There was much commotion last week on whether last April’s emergency SLR relief would be renewed before it’s expiration on March 31st. The “right” policy answer to this question is actually really easy, but politics matter as well. In this post I briefly explain what the SLR is, why a permanent exemption of reserves is the obvious policy answer (Treasuries are less clear), and the potential market impacts if the relief is not renewed.
What is the SLR?
The 2008 Financial Crisis was basically a giant bank run, where mortgage related assets held by the financial sector deteriorated rapidly and led to a series of panic withdraws by investors. There was a run on the primary dealers, money market funds, investment banks, commercial banks, and countless hedge funds. The official sector’s solution to this was to place limits on the size and composition of bank balance sheets. This was implemented via many layers of regulation, notably Basel III’s Liquidity Coverage Ratio, Tier 1 Capital Ratio, Net Stable Funding Ratio, and of course the Supplementary Leverage Ratio (“SLR”).
The SLR is an unweighted leverage ratio, where big banks must hold a minimum of 3% (5% for the GSIBs) Tier 1 Capital against their leverage exposure (basically, on and off balance sheet assets). “Unweighted” means SLR does not distinguish between risk free assets like reserves or risky assets like corporate loans – a bank must hold at least 3% capital against both exposures. In contrast, a risk weighted ratio like the Tier 1 Capital ratio would not force a bank to hold capital against risk free assets. An unweighted ratio is essentially a safeguard in case those risk weighted metrics do not accurately capture the risks of the asset.
Stylized example: if a mega bank has $10b in capital, it’s maximum balance sheet size under a 5% SLR would be $200b (in practice there are many details in how to calculate capital and assets). Given this constraint, the mega bank optimizes those $200b in assets to be as high return as they can be (subject to other regulatory constraints). As a result, the SLR has the perverse impact of discouraging banks from holding low yielding risk free assets like reserves.
Reserve Should be Excluded, but Maybe Not Treasuries
The purpose of all this regulation is to make banks safer by strengthening their solvency and liquidity. Central bank reserves are not only risk free, but they are literally the most liquid asset in the financial system. When everything in the financial system implodes, $1 of reserves is still a $1 of reserve and can be instantly transferred to make payments. A regulatory metric that has the effect of discouraging banks from holding reserves is at odds with the purpose of banking regulations.
In addition, a leverage ratio that includes reserves places a constraint on monetary policy. On-going QE presupposes a banking system is able to accommodate all the reserves that are created as a byproduct of asset purchases. In theory there is capacity on a systemwide basis, but the banking system will need to redistribute the surplus reserves from leverage ratio bound banks to other banks. That process may have undesirable rate impacts (discussed below). Note that this is a common problem among all central banks with massive QE: the ECB, SNB and BOJ have both temporarily exempted reserves from the SLR, while the BOE always exempted them.
The case for a renewal of the temporary SLR relief for Treasuries is less clear. Treasuries are also risk free, but though they are “money” in normal market conditions there is some question on their liquidity during a crisis. Last March the Treasury market broke and investors desperate for cash had trouble liquidating their Treasuries. This was not expected, and violates a key assumption of Basel III liquidity regulations – that sovereign debt is highly liquid. Traditional risk/liquidity metrics failed in that sense – exactly the scenario that unweighted ratios like the SLR are designed to backstop. In that context, it would make sense for Treasury holdings to incur some balance sheet cost to reflect that risk.
Non-Renewal Implication 1: Banks May Hold Fewer Treasuries
To be clear, the binding regulatory constraints each bank faces will differ depending on its business model. A large regional bank may find the Tier 1 Leverage Ratio more limiting, banks with lots financial institution clients may find the Liquidity Coverage Ratio more binding, while a big universal bank may find GSIB surcharges more binding etc. Some banks will be more constrained on the commercial bank (depository institution) level, rather than the bank holding company level (which would include other entities like its securities dealer).
The current temporary relief is on bank holding company level SLR. It’s not easy to see how binding that is because banks always manage their regulatory metrics to be comfortably above the minimum. However, the data does show that commercial banks as a whole increased their Treasury holdings from $900b last April to $1.2t (regulatory filings show JPM’s bank entity alone increased Treasury holdings by $50b). This could be a desire to increase duration amidst significant deposit growth (see TBAC presentation, slide 5), but it may also have been helped by SLR relief reducing balance sheet costs.
On the other hand, aggregate primary dealer net Treasury holdings have been stable since 2018. (Note that some banks under SLR have primary dealers subsidiaries. The Treasury holdings of the dealer subsidiary would affect the bank’s SLR. There are also primary dealers who are not under SLR). Furthermore – note that the aggregate Treasury holdings for all primary dealers are only about $250b. For context, JPM has over $3t in consolidated assets, while Citi and BofA each have around $2t. A bank holding company’s SLR is not meaningfully affected by the Treasury holdings of its dealer subsidiary. So it’s hard to see how any SLR news would meaningfully impact dealer intermediation or be related to this month’s Treasury market volatility. If anything, heightened price volatility itself should be feeding through internal risk models and mechanically limiting Treasury positions.
