A change in the underlying plumbing of the financial system is making it unlikely that QT can run its expected 2+ year course. An ideal QT would drain liquidity in the overall financial system while keeping liquidity in the banking sector above a minimum threshold. That is only possible if the bulk of the liquidity drained is sourced from the $2t RRP, which holds funds owned by money market funds. MMFs could facilitate QT by withdrawing funds from the RRP to invest in the growing supply of Treasury bills, but recent data suggests they have lost interest in bills. Households appear to have replaced MMFs as the marginal buyer of bills and are funding their purchases out of funds held in the banking sector. This suggests QT may lower banking sector liquidity below the Fed’s comfort level much earlier than anticipated. This post illustrates the emergence of households as the marginal investor in bills, suggests the change is due to high MMF fees, and discusses its implications on the path of QT.
Households Are the Marginal Buyer
The marginal buyer of short-dated Treasuries over 2022Q3 appears to surprisingly be U.S. households. Federal Reserve data show that household purchases of Treasuries surged to record levels on a seasonally adjusted annual basis as MMFs notably shrank their holdings. Note that the “household” category includes hedge funds, but circumstantial evidence suggest that they were not major investors in Treasuries. Hedge funds are usually not major investors in bills and tend to finance Treasury purchases in the repo market, where volumes were little changed over the quarter. Regardless, hedge funds account for less than 10% of household financial assets.
The overall household balance sheet suggests that households rebalanced out of their savings accounts and riskier investments into Treasuries. Treasury bills are substitutes for savings deposits and CDs, which both are offering returns much close to 0% than comparable Treasuries. A shift by households into bills is particularly plausible as MMFs notably decreased their bill holdings even as overall bill issuance rose. Interestingly, households passed over MMFs in their search for safety and yield.
Households likely decided to pass over MMFs due to their rising fees. MMF fees follow a cycle where managers waive fees when interest rates are low to retain clients, but charge full fees when interest rates are high. One $200b MMF is even charging a 0.59% fee ($1b+ a year) to press the buy button at the RRP Facility each day. Households could do much better by owning Treasuries directly.
Better than Bills
MMF’s loss of interest in bills will make an ideal QT more difficult. MMFs have steadily rotated out of Treasury bills and into the RRP and Agency discount notes throughout the year. The expected path of the RRP offering rate and Agency DN yields are both higher than bill yields. The supply of Agency discount notes has surged in recent months as Federal Home Loan Banks increased their annual debt issuance by $500b to record levels. This allowed MMFs to invest maturing bill proceeds into DNs and maintain their RRP allocation, but the historic surge in DN issuance is unlikely to be repeated next year. MMF may go back to rolling the proceeds of maturing bills into the RRP unless bill yields rise significantly.
With MMFs showing little interest for bills, repo lending is now the only pipeline through which the RRP can be drained. Recall, MMFs are highly constrained in their investment options and can only invest in very safe assets. Other than bills and DNs, they can also withdraw money from the RRP and lend in repo. Demand for repo financing has ticked up slightly over the past few months, but it will never come close to draining the $2t in the RRP. The RRP cannot be drained without a major allocation out of the RRP and into bills.
The emergence of households as the marginal bill investor suggests that future bill issuance will be an additional drain on banking sector liquidity. While bills purchased by MMFs would have been funded by withdraws from the RRP, households fund their investments out of funds held in banks. Households have $10t in savings deposits and their continual bid can keep bill yields unattractive to MMFs even as bill issuance rises. The marginal investor in Treasury coupons the past year already appeared to be financing their purchases from funds held in banks, now the marginal bill investor is as well.
The Fed’s aggressive $95b a month QT program was premised on tremendous amounts of excess liquidity in the $2t RRP keeping bank reserve levels above an estimated $2t minimum. Our understanding of the financial system always suggested draining the RRP was unlikely, but now it is looking very unlikely. QT is effectively proceeding with bank reserve levels ~$1t above target and no excess. In this scenario QT can only last several months.
