Deposit demolition derby

My fellow Americans, our long national regional-banking nightmare may not be over. It did look like the coast was clear on Tuesday, for a bit, after Treasury secretary Janet Yellen signalled willingness to support uninsured depositors at smaller banks.

But it’s another part of the US government that’s responsible for the scarcity of liquidity in the first place: the Federal Reserve.

The Fed has been draining reserves from the financial system by reducing the size of its balance sheet. And an aggressive pace of rate increases has pushed up yields on short-dated Treasuries and money-market funds, drawing what liquidity is left away from bank accounts.

So even though depositors appear to have called off the bank run (for now at least), banks’ profitability problems aren’t going away. Barclays laid out the challenge in a March 9 note, which (reasonably) got lost in the hubbub of the run on Silicon Valley Bank:

With market rates approaching 5% for the first time since before the GFC and opportunity costs at their highest level in decades, we expect flows out of deposits to accelerate this year. This is consistent with management commentary from banks (see below). The speed and size of the Fed’s 2022 rate hikes may have created a bit of a recognition lag for depositors who may have been somewhat slow to react. . . We think banks will need to compete more aggressively to retain deposits with market rates as high as they are currently.

The amount of deposits at US-chartered institutions dropped about 5 per cent between February 2022 and March 8th this year, totalling nearly $825bn, according to Federal Reserve data — and that’s before the run on Silicon Valley Bank. (These figures exclude certificates of deposit or CDs, which carry a higher interest rate and tie up depositors’ cash for a pre-determined amount of time.) Data for the week ended March 15 will be published Friday.

From Barclays:

The longer-term trend shows that bank stress is a natural side effect of the Fed’s policy stance, as Employ America argued Tuesday, even if panicky venture capitalists didn’t help in SVB’s case. As JPMorgan strategists pointed out last week:

The tightening of liquidity conditions in the US banking system has not only been the result of the Fed’s QT program, but also the result of the Fed’s rate hikes, as those rate hikes have already induced a large $400bn shift from bank deposits to money market funds over the past year.

It’s no surprise that the last comparable run on US banks was during the 1980s savings and loan crisis, said Greg Hertrich, head of US depository strategies at Nomura. After the Fed raised rates aggressively to fight inflation in the 1970s, the S&Ls’ fixed-rate mortgages experienced mark-to-market losses. Roughly one-third of S&Ls failed or merged between 1980 and 1988, in a slow-moving crisis. As Hertrich told Alphaville:

There were some similar concerns about asset-liability mismatch, certainly within the context of a Fed that was being very aggressive with rates . . . That is not a surprise. What might be a surprise is when there are deposit outflows as significant as has been experienced at Silicon Valley Bank, Signature Bank, and seemingly now First Republic. Usually it doesn’t move that quickly . . . 

If we were to use the S&L crisis as a historical reference point, I think one of the things that’s very likely, is you’d see more consolidation among the regional banks, and to a different extent, the super-regional and large community banks.

For those who want to read more on the S&L meltdown, the FDIC has a good list of resources. There’s also a muckraker’s history written by Steven Pizzo, Mary Fricker and Paul Muolo right after the crisis.

One point in the literature may be important here: the S&L crisis wasn’t resolved until years after regulators first noticed the problems with mismatched assets and liabilities. In fact, the crisis was made worse by regulators’ response, the FDIC argues, as the US deregulated the industry in what became an effort to let S&Ls make up for the losses with credit risk:

Most political, legislative, and regulatory decisions in the early 1980s were imbued with a spirit of deregulation. The prevailing view was that S&Ls should be granted regulatory forbearance until interest rates returned to normal levels, when thrifts would be able to restructure their portfolios with new asset powers . . . Legislative actions in the early 1980s were designed to aid the S&L industry but in fact increased the eventual cost of the crisis.

This time around, US regulators are addressing the central issue of bond losses, and simply letting banks of all sizes temporarily liquefy their safe-bond portfolios at par value. This isn’t really credit-risk forbearance, but it does provide significant relief from the double-digit losses those portfolios would have otherwise suffered because of the Fed’s recent rate hikes.

The outlook for the liability side of banks’ balance sheets is less certain. The government hasn’t directly pledged to cover all uninsured deposits, even if regulators have strongly hinted at it.

But the possibility of an unlimited deposit guarantee raises another question: what even is a deposit, anyway?

OK, yes, fine, it is money that someone puts in a bank. But as the financial system relearns every few decades, it’s also a maturity-free liability that can be redeemed on short notice.

There’s a reason First Republic shares didn’t respond much to last week’s “vote of confidence” from major banks in the form of $30bn in “uninsured deposits”: an uninsured deposit is functionally the same thing as a loan!

First Republic’s filing says the liquidity will be provided for 120 days, and “at market rates”, which could mean anywhere from its CD rates (which averaged 2.69 per cent in 4Q) to a spread above current short-dated Treasury yields, which are trading solidly above 4.5 per cent.