But even if banks end up buy fewer Treasuries it likely won’t have too much of an impact on rates. Anyone watching the central banking community can see that a paradigm shift has taken place where not just front end rates are controlled by central banks: Australia pins their 3-year, Japan pins their 10-year. Central banks are beginning to think they can pin longer term rates without side effects. Private sector demand is not required – all rates are becoming policy rates. (Note there can still be swap spread implications, as bank Treasury buyers may hedge some duration by paying fixed – widening swap spreads – while a CB buyer would not hedge. Just like MBS purchases, the official sector takes duration out the market and structurally lowers the demand for hedging.)
Implication 2: Front End Rates Will Go Lower
Without SLR relief for reserves, the balance sheets of some banks will eventually become too big as endless QE pumps reserve assets/deposits liabilities into the banking system. To improve their leverage ratio, some banks will have to optimizing their balance sheet by reducing their low yielding reserves. This is done by pushing out depositors via fees, negative rates, or just refusing further deposits.
For example, a GSIB could charge its hedge fund clients -0.5% on their deposits. In response, one of the hedge funds may withdraw say $100mm of its money and place it elsewhere. This would shrink the GSIB’s balance sheet by reducing its reserve assets / deposit liabilities by $100mm. This is a common method of pushing out unwanted deposits. (see Switzerland, Germany)
In prior years foreign banks held almost half of the banking system’s reserves. This is because some foreign banking jurisdictions (Europe in particular) calculate their leverage ratio based on quarter-end snapshots rather than the average balance sheet over the entire quarter. As a result, foreign banks would expand their balance sheet in between quarter-end dates and then shrink them on quarter-end to make their leverage ratio look good for regulators. Essentially, the regulatory costs of foreign bank balance sheets was lower than U.S. banks, so they were more willing to hold low yielding assets like reserves. (Note: foreign banks also don’t have to pay FDIC insurance fees, so that saves them a few bps and also allows them to more easily take in deposits).
However, the mechanism that channeled money into the foreign banks is broken. Back then, the mechanism was for money to flow into prime money market funds, who would then deposit it at foreign banks via time deposits for a rate slightly below interest on reserves. The foreign bank would earn the difference between IOR and the rate offered to the prime fund.
But today the prime fund complex is a shadow of its former self. Money Market Reform in 2016 essentially destroyed the industry, shrinking its AUM from $1.5t to $600b. Money market reform gave prime funds the power of limiting investor withdraws in times of market volatility, which was a deal breaker for many large investors (who wants to invest in a fund that might not return your money when you need it the most?). As a result, the pipes that could send excess deposits from balance sheet constrained domestic banks to foreign banks is broken. The banking system can be rewired, but it will take time and likely the push of negative rates.
As deposits get pushed out of SLR constrained GSIBs, they will flow throughout the system to other banks that can accommodate them. They should eventually find themselves back into foreign banks, pushing wholesale funding rates lower in the process.
Note that this is not in any way destabilizing – markets are destabilized if someone needs money and can’t find it. That investor may then may be forced to sell assets, which then puts downward pressure on asset prices, and then kickstarts a deleveraging panic. Our problem today is the opposite – there is too much money. If anything the portfolio balancing mechanisms in the process of avoiding negative deposit rates could be positive for risk assets. On the margins, someone will be pushed into IG, which may mean someone else being pushed into equities/HY.
5 comments On To SLR Or Not To SLR
interesting article, thanks for the explanation of the SLR regulation
What happens when a bank that tries to push out deposits to a money fund also banks that money fund?
Said another way, what if excess reserves that are pushed out end up coming back as a retail deposit, but still contributing to excess reserves? When the whole system has a reached its absolute limit and bumping against leverage constraints, and there is no way to invest in securities or push out deposits because everyone is full, what happens since reserves cannot be destroyed? Hot potato effect cannot be avoided… is that right way to think about it?
That is a good question and is solved by the RRP. The pushed out deposits go into a MMF, which then invests in the RRP. Reserves/deposits leave the banking system and go into the RRP. This shrinks the size of the banking system.
Thanks Joseph. However, the RRP mechanism only seems to work on a temporary overnight basis? So the next day, the reserves return to the bank banking the MMF and leverage problems for the bank continue, or if the whole system is full, then the system-wide problem is not solved. In other words, it does not seem to be a permanent reduction in reserves like QT that would alleviate the leverage constraints…
Can just roll RRP investments every day indefinitely