"So SVB was handcuffed, and when it needed money to pay depositors it could not sell the bonds that were in HTM. ... the Fed set up a loan fund for banks with HTM assets, so they will not have to sell them if they need cash."
When I heard about SVB and their HTM problem and how widespread it was, I came up with this idea as a cynical and snarky joke. "Mark to Government". But they actually did it!
This is a big deal. It neutralizes the immediate crisis. Someone else can deal with the resulting crisis down the line.
It means all the banks which face the same failure mode now have a formal backstop and bailout mechanism at the Fed. Of course the maximum amount of such bailouts is ... (checks notes about USG credibility and discipline in such situations) ... yeah, effectively unlimited. Intsead of just giving implicit unlimited guarantees to depositors, we can give more explicit unlimited guarantees to banks holding bonds too.
If (1) you are committed to maintaining the HTM accounting gimmick, but (2) lots of banks face rate risk and (3) might need to sell assets in the market to raise cash for withdrawls or runs, but (4) can't or won't sell HTM assets because of lower market rates and regulations, and (5) you don't want them to go bust, then (6) You let them do the equivalent of 'selling'* to the government directly instead, and for HTM notional valuations.
(*that is, using them as collateral for indefinite duration, low-interest loans of HTM-amounts of cash.)
Perma-TARP for bank-held bonds.
It's starting to feel like what those Latin American countries do when creating multiple official currency exchange rates for different purposes. "Well, it's 200 Argentine Pesos to the dollar on the market, but 150 if you are buying streaming services, or 300 if you are going to a concert ... "
I will restate my old and hopeless proposal to disaggregate finance and eliminate modern 'banks' as we know them.
If you want to hold cash deposits and just have access to the global fiat currency exchange, payments, and transactions system, that is a function which can be performed for a small fee by 'full reserve deposit' "narrow banks", or, heck, the Federal Reserve itself, with outsourcing of the annoying customer service to companies specializing in annoying customer service. "CashBase".
If you want to get a consumer loan, you can go to a consumer-loan-making finance company. That finance company raises the money to make those loans in the way any other company raises money, normal equity and debt. Same for business loans, whatever.
If you want yield, you need to invest, as in, go to a brokerage or hedge fund or whatever, and buy funds and stocks and so forth, and take your chances in the big casino. You don't want to take investment losses? Don't invest, use a "Payments-Only" institution. You want yield? Accept risk.
That is, banks are like the old newspapers which are still aggregating a bunch of separable functions together, and the aggregation of those functions creates terrible incentives and huge risks which are completely avoidable. Instead of Anti-Trust, we need Anti-Bank. "Split Them Up!"
Mar 18, 2023·edited Mar 18, 2023Liked by Arnold Kling
The bottom line on all this is that if the Fed tries to remove the inflationary stimulus that it created during the pandemic then the people that received that stimulus will take huge loses and in many cases be insolvent.
So it won't try to remove it. It will cry uncle. Inflation will see saw up and down while the baseline it regresses towards gradually increases (that's what happened in the 70s, it wasn't a straight line).
Had the Fed not monetized the pandemic spending such spending would have been unpopular and probably lost political support. A great deal of the spending happened on party line 51-50 votes, this isn't hard to imagine. Lockdowns also would have been much less popular if people weren't getting checks for doing nothing.
The Tyler's of the world will write articles about how a little inflation ain't so bad while they receive contract mandated COLA increases. Meanwhile, regular people will have to deal with the deadweight loss of the price signal breaking down as a way of coordinating economic activity.
Oops: rating agencies again behind the curve; and stress tests did/do not test for a scenario of rising interest rates, but whether a bank can still operate in an economic scenario of severe recession/depression( and that actually is hard to imagine, but...); but not one of stagflation.
Bet is bankers stress test their borrowers for ability to handle a rising interest rate environment and even require them to use interest rate swaps, hedges of various forms to box the risk; but seems banks not do that for their portfolios in too many cases.