41 comments On Trapped Liquidity
Why does the banking system need a minimum reserve balance of $2 T ? reserve requirements are currently at zero.
In 2007 ( pre-QE era) banking system had reserves of around $50 Bn. And that was with a 10% reserve reauirement.
Joseph wrote that this number was calculated based on some requirements by fed. More like a magic number.
Look up LCR
Thank you !
In theory, the banks could hold T-Bills instead of reserves ( for the HQLA, Tier 1 assets). I still think in the absence of reserve requirements, the only reason the reserves are there is the history of QE since 2009.
So why cant the Fed aggressively shrink reserves and let banks decide how much they want to own in T-Bills etc to meet the LCR.? Why this pussyfooting around?
Reserves have a floating rate of interest.
You also have to take RLAP into consideration. It supersedes LCR as it treats banks as a gone concern as opposed to a going concern. This is done to ensure a bank has enough reserves to meet its intraday outflows (LCR is based on an end of day balance sheet snapshot). So this forces banks to prefund their intraday outflows which can only be settled with reserves. So banks have a preference to hold reserves in their HQLA portfolios in order to meet resolution regulatory requirements. This was one of the key issues that exasperated the repo market in 2019. Banks with Primary Dealer subsidiaries and BoNY, who had large reserve outflows in the morning due to QT, are forced to pre fund their morning outflows regardless of the estimated afternoon inflows (unlike LCR). This increases the demand for reserves. Zoltan talks about this in his global money notes #22.
Because Obama. Not even being sarcastic
With something on the order of $10 T in Treasuries maturing over the next two years, the Fed can mostly rely on running them off their balance sheet rather than actively selling securities in the market,
The biggest issue is who is going to buy the new Treasuries that will need to be issued to finance the rolloffs + an additional $2T? or so to fund the fiscal deficit.
I don’t know where you got your figures from, but their latest financial statement says they have $5.6T total in Treasuries. $3.1T of that will mature over the next 5 years. They don’t break down the maturities into 0-2 year increments.
Sure. I was referring to the total outstanding Treasuries ( not just the Treasuries owned by the Fed). Thanks.
In short, the level of reserves in the banking system is not the primary issue. The primary issue is financing the US Government.
How would an ETF such as BIL held by a household be counted in those statistics?
Why can’t the FED lower RRP offering rates to force MMFs to switch into other pastures?
Amen — 2T of disintermediation — helps no one…. either lower the caps per participant – or lower the rate….
The banks suffer no disintermediation when: “The emergence of households as the marginal bill investor suggests that future bill issuance will be an additional drain on banking sector liquidity. ”
The commercial banks can force a contraction in the size of the non-banks, and create liquidity problems in the process, by outbidding the nonbanks for the public’s savings. This process is called “disintermediation”, an economist’s word for going broke. This process destroys money velocity. The reverse of this operation cannot exist (it is predicated on the credit worthiness of the U.S. Gov’t). Transferring saved demand or time deposits through the nonbanks cannot reduce the size of the payment’s system. Deposits are simply transferred from the saver to the nonbank to the borrower, etc.
…because the Fed is concerned with containing rates well within the stated policy range. Though SOFR isn’t an explicit policy target rate, there’s a preference to not let it trade outside of the range in efforts to demonstrate its stability to the public in efforts to use it in replacing LIBOR as the new benchmark rate.
Can you explain what will happen if the RRP rate is lowered such that: RRP Rate < Bill Rate? Wouldn't this drain the RRP (driving money from the RRP to Bills)? Right now there's no incentive for a MMF to put money in bills when the RRP rate is higher than bills (hence they are net sellers). Has the Fed contemplated this?
T-bills yields are priced by the market such that they are the daily average of the expected Fed Funds rate over the life of that T-bill, give or take some day-to-day effects from market forces. (FF and RRP rates are about the same.) There should be no difference in holding a MMF for the next 3 months versus buying a 3-mo T-bill today. If the Fed were to cut the RRP rate, T-bills yields would price that in. Therefore the Fed can’t make the RRP rate less than T-bills rates; the market would just equalize them.
