A bit off-topic, but I asked GPT-4 for a one paragraph summary of your last three posts:
In a series of posts, the author discusses the issues surrounding the U.S. financial system, highlighting the impacts of Silicon Valley Bank's failure and the subsequent actions by the Federal Reserve. The author argues that the Fed's focus on interest rates rather than balance sheet policy has left the financial system vulnerable to liquidity stress. Since the pandemic, Quantitative Easing (QE) has driven money growth, while Quantitative Tightening (QT) is now causing liquidity problems in the banking sector. To address these issues, the author proposes "Operation Squeeze," which involves forcibly reconnecting money market funds and banks by reducing the Reverse Repo facility counterparty limit, allowing the Fed to continue QT while maintaining or improving bank liquidity.
I, for one, feel like I have no choice but to welcome our new AI overlords
This BS AI stuff also missed that this article ignored analyzing a most important and major monetary flow loop - currency exchange rates and the flow activity from foreign central banks.
"The best solution is the simplest: forcibly reunite money market funds and banks"
That mechanic is already occurring naturally through the FHLBs. Why did RRP significantly drop 3/9 -> 3/14? Because FHLB issued a couple hundred bil of debt (to fund advances to probably mostly smaller banks looking to replace fleeing deposits). MMFs bought that debt (at the expense of RRP). Of course RRP has risen again since due to the inflows of new deposits/reserves into MMFs (117b last week) overwhelming any reallocation from RRP to FHLB debt
Mechanics aside.... I heartily disagree that there are too few aggregate reserves in the system, even when you just look at bank reserves (excluding RRP). Before SVB and the deposit runs on smaller banks there were 3T in bank reserves (~3.4T today). In Sept of 2019, when system legit snapped imo due to lack of aggregate reserves (given regulatory intraday liquidity constraints etc.) there were ~1.5T in aggregate reserves (and 0 in RRP). Sure, 30% money growth since then but also some auto-reserve printing capabilities like SRF as well. Anyways we shouldnt need x2 the bank reserves (again not even considering the 2.2T reserve tank we also have in the RRP). Bottom line, its not an aggregate reserve problem imo, its a deposit/funding problem for banks with problematic asset profiles and a collapse in trust from their depositors (particularly their uninsured ones) intermixed with a desire of those depositors to not get basically 0 on the bank deposits in excess of their operational liquidity need.
Regardless, still love this article and how it explains the setup, just disagree with what i think is a fundamental premise behind your conclusion and thought you might find my perspective thought provoking.
Anyways keep up the great work! It is much appreciated.
John - your thoughts are always welcomed, as you come to the table with some of the most data driven and informed perspectives out there. The mechanism around FHLB diverting cash from the RRP is interesting and frankly I need to explore it more.
The question of "adequate reserves" is tough to answer. Yes banks got down to $1.57tr in Sept 2019, but recall that the Fed thought that the adequate threshold reserves level should have been lower at that point. Even the bank manager survey suggested it should be well lower. And yet, we know what happened.
Extrapolating from the bank asset growth since 2019, you're right that we would expect the threshold to be in the low $2 trillions now. The number I had in my mind was ~$2.5tr. So I personally was a bit surprised at the level of liquidity stress we have already seen.
But I think you can look at it from a top-down or bottoms-up perspective. Bottoms-up, we can say that there "should" be adequate reserves, but top-down if banks are borrowing $392b from the Fed, its fair to say that there is a problem. I guess the point is that you find the level of reserves by hitting the wall.
Clearly the distribution of reserves (median vs. mean) is a huge factor here, and as reserves start to become more scarce you will see more hoarding of reserves. It's also not clear that cash squeezed from the RRP would go to the right places anyway. I meant to address this point in the article, but left it for another day (or the comments section :) ).
I still disagree about the characterization of the SRF and RPP. These are the "theoretical" sources of cash for the banking sector, but if they aren't providing liquidity in practice its not of much use. By this logic, the Fed can create unlimited reserves, and so there should never be a liquidity shortage ever. In the long run this is true, but it doesn't prevent crises from popping up to prompt intervention.
All of these points are totally valid, and admittedly the issue is more complex than one article can express. My big picture point I think is still valid. Why is the Fed borrowing from MMFs and lending to banks?
Could the answer to why is the Fed borrowing from MMFs and lending to banks be:
1. So that they can ensure transmission of their policy rates to the financial system (and hopefully influence credit creation to meet their inflation objectives and
2. Prevent liquidity issues at certain banks and the contagion issues that might bring
If they restricted access to the RRP, they could solve issue #2 but at the cost of #1.
Absolutely - for #1, the RRP has worked well at its primary objective, which is to ensure an overnight rate floor (at least until reason). And since their policy has been intentionally very restrictive, the RRP has removed a ton of liquidity, which brings about its own issues.
"Squeeze" would be the opposite. trading off #1 because #2 seems more pressing.
Agree Fed was caught off guard in sept 2019 but not everybody was. It is absolutely striking to me how well Zoltans Global Money Note 22 described what would happen (in all his gory/glorious detail) published May 31 2019, 3.5 months in advance of it happening.
It is certainly true that some banks are borrowing significant reserves from the FED. On the flip side though, some other banks (gsibs) are flooded with them. This is notably different than 9/2019 when all banks save jpm were merely adequate (to meet intraday liquidity requirements) at best and when jpm went from awash in reserves to just adequate (some say Dimon became stingy to prompt a response and they could be right but it would have happened eventually either way) and there was none in the RRP to tap, the cost to borrow them (effr, repo) skyrocketed. So in the aggregate, I believe reserves, both in the banking system and readily convertible into the banking system (RRP) are well in excess of whats needed for banks to settle the system each day.
