US economy crashes into recession: the first-quarter 2022 GDP report

On 28 April 2022 the Bureau of Economic Analysis reported that the American economy had contracted at an annualised rate of 1.4 per cent in the first quarter. The consensus forecast had been growth of about 1 per cent. The previous quarter had printed 6.9 per cent. A miss of 2.4 points against expectations and a swing of 8.3 points against the prior quarter is, outside the pandemic shutdown of 2020, close to the steepest one-quarter fall in the modern record. Whether to call it a crash is a matter of taste. Whether it was expected is a matter of record, and the record is what this page keeps.

Who expected what

Days before the release, Goldman Sachs was described as the most bearish house on Wall Street for putting the odds of a recession within a year at 35 per cent. Deutsche Bank had gone further and forecast one for late 2023. Those were the pessimists. This site had argued through the winter that a recession was not coming but already under way, and had put the probability of one beginning in the first quarter at 95 per cent in March, when the yield curve began to invert.

The reasoning did not rest on the curve. It rested on why the curve had been unable to speak.

The locked yield curve

An inverted yield curve, short rates above long rates, is the Federal Reserve’s own preferred recession signal, and it normally appears many months before a downturn. Through 2020 and 2021 it could not appear, because the Fed was buying Treasuries across every maturity in quantities large enough to set the curve wherever it chose. A gauge that is being held by the hand that reads it measures nothing. The argument made here was that when the Fed stopped buying, in March 2022, the curve would be released to price in both the inflation it had been unable to show and the recession it had been unable to show, and would do so in weeks rather than the usual months. That is what happened: the ten-year yield rose from 1.95 per cent on 9 March, the day of the last purchase, to 2.39 per cent within a fortnight, and the two-year briefly rose above it at the end of the month.

The second leg of the argument concerned the fourth quarter of 2021. Its 6.9 per cent headline was largely inventory: businesses restocking against shortages, which counts as growth when the goods arrive and as nothing when they sit. Strip out the inventory swing and the underlying economy had been slowing since the autumn. That is why the first-quarter number should not have surprised anyone who had looked at the composition rather than the headline.

What the war added

None of this anticipated the invasion of Ukraine, and it would be dishonest to pretend otherwise. The shortages described here in the autumn of 2021 became worse in March 2022 because of the war and, separately, because of the sanctions; net exports and inventories, the two lines that dragged the quarter negative, both took part of their fall from that shock. Even on the pessimistic view the winter was expected to be cold, not this cold. The correct conclusion is that the direction was foreseeable and the magnitude was not.

The definitional argument

The standard shorthand for a recession is two consecutive quarters of falling GDP. The National Bureau of Economic Research, which dates American recessions, has never used that rule; it looks at employment, income, industrial production and sales, and it dates recessions in retrospect, often a year or more later. Both facts were pressed into service in the argument that followed. One side said a single negative quarter proved nothing. The other said that a rule which cannot declare a recession until it is over is not much use to anyone living through one.

What the record shows since

The second quarter of 2022 also printed negative, at minus 0.6 per cent annualised, satisfying the two-quarter shorthand. The NBER did not declare a recession, then or later: employment kept growing throughout, unemployment stayed near a fifty-year low, and both quarters were subsequently revised, the first to minus 1.6 per cent and the second to minus 0.6, with later annual revisions lifting the second into positive territory. Growth resumed in the third quarter and the economy expanded through 2023.

So the call made on this page was right that the economy had stalled in the winter of 2022, right about why the bond market had been unable to say so, and wrong, or at least early, about what would follow. A contraction in output without a rise in unemployment was not the shape of any previous recession, and the arguments about whether it deserved the name are a record of how unusual the year was. The number itself stands: minus 1.4, against plus 1, in the quarter almost nobody had expected to be the one.