I think the best way to fund govt is for the treasury to issue the bills and the reserve bank of America buys the bills and then issues the digital currency to the USA treasury.
That way no interest is due and the debt can be written off in the near future.
Makes funding govt simple.
Dumb question – can’t the fed just lower the rate offered if they don’t like the flows?
How much of the RRP are deposited by Foreign institutions?
About 14%, second table shows foreign official and int. account of reverse repo. https://www.federalreserve.gov/releases/h41/20221215/
Ultimatly, it’s a question of who you believe. MMF believe in houses more than in government. Households now believe in government more than in companies. Will there be anything new on the menu that looks safe and not scarce before bill possiblely disppoint its marginal buyers?
Hey Joe, could you elaborate a bit more on how you think about the right level of reserves for smooth functioning of the banking sector? I’m under the impression that there’s ample of reserves and we are not even close to running out of them. Thanks!
Yes good one! The reserves of $3.2T are there at the banks because of QE. The banks have no control over the aggregate reserves – only the Fed does. The Fed buys securities from banks and pays for it with reserves ( which are just deposits at the Fed).
Only the banks and GSEs are allowed to own reserves ( ie. have deposits at the Fed). The RRP allows money market funds to also have “deposits” at the Fed.
The banks dont “lend out reserves” – what does that even mean? You and I cannot legally own reserves ( deposits at the Fed) – so when I get a loan from a bank , the bank does not loan me reserves – they merely credit my deposit at the bank.
In March 2020 the Fed reduced the “reserve requiremts” ( fraction of deposits that are required to be held as reserves) to Zero.
Now the banks have these reserves – they have no way to get rid of them – and why would they , after all the way the Fed “raises rates” is by simply paying the banks more money as Interest On Reserve Balances . So its nice. The Fed is paying the banks some $150 Bn/year as IORB. That covers at least some of the credit losses , bad loans , reckless speculations etc at the banks. Thats nice.
If the Fed does a lot of QT and extracts reserves from the banks, the banks will get less money from the Fed as IORB – and that may be bad since they need that money to cover all their other mar to market losses and speculative losses.
maybe that establishes a minimum reserve level?? The banks need that IORB to stay afloat?
The Fed doesn’t just give banks reserves, it gives reserves in return for banks crediting customer deposit accounts. Banks pay interest on these deposits accounts and so only earn the spread between the two rates (which admittedly can be large).
why is it considered to be of any importance whether public sector liabilities are Bills, Bonds, commercial banks’ reserves deposits at the central bank or MMFs’ deposits at the central bank (called RRPs, but effectively no different to reserves)?
They are all just Dollars issued by Uncle Sam in one form or another
I am not an expert, but presumably, it matters when the Federal Reserve is trying to shrink its balance sheets. The Fed can choose who to incentivize in order to achieve QT. Therefore, distinguishing the different forms of lending can help facilitate decision-making.
Presumably, it can make a difference in many scenarios, but those are outside of the scope of this article.
If the Fed wants to push cash out of RRPs into T-bills, then it merely needs to reduce the size of the overnight RRP operations. Fed could even end the program, forcing all $2.2 trillion into bills and bank deposits. In fact, when the time comes for the Fed to ease, this will probably be, if not its first step, certainly one of them.
As a trader and analyst observing the RRP levels, the main thing that’s important to me is that as long as these balances stay high, it’s a bearish sign for both stocks and bonds. Only if they start declining without being forced to, would it be a sign of increasing risk appetite.
Also key is that about 15% of the RRPs are held by dealers, traders, and others, who are not constrained to merely purchasing T-bills. When their share starts to drop, that would be another sign of bullishness.
None of that is happening yet, so I continue to favor shorting. However, I trade long when the technical side supports it.
Great point! Where can you see these balances as they change? Are they public information? TIA.
