An aggressive quantitative tightening (“QT”) pace would set the stage for another spike in rates, but this time further out the curve. During QT, the U.S. Treasury increases its borrowing from the private sector to repay Treasuries held by the Fed. While the Fed can be repaid with cash held in either the RRP or banks, the current issuance structure suggests repayment will largely come out of the banking system. The lesson of the prior QT was that reducing the cash balances of banks directly impacts markets that were recipients of that cash. In 2019, banks were pouring their extra cash into the repo market amidst surging demand for repo financing. The repo market broke when QT siphoned that extra cash away. This time around banks have poured their cash into Treasuries and Agency MBS amidst surging issuance. In this post we explain why QT will primarily drain bank cash balances, review the September 2019 repo spike and suggest that the stage is set for a potential spike in longer dated rates.
QT Will Drain Bank Cash, Not RRP Cash
QT will reverse the QE driven growth of RRP balances and bank cash balances, but the reduction will be tilted towards bank cash balances. The $1.5t in the RRP is deposited by money market funds (“MMFs”) and can only be accessed via certain “pipes.” MMFs are mandated by regulation in invest only in safe short-term assets that include Treasury bills and Treasury backed repo loans. This means QT can only drain RRP balances if either the Treasury significantly increases bill issuance or if investors purchase Treasury coupons with financing from a repo loan. Otherwise, QT will drain cash held by banks as the additional Treasuries would be purchased by Non-MMFs with money held in banks.
MMF cash will likely not be a major source of funding for QT. The bulk of issuance used to repay Fed Treasury holdings will be in tenors that MMFs cannot purchase. This is simply due to Treasury’s goal that short-dated debt comprise 15 to 20% of its debt. Investors have also not shown interest in buying coupon Treasuries on leverage, with repo volumes essentially flat the past year despite record issuance. Together this suggests RRP balances will remain high as QT will mostly drain cash held in the banking sector.
September 2019 Repo Spike
In 2019, demand for Treasury repo financing drove repo rates higher and enticed banks to enter the market as the marginal lender. Repo volumes rose from $800b a day in 2018 to almost $1.3t on the eve of the repo spike. The higher demand pushed repo rates above interest on reserves, which is the return banks were earning on their cash balances at the Fed. Noting that repo lending and cash have nearly identical regulatory treatment, banks shifted a few hundred billion of their cash balances into repo for a little extra return.
Banks were net lenders in the repo market, not borrowers. When repo rates spiked that meant repo borrowers (hedge funds and dealers) needed cash, but not banks. If banks were at any time short in cash should could just cut back their repo lending, but they did not. An extensive survey of banks consistently found that banks reported having far more cash than they needed. The impact of QT was felt not in the banking sector, but in the markets that grew reliant on extra bank cash. Repo rates spiked in September 2019 because demand for repo financing kept growing while QT was shrinking the supply of cash into the market.
The Next Spike
The mechanics behind the 2019 repo spike suggest that another spike in rates will eventually occur. Last time banks deployed their enormous QE cash balances into repo, but this time around they are pouring it into Treasury and Agency MBS securities to the tune of $1.5t. An aggressive QT will both rapidly increase the supply of duration to the market while at the same time rapidly reduce the cash balances of banks, a key marginal buyer. The combination of a positive supply shock and negative demand shock can eventually again lead to violent dislocations.
Note that QT would also be playing out in a context of rising rate hike expectations, rising energy prices, and a market depth that has not scaled anywhere near the surge in issuance outstanding. These factors all contribute to fragile market conditions that don’t matter until they do. The proposed $100b+ a month QT is very aggressive and its impact cannot easily be priced in advance. Risk markets had no idea what repo was, but a spike in rates would shake the foundations of the market.
This is great color. Thanks!
Just curious about when you worked at Fed. I worked Open Market Desk late 79 to May 81
2016-2021 – good to meet you!
Thanks, Joseph. If the Fed embarks on an aggressive QT ($100b per month let’s say), could the Treasury front load their T-bill issuance to facilitate this (while staying within the 15-20% range) and take the pressure off the private sector? Are the Treasury and the Fed coordinated enough to manage this?
You can answer that by looking at who is in charge of the Treasury.
This actually is a very good question. Why couldn’t the Treasury neuter QT by shifting all its net issuance to bills? Sure it would look bad, but it would stop long rates from rising. Aggressive “debt management” by Treasury is no weirder an idea than QE was 12 years ago.
I think they’ve already started this because the average duration of treasuries keeps dropping.
Wouldn’t that increase the yield on the front-end and potentially cause an inversion?
My doubt is quite simple: banks are drowning in reserves right now; their balance is huge, and part of their expositions needs to be “offset” by HQLA — which was and still is kind of scarce. So, in this scenario, why the QT would not help them?
“I am giving you more collateral in exchange for your reserves”. Sure, this is a flow, and we may trespass the optimal point, but so far this seems to be the scenario in the FED’s mind, no?
Thanks, and great post as always.
