The money supply is set to contract just as investors are clamoring for cash to hide from declines in both equities and bonds. A combination of increasing MMF allocation to the RRP and QT may drain ~$1t of bank deposits by the end of the year. The Treasury’s decision to further cut bill issuance will keep money market rates very low and likely push the RRP to over $2.5t by the end of the year. Furthermore, recent history suggests QT will largely be funded by deposits held in banking system rather than the RRP. The combination of these two mechanisms suggests a net contraction in bank deposits despite elevated bank credit creation. Investors looking to hide in cash will have to compete for a shrinking pool of cash by further lowering the asking prices of their assets. In this post we describe the mechanics behind the impending rapid withdraw of cash and suggest the market rout will continue.
QT To Drain Banking System
The current configuration of the financial system suggests that QT will be funded out of the banking system rather than the RRP. QT can proceed in a number of ways, where drains from the RRP are most neutral. That occurs when money market funds (“MMFs”) withdraw money they invested in the RRP and lend it to leveraged investors to purchase Treasuries. In that case, QT would leave the level of money like assets (bank deposits/MMF shares/reserves) in the financial system unchanged. As the financial system is opaque and constantly changing, it is not obvious beforehand how it will react to QT. The most straight forward way to predict its reaction is to look at recent history.
Treasury’s 2022Q1 build up of its Treasury General Account (“TGA”) is mechanically similar to QT and was funded almost entirely from the banking sector. Once the debt ceiling was resolved last December, the Treasury increased its balances in the TGA by $600b through debt issuance. That action has a similar effect to QT, except that QT would extinguish the funds rather than leaving them in the TGA. This natural experiment suggests that the financial system as currently configured will accommodate QT by draining bank deposits and reserves (see here for details). Note that the RRP not only did not decline, but actually rose throughout 2022Q1. Leveraged Treasury investors are not increasing their activity, so a smooth QT does not appear likely.
RRP to $2.5t and Beyond
The tightening effects of QT will likely be compounded by a significant rise in RRP participation in the coming months. The Treasury has indicated it intends to further cut bill issuance by ~$600b amidst unexpectedly high tax receipts. The decline in the supply of bills and a general ‘risk-off’ sentiment has pushed money market rates to trade significantly below the expected path of the RRP offering rate. Other than a small set of “Treasury Only” funds who can only buy Treasury securities, MMFs will continue to be attracted by RRP’s relatively high return and overnight tenor. A sizable fraction of the ~$2t in Treasury and Agency securities MMF currently hold are likely to be reinvested into the RRP as they mature.
MMFs reinvesting maturing investments into the RRP drains liquidity held in the banking system. A MMF receives bank deposits from maturing investments, but those deposit are destroyed when the MMF sends those deposits over to the RRP. At the end of the day, the transaction has a similar effect to QT in that it lowers the level of cash available to banks (reserves) and non-banks (bank deposits). Note that this is a different case from increases in the RRP from MMF investor inflows, which is more neutral as non-bank investors simply swap bank deposits for MMF shares.
Closing the Exits
Cash is the only asset that has been spared from the recent market route and it is set to become scarcer. Bonds are not acting as a hedge and appear to be becoming less ‘money’ like due persistent declines in price and elevated rate vol. Investors in both bonds and stocks are reaching for cash by selling their assets, driving further asset price declines. For non-bank investors, “cash” means bank deposits. A $500b increase in the RRP and $400b in expected QT would drain the pool of bank deposits by ~$1t by year end. As the pool of bank deposits declines or stagnates, investors may need to continue to lower their selling prices to compete for the cash they want.
Note that bank credit creation continues to expand, but it does not look like it can offset the rapid tightening. Recall, banks create bank deposits when they make a loan or purchase an asset. The pace of credit creation this year has so far been historically high, but the level of bank deposits has actually declined from the rising RRP and recent tax payments. QT and further increases in the RRP will continue to push against net bank deposit creation throughout the year. Unless credit creation in 2022 rivals the historic growth seen last year, bank deposits look set for the first annual decline since the early 90s.
51 comments On Turbo Tightening
Well thought out. Would like to hear more about the MBS/Agency impact as well. Thanks
The Fed owns $2.6 Trillion in agency MBS and the majority of that are 30-years already in negative carry. I can’t imagine the Fed holding all these until maturity.