But, thats cold comfort for an individual bank facing a shortage of reserves it owns because its spending them like mad to settle deposit flight and now it needs to generate more by either borrowing them or selling its assets (made worse by the fact those assets are MTM big down due to the rise in rates). So I think the real culprit are some individual bank balance sheet troubles which became liquidity problems (for those banks) due to depositors running. There is more to this (depositors profile, HTM accnt rules etc.) but I think this in essence explains why the Fed is lending reserves to banks. Its only some banks, and only because those banks needed it to meet deposit outflow/meet liquidity requirements and because those banks effectively could not borrow the reserves from other holders of reserves for various reasons, but those reasons related to the creditworthiness of the borrowing bank, not because the other reserve holders didnt have reserves to spare (like 9/2019)
Why is Fed borrowing from MMFs? Agree with you answer to Kevin. To police the floor of the range. For better or for worse its their framework.
I appreciate the dialog. Hopefully my end of it provides some value.
Its is absolutely helpful - and for the record, I think this take is probably more nuanced and accurate than the blanket way that I've described it in the article. Recall, prior to the SVB debacle, my intuition said that banks wouldn't be the problem this time because of the GSIB reserves that you're referencing.
I do think that QT mechanically shreds dollars, and those dollars have to come from somewhere. This will lead to higher leverage and lower liquidity in the banking sector.
Maybe ironically, if banks were allowed to fail, it would help address the gap between your explanation and my explanation. In other words, if the Fed was to take a hardline stance that there are adequate reserves, and if you don't have them that's too bad.
But our current leadership will not allow such a thing to happen. The expedient solution is to print liquidity for those who fucked up. therefore, if we are insuring the weakest link at a national level, then our assessment of adequate reserves must only focus on the weakest link.
In any case, I want to again stress how much I appreciate your feedback. I think we are of similar minds - neither of us with a ton of background in the topic but both eager to look at the data and *figure it out*.
Thanks for the perspective, appreciate the thought-provking experiment.
I think as you say, it's all about (finally) losing trust in Fiat. They see the harvest of what they have sown since 2008 - all the "experts" saying this time inflation and devaluation is different from 1929; 1970 might be right only in that it's way more central-bank-induced than anyone (in said central banks) admits. And the macro and economical backdrop is a very bad mixture of both of these very different periods.
After all I have read and researched, 2023 looks to be a mix of "roaring 20s ending and swinging into war economy with massive expansion" - then 1929-33 recession / depression with rapid expansion of prices runaway inflation (already happening) and potentially more than that.
To this, add the 1970s Energy Crisis/Oil Shock energy price inflation due to supply congestions in Europe with to bad political actors all around, who proved they cannot handle an energy crisis because they are pressuring to "go green" despite this scenario, shooting your own or rather the public's foot - and knee, and ass - in the process.
These factors cannot be solved by Fed QT - it would actually need massive expansion and QE to plaster over it, again.
Which would make the whole scenario worse, again.
This is not a rock and a hard place, it's a vast chasm in front and a lava stream advancing behind you.
As a starter, I appreciate the visual metaphor of the lava and a chasm! I may steal this for a future post ;)
Devaluation is a feature of the system - anyone who wants to deny that must contend with the near total loss of purchasing power of a dollar over the past century. The entire notion of "sticky" inflation is kind of ironic, because we have sticky inflation for centuries! Prices almost never go down, and if they do its not for long.
The question then, is the *pace* of devaluation (i.e. the rate of inflation) - collectively we've kind of decided that 2% per year is okay but 10% is not.
But controlling inflation involves long and lagging tradeoffs, with "economic crises" demanding immediate attention and throwing caution to the wind. For many years we have been gorging on sovereign debt because it has been so cheap. As the cost of servicing that debt increases, you can cut back on spending and raise taxes (hurting the economy, to preserve currency) or you can print more to cover it, even as that exacerbates inflation and drives debt service costs even higher. I'm not sure that we are there yet, but its a risk - and a much bigger one than the 1970s when public debt represented 40% of GDP.
Now, we can protest this, but its hard to see what success you will have. Individuals can look to alternative financial commodities (gold/bitcoin etc) or inherently valuable and inflation protected assets like stocks and real estate, if they believe that policy makers will allow inflation to run.
Kind of a winding response on my part, but thanks as always for reading and sharing your thoughts!
Hey Bear, appreciate the long winding response ;-)
Since mine was just a "comment" and a nutshell version of what I've been personally reading - mainly a mix of Ray Dalio ("Big Debt Crises" & his recent one "Changing World Order"); Zeihan's latest one "End of the World..." & after that "Disunited Nations" ... a bit late to that party); also Ben Bernanke's memoirs and many other things, this is just condensing what I believe to be the lead-up to come to its crescendo in a "big unwind" - in some way or other.
Sticky inflation, true, has always been intrinsically set within the Keynesian model; but this agreed-upon 2% target has been arbitrarily set of course - it's just too low to not make people pick up pitchforks while it is high enough to satisfy the "growth" rates. Unhinged, it wrecks havoc as we can see since 1980's (where it was caught by Volcker rate increases) 2000 (where it was contained by a deflation of stock bubbles) - but now in 2008, I kind of see they fight fire with hoses spewing oil. Bascially in 2008 they did justify QE with "low inflation" - while in 2020, they state "no other choice" - both of these instances contain vast amounts of moral hazard monetary policies.
That's what drove the frothy 2020-2022 market action. The slide downwards will be into unknown territory. With, as you also mentioned, debt now being a multiple of GDP rather than what is considered a "safe" percentage pretty much across the developed world monetary policy is bascially out of options. Turn on the hose again you will create very real inflation risks and fire up the furnace of the yield-seekers again, try to turn it off - you have what we have now.
In 2023 the moral hazard is being repeated - Swiss is being made a cuplrit for writing down A1 securities to zero, while the risk in getting these high-yield debt instruments is clear - "a full loss of principal is possible". The saqme people cry foul because they got tackled, while they just wanted to play ball without acknowledging the (however low % calculated) risks.
Gold and Bitcoin are just as flawed when viewed in context of preserving principal - the latter even more so just now in light of recent action of regulators - that would be another whole posting for sure in its own right. Protesting won't help any of that (in the short term), I agree.
The conclusion of my long-winding remark to a remark:
I agree that a mix of assets is the best bet here, real estate also has its pitfalls, but likely something you can touch beats bitcoin in 2023.
