Recent gyrations in global equity markets has stem from the gradual (is it gradual?) steepening of the US Treasury curve. Steepening started in November 2020. In our view, the steepening stems from:
- Upcoming new supply (Biden’s US$1.9trn stimulus package) and the US Treasury’s new supply extended their weighted average maturity.
- Upcoming data around inflation and GDP growth points to a ‘hot’ economy, one in which a potential rate hike is on the horizon (rather than being on the verge of infinity).
- Potential narrative around the Fed’s tapering of quantitative easing, which creates fear of a subsequent Fed Funds Rate hike. [QE would need to end before any rate hike.]
That said, there are several other theories floating around:
- China government not buying (or buying less) long-end Treasuries [low likelihood], while Japanese investors have (reportedly) been buying long-end Treasuries [medium likelihood]. Looking at the relative spread between US and Japanese 10-year bonds, the UST10yr provides a decent yield / carry pick-up of ~150bps over Japan’s, yet still not as wide as pre-Covid (200-300bps) as historical context (shown below);
- Convexity hedging (which we covered in 26-Feb-2020 update) results in MBS investors selling their long-end Treasuries (which were hedging the negative convexity of MBS), and as long-end yields rise, the incentive for homeowners to refinance their mortgage (likely at a higher rate than previously) decreases and therefore the duration of such MBS (holding the current, non-refinanced mortgages) would increase (given lower early repayments), which reduces the MBS investor’s need to hedge duration using long-end Treasuries. [medium likelihood];
- US banks (JPM, C, BAC) concerned about the Supplementary Leverage Ratio (SLR) exemption renewal and reducing their USTs holdings before a potential non-extension announcement by the Fed [medium likelihood, discussed below].
The SLR is a leverage capital constraint on banks regulated by the Fed. The SLR was adopted by the Basel Committee, under the Bank of International Settlements, as part of the Basel III revisions. The SLR deals with capital adequacy, which is a bank’s ability to absorb unanticipated losses and declines in asset values that could otherwise cause a bank to fail. This in contrast to the traditional view of bank leverage, which is bank capital divided by total assets.
The SLR is more encompassing and calculated as:
- SLR = Tier 1 Capital / Total Leverage Exposure, where:
- Tier 1 Capital = CET1 + AT1, as defined by U.S. Basel III
- Total Leverage Exposure = on- and off-balance sheet exposures, which includes Treasuries & Fed Reserves.
The SLR views bank leverage by combining on- and off-balance sheet exposure:
Source: Federal Financial Institutions Examination Council [Data Point: FFIEC101]
The minimum SLR for G-Sib banks is 5%. Below 5%, the Fed will put restrictions on the bank’s ability to make capital distributions to equity shareholders (dividends, share buybacks) and restrict discretionary bonuses to bank employees. On the flip side, a SLR above 6% would be considered “well capitalised”. Thus, the margin between “good” and “bad” is relatively tight. In terms of sensitivity, the SLR ratio is effectively a cap rate whereby your denominator continues to grow faster than your numerator, compressing your ratio and risking a breach below the 5% threshold.
On 1 April, 2020 the Fed announced a temporary exemption by excluding Treasuries & Fed reserves from the TLE calculation, effectively shrinking the SLR’s denominator and providing US banks with a (temporary) buffer to the leverage capital constraint. However, no update has since been made regarding this exemption, which is scheduled to expire on 31 March, 2021. The Federal Reserve may discuss the SLR issue in its upcoming FOMC announcement scheduled for 17 March, 2021. If not, we would expect a clarification shortly thereafter and before the exemption’s 31 March 2020 expiry date.
The SLR exemption allowed US banks to expand their balance sheets by participating in (i) US Treasury auctions, as a primary dealer, and (ii) the Fed’s quantitative easing programme, which without the SLR exemption would have required them to increase their Tier 1 Capital. During the pandemic, the banks played a vital role in ensuring sufficient liquidity in the financial system (recall that a solvency crisis is preceded by a liquidity crisis) and continued to make loans. This ensured that sufficient liquidity in the system prevented widespread solvency issues.
If the SLR exemption is not extended by the Fed, either in full or in part (UST or reserves), US banks would need to hold more capital against the USTs and reserves (after selling USTs & MBSs into the Fed’s QE programme) accumulated in the last 12 months and on a forward looking basis. Therefore, absent any exemption, one would expect to see US banks lower their demand for USTs as they would need to increase Tier 1 Capital to maintain compliance with SLR, before considering any further increases to their TLE.
Therefore, the SLR exemption enables primary dealers to accumulate and sell USTs into the Fed’s QE, which has been predominantly the short-end tenors the Fed NY (whose ‘Trading Desk’ is responsible for open market operations) is looking to purchase. From this perspective, the impact of SLR on long-end USTs is minimal.