Looks like The FED Guy was right about liquidity
While Households do purchase Bills out of their bank savings accounts, the money spent on the purchase will presumably go into the sellers bank account, so no net change to deposits in the banking system. Exceptions to this would be if Households purchase from MMFs (who then place funds in the RRP) or if the Treasury keeps the money raised in the Treasury Government Account at the Fed, but the article does not indicate either to be the case.
My understanding is that money gets destroyed when a security held by the FED matures. The Treasury transfers the matured amount to the Fed’s account. The Fed’s balance sheet shrinks by that amount. Poof money gone.
Where did the Treasury get money to pay back the fed? The Treasury borrowed it by issuing a new government security. In case the buyer was a household using savings held in a bank to make the purchase, the money came of the buyer’s bank account went to the Treasury then went to the Fed then disappeared.
The money can reappear sometime in the future if the Fed decides to start QE.
If bills are interesting to the HH sector, why won’t the HH sector do an asset swap out of MMMFs and into bills? This would cause the MMMFs to liquidate their RRP holdings due to HH redemptions.
I can’t speak for all households. However, I have some money in an MMF. I treat it as my emergency fund since pulling money out of it is fairly easy and filing taxes for it is simple enough as well. It’s almost like a savings account but with a much higher interest rate.
The treasuries I own are liquid as well. However, selling them in a rising rates environment could result in losses. Selling them early would slightly complicate tax filing as well. I think of them like CDs. I buy them and wait for them to mature.
I’m guessing most households are thinking along similar lines. MMFs for emergency money, or almost cash waiting to buy stocks and treasuries for longer term savings.
Banks could increase their interest rates (pass on more of what they’re getting from the fed) to compete with MMFs. However, they won’t until they think they need the reserves.
Household’s might start doing that once the interest rate peaks (no more hikes or rate cuts).
A stable rate would keep the prices of short term treasuries stable as well.
Declining rates would result in a rise in the market price of existing short term treasuries making them an attractive buy if rates are expected to fall further (lock in a higher rate and not lose money in case of a sale)
I disagree with the notion that the MMF’s have permanently lost interest with Treasury Bills. Every player in the money market lost interest in Bills this year because they are rich due to supply dynamics (record tax receipts from funny money driving down issuance). That does not mean that dynamic will continue. A lot will depend on the path of the fiscal authorities and the debt ceiling but if the Treasury increases Bill issuance like they should, we could see a violent repricing of Bills and a return of interest from the MMF’s. Bills still have cash management advantages and a duration component that could come in handy as we enter late cycle.
Your charts from the HH sector have left out the increase in Checkable Deposits, which basically is the same as the Time Savings Deposits represented here. I’m curious why you left off this important category in your selection?
Why can’t the Fed reduce RRPs? Because RRPs are the means to control the floor for interest rates. The Fed will do/must do unlimited amounts of RRPs to stop rates falling.
Some observers suggest that interest on reserves will stop rates falling. But the problem there is that the amount of reserves is not unlimited. They can only be increased by the Fed buying more treasurys.
In the old days, the reserve shortage system combined with zero interest rate on reserves was very effective for controlling rates with a small balance sheet.
The new system of excess reserves needs unlimited balance sheet to stop rates falling. If the Fed wants to keep rates high when the rest of the market thinks they should be lower then, at the limit, they become the only borrower at the high rate and, theoretically, would have to RRP the whole multi trillion cash market to keep rates high.
Ps. Once the Fed has reduced reserves to a minimum to make way for increasing RRPs, then the Fed has to start buying more treasurys. I’ve always wondered if the Fed, under this new system, can actually win the battle with the markets to raise rates. Maybe the FedGuy can tell us!
Hi Joseph, the article begins by saying RRP is liquidity in the Financial system. However, the basic liquidity function floating around = WALCL (All Liabilities) – WLRRAL (RRP) – WDTGAL (TGA), which suggests the RRP is not part of the financial liquidity, ie its already locked up. Can you help clarify this pls?