The difficulty with depending on regulators to take the appropriate steps is that they don’t want to get blamed for anything. I’m reminded of Merton Miller’s response to the idea of a circuit breaker on the stock market after the 1987 meltdown. “No regulator is going to pull the big red lever. Daddy, what did you do in the Great War on speculation? I wiped out a trillion dollars of wealth. It isn’t going to happen.” This is why market-based tripwires are necessary. They are easily observable by all. They aren’t susceptible to manipulation. The regulators don’t get blamed. The only downside is it reduces the opportunity for rent-seeking.
Arnold: Analysis and balanced commentary like this is why financial tyros like me should become subscribers. I learn something almost every post. Thanks.
Agree it is point of view. Causes: 1) SVB management (permitted by stockholders) to take on the risk of investing _so much_ of uninsured deposits in _fixed rate_ securities. [This more or less correspond to the "too fast" growth story as a less risky investment strategy, deposit attraction strategy would have constrained growth.] 2) Regulators must have been aware of the interest rate mismatch, fast growth and, notwithstanding the 2018 exemption from regulation like larger banks, _could_ have said "boo." 3) The 2018 exemption, which I believe SVB was an active lobbyist for, signaled the regulators that Congress wanted to allow such banks to take on more risk. 4) The Fed for allowing inflation, useful for adjusting to supply chain and oil price shocks, to go on too long before starting to take action (at least when TIPS moved above target) 5) Structural deficits (which _I_ blame more on too little and to high DWL taxation) slowed growth of the economy over the last 75 years.
Re: it is difficult to nail down a causal story that everyone can accept. [...] I am more inclined to blame government actors than private actors. But you can make the call the other way and not be definitively wrong."
A full narrative explanation of the SVB disaster would include each and every cause that increased the probability of disaster.
More ambitiously, a full narrative explanation might provide an ordinal ranking of these various causes according to their impacts on the probability of disaster. For example, perhaps loose fiscal and monetary policy had more impact than did any lack of due diligence ex ante by SVB depositors.
Even more ambitiously, a full narrative explanation would estimate how much each cause increased the probability of disaster. For example, perhaps loose fiscal and monetary policy increased the probability of disaster 40%; and lack of due diligence ex ante by SVB depositors increased the probability of disaster 5%; and so on, for each cause.
Conceivably, a narrative explanation might identify a set of causes, which, taken together, were *sufficient causes* for the disaster to occur, even if none of the causes was necessary.
Alternatively, a narrative explanation might also identify *necessary causes* of the disaster. For example, perhaps, absent loose fiscal and monetary policy, the disaster could not have occurred; whereas, perhaps, disaster could have occurred regardless of depositor due diligence (even though some lack of due diligence perhaps did increase the probability of disaster.)
I have no expertise in financial institutions, regulations, or economics. I have come to trust Arnold Kling's expertise and judgment in each of these areas. His narrative explanation of the SVB disaster strikes me as comprehensive, as a list of causes and mechanisms. I wonder how he might rank the various causes, according to their impacts on the probability of disaster. And I wonder if he would find it coherent and worthwhile to try and draw subtle distinctions between necessary causes and sufficient causes.
Blame is conceptually distinct. It addresses the question: Did persons—especially persons in positions of authority or power—fail to meet standards of appropriate motivation, cognition, and/or rationality? A most relevant criterion is Henry Sidgwick's: "My station and its duties." On the depositor side, there are collective-action problems and perhaps also sub-culture issues in tech venture capital.
Here is a paper I wrote a couple of years ago influenced in part by the "easy to fix / hard to break" distinction, an analytical idea worth Klinging to:
To the best of my knowledge, all banks borrow short and lend long. I've never understood how that works as well as it does. I suppose I'd need much more involvement in banking to have much chance of really understanding. Maybe not then.
Thanks for sharing a link to your plan. I like some ideas better than others. Without going into the details, I see no reason to think it would work better than what we have now. I'll note that your mention of vouchers moved me a little further towards a basic income. I haven't been following you that long and wonder if there's a link to your thoughts on that topic.
While not a solution, your mention of the dog that barked made me think of The Checklist Manifesto. It seems that might have some value in forcing a deeper consideration of concerns about a troubled bank. Make the red flags hard coded. Of course then banks in trouble but not raising red flags would be almost certain to be missed. Oh well.