Great article. Why do you think QE is the key driver of the inflationary impulse? I think the lesson since the GFC is that credit growth (in both the public and private sector) is the main driver of inflation (see the last 12 months when inflation has been high despite QT and aggregate money supply declining). If restricting RRP means the Fed's target rates are not transmitted to the real economy, that would seem to encourage credit growth and go against their inflation goals. They don't want a banking crisis either, but it seems like the Fed is in a very tricky spot.
So, clearly its a bit more complicated that just money growth - but I think money growth is the most "obvious" factor that nevertheless remains underrated, when compared to specific micro drivers. The 2020 QE funded enormous fiscal spending that put new cash directly into folks hands. This made QE more stimulating and more inflationary.
Yes, squeezing the RRP would but inflationary, other things equal, by giving banks more capital and potentially allowing for more credit growth. My argument though is that credit growth has already slowed considerably, and wasn't the main source of money growth anyways.
The Fed is always balancing tradeoffs - often it is when they try to only solve for one problem that creates issues. This was the case with excessive QE, but could also now be the case with misdiagnosing the cause and remedy for inflation.
How quickly things have changed! The whole purpose of the ON-RRP was to create a reserve-like account for MMFs, since banks didn't want and couldn't take all the excess reserves. The main reason was SLR, it would be forcing banks to raise CET1 only to meet their SLR.
Now that banks are bleeding deposits, I think you're right that reversing that flow by restricting RRP would alleviate some of the stress.
It is certainly a head-spinning 180 from absolutely excessive reserves to reserve-scarcity in banking. Ultimatley this is a result of the Fed's excessively loose, and belatedly restrictive policy, combined with their cavalier use of their new balance sheet tools that are much more complicated than many have realized.
Even today we are now hearing calls for people to have direct access to banking from the Fed. While I understand the motivation because it removes the "FDIC/backstop" question - but what would happen to bank deposits and bank liquidity if people had the option to park money at the Fed? It would be a disaster - effectively the RRP dilemma but for retail deposits.
It’s an interesting idea. Bank deposits would evaporate, yes, so that lost bank funding would have to come through Fed borrowing. In the past the Fed would not have the capacity to manage retail deposits, clear checks, etc. With technology development, digital currencies, etc. that doesn’t seem too far fetched.
It will be a world where bank’s value would come entirely from its proprietary lending, funded by Fed funds, and not from proprietary depositor funding. We’re slowly moving in that direction anyway. SVB showed that quite convincingly: they based their franchise on the deposit side, wooing the tech industry to bank with them, while their assets were mostly plain vanilla bonds and MBSs. Didn’t work: despite their vaunted "relationships" depositors fled in a hurry. There’s just not too much value in proprietary relationship with depositors, and there is huge behavioral risk.
Fed funding is more predictable. There is interest rate risk, but it is manageable, as they generally try to telegraph their intentions. You’re right that there may likely be some unintended consequences, plus it is not clear that the Fed can be easily setup to manage deposits at mass scale. But it’s an interesting idea, I can see the world going there eventually.
You may be interested in a recent Odd Lots podcast that touched on the concept of the Fed stepping into the role your describing.
While I understand the sentiment, and agree that the Fed is moving towards more direct management of the banking sector, I am very against the notion of the Fed providing a direct depository function to individuals and then doling out lumps of capital for banks to go and make loans.
Effectively, it would let the federal government decide which industries or people deserve credit, and I don't think that the government should play that role and don't think they would do a good job of it. Under either party, the credit would flow inefficiently and towards pet projects and away from politically undesirable projects, that may still have a lot of importance.
As it is, the Fed has already stepped into the role of direct money creation and destruction via QT/QE, which has sidelined the role of interest rates, as described above. I don't think that they have proven particularly efficient or prudent in that function and doubt that their delegation of loan capital would be efficient or prudent either.
The best version of these plans is that the Fed provides a "sleeve" or backstop to depositors and everything else sort of continues as it has.
Credit allocation by the government is not a good idea, absolutely. But it doesn't have to be. Banks can have access to Fed funds on general terms, limited by the regulations that are currently in place: capital cushions, stress tests, no self-dealing, etc...
Bear, solid gold. This one really draws some things together. Thank you!! And the foot notes, I had a question about a data source -- right there in the foot notes. Top notch work, sir!
I should be better at citing sources - generally, most bank data is sourced from FRED unless otherwise noted. If taken from another source I try to provide a link back.
Remember, the repo market is divided equally between tri-party repo (centrally cleared) and bilateral repo (no clearing house). Banks tend to use centrally cleared repo. Regulators have visibility into the centrally cleared part of the repo market. So your proposed solution might work well for centrally cleared repo.
But what about the bilateral repo side of the repo market--used by non-banks (hedge and PE funds)? Regulators have no visibility into these repo transactions. So would your solution work for non-bank transactions in the bilateral side of the market?
This is a good distinction to draw. My thought is that improving bank liquidity in tri-party repo would have a downstream effect to bi-lateral repo as well as other sorts of securities lending and leveraging. It may not be a direct mechansim, but it would be pushing cash in the right direction.
Prior to the recent bank crisis, my bet was on non-bank entities to suffer a cash crunch first, because they are further away from the Fed's spigot, though clearly the issue has spilled over in banks first.
As a simplifying theorem, getting money out of the Fed and back into the private markets is the key point.
Well, to be fair, I don't think I'll be right, because I don't think the Fed will end up taking this approach.
My best guess is no change to current policy, which I expect to result in continued stress and emergency intervention, rather than root cause fixes. Hopefully I will be wrong!
Very good point. The RRP facility should be dismantled in time because it wasn't meant to be a long-term thing (much like the BTFP, right). The cash sitting there shouldn't exist at all since it embodies money-like instruments that create more money out of thin air.
QT isn't the main problem why banks are losing deposits; the problem is that banks have not raised deposit rates in order to safeguard thier NIMs (profitability), knowing that the Fed will sooner or later blink, and the Fed did blink with the BTFP facility. Financial institutions know that there is no political will in punishing grift, so they will keep pushing until the Fed has to step in more and either lower rates or make up another facility.
