The Fed's banking lifeline expires

The Federal Reserve stopped making new loans under its Bank Term Funding Program (BTFP) on 11 March, closing the emergency facility it opened during the regional banking panic of 2023. The scheme did its job by letting banks pledge underwater securities at par. Its closure removes an important safety net for bank balance sheets.

The facility was built in a weekend. Silicon Valley Bank failed on 10 March last year, and Signature Bank followed two days later, when the Fed announced the programme alongside the Treasury, which set aside $25 billion from its Exchange Stabilization Fund as a backstop. Banks could borrow for up to a year against Treasuries and agency securities valued at face value, however far their market prices had fallen. That promise, more than the money actually lent, helped halt the run on regional deposits.

An emergency turned arbitrage

By late last year the programme had become a trade. The BTFP charged the one-year overnight index swap rate plus a tenth of a percentage point, and as markets priced in rate cuts that fell below the rate the Fed pays on bank reserves. Banks could borrow from the facility and park the money at the Fed for a profit, and usage hit a record in January. In late January the Fed ended the trade by setting a floor on the lending rate at the interest rate on reserves, and confirmed that new lending would stop in March as planned. Existing loans run until they mature, up to a year after they were taken out.

Back to the discount window

Banks holding Treasuries and agency mortgage securities worth less than face value can no longer borrow against them at par. They must use the discount window, where collateral is taken at market value less a haircut. Wholesale funding has calmed down, but regional banks with large unrealized losses on securities are still exposed to a sudden liquidity shock. The Fed has taken the stabilisers off.

What the haircut costs

The difference is in the arithmetic. A bond bought at par in 2020 or 2021 with a low coupon may now trade well below face value, and the discount window lends against that lower market price, less a further haircut. A bank that could have raised a dollar against each dollar of face value under the BTFP will raise noticeably less. The window's primary credit rate sits at the top of the Fed's target range, 5.5 per cent, so the money is also dear.

American banks still carry hundreds of billions of dollars of unrealized losses on their securities, most of it the result of rates rising after they bought low-yielding bonds during the pandemic. Those losses do not matter while deposits stay put. They matter a great deal when depositors leave and the bank must sell or pledge the bonds, which is exactly the chain that sank Silicon Valley Bank. The turmoil at New York Community Bancorp in recent weeks has shown that confidence in mid-sized lenders can still wobble.

Where the stigma comes from

The Fed has fought the discount window's reputation before. In December 2007, as the subprime crisis spread, banks avoided the window for fear that borrowing would be seen as a confession of trouble, even as funding markets seized up. The Fed responded by creating the Term Auction Facility, which lent through anonymous auctions so that no single bank stood out. The lesson was that a facility banks will not use provides no protection, however generous its terms. Supervisors have since urged banks to test their access to the window and to pledge collateral in advance, but a habit of decades is hard to break.

The Federal Home Loan Banks, which lent heavily to troubled regional lenders last spring, are an imperfect substitute. They were created to support housing finance, and their regulator said in a review published last November that they should not act as a lender of last resort to failing banks. Their advances also need collateral and take time to arrange, which matters in a run that unfolds within hours.

The case for closing it

The Fed has good reasons. Lending at par against underwater bonds invites moral hazard, since it spares banks the consequences of their interest-rate bets, and an open-ended facility would have encouraged more of them. Deposits have stabilised since last spring, supervisors have pushed banks to test the discount window and pre-position collateral there, and the Federal Home Loan Banks still provide funding in ordinary times. Rate cuts, which markets expect this year, would shrink the paper losses and ease the problem.

The Fed could also reopen an emergency facility under its section 13(3) powers, with Treasury approval, if another panic struck. That option exists, but it would arrive after a run had started rather than before. The stigma attached to the discount window, where borrowing has long been read as a sign of weakness, may make banks slow to use it until it is too late.

The winners from the closure are the largest banks, whose deposit bases are stickier and whose securities losses are small relative to capital, and which will look safer by comparison. The losers are mid-sized lenders with concentrated deposits and heavy bond losses, which may have to pay more for deposits or hold more cash to reassure markets, squeezing already thin margins.

The indicators to watch are discount-window borrowing in the Fed's weekly balance-sheet data, deposit flows at small and mid-sized banks, Federal Home Loan Bank advances and bank share prices after each quarterly results season.

Regional banks with big paper losses must now borrow against their bond portfolios on ordinary terms. Any lender that faces a run this year will find help from the Fed a good deal more expensive, and the weakest mid-sized banks will pay for that through higher deposit costs well before any run begins.