... And that somehow leads me to shuttle disasters. The problems were known and dismissed as not important, just like with SVB. It raises the question whether humans are equipped to deal with precursors to what are in the end relatively rare events (I'd bet most regulators have never been directly involved in overseeing a bank that failed) and the observation that most banks that look at risk probably in the end don't fail. Where would we be if inflation had dropped way more in the last few months? I suspect correctly identifying the ones that will fail early enough but not too early is more difficult than it looks from the outside.
I would also add that mark to market accounting can be nefarious in not differentiating between the different assets a bank may hold. There is a difference between the US gov as an obligor vs some commercial enterprise, and M-T-M accounting does nothing more than address what a banking company might be worth on any given day were it to be liquidated and market prices for assets reflective of liquidation prices. For a going concern this is problematic if creditworthiness of the obligor is unimpaired.
Agree. The system has no/ not sufficient feedback mechanisms to encourage learning.
History repeating. In Europe it is CS (Credit Suisse). People knew at least in November that it would not lead to a happy end. Hans-Werner Sinn calls it Casino-Capitalism, a somehow unlucky combination of words addressing both sides though. It depends on your point of view what you can see and what you want to emphasize. Incentives matter, always.
I still have a doubt on this story. The asset side looked healthy with Treasuries and Agencies so why the repo window was not exploited? The haircut would have been severe due to mark to market but still...
I was lucky to grow up in what came to be known as Silicon Valley, part of the wave of families of engineers who flooded the suburbs beginning in the late 1950s. To me, the SVB story is yet another sign that Silicon Valley is well past its heyday. I started studying Russian in school there, partly to escape the boredom, which brings me to your jarring analogy between SVB's hiring and promotion practices and 'the way Russians are throwing raw recruits into the war against Ukraine.' Earlier this week, the WaPo ran an article reporting with eyewitness details that it is Ukraine that is running short of skilled troops as well as munitions, and is therefore the one that is 'throwing raw recruits' into the war with Russia. Of course, this was no surprise to anyone reading alternative sources, but when a story like that runs in the WaPo, it is a sign that part of the official narrative is changing. Don't be surprised if the war turns out to be another shoe waiting to drop.
"So SVB was handcuffed, and when it needed money to pay depositors it could not sell the bonds that were in HTM. ... the Fed set up a loan fund for banks with HTM assets, so they will not have to sell them if they need cash."
When I heard about SVB and their HTM problem and how widespread it was, I came up with this idea as a cynical and snarky joke. "Mark to Government". But they actually did it!
This is a big deal. It neutralizes the immediate crisis. Someone else can deal with the resulting crisis down the line.
It means all the banks which face the same failure mode now have a formal backstop and bailout mechanism at the Fed. Of course the maximum amount of such bailouts is ... (checks notes about USG credibility and discipline in such situations) ... yeah, effectively unlimited. Intsead of just giving implicit unlimited guarantees to depositors, we can give more explicit unlimited guarantees to banks holding bonds too.
If (1) you are committed to maintaining the HTM accounting gimmick, but (2) lots of banks face rate risk and (3) might need to sell assets in the market to raise cash for withdrawls or runs, but (4) can't or won't sell HTM assets because of lower market rates and regulations, and (5) you don't want them to go bust, then (6) You let them do the equivalent of 'selling'* to the government directly instead, and for HTM notional valuations.
(*that is, using them as collateral for indefinite duration, low-interest loans of HTM-amounts of cash.)
Perma-TARP for bank-held bonds.
It's starting to feel like what those Latin American countries do when creating multiple official currency exchange rates for different purposes. "Well, it's 200 Argentine Pesos to the dollar on the market, but 150 if you are buying streaming services, or 300 if you are going to a concert ... "
I will restate my old and hopeless proposal to disaggregate finance and eliminate modern 'banks' as we know them.
If you want to hold cash deposits and just have access to the global fiat currency exchange, payments, and transactions system, that is a function which can be performed for a small fee by 'full reserve deposit' "narrow banks", or, heck, the Federal Reserve itself, with outsourcing of the annoying customer service to companies specializing in annoying customer service. "CashBase".
If you want to get a consumer loan, you can go to a consumer-loan-making finance company. That finance company raises the money to make those loans in the way any other company raises money, normal equity and debt. Same for business loans, whatever.