Thanks Joe, very interesting. Perhaps another channel to be watch is the velocity of reserves in the FedWire system which increasingly has been dominated by facilitating ON RRP for money market funds. This leaves less lubrication for securities settlement outside of repo – placing downward pressure on prices. I think this is restatement if your point about demand for cash – though this version applies only to “inside money”.
If banks wanted to keep the deposits for, maturing bills/treasury couldn’t they just raise their deposit rate > RRP rate? In this case, wouldn’t your recent history example be less relevant, and conditions less dependent on treasury issuance? But since large banks still have so much excess liquidity above regulatory requirements I guess they don’t need to? Changes to SLR or GSIB would have implications too.
Thanks Joseph, cool take on the RRP. I was wondering why increases in the RRP from MMF inflows is more neutral in terms of banking sector liquidity? Even though the non-bank investor swaps deposits for MMF shares, doesn’t the MMF still end up investing its deposits into the RRP thus forcing the bank to send reserves to the RRP?
“For non-bank investors, “cash” means bank deposits. A $500b increase in the RRP and $400b in expected QT would drain the pool of bank deposits by ~$1t by year end”
a 500bn increase as in 500bn RRP reinvestement correct (?) since that drains bank deposits only where new MMF inflows are more “neutral”?
Thank you!
Joe – excellent writeup. Where do you see long term treasury rates at the end of this year? How likely is 5% in your opinion?
I think Joe is thinking 4% for 10yr treasury at the end of year.
There dont seem to be a lot of MBS maturing in the near term. How is the Fed going to “roll off” MBS? Looks like they will have to sell them – no?
Yes.
If the Fed actually has to sell longer duration MBS as part of its QT , it would have a disproportionately bigger impact on the bond markets than letting near maturity Treasuries roll off ?
Absolutely. Not buying and Selling are very different and have disproportionate impact. The former mild, the latter severe (of course volumes are critical too)
When does the FED freak out at the loss of tax receipts adding to wider deficits (congress could cut spending, unlikely)? Normally large deficits provides an over supply of USTs, to sell this supply the price must come down (much lower USD). Is Fiji, Belgium or XXXstan going to buy US debt? Who then?
This is a trade balance question. China HAS to buy US Treasuries, because they have a trade surplus with the US. Replace “US Treasuries” with gold and it’s the same system as the old gold standard, it just works off collateral including but not limited to gold.
Question: If MMF holds a treasury security that matures, and UST then pays back principal, doesnt that lead to a reduction in the TGA account and not the banking system? Therefore, if MMF reinvestes monies into RRP, isnt it +RRP and -TGA ? The banking system maybe a conduit to facilitiate this, but its ultimately a wash on it? What am i missing ? Thx
I don’t think the treasury ever actually repays maturing debt, they simply roll it over, issuing new debt to replace it. the reduction in issuance we have been hearing about just means they don’t have to issue as much new debt as previously anticipated. the TGA is for other expenditures.
I don’t think they issue new debt to replace the MBS they are allowing to mature. This is what would enable the Fed’s Bal sheet to reduce, which wouldn’t happen if new debt were created. (In effect printed money would be cancelled. A bit like a tax hike, or budget surplus scenario.)
Along with questions by mutang, Jon, and MaxG above, I further ask why is RRP interest paid ignored?
FedGuy writes:
《 A MMF receives bank deposits from maturing investments, but those deposit are destroyed when the MMF sends those deposits over to the RRP. 》
Is this language calculated to press fear buttons in the reader? Don’t the MMFs receive daily interest, which increases shares (completely ignored in FedGuy’s diagrams)? And can’t MMF share holders withdraw their shares as cash on demand, pretty much?
Is the use of the word “destroy” hype?
FS Technology is a company dedicated to pcb manufacturing and pcb assembly, located in Shenzhen, China. Here, we will introduce some information about our company and our view on the circuit board industry. If you are interested in us, welcome to our website.
QT should be slowing the growth of the money supply. The “reach for cash”, as you suggest, should also cause much weaker economic conditions.
It should be reasonable to expect lower yields as in previous business cycles under this scenario.
What am I missing?
Thank you for you reply.
George Dagnino
I don’t think the increase in TGA in 1Q2022 was due primarily to debt issuance.
I think it was mainly due to sharply higher than expected tax receipts.