Nonetheless, I agree. The Fed broke one of the key mechanisms of monetary transmission with their freevolous use of QE/ZIRP. They must now fix that mechanism or we just keep digging ourselves deeper into the mess...
So I would push back a bit on the driver of reserve outflows, though I think you raise good points. Could banks have been more aggressive and forward thinking in protecting deposit outflows? perhaps. But there a lot of it is simply out of their control. QE stuffed them full of reserves, and now QT and RRP are draining them. If the Fed is shredding ~$75bn a month, it has to come from somewhere. Banks don't control MMFs allocation process, and it is hard for private entities to compete with the risk-free Fed.
The BTFP facility of course is "helpful" but for banks, but any direct Fed borrowing is still a measure of last resort, and is likely only being used by banks without alternatives.
Your point is well taken that in general banks expect the Fed to accommodate their liquidity needs and I'm sure this factors into their decision making. In the end, there will never be a single answer to explain what's happening in a complex system.
Yes, precisely. As you noted, everything started to unwind since GFC and the measures the Fed took then. The more they intervene and the more 'instruments' they create, the more the system gets distorted, its long-standing mechanisms lose power, and it becomes impossible to find equilibrium without new intervention. A self-fulfilling doomloop.
Your solution to slowly phase out the RRP is very elegant and logical. It might not solve the problem instantly but over time it will help. This is why it probably won't be implemented. I think Doomberg's theory of Antilogic (governments and regulators will always follow the least logical means to an end) still rules supreme.
I like this plan. I also think the Fed could just admit its error and fix it. Obviously switching gears from negative interest rates to aggressive QT was going to make a heap of older T-bonds from the QE era next-to-worthless. Who wants an asset earning 1.3% with a 10-year maturity when a savings account is paying 3.5%? The Fed couple simply... buy the old bonds back at face value. Voila.
Yes, it seems like many of the "problems" we see today should have been foreseeable. In part, I think they were foreseen - the Fed was aware it was blowing a duration bubble, and intended to pop it when the time was appropriate. The problem is that popping it in practice is far less palatable than it seems on paper. I also think the Fed was just late to recognize and adjust to inflation in real time which exacerbated the issue.
To your last point, this is effectively what the Fed is doing in the BTFP facility - albeit loaning against them, not buying them, and only for a year.
Thanks for exploring this - really interesting. A few thoughts I'd be curious to hear your reaction to:
-- You suggest that limiting access to RRP could limit the Fed's ability to keep policy rates as high as it would like/need. Isn't this a great indicator to use to shift QT from autopilot to active management?
-- If there are too many reserves in the system, limiting access to RRP should pressure short rates below the RRP rate, indicating you can sell more securities from the Fed balance sheet
-- If/when there are too few reserves in the system, you would see short rates creep above the RRP rate, indicating it's time to slow down selling securities
-- All this while having pretty limited incremental impact on real economic activity beyond that implied by policy rates around current levels (inverse of your point that organic credit expansion played a limited role in money supply growth)
thanks for reading ! to explore your thoughts in the order they were presented
-- Yes, the Fed could absolutely take an "active management" approach to QT. Though this doesn't seem in line with any of their current thinking on the topic. All that they have said so far is that they expect QT to run in the background for years, and they have not adjusted the run-off schedule since they laid it out in June. Plus I think it may be harder to actively manage QT if you're not sure about the adequate level of reserves necessary. Its one of those things that easier to determine in hindsight.
2/3 --- Yes I think you're describing how the RRP works on paper - when there are too many reserves, short rates fall below RRP and money flows to the RRP. To be fair this has worked well to ensure short-term rates follow the Fed's path. In fact, I would argue that removing a ton of reserves is what's required to actually achieve the massive rate hikes that we have seen.
The problem though, with this idea, is oversimplfying the financial actors involved, and their relative motivations. As I've blabbered on about for the past two years, MMFs have no incentive or fiduciary duty to ensure that banks have adequate liquidity. Their only concern is protecting their downside and finding the highest return. This is the fundamental problem - RRP is inherently more attractive than the alternatives, and it becomes *even more attractive* in times of stress.
-- I agree that this discussion is more of a plumbing discussion that one that has a material real economic impact in the short term. BUT, I think over the long term, it makes much more sense to use existing liquidity to shore up banks rather than injecting NEW liquidity. The money that MMFs park at the Fed is their money - and their investors money. The Fed can't simply decide to let that facility mature, as it could the BTFP, and make that money go away. They aren't the lender, they are the borrower. By injecting newly printed reserves via the discount window and BTFP facility, they are adding more money to the system, full stop. This doesn't make sense to me when there is a more obvious solution of re-wiring the existing money in circulation
The article presents a new Federal Reserve scheme but does not identify its impact on a most important flow loop - international currency exchange rates.
As Foreign Central Banks de-dollarize, the NY Fed is quite busy managing international exchange rates. The article ignores this most important flow loop and is therefore an incomplete analysis/proposal. After all, who is more important today, the hoi polloi in Des Moines or a Fat Man in China?
A bit off-topic, but I asked GPT-4 for a one paragraph summary of your last three posts:
In a series of posts, the author discusses the issues surrounding the U.S. financial system, highlighting the impacts of Silicon Valley Bank's failure and the subsequent actions by the Federal Reserve. The author argues that the Fed's focus on interest rates rather than balance sheet policy has left the financial system vulnerable to liquidity stress. Since the pandemic, Quantitative Easing (QE) has driven money growth, while Quantitative Tightening (QT) is now causing liquidity problems in the banking sector. To address these issues, the author proposes "Operation Squeeze," which involves forcibly reconnecting money market funds and banks by reducing the Reverse Repo facility counterparty limit, allowing the Fed to continue QT while maintaining or improving bank liquidity.
I, for one, feel like I have no choice but to welcome our new AI overlords
Wow. I mean yea, that’s it.
Pretty crazy!