If you want yield, you need to invest, as in, go to a brokerage or hedge fund or whatever, and buy funds and stocks and so forth, and take your chances in the big casino. You don't want to take investment losses? Don't invest, use a "Payments-Only" institution. You want yield? Accept risk.
That is, banks are like the old newspapers which are still aggregating a bunch of separable functions together, and the aggregation of those functions creates terrible incentives and huge risks which are completely avoidable. Instead of Anti-Trust, we need Anti-Bank. "Split Them Up!"
The bottom line on all this is that if the Fed tries to remove the inflationary stimulus that it created during the pandemic then the people that received that stimulus will take huge loses and in many cases be insolvent.
So it won't try to remove it. It will cry uncle. Inflation will see saw up and down while the baseline it regresses towards gradually increases (that's what happened in the 70s, it wasn't a straight line).
Had the Fed not monetized the pandemic spending such spending would have been unpopular and probably lost political support. A great deal of the spending happened on party line 51-50 votes, this isn't hard to imagine. Lockdowns also would have been much less popular if people weren't getting checks for doing nothing.
The Tyler's of the world will write articles about how a little inflation ain't so bad while they receive contract mandated COLA increases. Meanwhile, regular people will have to deal with the deadweight loss of the price signal breaking down as a way of coordinating economic activity.
Oops: rating agencies again behind the curve; and stress tests did/do not test for a scenario of rising interest rates, but whether a bank can still operate in an economic scenario of severe recession/depression( and that actually is hard to imagine, but...); but not one of stagflation.
Bet is bankers stress test their borrowers for ability to handle a rising interest rate environment and even require them to use interest rate swaps, hedges of various forms to box the risk; but seems banks not do that for their portfolios in too many cases.
The difficulty with depending on regulators to take the appropriate steps is that they don’t want to get blamed for anything. I’m reminded of Merton Miller’s response to the idea of a circuit breaker on the stock market after the 1987 meltdown. “No regulator is going to pull the big red lever. Daddy, what did you do in the Great War on speculation? I wiped out a trillion dollars of wealth. It isn’t going to happen.” This is why market-based tripwires are necessary. They are easily observable by all. They aren’t susceptible to manipulation. The regulators don’t get blamed. The only downside is it reduces the opportunity for rent-seeking.
Arnold: Analysis and balanced commentary like this is why financial tyros like me should become subscribers. I learn something almost every post. Thanks.
Agree it is point of view. Causes: 1) SVB management (permitted by stockholders) to take on the risk of investing _so much_ of uninsured deposits in _fixed rate_ securities. [This more or less correspond to the "too fast" growth story as a less risky investment strategy, deposit attraction strategy would have constrained growth.] 2) Regulators must have been aware of the interest rate mismatch, fast growth and, notwithstanding the 2018 exemption from regulation like larger banks, _could_ have said "boo." 3) The 2018 exemption, which I believe SVB was an active lobbyist for, signaled the regulators that Congress wanted to allow such banks to take on more risk. 4) The Fed for allowing inflation, useful for adjusting to supply chain and oil price shocks, to go on too long before starting to take action (at least when TIPS moved above target) 5) Structural deficits (which _I_ blame more on too little and to high DWL taxation) slowed growth of the economy over the last 75 years.
Re: it is difficult to nail down a causal story that everyone can accept. [...] I am more inclined to blame government actors than private actors. But you can make the call the other way and not be definitively wrong."
A full narrative explanation of the SVB disaster would include each and every cause that increased the probability of disaster.
More ambitiously, a full narrative explanation might provide an ordinal ranking of these various causes according to their impacts on the probability of disaster. For example, perhaps loose fiscal and monetary policy had more impact than did any lack of due diligence ex ante by SVB depositors.
Even more ambitiously, a full narrative explanation would estimate how much each cause increased the probability of disaster. For example, perhaps loose fiscal and monetary policy increased the probability of disaster 40%; and lack of due diligence ex ante by SVB depositors increased the probability of disaster 5%; and so on, for each cause.
Conceivably, a narrative explanation might identify a set of causes, which, taken together, were *sufficient causes* for the disaster to occur, even if none of the causes was necessary.
