The Fed's quiet bailout of banks' bad bond bets

Banks borrowed a record $152.9 billion from the Federal Reserve's discount window, according to the H.4.1 balance-sheet release published on Thursday, 16 March. That beat the peak of the 2008 financial crisis, when borrowing topped out at about $111 billion. The new Bank Term Funding Program (BTFP) lent another $11.9 billion in its first four days.

Officials insist this is not a bailout, since shareholders were wiped out. Corporate finance analysts see it differently. In their view the Fed has bailed out the banks' disastrous bond portfolios, and done so with remarkably little fuss.

A week of emergency borrowing

For years banks avoided the discount window for fear of the stigma attached to using it, and the record was set within a week of Silicon Valley Bank's failure. More Fed credit went to the bridge banks that the FDIC created to run SVB and Signature Bank. On Thursday eleven of the largest American banks also agreed to deposit $30 billion with First Republic to steady it. Little of that official borrowing would be so cheap if the collateral were valued at what it would fetch in the market.

The speed of the switch says something about how deep the fear ran. Banks that would once have paid almost any price to avoid being seen at the window queued up within days, because the risk of being unable to meet withdrawals had come to look worse than the stigma. Primary credit at the window costs 4.75 per cent, the top of the Fed's target range, for loans of up to ninety days.

A bailout happens when the state takes on private risk. In normal markets, a bank holding a 10-year Treasury yielding 1.5 per cent when rates are 4.5 per cent must accept a 20 per cent markdown if it needs cash in a hurry.

Lending at par is a subsidy

By lending against these underwater bonds at par, the Fed has handed banks a free option on interest-rate risk. The central bank takes the economic loss onto its own books and shields the banks from the cost of their unhedged bets.

The discount window now offers much the same treatment. When it launched the BTFP, the Fed said it would apply the programme's generous collateral margins at the window as well, which helps explain why borrowing there jumped so fast. In the example above, a bank pledging bonds worth 80 cents on the dollar receives 100 cents in cash, and the 20-cent gap is a loan against value that does not exist.

The BTFP adds a year of breathing room. Its loans run for up to one year, against Treasuries, agency debt and agency mortgage-backed securities that a bank already owned when the programme was announced on 12 March. That last condition stops banks from buying cheap bonds in the market in order to pledge them at par, but it does nothing to change the treatment of the bonds they already hold.

That warps the pricing of credit. Lenders that botched the basic matching of assets and liabilities escape the market discipline that sank their regional peers, while the cost of the BTFP lands on the central bank's growing operating deficit. Since last autumn the Fed has paid more in interest on bank reserves than it earns on its bond holdings, and it books the shortfall as a deferred asset rather than seeking money from the Treasury.

Who holds the loss

The defence is that par lending costs nothing if the borrowers survive. A Treasury held to maturity does repay 100 cents, the BTFP charges the one-year overnight index swap rate plus 10 basis points, and the Treasury has put $25 billion from its Exchange Stabilization Fund behind the facility. All of that holds for solvent banks. The option pays off precisely when a bank fails with its bonds still underwater, and then the loss falls on the Fed and the Treasury rather than on the bank's creditors.

The deposit guarantees for SVB and Signature carry a separate bill. Under the joint statement of 12 March, any loss to the Deposit Insurance Fund from protecting uninsured depositors will be recovered through a special assessment on banks. That means healthy lenders, including those that hedged sensibly, will pay part of the cost of rescuing the depositors of those that did not.

Where forbearance led before

America has tried letting banks sit on underwater fixed-rate assets before. In the late 1970s and early 1980s savings and loan associations held long-dated mortgages funded by short-term deposits, and the jump in interest rates left many of them insolvent on a market basis. Regulators chose forbearance, loosening capital rules and letting the thrifts keep operating in the hope that rates would fall. Many used the time to take bigger risks, and the eventual clean-up cost far more than prompt closure would have. The BTFP is shorter and better secured, but it rests on the same wager that time will heal a duration mismatch.

Moral hazard returns

The official justification is financial stability and the need to stop contagion spreading. The moral hazard created by lending at par is large all the same.

The phrase too big to fail entered American politics in 1984, when the Comptroller of the Currency told Congress, after the rescue of Continental Illinois, that the government would not let the biggest banks go under. The rules written after 2008 tried to push losses back onto shareholders and creditors. On interest-rate risk, the BTFP moves in the opposite direction, and it covers banks of every size rather than a handful of giants.

The case for the Fed is not trivial. A run feeds on itself, and a bank forced to sell bonds at a loss to meet withdrawals crystallises the very insolvency that depositors fear, which prompts more withdrawals. Breaking that loop quickly probably saved several lenders that were sound on a hold-to-maturity basis. The objection is to the terms, which could have included a haircut or a higher rate without leaving banks exposed to fire sales.

If banks know the Fed will always accept underwater collateral at face value in a crisis, they have much less reason to hedge the interest-rate risk on their government bonds. The Fed's review of its supervision of SVB, led by Michael Barr and due by May 1st, will show whether officials plan to offset that with tougher rules on interest-rate risk. Weekly balance-sheet data will show whether banks keep leaning on the facility once the panic eases. The Fed has in effect taken the banks' interest-rate losses onto the public balance sheet in the name of stability. Banks will probably hedge less as a result, and the next rate cycle is likely to find them just as exposed.