This BS AI stuff also missed that this article ignored analyzing a most important and major monetary flow loop - currency exchange rates and the flow activity from foreign central banks.
"The best solution is the simplest: forcibly reunite money market funds and banks"
That mechanic is already occurring naturally through the FHLBs. Why did RRP significantly drop 3/9 -> 3/14? Because FHLB issued a couple hundred bil of debt (to fund advances to probably mostly smaller banks looking to replace fleeing deposits). MMFs bought that debt (at the expense of RRP). Of course RRP has risen again since due to the inflows of new deposits/reserves into MMFs (117b last week) overwhelming any reallocation from RRP to FHLB debt
Mechanics aside.... I heartily disagree that there are too few aggregate reserves in the system, even when you just look at bank reserves (excluding RRP). Before SVB and the deposit runs on smaller banks there were 3T in bank reserves (~3.4T today). In Sept of 2019, when system legit snapped imo due to lack of aggregate reserves (given regulatory intraday liquidity constraints etc.) there were ~1.5T in aggregate reserves (and 0 in RRP). Sure, 30% money growth since then but also some auto-reserve printing capabilities like SRF as well. Anyways we shouldnt need x2 the bank reserves (again not even considering the 2.2T reserve tank we also have in the RRP). Bottom line, its not an aggregate reserve problem imo, its a deposit/funding problem for banks with problematic asset profiles and a collapse in trust from their depositors (particularly their uninsured ones) intermixed with a desire of those depositors to not get basically 0 on the bank deposits in excess of their operational liquidity need.
Regardless, still love this article and how it explains the setup, just disagree with what i think is a fundamental premise behind your conclusion and thought you might find my perspective thought provoking.
Anyways keep up the great work! It is much appreciated.
John - your thoughts are always welcomed, as you come to the table with some of the most data driven and informed perspectives out there. The mechanism around FHLB diverting cash from the RRP is interesting and frankly I need to explore it more.
The question of "adequate reserves" is tough to answer. Yes banks got down to $1.57tr in Sept 2019, but recall that the Fed thought that the adequate threshold reserves level should have been lower at that point. Even the bank manager survey suggested it should be well lower. And yet, we know what happened.
Extrapolating from the bank asset growth since 2019, you're right that we would expect the threshold to be in the low $2 trillions now. The number I had in my mind was ~$2.5tr. So I personally was a bit surprised at the level of liquidity stress we have already seen.
But I think you can look at it from a top-down or bottoms-up perspective. Bottoms-up, we can say that there "should" be adequate reserves, but top-down if banks are borrowing $392b from the Fed, its fair to say that there is a problem. I guess the point is that you find the level of reserves by hitting the wall.
Clearly the distribution of reserves (median vs. mean) is a huge factor here, and as reserves start to become more scarce you will see more hoarding of reserves. It's also not clear that cash squeezed from the RRP would go to the right places anyway. I meant to address this point in the article, but left it for another day (or the comments section :) ).
I still disagree about the characterization of the SRF and RPP. These are the "theoretical" sources of cash for the banking sector, but if they aren't providing liquidity in practice its not of much use. By this logic, the Fed can create unlimited reserves, and so there should never be a liquidity shortage ever. In the long run this is true, but it doesn't prevent crises from popping up to prompt intervention.
All of these points are totally valid, and admittedly the issue is more complex than one article can express. My big picture point I think is still valid. Why is the Fed borrowing from MMFs and lending to banks?
Again, thanks as always for your feedback.
Could the answer to why is the Fed borrowing from MMFs and lending to banks be:
1. So that they can ensure transmission of their policy rates to the financial system (and hopefully influence credit creation to meet their inflation objectives and
2. Prevent liquidity issues at certain banks and the contagion issues that might bring
If they restricted access to the RRP, they could solve issue #2 but at the cost of #1.
Absolutely - for #1, the RRP has worked well at its primary objective, which is to ensure an overnight rate floor (at least until reason). And since their policy has been intentionally very restrictive, the RRP has removed a ton of liquidity, which brings about its own issues.
"Squeeze" would be the opposite. trading off #1 because #2 seems more pressing.
Always a trade-off and this one isn't perfect.
Thanks TLBS. I appreciate the kind words.
The FHLB/RRP dynamic can be seen pretty well
in the Dreyfuss govt MMF holdings if you compare the 3/8 holdings to the 3/15 holdings (Its the only MMF I know about that publishes holdings on a daily basis, one of those hidden data gems) https://www.dreyfus.com/products/mm/fund/dreyfus-government-cash-management.shareclass.Wealth-Shares.html#?section=expenses
Agree Fed was caught off guard in sept 2019 but not everybody was. It is absolutely striking to me how well Zoltans Global Money Note 22 described what would happen (in all his gory/glorious detail) published May 31 2019, 3.5 months in advance of it happening.
It is certainly true that some banks are borrowing significant reserves from the FED. On the flip side though, some other banks (gsibs) are flooded with them. This is notably different than 9/2019 when all banks save jpm were merely adequate (to meet intraday liquidity requirements) at best and when jpm went from awash in reserves to just adequate (some say Dimon became stingy to prompt a response and they could be right but it would have happened eventually either way) and there was none in the RRP to tap, the cost to borrow them (effr, repo) skyrocketed. So in the aggregate, I believe reserves, both in the banking system and readily convertible into the banking system (RRP) are well in excess of whats needed for banks to settle the system each day.
But, thats cold comfort for an individual bank facing a shortage of reserves it owns because its spending them like mad to settle deposit flight and now it needs to generate more by either borrowing them or selling its assets (made worse by the fact those assets are MTM big down due to the rise in rates). So I think the real culprit are some individual bank balance sheet troubles which became liquidity problems (for those banks) due to depositors running. There is more to this (depositors profile, HTM accnt rules etc.) but I think this in essence explains why the Fed is lending reserves to banks. Its only some banks, and only because those banks needed it to meet deposit outflow/meet liquidity requirements and because those banks effectively could not borrow the reserves from other holders of reserves for various reasons, but those reasons related to the creditworthiness of the borrowing bank, not because the other reserve holders didnt have reserves to spare (like 9/2019)
Why is Fed borrowing from MMFs? Agree with you answer to Kevin. To police the floor of the range. For better or for worse its their framework.
