Why regional banks lost a third of their value

The KBW Regional Banking Index (KRE) has fallen by more than 35 per cent since Silicon Valley Bank failed on 8 March. The FDIC has since resolved SVB, Signature Bank and First Republic, and still shares in PacWest, Western Alliance and Comerica trade like distressed options, dropping twenty to forty per cent within a single session. Analysts who blame short sellers or panic on social media are missing the arithmetic. Investors are repricing regional banks for three things at once: deposits leaving, funding costs rising, and a large pile of commercial property loans.

The way the market valued these banks rested on one assumption that no longer holds, namely that cheap deposits would stay put. For fifteen years regional lenders paid little extra on deposits when rates rose, and collected steady net interest margins in return.

How the run spread

Each failure taught depositors something about the next bank. SVB showed that a lender funded by a narrow group of large, uninsured depositors could lose most of its funding within days, once those depositors talked to each other and moved money by phone. Signature showed the same weakness in a different client base. First Republic showed that even a bank with high-quality loans could not survive once the run had started, and that the answer from regulators would be a sale to JPMorgan Chase rather than a rescue of shareholders.

The lesson depositors drew was simple. Money held above the $250,000 insurance limit at a mid-sized bank carries a risk that the same money at one of the largest banks does not, and moving it costs nothing. Assets in money market funds have risen above $5 trillion, a record, as savers have chased higher yields and the implicit safety of government paper.

Deposits and offices

That franchise is gone. Regional lenders must now pay 5 per cent on deposits to keep them, or watch customers move their money to the money-centre banks that are seen as too big to fail.

They also carry a heavy, concentrated exposure to commercial real estate (CRE), holding roughly seventy per cent of all outstanding office and retail debt in the United States. Office vacancies are at record highs and property values have fallen by thirty per cent. Defaults on commercial mortgages are picking up.

The office problem is partly permanent. Remote and hybrid working have cut the space many employers need, and leases signed before the pandemic are only now coming up for renewal. A building that loses tenants when its loan matures must refinance at far higher rates on lower income, and the bank that holds the loan faces a choice between extending it on generous terms and taking a loss.

Losses already on the books

The third pressure comes from the securities that banks bought with the deposit flood of 2020 and 2021. The FDIC estimated unrealised losses on banks' securities holdings at $620 billion at the end of 2022. Accounting rules let banks ignore most of those losses as long as they hold the bonds to maturity, but a bank that loses deposits may have to sell, and SVB's collapse began when it did exactly that.

The Fed's Bank Term Funding Program, launched in March, lets banks borrow against Treasuries and agency bonds at face value for up to a year. That removes the need to sell at a loss. It does not remove the loss, since the bonds still earn low yields while the borrowing costs today's rates.

Not much left to earn

Put a collapsing net interest margin together with the loan-loss provisions that commercial property will soon demand, and the earnings power of a regional bank falls close to zero.

The closest precedent is the commercial property bust of the early 1990s, when falling office and development values in New England, Texas and California brought down hundreds of banks, including the Bank of New England. That episode was driven by bad lending more than by interest rates, but it showed how long property losses take to work through: banks recognised them over several years as loans came due, and bank shares stayed depressed until the losses had been absorbed.

The case for buying

The strongest counterargument is that SVB was an outlier. The great majority of its deposits were uninsured, and most regional banks have a broader base of small customers whose balances are fully covered. Many regional bank shares now trade below the book value of their equity, and the Fed's lending programme has removed the risk of fire sales. On this view the sell-off has run ahead of any realistic loss.

That argument holds for the best-run banks, and some will prove to have been bargains. It does not answer the earnings problem. A bank that keeps its deposits by paying far more for them, while its loans and bonds still yield what they did two years ago, earns little even if it never fails.

The signals to watch are the Fed's weekly figures on deposits at small banks, the survey of senior loan officers due next week, and the FDIC's proposals on reforming deposit insurance, published on 1 May, which could change how safe uninsured balances feel. Delinquency rates on office loans will show how quickly the property losses arrive.

Investors selling these shares are doing sums, and the sums say that losses on long-dated bonds and bad loans have permanently eaten into tangible common equity. The 35 per cent fall in regional bank shares is a fair price for a business whose cheap deposits have gone and whose main lending book faces years of restructuring. Shareholders should not count on a rebound before office values find a floor.