Can’t say we weren’t warned. Money and bond curves have been screaming for months trouble was lurking. A decent bout of it had erupted last fall, though hardly anyone saw it for what it was. Something about the UK, a bit of dollar activity in Switzerland. So what?
Bank runs are rather more difficult to ignore, less likely to be confused for politics or whatever other distraction. However, Silicon Valley Bank (SVB) like its compatriot Signature Bank are only a small piece of this larger story. The more recent symptoms of serious and growing inelasticity.
We have in our minds – because it is put there purposefully – to associate a bank run with something out of the 1930s. It’s not just in the Economics textbook, it also takes up the bulk of today’s “central bank” manual, too.
The term run used to mean you’d literally have to run to the bank and get cash first before that bank might run out. Once it did, you were screwed; the bank would be closed and as a depositor you’d become an unsecured creditor merely hoping there’d be pennies left on your long-parted dollars.
It’s the scene staged vividly in the Christmas classic It’s A Wonderful Life. Heroes George and Mary Bailey selflessly sacrifice their honeymoon kitty (we have to assume as a loan) to the Building & Loan to bolster its physical reserves after panic sweeps Bedford Falls. This infusion of cold hard cash turned out to be just enough, leaving only those two celebrated national notes to keep the building from being completely empty.
Panics like that used to dot the landscape of monetary history, yet those are drastically different from what we experience nowadays. Back then, banks operated as literal warehouses of actual cash.
Once taken into its vault, the bank would lend it out to someone else creating elasticity via competing claims on the same currency (intermediation). To stay afloat, the bank had to carefully manage its “liquidity” in the form of cash in the vault. Thus, the amount of physical dollars demanded by the public was the primary aspect to these runs.
As Milton Friedman and Anna Schwartz showed in excruciating detail, when the amount of currency demanded by the public surged, it drained the banking system of available cash leaving its weakest members exposed – absent George and Mary – to being unable to meet higher-than-usual demands for what might be left sitting idle in its vault.
And when one bank closed and defaulted, others would follow as fear spread and the public withdrew physical funds to hold in safety in other decidedly non-bank places.
The original and primary task of the Federal Reserve was to step in at times like these (elasticity). The more cash withdrawn and held by the public, the greater the danger to the banking system, the more the Fed was supposed to open its books and supply its forms of liquidity to replace enough of what might’ve been lost to “flighty” depositors.
The FOMC is meant to be George and Mary Bailey, shorn of any charm or equal to their effectiveness.
While this was a close enough depiction all those years ago, it hasn’t been this way really since then. This was not the mode of failure in 2008 even though that monetary crisis bore all the same hallmarks of a systemic monetary crisis. Though it was a bank panic, only banks had panicked.
This is not just some cute play on words, rather it fully recognizes how the modern monetary system truly operates. Hand-to-hand currency is no longer more than a negligible piece of the commercial economy. Therefore, demand for currency – physical currency – by the public is insignificant.
What happens when anyone grows suspicious of an individual depository is very different. Rather than running down to the local branch and demanding the teller deliver funds from the vault, should you or I become uncomfortable about the institution we’d instead electronically open an account with one we are comfortable with and then initiate an electronic transfer from the one to the other.
No real cash changes hands. Vaults are not drained when currency is transferred outside the banking system to the public as they once had been.
Today, book entries must be made to balance the numbers on both sides of the transfer. The troubled institution is hit with a debit notice (either correspondent or via some interbank utility like Fedwire) it must somehow settle. On the other side, the receiving bank credits your new account and its own books, the whole thing “made real” by how both participants match and transit through the interbank plumbing.
Instead of tangible currency removed from the system while being hoarded under peoples’ mattresses, intangible book entries get redistributed all staying within the banking system.
This means liquidity isn’t what’s available in any individual bank vault, it is what might be provided by redeployment via what are called wholesale means. If the bank suffering the withdrawal faces a shortfall in its books rather than its vault, it can go into that wholesale market – in theory – to somehow make up for it.
Under normal and ordinary circumstances, this isn’t much of a problem. There used to be unsecured lending channels, legacies of old interbank methods under the ancient correspondent system (even in the thirties, the monetary framework was quite a bit more complicated than it has been portrayed). Failing that, repo, especially when collateral is widely and readily available for cheap.
Dealers will intermediate between pools of virtual cash, those who have a surplus of book entries delivering to those who are in deficit. They might at times extend their own resources, creating new liabilities out of thin air (or even more imperceptibly in the case of derivatives). Even when playing matchmaker, the dealer undertakes some risk.
The entirety of this redistribution is governed by dealer perceptions of what are essentially arbitrage opportunities.
The term arbitrage is somewhat misunderstood in this context. It suggests a no-loss possibility when risk consideration is very much apart of whether a dealer will choose to pursue one. A spread (reward) might become huge and still go unfilled if or when money dealers perceive too much risk whether you or I see it the same way – or see it at all.
Systemic liquidity, therefore, is no longer about making available cash for public demand as the appeal of mattresses for currency storage overtakes the bank vault down the street, it is entirely the willingness of the wholesale network to redistribute book entries under conditions of arbitrage that could very well be very far from risk-free.
Beginning in 2007, book entries of “cash” were pulled from institutions increasingly regarded as “troubled”, and not strictly those located in the US, either. In fact, in the earliest days of the crisis European firms were more likely than not to be regarded as suspect. The redistribution of book entries, therefore, transiting through eurodollar plumbing and guided by eurodollar arbitrage conditions as governed by eurodollar money dealers.
This meant eurodollar banks unwilling to undertake the risks of arbitrage left these firms exposed to being on the wrong side of book entry withdrawals. Without any other recourse, they’d have to sell some assets to raise some “cash.”