I appreciate the dialog. Hopefully my end of it provides some value.
Its is absolutely helpful - and for the record, I think this take is probably more nuanced and accurate than the blanket way that I've described it in the article. Recall, prior to the SVB debacle, my intuition said that banks wouldn't be the problem this time because of the GSIB reserves that you're referencing.
I do think that QT mechanically shreds dollars, and those dollars have to come from somewhere. This will lead to higher leverage and lower liquidity in the banking sector.
Maybe ironically, if banks were allowed to fail, it would help address the gap between your explanation and my explanation. In other words, if the Fed was to take a hardline stance that there are adequate reserves, and if you don't have them that's too bad.
But our current leadership will not allow such a thing to happen. The expedient solution is to print liquidity for those who fucked up. therefore, if we are insuring the weakest link at a national level, then our assessment of adequate reserves must only focus on the weakest link.
In any case, I want to again stress how much I appreciate your feedback. I think we are of similar minds - neither of us with a ton of background in the topic but both eager to look at the data and *figure it out*.
Cheers.
Thanks for the perspective, appreciate the thought-provking experiment.
I think as you say, it's all about (finally) losing trust in Fiat. They see the harvest of what they have sown since 2008 - all the "experts" saying this time inflation and devaluation is different from 1929; 1970 might be right only in that it's way more central-bank-induced than anyone (in said central banks) admits. And the macro and economical backdrop is a very bad mixture of both of these very different periods.
After all I have read and researched, 2023 looks to be a mix of "roaring 20s ending and swinging into war economy with massive expansion" - then 1929-33 recession / depression with rapid expansion of prices runaway inflation (already happening) and potentially more than that.
To this, add the 1970s Energy Crisis/Oil Shock energy price inflation due to supply congestions in Europe with to bad political actors all around, who proved they cannot handle an energy crisis because they are pressuring to "go green" despite this scenario, shooting your own or rather the public's foot - and knee, and ass - in the process.
These factors cannot be solved by Fed QT - it would actually need massive expansion and QE to plaster over it, again.
Which would make the whole scenario worse, again.
This is not a rock and a hard place, it's a vast chasm in front and a lava stream advancing behind you.
As a starter, I appreciate the visual metaphor of the lava and a chasm! I may steal this for a future post ;)
Devaluation is a feature of the system - anyone who wants to deny that must contend with the near total loss of purchasing power of a dollar over the past century. The entire notion of "sticky" inflation is kind of ironic, because we have sticky inflation for centuries! Prices almost never go down, and if they do its not for long.
The question then, is the *pace* of devaluation (i.e. the rate of inflation) - collectively we've kind of decided that 2% per year is okay but 10% is not.
But controlling inflation involves long and lagging tradeoffs, with "economic crises" demanding immediate attention and throwing caution to the wind. For many years we have been gorging on sovereign debt because it has been so cheap. As the cost of servicing that debt increases, you can cut back on spending and raise taxes (hurting the economy, to preserve currency) or you can print more to cover it, even as that exacerbates inflation and drives debt service costs even higher. I'm not sure that we are there yet, but its a risk - and a much bigger one than the 1970s when public debt represented 40% of GDP.
Now, we can protest this, but its hard to see what success you will have. Individuals can look to alternative financial commodities (gold/bitcoin etc) or inherently valuable and inflation protected assets like stocks and real estate, if they believe that policy makers will allow inflation to run.
Kind of a winding response on my part, but thanks as always for reading and sharing your thoughts!
Hey Bear, appreciate the long winding response ;-)
Since mine was just a "comment" and a nutshell version of what I've been personally reading - mainly a mix of Ray Dalio ("Big Debt Crises" & his recent one "Changing World Order"); Zeihan's latest one "End of the World..." & after that "Disunited Nations" ... a bit late to that party); also Ben Bernanke's memoirs and many other things, this is just condensing what I believe to be the lead-up to come to its crescendo in a "big unwind" - in some way or other.
Sticky inflation, true, has always been intrinsically set within the Keynesian model; but this agreed-upon 2% target has been arbitrarily set of course - it's just too low to not make people pick up pitchforks while it is high enough to satisfy the "growth" rates. Unhinged, it wrecks havoc as we can see since 1980's (where it was caught by Volcker rate increases) 2000 (where it was contained by a deflation of stock bubbles) - but now in 2008, I kind of see they fight fire with hoses spewing oil. Bascially in 2008 they did justify QE with "low inflation" - while in 2020, they state "no other choice" - both of these instances contain vast amounts of moral hazard monetary policies.
That's what drove the frothy 2020-2022 market action. The slide downwards will be into unknown territory. With, as you also mentioned, debt now being a multiple of GDP rather than what is considered a "safe" percentage pretty much across the developed world monetary policy is bascially out of options. Turn on the hose again you will create very real inflation risks and fire up the furnace of the yield-seekers again, try to turn it off - you have what we have now.
In 2023 the moral hazard is being repeated - Swiss is being made a cuplrit for writing down A1 securities to zero, while the risk in getting these high-yield debt instruments is clear - "a full loss of principal is possible". The saqme people cry foul because they got tackled, while they just wanted to play ball without acknowledging the (however low % calculated) risks.
Gold and Bitcoin are just as flawed when viewed in context of preserving principal - the latter even more so just now in light of recent action of regulators - that would be another whole posting for sure in its own right. Protesting won't help any of that (in the short term), I agree.
The conclusion of my long-winding remark to a remark:
I agree that a mix of assets is the best bet here, real estate also has its pitfalls, but likely something you can touch beats bitcoin in 2023.
Great article. Why do you think QE is the key driver of the inflationary impulse? I think the lesson since the GFC is that credit growth (in both the public and private sector) is the main driver of inflation (see the last 12 months when inflation has been high despite QT and aggregate money supply declining). If restricting RRP means the Fed's target rates are not transmitted to the real economy, that would seem to encourage credit growth and go against their inflation goals. They don't want a banking crisis either, but it seems like the Fed is in a very tricky spot.