In any case, both scenarios result in a drain on reserves, whether be it through debt issuance or increased tax receipts.
However, coupon Issuance will probably be lower at the margin, although absolute quantum of financing needs remains very high relative to history.
Great article. Can you publish a glossary of acronyms?
To me it seems, that we’re sleepwalking into a systemic Lehman Moment, as the banking system is drained of deposits as the interest rates rise and the big money is lured into the Fed balance sheet via the MMFs.
Or am I getting it all wrong?
So – lets say the FF target gets to 2% .
So the Fed will be paying the banks ( IORB) and Money Market Funds (RRP) 2% on something like$6 T ? Thats $120 Bn/yr ?
In 2021 the Fed net interest was around $110 Bn. ( which by law it returned to the US Treasury – AKA “taxpayer”).
So in the coming 12 mos, the Fed will not be sending any money to the US Treasury , because it would have paid all its net income and then some to the banks.
So, in effect in the Post-Bernanke era , “raising rates” is sort of like a fiscal stimulus targeted at banks, with some stimmy for the Money Market Funds.
Why this is supposed to crush inflation is beyond me.
There is zero evidence ( Economists are not bothered by things like empirical evidence – they are not biologists!) for the efficacy of “rate increases” under the post-bernanke scheme – of the Fed simply writing the banks a big check and calling it IORB.
I guess we’ll find out – if the Fed paying the banks more money has the same effect as tightening in the pre-Bernanke ( reserve constrained) banking system – where the Fed would suck reserves out forcing FFR rates up and crushing business lending by the banks.
The market is behaving as if this new age approach to “tightening” is just like tightening cycles in the 1980s or 1990s. It is clearly not.
This is the first real test of the post-Bernanke Fed system. It has never been tested with 8.5% inflation etc .
The scary part if not the Fed paying the banks more money ( and calling that a “rate increase”).
The scary part is the Fed may fail to contain runaway inflation using this laughable tool of IORB.
What if the Fed paid the inflation rate as interest on individual CBDC accounts, to transmit monetary policy directly by encouraging individuals to save as inflation rises?
There is also the small issue of the Fed maintaining credibility as it runs a Negative net interest margin and racks up losses. iam sure some cute new accounting scheme will be invented to account for this ( “Contra-Asset”?).
Add to that the huge mark-to-market losses on the Fed balance sheet – that the whole world can see and easily calculate ( even if the Fed does’nt actua;lly mark-to-market).
There is no urgent cash flow crisis – since by design the Fed cant go belly up. And the accounting can always be …umm… “finessed”.
So its more a question of how far into the absurd can they push on and have everyone view them as sane.
And would major countries trust their precious foreign reserves to a system with a Central Bank with a huge multi trillion dollar negative “net worth” running huge negative multi hundreds of billions in “net income”. ?
Dont know .
On a recent RV interview you mentioned that QT started on June 15th and the next will be on the 30th. Basically mid and end of each month. Where can I find this data online or FED QT schedule?
you can read the fed h.4.1 it is posted every thursday at 430 pm eastern,
The problem is they have so far not done any QT, in fact if you read all their h.4.1 releases they have so far done QE increasing the balance sheet by 19 billion,I would love for Joseph to adress this
Except for the 2 year FRN, almost all notes and bonds (=>2Y) are issued mid month or end of the month. Therefore, they also mature mid month or end of the month. Hence, the comment June 15th and 30th.
If you want a finer (day to day running balance) understanding about the state of QT, I propose you add up all the maturities for the month and subtract all the SOMA Add On purchases for the same month. You might be surprised what you will get for June (as only 2 more auctions are left, 4 and 8 weeks in June 30, but they will settle in July).
If you go by H 4.1 remember those are Thursday snapshot figures only.
In my opinion, it is an oversimplification to track only notes and bonds as the Fed SOMA account is quite active in rolling over and buying T bills.
You can get the maturities by downloading from here: https://www.newyorkfed.org/markets/soma-holdings
Then you can look at what the Fed SOMA Rollover amounts are in TreasuryDirect: https://www.treasurydirect.gov/instit/annceresult/press/press_auctionresults.htm
Just match the dates. Good Luck.