So QT makes rates go higher. But as rates go higher our overleveraged economy will crash/correct 20%. A crashing stock market makes the fed execute their fed put and may even start to cut rates.
Seems silly to me, but when do you think they stop? At around 8 rate hikes or less plus full taper?
Their hands are tied. Inflation is out of control…
This inflation isn’t coming from the money side of the equation. Velocity of money is all time lows. https://fred.stlouisfed.org/series/M2V
It’s coming from the supply chain being broken. And from wage and rent increases from perceived inflation from the supply chain being broken.
Hiking rates could break the supply chain even more with layoffs and more disfunction. The relationship between rates and inflation is not linear, it’s multifaceted and in this case as Joseph has pointed out in one of his previous articles, higher rates can actually stoke more inflation.
And once velocity normalizes we’ll have runaway inflation.
Velocity can’t normalize because ordinary people really don’t have much money to spend, they are mostly trapped in debt. If you give money to normal working americans, they spend it quickly, but if you give to the rich they mostly hoard and buy stocks and bonds.
The “wealth” that the Fed creates with QE is a benefit to people who are already affluent and that’s why the velocity of money is trending down.
This is also why you see all these people dropping out from working – they gave up, nothing is affordable anymore. The affluent who benefit from low rates and QE buy up all the houses, and people who buy from income are priced out.
But I do agree with you that the system in it’s current setup is not stable and will end up in hyperinflation or a deflationary great depression style crash. It’s going to be one of them.
Velocity can’t for equality reasons and doesn’t have to because broad money supply would normally do the job to cause runaway inflation which we’ve been having for sometime now….meanwhile supply is a problem but it’s not weakening it’s just that the demand is too strong to be caught up with.
Supply chain isn’t broken its being stretched by to much demand. Ports are currently processing 15% more containers than pre Covid
Aggregate demand is the same. People are consuming more goods than services because of Covid. This change might even be structural because of the shift to remote work.
The subway here is still half empty most of the time and so are the restaurants nearby during working hours. The gyms are still empty. A lot of these businesses have gone bankrupt.
Velocity can be increased. It’s called public policy by way of universal basic income.
This is well presented
It is well written article and provides good overview.
In reality- Fed does not care about inflation, it all political ploy. After some months this elevated inflation will be norm and party will continue. No one can afford to spoil the party.
I am not a bond guy, but it strikes me that observing Fed actions through the years, they are playing with fire that they have convinced themselves is not a problem. History shows, however, that it is a problem and that we can expect a significant dislocation in some markets going forward. Not only that, but I am sure that when things go pear-shaped, the Fed will claim there was no way they could foresee the problem.
“In 2019, demand for Treasury repo financing drove repo rates higher and enticed banks to enter the market as the marginal lender…..” What caused this higher demand for Treasury repo financing? I assume it was a collateral repudiation of some kind. It is often pointed out by other authors that the cash side of repo gets too much attention. What was happening in the wider world of collateral leading up to the repo spike? Could it have been that globally synchronised growth was not what it appeared to be and sec lending, collateral transformation etc was reigned in causing demand for the best kind of collateral – on the run Treasury Bills?
Tax collection and new issuance, weaker liquidity. combined with growing speculation (gradual increase volume in repo)
I think this article by Jeff Snider provides some insight into one reason why heavy demand for Treasury Repo can dramatically increase. He has provided many others over the years which look at problems with collateral that start the process that the Fed later responds to.
An aggressive QT will both rapidly increase the supply of duration to the market while at the same time rapidly reduce the cash balances of banks, a key marginal buyer.The combination of a positive supply shock and negative demand shock can eventually again lead to violent dislocations.
Could you a little bit more explain why increasing in duration of TSE leads to positive supply shock?
the Fed has been mainly buying long end of the YC since last year to bring the yields down so companies can “survive” and get cheaper debt financing etc. When QT gains pace, then this will not be the case anymore, which is why this will be a burden on private sector – the opposite of the ” crowding out” effect. That s why yields might rise from a technical standpoint, but really if the growth stalls in the next year and inflation calms down.
I meant since Covid crash
You fail to mention the Standing Repo Facility (SRF)
The SRF will only cap short rates. One reason why another September 2019 like spike in repo is unlikely. I could be wrong, but I dont think the SRF will help cap long end yields, which is where FEDGUY is suggesting the next spike in rates will be felt .
Let’s say investors want to buy long bonds. Why not right, prices down and yields up, looks good. Plus they can hit up the SRF anytime to convert those bonds to cash. So now the Ts r cash w a yield. Issue is Basel regulations. They can’t hold the Tys on their sheets, same reason rrps blew up. Banks had to move the cash to mmfs. They can’t hold the Tsys, that’s why srfs rnt rly relevant here unless, Tsys r excluded from slr calculations.
If long term rates rise (as term premia widens out), wouldn’t it be more attractive for institutional (HFs) to receive Treasury spreads? I think your assumption that there will be no leverage demand for Tsy because there hasn’t been recently won’t hold. You’ll then see overnight repo become more expensive, and that will draw cash away from RRP.
