could move up a level at the end of the year, which would result in a 50-basis-point increase in their surcharges (from 3.5% to 4% for JPM, 3% to 3.5% for Citigroup and 2.5% to 3% for BoA and GS; Figure 3). Only two basis points separate State Street’s current score from the threshold of the upper tranche (at 1.5%).

There is scope to avoid an increase in surcharges between now and 2023

In practice, US institutions have two options to avoid an increase in their G-SIB surcharge^[12]. The first would be to reduce their overall score over the course of the final quarter of 2020. In effect, the score for the fourth quarter determines the level of the G-SIB surcharge. This could be a possible option as indicators of complexity, interdependence and cross-border business are assessed on the basis of the position at 31 December of the year in question^[13].

In the light of past experience, most G-SIBs should be able to avoid an increase in their surcharge. Only JP Morgan seems too far above the threshold for the next band down (by 65 basis points, whereas the biggest reduction over a single half-year period in the last three years has been no more than 42 basis points). Indeed, JPM has already announced that its surcharge will probably be higher at the year end. The reclassification of part of its securities portfolio as “held to maturity” should, it claims, help it minimise its CET1 capital requirement^[14] by reducing its Stress Capital Buffer^[15] and thus offsetting any possible increase in its G-SIB surcharge.

The second solution, for US banks, would be to minimise their overall score at the end of 2021. When the overall score for a G-SIB increases, taking it over the threshold between two bands (on the basis of scores at the end of year n), the new surcharge is only applied two years after the announcement that the threshold has been exceeded (announced in November of year n+1, applied on 1 January of year n+3). In other words, the possible increase in the G-SIB surcharge for JP Morgan to 4%, based on 2020 data, would not be effective until 1 January 2023. However, when the overall score falls to an extent that warrants a reduction in the surcharge, the new lower surcharge is applied a year after the announcement that the threshold has been crossed (announced in November of year n+1, applied on 1 January in year n+2). Thus, targeted measures to reduce its overall score at the end of 2021 would allow JP Morgan to retain an unchanged G-SIB surcharge of 3.5% from 1 January 2023.

^[1] https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32020D1306

^[2] The new definition of leverage exposure will also apply to the calculation of total loss-absorbing capacity (TLAC).

^[3] The volume of reserves excluded from leverage calculations in Q2 2020 is approximated on the basis of outstanding current accounts and the deposit facility at end-June 2020.

^[4] The analysis is focused on the 8 US G-SIBs, which have communicated more broadly on the effect of the April 2020 rule (temporary exclusion of reserves at the Fed and Treasury securities for the calculation of the leverage ratio).

^[5] See Choulet C (2020), US banks: leverage ratios under pressure, Eco Conjoncture, June 2020, for a summary of leverage ratios in force in the USA and the relaxations introduced over the past year.

^[6] The three banking regulators have also been able to neutralise the impact of participation on two specific schemes introduced in response to the pandemic: the Money Market Mutual Fund Liquidity Facility (MMLF) and the Paycheck Protection Program Lending Facility (PPPLF). Under temporary rules, published respectively on 19 March and 9 April 2020, assets used as collateral for the MMLF and the PPP loans used as collateral for the PPPLF can be excluded from the calculation of leverage ratios. According to the rather fragmented information released by banks, these measures have not allowed a noticeable improvement in leverage ratios. Of the G-SIBs, JP Morgan has indicated that it has subtracted USD7.8 billion of assets purchased from money market funds and used as collateral under MMLF (giving a 2 basis point improvement in its leverage ratio).

^[7] Figures for reserves held in excess of the required minimum no longer have meaning as the Fed announced the removal of its minimum reserve requirement as part of its updated monetary policy of 15 March (reduction in required reserve coefficient to 0%, effective from 26 March).

^[8] Wells Fargo and BONY have not reported the effect of the April 2020 rule on their Basel ratios. We have estimated the impact on the basis of balance sheet data to end-March and end-June 2020 (interest-earning deposits at US depository institutions and portfolios of Treasuries).

^[9] Moreover, since Q3 2020, the Fed has imposed restrictions on dividend payments on holding companies with balance sheets over USD100 billion.

^[10] In February 2018, the Fed prohibited Wells Fargo from increasing its balance sheet beyond its end-2017 level (USD1,951 billion) until it sufficiently improves its governance and controls. The bank has, however, been authorised to allow its balance sheet to expand where this growth comes from its participation in government support schemes (Paycheck Protection Program and Main Street Lending Program) introduced in response to the pandemic.

^[11] Notional value of OTC derivatives, level 3 assets, value of portfolios of securities held for trading or available for sale.

^[12] European G-SIBs seem to have less room for manoeuvre, however, as their systemic importance scores are evaluated by comparing their exposures to those of other banks (relative scoring).

^[13] The non-renewal, at the year end, of overnight lending and borrowing on the repo markets and forex swap lines is generally an effective means of reducing these indicators. The size indicator corresponds to average leverage in the fourth quarter (daily average of balance sheet exposure over the fourth quarter and monthly average of non-balance sheet exposure over the fourth quarter). Only the indicator of dependence on short-term market borrowing is calculated on the basis of average daily values over the previous 12 months.

^[14] CET1 requirements include a minimum of 4.5% of risk-weighted assets + the G-SIB surcharge + the countercyclical capital buffer + the Stress Capital Buffer (the level of which is set by the Fed following CCAR stress tests).

^[15] And, most likely, its complexity score as well.