Looking ahead at the Fed’s FOMC announcement, we expect SLR exemption being extended as “hopeful”, however the exemption would likely apply to Fed reserves only, rather than include USTs as well.
Several reasons for this view: part theory, part practical.
- Theory (Part A): when the Fed conducts QE and buys an asset (UST, MBS) from a bank, the bank receives a Fed Reserve Deposit (effectively a digital token of equivalency to cash, against which can only be used in the Fed Funds Market) in return for selling the UST/MBS to the Fed. As explained above, a bank’s TLE includes both Treasuries & Fed Reserves. Therefore, as banks enlarge their balance sheets (shown in below table from Bloomberg) when participating in the Fed’s QE, they will need an extension to the SLR exemption to continue participating in the QE programme (unless their numerator grows proportionately, though this is unlikely). Put another way: from the bank’s perspective, Fed reserves are a negative externality from participating in QE.
- Theory (Part B): customer deposits held with banks can either be used to underwrite loans or held as USTs or Fed reserves, all three of which generate positive carry. However, without SLR exemption being extended, the additional deposits coming from (i) President Biden’s US$1.9trn fiscal package and (ii) US$1.2 trillion decline in the Treasury General Account (TGA), will mean banks will either need to (a) charge negative interest on deposits (to maintain their ROTCE) or (b) turn away deposits, which will mean the only place left are (i) depository institutions (State Street, Bank of NY Mellon) or (b) money market funds, which are capped to US$30bn on direct inflows into the overnight RRP facility. In fact, uncapping direct inflows (from $30bn) on overnight RRP facility and hiking the overnight RRP rate to just above 0% would go along way towards alleviating the balance sheet pressure on commercial banks with or without extension to the SLR exemption. The mechanics of the US$1.2trn TGA drawdown is shown in the below table from Credit Suisse, and illustrates the reserve constraint that banks will face.
- Practical (Part A): banks simply have no practical use for Fed reserves on their balance sheet (except for the Fed Funds Market) given the Fed eliminated the reserve requirements in March 2021 (and hence IORR = IOER). Reserves can be turned into loans but isn’t a prerequisite for credit creation (when loans are ‘created’ as assets, corresponding ‘deposits’ are created as liabilities). As reserves are truly agnostic (as liabilities of the Fed), their exemption from the SLR calculation would make sense given they carry no risk (and in reality, limited reward of 0.10% per annum). Whereas for Treasuries, assuming they are still considered risk-free assets, have a component of interest rate risk.
- Practical (Part B): the market forecasts a sharp economic recovery in the US, thanks to the continued rollout of vaccines and stimulus checks supporting consumption. However, the Fed needs to ensure there is sufficient liquidity (via ample reserves) in the system, which means keeping the FFR pinned within the ZLB and short-end curve (0-5yrs) on our path to ‘recovery’. That path will require the Fed to continue (and eventually taper) the quantity of QE (currently US$80bn for Treasuries & US$40bn for MBS), which requires the banks’ primary dealers to have balance sheet room for reserves expansion.
- Practical (Part C): the Fed needs excess reserves in the banking system to facilitate the financial system’s liquidity. While we currently face an opposite constraint to the Sep-2019 cash crunch (Bloomberg article extracts below), the Fed will be mindful so as to maintain an ‘ample’ level of reserves in the system as it eventually tapers QE (with a view to normalising via FFR hikes). From a simple perspective, as the financial system’s value has grown, the amount of reserves in the system that would constitute ‘ample’ would also have risen against previous levels.
- “The repo market is where the cash-rich of the financial system lend to the cash-poor, with banks, money-market funds, hedge funds, broker-dealers, asset-managers, and others borrowing and lending to each other short-term. While borrowers in this system have plenty of longer-term assets, on a day-to-day basis they may not have the liquidity they need. Repo lets them borrow cash against those assets to tide them over.”
- “After the financial crisis, the Fed began stimulating the economy by buying bonds. The money it spent to do so added to the excess reserves held by banks. After years of QE, big banks became accustomed to always having cash to lend in the repo market—they were swimming in excess reserves. But then the Fed began tapering off QE, buying fewer bonds and reducing the excess reserves in the system. There’s still a lot, but this is where the regulatory experiment comes in. To prevent a repeat of the 2008 crash, bank watchdogs have tightened rules in such a way that the biggest lenders feel they have to keep more reserves on hand. So falling reserves and banks’ desire to hold on to more began to pinch.”
- “In September [2019] all these forces combined with a few other quirky events—such as corporations needing cash to settle quarterly tax bills—to create a brief systemwide shortage of ready cash.”
