After SVB, the biggest bond rally since 1987

The two-year Treasury yield fell by more than 100 basis points in three days, from above 5.07 per cent to below 4.00 per cent, as the collapse of Silicon Valley Bank set off the biggest rush into government bonds since the Black Monday crash of October 1987. Trading algorithms scrambled to price out further rate rises and price in emergency cuts. Over the weekend of 12 March, the Federal Reserve, the Treasury and the FDIC invoked a systemic risk exception to guarantee all uninsured deposits at SVB and Signature Bank, and set up the Bank Term Funding Program (BTFP) to stop the bleeding.

The BTFP was a bold piece of engineering. Banks could pledge Treasuries and agency mortgage-backed securities as collateral for one-year loans from the Fed, and the collateral would be valued at par rather than at market prices.

From 5 per cent to fear

A week ago the argument was about how fast the Fed would climb. Jerome Powell had just told Congress that half-point increases were back on the table, and the two-year yield had pushed above 5 per cent for the first time since 2007. Then SVB failed on Friday, New York regulators closed Signature Bank on Sunday, and the market stopped asking how high rates would go and started asking what else might break. A move of that size in the most policy-sensitive part of the curve is a verdict on the Fed's path.

Why the front end moved most

Two forces pushed in the same direction. The first was a straightforward flight to safety, as companies and funds pulled cash from banks they no longer trusted and parked it in Treasury bills and short-dated notes, which carry no credit risk. The second was a repricing of the Fed. The two-year yield is in effect the market's forecast of the average policy rate over the next two years, and when traders stopped expecting the rate to reach 5.5 per cent or more, the yield had to fall.

Positioning made the move more violent. After Powell's testimony many funds had bet on higher short-term yields, and a sudden rally forced them to buy back those positions at a loss, which pushed prices up further. Liquidity in the Treasury market, never deep in a panic, thinned as dealers widened their quotes. A move that might have taken weeks in calmer conditions was squeezed into three sessions.

A dollar for every discounted dollar

By counting bonds that had lost value at 100 cents on the dollar, the Fed made the $620 billion of unrealised losses that threatened regional banks disappear, at least for the purposes of borrowing. Banks no longer had to dump securities into a falling market to meet withdrawals. They could hand the discounted paper to the Fed and get 100 cents back in fresh reserves. The facility was an emergency bridge, meant to stop a run on regional banks from spreading nationwide.

The terms show who carries the risk. Loans are priced at the one-year overnight index swap rate plus 10 basis points, and the Treasury has put up $25 billion from its Exchange Stabilization Fund as a backstop against losses. Shareholders and some unsecured creditors of the two failed banks will not be protected, but every depositor at them will be, which removes the market discipline that the $250,000 insurance limit was meant to preserve.

The bond rally itself gave banks some relief. Falling yields push up the price of fixed-rate securities, so the same move that signalled fear also shrank the paper losses on bank balance sheets. The relief is partial and reversible, since a return to higher yields would open the hole again.

Tightening and rescuing at once

The contradiction is hard to miss. For twelve months the Fed had been shrinking its balance sheet and raising rates to tighten financial conditions and cool demand.

Since September the runoff of maturing bonds has been capped at $95 billion a month, draining reserves from the banking system. In a single weekend the Fed created an open-ended facility that pumps central bank credit straight into the banks. One arm of the central bank is now adding the reserves the other is withdrawing.

Officials will argue that the two tools do different jobs. The policy rate sets the price of money across the economy, while emergency lending addresses the plumbing of particular institutions. The distinction holds in textbooks, but markets do not treat them separately. A central bank that has just shown how quickly it will protect depositors invites investors to bet that it will also blink on rates when the next strain appears.

What 1987 suggests

The October 1987 precedent points one way. After Black Monday the Fed under Alan Greenspan promised to supply liquidity to the financial system and nudged rates lower, even though it had been tightening in the months before the crash. The economy did not slide into recession, and the Fed resumed raising rates the following year once markets had settled.

The autumn of 1998 tells a similar story. After Russia's default and the near-collapse of Long-Term Capital Management, the Fed cut rates three times between September and November, although the domestic economy was growing strongly. It began raising rates again in the middle of 1999. In both episodes a financial shock bought a pause, not a change of direction, and inflation pressure was weaker then than it is today.

The case for pressing on

The case against a pause is still strong. February's consumer price report, released on Tuesday, showed core inflation running at 5.5 per cent a year, and the labour market has shown little strain. If the BTFP and the deposit guarantee calm depositors, officials may argue that they can handle financial stability with lending facilities and inflation with rates.

Several markers will show which way the committee leans. The European Central Bank meets on Thursday and had signalled another half-point increase before the banking scare, so its choice will show whether policymakers elsewhere think the turmoil is contained. Weekly Fed balance sheet data will show how much banks borrow, and the committee meets on March 21st and 22nd. Its projections will reveal whether officials still expect rates to rise above 5 per cent this year.

The bond rally after SVB showed the Fed's two jobs in open conflict. It had to build a lifeline so that its own tightening did not bring down the banks, and it will probably slow its rate rises as a result.