This makes First Republic’s liquidity infusion an excellent example of the problems with treating depositors differently to other unsecured lenders (whether by design or accident). For example: If banks decide to keep making giant interest-bearing unsecured loans deposits with one another, will those all be guaranteed too?

In the wake of the Great Financial Crisis, global banking regulators took the potential flightiness/risk of deposits into account in banks’ Liquidity Coverage Ratio or LCR requirements, as Credit Suisse UBS’s Zoltan Pozsar wrote in 2016.

The largest global banks are required to hold a reserve of high-quality liquid assets to back their deposits, as Dan Davies wrote for Alphaville last week, with those deposits weighted according to their “heat”. “Hotter” deposits, which are expected to flee faster and in greater numbers in a bank run, require a greater amount of liquid assets. (Dan also noted these requirements don’t apply to small and midsized US banks.)

Those rules provide helpful guidelines for those of us — Alex, Louis and every other millennial — who are still adjusting to a world where deposits yield anything above zero.

So, below, find a brief and incomplete list of US deposit types, rated on the highly unscientific scale of 🔥 (not hot / low run risk) to 🔥🔥🔥 (hot / high run risk).

We have included LCR designations — drawing on this handy chart from a UBS note — with a few extras, and are excluding uninsured deposits for now. That’s because uninsured corporate deposits were 🔥🔥🔥 before Silicon Valley Bank’s failure, but now that regulators are hinting at full deposit coverage, who knows?

Do email and/or let us know in the comments which categories we should have included, or if you disagree with our ratings. (Really feel free, because the author graduated university in 2009 and thus spent a full decade of adult life without ever Earning Safe Yield.)

---

Retail deposits:

This is money that a normal person would put in the bank — or in First Republic’s slightly riskier case, a wealthy person working with an adviser. Still, retail deposits generally aren’t especially run-prone, as most fall below the $250,000 FDIC insurance limit.

Rating: 🔥

Wholesale deposits:

These are deposits from companies, money managers and other institutions. They are rather more prone to runs, but regulators say their heat depends on whether the cash is going to be used for:

1) Operations and other day-to-day needs: 🔥

2) Investment, savings, nothing, etc: 🔥🔥🔥

Time deposits:

This is a category that includes things like CDs, where depositors agree to tie up their money for a short period of time in exchange for a higher interest rate. That means the funding is fully stable, but only until the CD matures, at which point there’s a decent chance the depositor will withdraw the cash instead of letting the bank reinvest.

Rating: 🔥🔥

Brokered deposits:

This one is kind of a puzzle.

A “brokered deposit” often calls to mind a CD from a treasury-management firm like IntraFi (formerly Promontory Interfinancial Network). That would fit into the “time deposit” category, though brokered CDs are seen as slightly less stable, because there is no direct relationship between the bank and the depositor, to the extent that such a relationship even helps.

Simple enough. But then we find footnotes like this one, in the 10-K of Western Alliance Bancorp, another regional bank that’s gotten shellacked by its shareholders:

WAB is a participant in the IntraFi Network, a network that offers deposit placement services such as CDARS and ICS, which offer products that qualify large deposits for FDIC insurance. At December 31, 2022, the Company had $683 million of CDARS deposits and $2.1 billion of ICS deposits . . . At December 31, 2022 and 2021, the Company also had wholesale brokered deposits of $4.8 billion and $1.8 billion, respectively.

So they use the IntraFi network’s CD placements, and also “wholesale brokered deposits”? Bit confusing. That $4.8bn is around 9 per cent of Western Alliance’s $53.6bn total deposits, according to the 10-K.

Here’s how Barclays describes the rise in brokered deposits since the Fed started tightening policy:

These interest-rate-sensitive balances rose $150bn in Q4 22 and are up nearly 50% ($296bn) since lift-off. Combined, time and brokered deposits have grown to 13.7% (from 6.9%) of bank liabilities, while non-interest deposits have shrunk from 25.4% to 22.6%.

Partly for confusion reasons, we will give them a rating of 🔥🔥🔥

---

To move on from run risk: Non-interest-bearing deposits are better for banks than interest-bearing deposits, for obvious reasons (the bank doesn’t have to pay depositors to hold their money).

So bank analysts will be watching closely to see how many depositors start demanding interest on their cash as short-term rates rise. From UBS:

 . . . we looked back as far as we could to see where NIB has troughed, relative to Fed funds (see Figure 1). Going back to 1984 (10% fed funds), our data suggests that NIB troughed at 13% of total deposits in 1Q89 — when Fed funds was at ~9%. Our coverage is at 31%, and even relative to the ‘07 trough of 17%, this suggests mix shift will continue.

As we said previously, it’s difficult to know exactly how quickly deposits are leaving US bank accounts. As our colleagues at mainFT reported last week, money-market funds saw the biggest inflows since 2020 during the run.

But the full extent of the deposit flight won’t be clear until the Fed’s next H.8 report, released Friday. But that will come two full days after the Fed’s policy statement, so the entire market could be different by then.