The deficiency did not go unnoticed among authorities (which makes what followed all the more criminal). On September 18, 2007, the day the FOMC voted to stem the growing crisis with nothing more than a 50-bps rate cut, the discussion turned to arbitrage:
“MR. ROSENGREN. Over the past month, the funds for overnight Eurodollars have frequently been trading much higher than overnight fed funds. This highlights the difficulty that some European financial institutions are having borrowing in dollars and the unwillingness of many financial institutions to arbitrage these spreads if they involve credit exposures to European financial institutions, even for overnight.”
What possible good would a cut to fed funds rates do here? It would cheapen the rate at which firms already surplus book entry cash might borrow, adding a bit more to the arbitrage potential, but then, as Eric Rosengren admitted, nothing would happen anyway because dealers were refusing from the start. Fifty bips in additional spread wasn’t going to suddenly change anyone’s mind.
In fact, the longer it went on without some useful answer, the worse it would become. Again, September 2007:
“MR. KOHN. As is typical in a financial crisis or panic, people have fled toward liquidity and safety in Treasury bills and overnight lending, and the normal arbitrage that happens across markets just isn’t happening. It’s great that the markets seem to be getting better, but if they continue to malfunction or if it gets worse again, I think there’s a serious problem.”
The monetary condition only appeared to get better in short run fits and starts. Even the worst monetary crises like the Great Collapse in the early thirties or the Global Monetary (not financial) Crisis of 2008 don’t go in a straight line. They come on incrementally, in stages (keep this in mind for the rest of March and the next few months).
No one could provide an answer because money didn’t then – and doesn’t now – work outside this virtual framework. Authorities can’t replace cash that might have been to stash somewhere because no cash had been withdrawn anywhere; it is entirely book entries at the mercy of (eurodollar) money dealers to redistribute.
Their willingness is guided and motivated by other factors, balance sheet constraints which officials are powerless to affect directly. Instead, they’re left attempting to influence psychology through indirect means such as rate cuts. No wonder they aren’t effective.
The latest instance of this modern version of a bank run, the one which took down Silicon Valley Bank and Signature is largely of the same “substance.” Virtual cash was drained from both for various reasons, but it did not disappear into public hands. Rather, it migrated to others in the system as more book entries.
Therefore, the failures of SVB and Signature are uninteresting and pretty much irrelevant. The more important question is the same one about arbitrage; how there wasn’t any to help either of them. What must be inhibiting wholesale redistribution from those who have benefited from this migration, and was it specific to these two banks?
While we have no way of knowing for sure, it’s not as if we don’t have useful clues. Money market prices such as eurodollar futures give us a systemic sense of arbitrage potential, more so the risks dealers might be perceiving when considering them. And it is here where trading over the past week has been both illuminating and frightening.
Single-day moves in eurodollar futures contracts were historic. The price for the September 2023, for example, gained nearly 45 bps last Friday, March 10, in the wake of SVB’s seizure. That was almost the same as the March 2009 contract had been moved on September 15, 2008, the day Lehman Brothers’ insolvency was made public.
But then on Monday, March 13, a buying panic in these hedging instruments became so intense, the same September 2023 added an unthinkable 101 bps! In a single day, more than double Lehman. After a relatively small reverse on Tuesday, massive buying stuck again on Wednesday with the contract rising another 35 bps.
What do these moves tell us about perceptions of systemic risk? Bill Dudley, then the Federal Reserve System’s Open Market Manager, spilled the secret on August 7, 2007:
“MR. DUDLEY. That said, the Eurodollar market is a very deep market, and if one thought that the Fed was not going to do what the market priced in, there certainly would be the ability of people to take the other side of the bet…In the short run, that kind of thing certainly goes on. If I can’t sell the bad asset that I hold, then I will buy something that will perform well if the bad asset deteriorates.”
OK, what Mr. Dudley was saying was as liquidity dried up because dealers weren’t willing to intermediate these asset markets, either, rather than selling assets losing value people with European accents instead bought hedging instruments like eurodollar futures because those would rise in price should everything continue going wrong (rates go down when things go bad).
In other words, we can and do get a sense of systemic liquidity from how intense the demand for these hedges might become. So, yeah, those historic moves in eurodollar futures over the past week absolutely do tell us there is little faith in the monetary system’s ability to undertake dealer-driven book entry cash redistribution – nor much faith in the Fed to do much to increase or at least modestly influence it.
SVB was a symptom of a much bigger potential problem, basically the same one as fifteen years ago.
Primary among those, collateral scarcity inhibiting repo. The best collateral can suddenly become unavailable for its own wide and necessary redistribution (wholesale book entry money is, shall we say, complicated). At the same time hedging in eurodollar futures this week became extreme, so, too, were indications of systemic collateral shortfalls; the two very much tied together.
In fact, on Monday and again Wednesday, they became straight-up collateral runs. Treasury bill prices skyrocketed, where at one point on Wednesday the 4-week T-bill rate had fallen more than 60 bps! For reference, five or six bps would indicate serious strain.
All the while politician after politician and Fed official after Fed official reassured the public there was nothing to this SVB business.
Markets not only say otherwise, they’ve been positioning for this – therefore warning us - for months. That means this is far from over. As Donald Kohn had worried all those years before, then did nothing particularly productive about it, “if they continue to malfunction or if it gets worse again, I think there’s a serious problem.”
You don’t say.
This isn’t over. And that’s not my verdict, either. It is just what’s clearly being priced right now after SVB where it comes to dangerous deficits in the book entry money of modern systemic runs. If George and Mary Bailey are out there somewhere, we’ll need stacks of T-bills and an angel’s arbitrage guarantee from them this time.