I’m agree with Mr Wang
Hi Joseph
I really enjoy your interviews on blockworks macro , and the knowledge you bring is impressive, I wanted to ask you what you make of the fact that the fed was planning to reduce the balance sheet in June by 47 billion , but if you read the h.4.1 as of june 23 they have actually increased the balance sheet , doesen’t this show that they are not as serious as they claim to be ?
you can read the fed h.4.1 it is posted every thursday at 430 pm eastern,
The problem is they have so far not done any QT, in fact if you read all their h.4.1 releases they have so far done QE increasing the balance sheet by 19 billion,I would love for Joseph to adress this
When Bernanke got Congressional approval to pay IOER in 2008 , he was utterly convinced that inflation was never going to be an issue ( demographics, globalization etc).
The new Fed approach since 2008 is radically, qualitatively different that in 2013-2007.
Recall in 2007, bank reserves were in the $20 Bn range. ( and we had reserve requirements).
Now we have $3.1 Trillion in bank reserves and $2.5 T in RRP..
There is no , repeat no, empirical evidence that paying banks and money market funds more money, in a system with zero reserve requirements is going to anything at all to reduce inflation.
There is only one data point of a “tightening cycle” in the post-2008 period, the period from 2016-2019. If you examine GDP growth, Inflation and the stock market – they were all higher in 2019 than in 2016.
oops : that should be “1913-2007” !!
“rates” ( IORB and RRP rates) at 2% is a key thereshold. With a combined $5.5 T in reserves/RRP , a 2% rate paid by the Fed is about $100 Bn/yr which will turn the Fed net income negative.
Thats why any excursions above 2% are likely to be very short lived. I dont think the Fed wants to run a huge deficit year after year – for the first time in its history.
As far as shrinking the balance sheet – I think that too will have a very short ( months) life. The Treasury will need to start issuing large amounts of Treasuries – as tax revenues decline and their balances in the TGA get drawn down. Foreign demand for Treasuries does not seem healthy. So the only buyer is the Fed.
I have a different take which might be interesting.
So, you’re saying that high fed payout of interest will prevent the o/n rate from rising too much and foreign buying dying out will stop QT from going on for long.
My take is as follow:
1. Treasury issuance increases as you say over next several months
2. Fed isn’t buying much and foreigners slow down too
3. As a result, mid to long duration rates start bumping up (Infact the Fed could help this by reinvesting in short duration bonds (0-3 yrs for example)
4. Now, all that RRP money starts buying duration bonds since there’s finally a decent reward (i.e. yield)
5. Basically, imo RRP is money just waiting around because QE has kept duration yields too low. Higher yields will give it somewhere to go and financial conditions tighten too (which helps the Fed)
6. Of course even if what I describe actually happens, finding that “clearing” yield is going to be a roller coaster and we may not like where we end up – too high yields causing a bad recession.
Great discussion forum and excellent blog by Joseph!
Thanks. Actually I was not saying the Fed paying out higher IORB/RRP rates will prevent o/n rates from rising ! Obviously o/n rates will mirror the IORB rate.
I was simply pointing out that at an IORB higher than 2% , the Fed will be running a negative interest rate margin: ie. its payments on the reserves and RRP will exceed its receipts from its bond holdings.
This will reduce its annual payments to the US Treasury ( $107 Bn in 2021) to Zero , slightly increasing the Fiscal Deficit !
Moreover, a lot of people with cash balances will be getting a nice 2% return on cash – compared to zero a few months ago. Thats a kind of stimulus.
So – I am questioning whether these increases in IORB and RRP rates will have any effect whatsoever on inflation.
Is the Fed “hawkish” when it takes the money it would normally pay to the US Treasury ( AKA “taxpayer”) – and pays it to JP Morgan and other banks?
Sure – words can mean whatever you want them to mean!
The Fed Funds market – Before- Bernanke and Anno -Domini
Pre 2008:
bank reserves were around $20 Bn. There were reserve requirements that required banks to have reserves at least at 10% of current account deposits. When the Fed “tightened policy” – it would sell T-Bills to the banks – thereby reducing bank reserves. Banks would then have tro scramble for reserves – bidding up the FF rate. And on the margin, bank lending would decrease – because banks could not create a matching deposit to a loan – unless they had 10% in reserves. The Fed by law could not pay the banks any interest on the reserves.
Post-2008 ( like .. 2022):
Bank reserves are at $3.1 Trillion. There are NO reserve requirements. When the Fed “raises rates” – it simply pays the banks more money as IORB on the reserves. The banks say – Thank you very much. No one is scrambling for reserves – why? No one needs reserves. Bank lending in unaffected by reserves. Nothing changes.