- Practical (Part D): common stockholders of the big US banks are (arguably) being unfairly treated as any dividends and/or share buybacks would dilute the CET1, the larger component of the SLR’s numerator. In fact, the other numerator, AT1, is also dilutive to CET1 since it carries a non-deferrable coupon. The denominator (as discussed above) continues to rise as the big US banks ensure ‘smooth functioning’ and ‘orderly market conditions’ via the quantitative easing programme. This is where US politicians have (in our view) unfairly targeted the ‘big banks’ which have, since the Fed’s Dec-2020 stress test (on page 143), been permitted banks to “pay common stock dividends and make share repurchases that, in aggregate, do not exceed an amount equal to the average of the firm’s net income for the four preceding calendar quarters”. The position of US (democratic) senators has been to urge the regulators to “reject the coordinated lobbying efforts” in regards to extending the SLR exemption.
Therefore, the Fed extending SLR to exclude reserves would ensure banks have sufficient space on the balance sheet to take up & store Fed reserves, while also excluding Treasuries would likely prevent a selloff (in order to comply with the SLR rule + a management buffer).
Given the above, it’s clear that the SLR, in its non-exempted form, doesn’t suit the practical nature of how the Fed relies on banks’ primary dealers to operate monetary policy. Instead of an SLR exemption extension, we could even see a revised SLR threshold, although this would be extremely unlikely in the near term, as the SLR forms a buffer (or an early indication of stress) over a bank’s minimum capital requirements, as Randal Quarles, a Fed Governor, highlighted in November 2018,
“As the Federal Reserve has long maintained, leverage requirements are intended to serve as a backstop to the risk-based capital requirements. By definition, they are not intended to be risk-sensitive. Thus, I am concerned that explicitly assigning a leverage buffer requirement to a firm on the basis of risk-sensitive post-stress estimates runs afoul of the intellectual underpinnings of the leverage ratio, and I would advocate removing this element of the stress capital buffer regime. Of course, leverage ratios, including the enhanced supplementary leverage requirements, would remain a critical part of our regulatory capital regime, and we will maintain the supervisory expectation that firms have sufficient capital to meet all minimum regulatory requirements.”
An alternative view on a non-extension of the SLR exemption is that the Fed may have (preemptively) determined its QE tapering schedule. Recall that excess reserves are required in the banking system to supply short-term funding. If the Fed expects to taper (e.g. by 2022), the accumulation of USTs & Fed reserves pushing the banks’ SLR ratio thresholds would be alleviated. The SLR relief would have limited / no impact on the TGA decline and fiscal stimulus checks.
In any case, Bloomberg’s Weekly FIX sums up the Fed’s dilemma nicely,
“Fed policymakers have largely been mum about their plans — and probably for good reason. Senators Elizabeth Warren and Sherrod Brown said in a letter to U.S. regulators this month that extending relief to banks would be a “ grave error.” As my Bloomberg Opinion colleague Brian Chappatta notes, this puts the Fed in a somewhat impossible position — the central bank can renew the exemption and appear beholden to the banks, or let it lapse and risk destabilizing the $21 trillion Treasury market.”
Looking ahead at the FOMC announcement on 17-March-2021, we expect the following:
- No change to the Fed’s FFR or QE volume (nor any commentary on ‘twist’).
- 5-10bps hike to IOER (a technical adjustment to pull the effective FFR towards its mid-band)
- Uncap the o/n RRP inflow limit (current: $30bn per MMF), with possible hike by 5bps (to ensure GC / SOFR doesn’t go & stay negative)
- SEP revised steeply higher, with inflation running just above 2%
- Immaterial changes to dot plot in pointing to a rate hike
- Revision to dot plot median of longer-run projection to 3% (raising the UST30yr yield threshold from 2.5%)
Clearly, our focus is on whether the Fed extends the SLR exemption.
If the Fed doesn’t extend the exemption, we believe this would be hawkish. The SLR exemption rollover debate deals with US bank balance sheets accumulating reserves (and USTs) and being included in the SLR’s denominator. If the Fed plans to taper sooner, rather than later, the accumulation of reserves on the bank’s balance sheet will begin to end. Recall: Fed QE needs to taper to $0 before rates can hike. However, if the recovery is choppy & uneven, we may see a ‘twist’ (like the Bank of Canada) or ‘taper and twist’ with the Fed lowering QE size but focus more on long-end – this could also be a technical hike as short-end would rise gradually.
The TGA’s drawdown ($1.2rn by Aug-2021) can be solved without SLR relief. If the Fed uncaps the o/n RRP inflow limit, the incoming deposits can be absorbed by MMFs, rather than bank balance sheets. In fact, this would draw more deposits away from bank balance sheets, which would alleviate their SLR constraint (since deposits are included in the SLR denominator).
Of course, we would need to hear from the Fed on its reasoning for / against extending SLR exemption before drawing concrete conclusions.
Implications?
- UST long-end continues to steepen: UST10yr to 2.25%, UST30yr to 2.85%;
- DXY recovers from 90-92 and stabilises ~95;
- Value continues to outperform growth; and
- Fund flow prefers DM over ex-China EM.