Protecting their business from interest-rate risk is supposed to be what banks specialize in. But a report from Goldman Sachs suggests the calculation most banks use to manage the process could be way off, due to an overestimation in how resilient their ‘deposit franchises’ are to rising interest rates.
In banking parlance, the deposit franchise reflects the base of depositors who value a bank’s services enough not to move their funds in search of a higher interest rate.
For the most part, banks assume their deposit franchise is quite “sticky” — meaning banks can treat the funding as if it were long-term in most of their calculations, as well as cheaper relative to other alternatives. This, as New York University’s Bruce Tuckman told POLITICO, has historically always made sense because deposits aren’t really just a form of funding.
“People really want the deposit product because it offers various services the bank does for them, and they can put in money and take it out on demand,” he said, noting such services rightly carry costs that must be recouped by banks.
Even so, some of their assumptions, as Goldman’s David Mericle explained in a note this week, may be outdated due to erroneous calculations of ‘deposit beta’ — the formula which dictates how much interest banks should pay customers to make sure they don’t scarper. In general, the higher the deposit beta, the more interest banks have to pass on to customers.
If Mericle is right, banks could be broadly overvaluing their deposit franchises. And if that’s the case the banking industry as a whole could be more exposed to rising interest rates than regulators appreciate — and in ways that might be hard to course-correct.
“If banks aim to shorten the average maturity of their lending because they are worried about their deposit beta or deposit outflows could turn out to be higher than expected, they might pull back on longer-maturity commercial and residential real estate loans,” the economist wrote on Monday.
The paradox of speed
As to what’s driving the miscalculation, Mericle identifies three key factors. The abrupt nature of the Fed’s rate-hiking move, technological changes facilitating the speed of withdrawals and a larger than-appreciated share of run-prone deposits in the system.
All these changes, he says, have made “the large unrealized losses caused by rapid increases in interest rates more dangerous” and contributed to recent bank failures in underappreciated ways.
But even then, that’s just half the story.
The other pertains to the fixed-rate long-duration safe assets, such as Treasury bonds, that banks have in the meantime accumulated to improve the overall liquidity of the assets on their balance sheets. In many cases, these are long-term bonds, whose higher yields served to prop up margins while official interest rates were low.
“The textbook model of banks as institutions that borrow short and lend long suggests that they should experience a decline in net worth when interest rates rise and a decline in profitability when short-term rates exceed long-term rates,” wrote Mericle. But this, he explained, “is not quite right”.
As research of Itamar Drechsler, Alexi Savov and Philipp Schnabl corroborates, banks have historically coped very well even in challenging conditions by matching their sensitivity to market interest rates, usually by holding assets with a longer average maturity if their deposit beta is low and a shorter maturity if it is high.
So, while in theory the over-accumulation of long-dated bonds at fixed rates puts even more pressure on banks during challenging periods, such as when it costs more to borrow short-term than long-term (as it does now), in practice, it doesn’t. Banks simply rely on interest expense adjustments to keep their margins intact.
There is, however, a significant trade-off with the arrangement. Interest-rate risk is transformed into liquidity risk, which puts even more pressure on the accuracy of deposit franchise valuations. In the case of Silicon Valley Bank (SVB) these proved spectacularly wrong and led to the bank’s failure in March.
But while SVB’s failure has largely been dismissed by regulators as being driven by uniquely bad practice at one institution, Mericle’s concern is that SVB’s practice of hedging interest rate risk with its deposits was not fundamentally abnormal and that structural factors which influenced the error may also be affecting other institutions.
Deposits as discounted bonds
The way to think of the balancing act banks engage in when they hedge interest rate risk, says Tuckman, is that banks treat deposits more like bonds that pay variable interest rates at below-market rates than as pure deposits. That, for the most part, is perfectly good practice. Somewhat surprisingly, a related paper that Tuckman authored with Schnabl found that banks rarely use derivative products for this purpose, despite the common perception that that is what derivatives are supposed to be for.
“I don’t think that the SVB experience necessarily negates the whole idea of the way banks are [currently] hedging their interest rate risk,” Tuckman said. “What it does point out I think is that in general, it is really hard to model that relationship … especially if you’ve got losses on your asset side and your deposit franchise is not going to stay around.”
Regulators, though, have finally started to pick up on the risk that broader structural forces could be skewing deposit beta assumptions. They’re also growing concerned that quantitative tightening (QT), when it begins in earnest, could make it even harder for banks to liquidate their longer-duration assets to offset any unexpected withdrawals.
The European Central Bank noted on Wednesday in its Financial Stability Review, for example, that once QT gets underway, banks will struggle to match the maturity of their liabilities quickly, making them vulnerable to runs by their largest clients, most of whom are non-banks and non-financial corporation customers.
“As such, both market liquidity and funding liquidity conditions might be more fragile and flightier than the aggregate measures for liquidity suggest, and they therefore warrant continuous monitoring,” the ECB said.
But one new problem facing central banks is that, in the event of a renewed panic, even interest rate cuts might not be enough to restore confidence.
“Cuts would reduce unrealized losses on banks’ assets,” noted Mericle in his report, but they could also “hurt banks that respond to recent events by shortening the duration of their asset portfolios to reduce liquidity risk at the expense of hedging their interest rate risk.”
That could leave a large part of the financial system pretty snookered.