So – at the very least people need to stop looking at the 1990s to say – well during the last 3 tightening cycles this or that happened. This is completely and utterly different to how things were prior to 2008.
Now – lets consider what is driving T-Note/Bond yields higher. its simple and nothing to do with the Fed “interest rate increases” – as described above. It has to do with the Fed promising to no longer be net-buyer of Treasuries and MBS . So bond investors have to confront the novel idea of thinking like investors – and asking what is an adequate compensation for locking your money up at a fixed rate for 10 years. Thats why.
I see serious economists wondering if the Fed would have to raise rates above the current inflation rate in order to get control of inflation .
Lets think about this.
So we have reserves + RRP at $5.5 Trillion.
If the Fed were to pay , say, 8% interest on that $5,5 Trillion, that would be $440 Bn/yr.
Where does the Fed get that money?
They get something like $100 Bn interest on the bonds they own. So they will have a net loss of $340 Bn/yr. I assume that would simply reduce their “equity” by $340 Bn while the Reserves would increase by $340 Bn.
Then the next year they would have to pay a large interest on $5.5 T + $340 Bn=$5.84 T .Which will further expand the reserve balance. You can see where this goes.
And all that money flowing to banks and money market funds – thats a massive stimulus – Banks would double or triple their dividend payments, Money Markets would pay somewhere around 8% to their retail customers. That would be a bigger stimulus than anything in 2021. Driving up the price of everything that people with Bank Stocks and Money Market Funds buy.
And … no … it wont reduce inflation – it would usher in hyperinflation if the Fed started paying 8% IORB and RRP rates.
But that money flowing into banks and mmf’s from RRP interest is at the expense of having those funds deployed elsewhere, like stocks, real-estate, etc. If you park your money in a 8% CD, you’re not spending it on housing or goods – so how would that cause hyperinflation?
So – “Raising rates” above the inflation rate is a non starter. So what about QT?
So – yeah if the Fed actually did QT in any significant amount – it would drive up bond yields and drive euity prices into the mud. And usher in a great depression.
tax revenues would plummet. Stock and House prices would collapse.
And how id the Govt supposed to finance itself? Default on Treasuries? Collapse the Dollar?
So – yeah .. no , QT may happen in baby steps for a few weeks . Then it will stop. The Fed’s primary unstated “mandate” is to keep the govt functioning and their own institution intact.
Collapse the dollar against what?
Thats always the question isnt it?
Well one scenario is the world splits into two currency blocks :
A, USD, EUR, GBP, JPY, AUD, CAD
B. CNY, INR, RUB + various Asian and middle east currencies – possibly pegged to gold or oil or a basket of commodities.
Then we might see Block A currencies depreciate against Block B currencies if they get into a spiral of govt spending enabled by central bank expansion.
Look – we are not in Kansas anymore. I have no clue where this all ends up . Just guessing.
Block B is seeing more participants! Brazil, Argentina and now possibly Saudi Arabia and Iran. If SA joins the BRICS block that would make the Brics a pretty formidable economic block.
If you look at page 15 in Chicago’s “Modern Money Mechanics” you will see that “sells of securities” decrease both the assets and liabilities of “Factors Changing Reserve Balances”
“In concept, the ON RRP facility acts like IORB for a set of nonbank money market participants. Through the ON RRP facility, eligible institutions—money market funds, government-sponsored enterprises, primary dealers, and banks—can invest overnight with the Fed through a repurchase agreement (repo).”
https://libertystreeteconomics.newyorkfed.org/2022/01/how-the-feds-overnight-reverse-repo-facility-works/
Thus, the O/N RRP facility is primarily used by the nonbanks (draining the money stock). So, the FED’s operations are not transparent. Powell is making a huge mistake.
The money supply can never be properly managed by any attempt to control the cost of credit.
Well after 2 months or so of Turbo tightening the Fed’s balance sheet has only shrunk by roughly 35 billion dollars. The monthly targets for balance sheet rolloff for June and July were 47.5 billion dollars each month. So the Fed has only achieved roughly 1/3rd of their QT target.
What can we infer about the future of QT from the results so far?
Hi colleagues, its fantastic post about teachingand completely defined, keep it up all the time.