Agreed, if long end yields creep up enough relatively (even with rate hikes), funds use repo to buy longer maturities. Plus if look back to 2015, after initial rise long end falls when QT and rates increase, as poor equity and economic outlook. So although long end DnS might be the cause the effects won’t start to show until repo volume and short end go much higher. Based on this struggle to see how a rate spike could happen on 10y, more likely usually flattening of curve happens. Love the articles Joseph, would love some clarification on this
Joseph, can you help me understand this line “ Last time banks deployed their enormous QE cash balances into repo.” I am having trouble visualizing this with a balance sheet example.
It was my understanding that banks don’t lend/reserves when they make a loan so what do you mean “deploy”? The aggregate reserves in the system shouldn’t change correct? Thanks!
The rates were so high, so they wanted to take advantage of those rates and participate in the massive speculation by using all of their arsenal… a calculated, low risk gamble.
HI Joseph, I have read your book and it is a fantastic introduction to IR markets and the US Monetary System. Thanks for that!
Maybe I am not reading things correctly but this latest post seems a bit contradictory to your previous post, “The QE Afterparty”, where you implied that 1 trillion USD would flow to the banking system through increased bill issuance being bought by MMFs. This would result in a drawdown of the RRP, with the proceeds ending up in the TGA and then banking system, as those funds are spent. This would be positive for risk assets.
It seems that you are now thinking that the opposite will take place, as bill issuance will not be enough to release this money from the RRP and non-MMF investors will need to use their bank reserves to purchase coupon issuance, thus draining reserves from the banking system. This will coincide with banks having to absorb the QT flows as well. It seems that this situation would be very negative for risk assets. Am I reading this correctly?
Yes – I wrote the earlier post before the Fed changed its tune and started pushing super QT. That is bad for risk.
I have been wondering recently 2 things, 1, why is the DXY not going up and why is the dollar not going up in the FX market and 2; why are the Fed falling over themselves to tighten as if their life depended on it? Could it be that the dollar has been so debased these past 2 years that there the demand for dollars has decreased?
Thanks for the clear exposition Joseph. I found the comments by FOMC member Raphael Bostic last week surprising. His suggestion that $1.5 tln represents pure excess reserves seems to equate RRP with ‘pure excess’. That is far from clear to me that this is the case especially in light of the distributional issues you highlight. I do understand the Fed is concerned that it may have to raise rates faster than they had expected and that accordingly they will pay more in IOER than expected – adding to the losses associated with the QE programme. If that is their priority then it explains the alacrity with which they seem to approach sharp reserve reduction. But to my eyes, IOER payments are a small political cost and the potential QE losses appear manageable with seigneurage and organic growth of currency and are spread over a large number of years. In contrast, the difficulties of assessing the boundary of a ‘surplus’ system where a breach of the unknown and unknowable limit to reserves required to maintain a functioning financial system would seem to argue for a high degree of caution.
hi Fedguy, you had a post on Aug 25 2020 on the mechanics of QE vs. M2. Do you have any views on how QT at $100b/month will affect M2?
I have a question that has arisen while reading your book: You say that Primary Dealers fall into the category of shadow banks (non-commercial-bank businesses), so I understand that they cannot hold Fed reserves on their balance sheet, but at the same time they are the entities that have the privilege of trading directly with the Fed. How does the Fed buy assets from them if Primary Dealers cannot hold central bank reserves? Thanks!
At the risk of being indelicate and/or accusatory (I don’t mean to), but what prevents the Fed, apart from it being illegal, from directly buying equities? I have zero evidence for this, of course, and I would assume that the good people at the Fed follow the law and would never violate the trust of the American people, but is the Fed setup like other secretive government agencies where things are compartmentalized and “need to know” where possible conspiracies like this be plausible, if possible at all?
although illegal to directly buy them there are loopholes around it through special purpose vehicles. Quite likely in the next crash fed will buy equities, just like Japan did, fed already bought corporate bonds this time around
Hi Joseph, I am a new reader of your blog and I would like to thank you for the fantastic analyses – I’ve a lot to learn and this has given me a great starting point in understanding more about a critical but often misunderstood segment of the markets.
My question is – there has been some concern about the flattening of the yield curve arising from the increase in short-end yields off the expectations of a more aggressive Fed tightening cycle. From my (potentially erroneous) understanding of your article, you seem to have the view that the impact of QT will primarily be on the long-end of yields? If that is the way that QT is expected to impact the market, is the implication then that market participants are pricing the impact at the wrong-end of the curve?
Thanks in advance!
What would be the expected policy response from the Fed if bank cash gets drained while RRP cash remains undeployed and rates (esp on long end) spike?
What is wrong with using the Fed’s Discount Window to fill the gap? What am I missing here?
What does RRP mean?
Powell “did it yet again” by raising the remuneration rate on interbank demand deposits. That Romulan cloaking device (payment of interest on interbank demand deposits, on a “Master Account”), vastly exceeds the level of short-term interest rates which was explicitly illegal per the FSRRA of 2006. That destroys the carry trade.