So, clearly its a bit more complicated that just money growth - but I think money growth is the most "obvious" factor that nevertheless remains underrated, when compared to specific micro drivers. The 2020 QE funded enormous fiscal spending that put new cash directly into folks hands. This made QE more stimulating and more inflationary.
Yes, squeezing the RRP would but inflationary, other things equal, by giving banks more capital and potentially allowing for more credit growth. My argument though is that credit growth has already slowed considerably, and wasn't the main source of money growth anyways.
The Fed is always balancing tradeoffs - often it is when they try to only solve for one problem that creates issues. This was the case with excessive QE, but could also now be the case with misdiagnosing the cause and remedy for inflation.
Great article as always.
How quickly things have changed! The whole purpose of the ON-RRP was to create a reserve-like account for MMFs, since banks didn't want and couldn't take all the excess reserves. The main reason was SLR, it would be forcing banks to raise CET1 only to meet their SLR.
Now that banks are bleeding deposits, I think you're right that reversing that flow by restricting RRP would alleviate some of the stress.
Thanks Krassen!
It is certainly a head-spinning 180 from absolutely excessive reserves to reserve-scarcity in banking. Ultimatley this is a result of the Fed's excessively loose, and belatedly restrictive policy, combined with their cavalier use of their new balance sheet tools that are much more complicated than many have realized.
Even today we are now hearing calls for people to have direct access to banking from the Fed. While I understand the motivation because it removes the "FDIC/backstop" question - but what would happen to bank deposits and bank liquidity if people had the option to park money at the Fed? It would be a disaster - effectively the RRP dilemma but for retail deposits.
It’s an interesting idea. Bank deposits would evaporate, yes, so that lost bank funding would have to come through Fed borrowing. In the past the Fed would not have the capacity to manage retail deposits, clear checks, etc. With technology development, digital currencies, etc. that doesn’t seem too far fetched.
It will be a world where bank’s value would come entirely from its proprietary lending, funded by Fed funds, and not from proprietary depositor funding. We’re slowly moving in that direction anyway. SVB showed that quite convincingly: they based their franchise on the deposit side, wooing the tech industry to bank with them, while their assets were mostly plain vanilla bonds and MBSs. Didn’t work: despite their vaunted "relationships" depositors fled in a hurry. There’s just not too much value in proprietary relationship with depositors, and there is huge behavioral risk.
Fed funding is more predictable. There is interest rate risk, but it is manageable, as they generally try to telegraph their intentions. You’re right that there may likely be some unintended consequences, plus it is not clear that the Fed can be easily setup to manage deposits at mass scale. But it’s an interesting idea, I can see the world going there eventually.
You may be interested in a recent Odd Lots podcast that touched on the concept of the Fed stepping into the role your describing.
While I understand the sentiment, and agree that the Fed is moving towards more direct management of the banking sector, I am very against the notion of the Fed providing a direct depository function to individuals and then doling out lumps of capital for banks to go and make loans.
Effectively, it would let the federal government decide which industries or people deserve credit, and I don't think that the government should play that role and don't think they would do a good job of it. Under either party, the credit would flow inefficiently and towards pet projects and away from politically undesirable projects, that may still have a lot of importance.
As it is, the Fed has already stepped into the role of direct money creation and destruction via QT/QE, which has sidelined the role of interest rates, as described above. I don't think that they have proven particularly efficient or prudent in that function and doubt that their delegation of loan capital would be efficient or prudent either.
The best version of these plans is that the Fed provides a "sleeve" or backstop to depositors and everything else sort of continues as it has.
Credit allocation by the government is not a good idea, absolutely. But it doesn't have to be. Banks can have access to Fed funds on general terms, limited by the regulations that are currently in place: capital cushions, stress tests, no self-dealing, etc...
Bear, solid gold. This one really draws some things together. Thank you!! And the foot notes, I had a question about a data source -- right there in the foot notes. Top notch work, sir!
Thanks MrDustan! Glad you enjoyed this piece!
I should be better at citing sources - generally, most bank data is sourced from FRED unless otherwise noted. If taken from another source I try to provide a link back.
Remember, the repo market is divided equally between tri-party repo (centrally cleared) and bilateral repo (no clearing house). Banks tend to use centrally cleared repo. Regulators have visibility into the centrally cleared part of the repo market. So your proposed solution might work well for centrally cleared repo.
But what about the bilateral repo side of the repo market--used by non-banks (hedge and PE funds)? Regulators have no visibility into these repo transactions. So would your solution work for non-bank transactions in the bilateral side of the market?
Thanks.
This is a good distinction to draw. My thought is that improving bank liquidity in tri-party repo would have a downstream effect to bi-lateral repo as well as other sorts of securities lending and leveraging. It may not be a direct mechansim, but it would be pushing cash in the right direction.
Prior to the recent bank crisis, my bet was on non-bank entities to suffer a cash crunch first, because they are further away from the Fed's spigot, though clearly the issue has spilled over in banks first.
As a simplifying theorem, getting money out of the Fed and back into the private markets is the key point.
Thanks for the clarification of your thinking. Hopefully you will be right!
Well, to be fair, I don't think I'll be right, because I don't think the Fed will end up taking this approach.
My best guess is no change to current policy, which I expect to result in continued stress and emergency intervention, rather than root cause fixes. Hopefully I will be wrong!
Hey, you threw out a viable course of action—so please keep talking about it and let’s hope for the best!
Very good point. The RRP facility should be dismantled in time because it wasn't meant to be a long-term thing (much like the BTFP, right). The cash sitting there shouldn't exist at all since it embodies money-like instruments that create more money out of thin air.
QT isn't the main problem why banks are losing deposits; the problem is that banks have not raised deposit rates in order to safeguard thier NIMs (profitability), knowing that the Fed will sooner or later blink, and the Fed did blink with the BTFP facility. Financial institutions know that there is no political will in punishing grift, so they will keep pushing until the Fed has to step in more and either lower rates or make up another facility.