Alternatively, a narrative explanation might also identify *necessary causes* of the disaster. For example, perhaps, absent loose fiscal and monetary policy, the disaster could not have occurred; whereas, perhaps, disaster could have occurred regardless of depositor due diligence (even though some lack of due diligence perhaps did increase the probability of disaster.)
I have no expertise in financial institutions, regulations, or economics. I have come to trust Arnold Kling's expertise and judgment in each of these areas. His narrative explanation of the SVB disaster strikes me as comprehensive, as a list of causes and mechanisms. I wonder how he might rank the various causes, according to their impacts on the probability of disaster. And I wonder if he would find it coherent and worthwhile to try and draw subtle distinctions between necessary causes and sufficient causes.
Blame is conceptually distinct. It addresses the question: Did persons—especially persons in positions of authority or power—fail to meet standards of appropriate motivation, cognition, and/or rationality? A most relevant criterion is Henry Sidgwick's: "My station and its duties." On the depositor side, there are collective-action problems and perhaps also sub-culture issues in tech venture capital.
Here is a paper I wrote a couple of years ago influenced in part by the "easy to fix / hard to break" distinction, an analytical idea worth Klinging to:
https://sites.krieger.jhu.edu/iae/files/2020/12/Financial-Firefighters-Manual.pdf
To the best of my knowledge, all banks borrow short and lend long. I've never understood how that works as well as it does. I suppose I'd need much more involvement in banking to have much chance of really understanding. Maybe not then.
Thanks for sharing a link to your plan. I like some ideas better than others. Without going into the details, I see no reason to think it would work better than what we have now. I'll note that your mention of vouchers moved me a little further towards a basic income. I haven't been following you that long and wonder if there's a link to your thoughts on that topic.
While not a solution, your mention of the dog that barked made me think of The Checklist Manifesto. It seems that might have some value in forcing a deeper consideration of concerns about a troubled bank. Make the red flags hard coded. Of course then banks in trouble but not raising red flags would be almost certain to be missed. Oh well.
... And that somehow leads me to shuttle disasters. The problems were known and dismissed as not important, just like with SVB. It raises the question whether humans are equipped to deal with precursors to what are in the end relatively rare events (I'd bet most regulators have never been directly involved in overseeing a bank that failed) and the observation that most banks that look at risk probably in the end don't fail. Where would we be if inflation had dropped way more in the last few months? I suspect correctly identifying the ones that will fail early enough but not too early is more difficult than it looks from the outside.
I would also add that mark to market accounting can be nefarious in not differentiating between the different assets a bank may hold. There is a difference between the US gov as an obligor vs some commercial enterprise, and M-T-M accounting does nothing more than address what a banking company might be worth on any given day were it to be liquidated and market prices for assets reflective of liquidation prices. For a going concern this is problematic if creditworthiness of the obligor is unimpaired.
Agree. The system has no/ not sufficient feedback mechanisms to encourage learning.
History repeating. In Europe it is CS (Credit Suisse). People knew at least in November that it would not lead to a happy end. Hans-Werner Sinn calls it Casino-Capitalism, a somehow unlucky combination of words addressing both sides though. It depends on your point of view what you can see and what you want to emphasize. Incentives matter, always.
I still have a doubt on this story. The asset side looked healthy with Treasuries and Agencies so why the repo window was not exploited? The haircut would have been severe due to mark to market but still...
I was lucky to grow up in what came to be known as Silicon Valley, part of the wave of families of engineers who flooded the suburbs beginning in the late 1950s. To me, the SVB story is yet another sign that Silicon Valley is well past its heyday. I started studying Russian in school there, partly to escape the boredom, which brings me to your jarring analogy between SVB's hiring and promotion practices and 'the way Russians are throwing raw recruits into the war against Ukraine.' Earlier this week, the WaPo ran an article reporting with eyewitness details that it is Ukraine that is running short of skilled troops as well as munitions, and is therefore the one that is 'throwing raw recruits' into the war with Russia. Of course, this was no surprise to anyone reading alternative sources, but when a story like that runs in the WaPo, it is a sign that part of the official narrative is changing. Don't be surprised if the war turns out to be another shoe waiting to drop.