Nonetheless, I agree. The Fed broke one of the key mechanisms of monetary transmission with their freevolous use of QE/ZIRP. They must now fix that mechanism or we just keep digging ourselves deeper into the mess...
Thanks SK.
So I would push back a bit on the driver of reserve outflows, though I think you raise good points. Could banks have been more aggressive and forward thinking in protecting deposit outflows? perhaps. But there a lot of it is simply out of their control. QE stuffed them full of reserves, and now QT and RRP are draining them. If the Fed is shredding ~$75bn a month, it has to come from somewhere. Banks don't control MMFs allocation process, and it is hard for private entities to compete with the risk-free Fed.
The BTFP facility of course is "helpful" but for banks, but any direct Fed borrowing is still a measure of last resort, and is likely only being used by banks without alternatives.
Your point is well taken that in general banks expect the Fed to accommodate their liquidity needs and I'm sure this factors into their decision making. In the end, there will never be a single answer to explain what's happening in a complex system.
Yes, precisely. As you noted, everything started to unwind since GFC and the measures the Fed took then. The more they intervene and the more 'instruments' they create, the more the system gets distorted, its long-standing mechanisms lose power, and it becomes impossible to find equilibrium without new intervention. A self-fulfilling doomloop.
Your solution to slowly phase out the RRP is very elegant and logical. It might not solve the problem instantly but over time it will help. This is why it probably won't be implemented. I think Doomberg's theory of Antilogic (governments and regulators will always follow the least logical means to an end) still rules supreme.
As always, great work!
I like this plan. I also think the Fed could just admit its error and fix it. Obviously switching gears from negative interest rates to aggressive QT was going to make a heap of older T-bonds from the QE era next-to-worthless. Who wants an asset earning 1.3% with a 10-year maturity when a savings account is paying 3.5%? The Fed couple simply... buy the old bonds back at face value. Voila.
Yes, it seems like many of the "problems" we see today should have been foreseeable. In part, I think they were foreseen - the Fed was aware it was blowing a duration bubble, and intended to pop it when the time was appropriate. The problem is that popping it in practice is far less palatable than it seems on paper. I also think the Fed was just late to recognize and adjust to inflation in real time which exacerbated the issue.
To your last point, this is effectively what the Fed is doing in the BTFP facility - albeit loaning against them, not buying them, and only for a year.
Thanks for exploring this - really interesting. A few thoughts I'd be curious to hear your reaction to:
-- You suggest that limiting access to RRP could limit the Fed's ability to keep policy rates as high as it would like/need. Isn't this a great indicator to use to shift QT from autopilot to active management?
-- If there are too many reserves in the system, limiting access to RRP should pressure short rates below the RRP rate, indicating you can sell more securities from the Fed balance sheet
-- If/when there are too few reserves in the system, you would see short rates creep above the RRP rate, indicating it's time to slow down selling securities
-- All this while having pretty limited incremental impact on real economic activity beyond that implied by policy rates around current levels (inverse of your point that organic credit expansion played a limited role in money supply growth)
thanks for reading ! to explore your thoughts in the order they were presented
-- Yes, the Fed could absolutely take an "active management" approach to QT. Though this doesn't seem in line with any of their current thinking on the topic. All that they have said so far is that they expect QT to run in the background for years, and they have not adjusted the run-off schedule since they laid it out in June. Plus I think it may be harder to actively manage QT if you're not sure about the adequate level of reserves necessary. Its one of those things that easier to determine in hindsight.
2/3 --- Yes I think you're describing how the RRP works on paper - when there are too many reserves, short rates fall below RRP and money flows to the RRP. To be fair this has worked well to ensure short-term rates follow the Fed's path. In fact, I would argue that removing a ton of reserves is what's required to actually achieve the massive rate hikes that we have seen.
The problem though, with this idea, is oversimplfying the financial actors involved, and their relative motivations. As I've blabbered on about for the past two years, MMFs have no incentive or fiduciary duty to ensure that banks have adequate liquidity. Their only concern is protecting their downside and finding the highest return. This is the fundamental problem - RRP is inherently more attractive than the alternatives, and it becomes *even more attractive* in times of stress.
-- I agree that this discussion is more of a plumbing discussion that one that has a material real economic impact in the short term. BUT, I think over the long term, it makes much more sense to use existing liquidity to shore up banks rather than injecting NEW liquidity. The money that MMFs park at the Fed is their money - and their investors money. The Fed can't simply decide to let that facility mature, as it could the BTFP, and make that money go away. They aren't the lender, they are the borrower. By injecting newly printed reserves via the discount window and BTFP facility, they are adding more money to the system, full stop. This doesn't make sense to me when there is a more obvious solution of re-wiring the existing money in circulation
WoW!! Thank you for this! We need TLBS as the new chairman of the Federal Reserve.
Somebody start a petition!
Hi TLBS,
Great article that I was re-reading now. Can I ask about the variables and sources you use for the series on the chart on Bank Deposit Creation?
Thank you in advance. Keep up these posts!!
Best,
Christany
I should have noted the source!
Credit expansion: Total loans and leases in bank credit - https://fred.stlouisfed.org/series/TOTLL
QE/QT: Fed SOMA portfolio https://fred.stlouisfed.org/series/WSHOSHO
Bank Deposits: https://fred.stlouisfed.org/series/DPSACBW027SBOG
Thank you so much!!
Best
The article presents a new Federal Reserve scheme but does not identify its impact on a most important flow loop - international currency exchange rates.
One of the most brilliant insights so far.
Have you applied for Governor of NY Fed or Chairman?
Fed.com/careers
Thanks Sam! If this whole substack thing doesn't work out, Chairman of the Fed is a good backup.
Thanks as always for reading!
As Foreign Central Banks de-dollarize, the NY Fed is quite busy managing international exchange rates. The article ignores this most important flow loop and is therefore an incomplete analysis/proposal. After all, who is more important today, the hoi polloi in Des Moines or a Fat Man in China?