A recessional for the recession: everyone sings the strong-labour-market tune in unison, and the tune is wrong

A recessional is the music played while people leave a service. By September 2022 the whole of official and financial opinion was singing one, in unison, to make the recession leave the room. Two consecutive quarters of falling real output had been recorded. The president said it was not a recession. The Fed chair said it beside him. A Nobel laureate in economics called it “just funny” to use the word. The chorus in every case was the same line: because the labour market is strong. This page, expanded from a note shared privately two weeks earlier, was about why that line was the error on which the Fed’s whole plan rested.

The chorus

The examples accumulated. Senator Elizabeth Warren went on television to warn that the Fed’s rate rises might tip the economy into recession, which one commentator quoted here found absurd, since the thing she feared had already happened. Economists expected another strong payrolls print, which would justify another three-quarter-point rate rise. Bank of America’s strategist calculated that the stock market was now pricing a 20 per cent probability of recession, down from 36 per cent in June and 75 per cent in March: the further the economy went into recession, the less the market believed in one. Richard Thaler, the laureate, said real GDP had fallen “a little bit” but that “it’s not like any recession we’ve seen in my rather long lifetime”, and offered record-low unemployment as the proof.

On that last point Thaler was right, and it was the point. It was not like any previous recession. Previous recessions had not followed a pandemic.

The Fed’s deaf zone

The Fed’s model says that low unemployment means demand for labour exceeds supply, that this pushes wages and prices up, and that the remedy is to reduce demand until unemployment rises. Every part of that is correct if the supply of labour is normal. The argument made here was that it was not. Over a million Americans had died; several million had retired early; and a large number were, in the words of a Bloomberg Law report cited at the time, caught in long Covid’s catch-22, too sick to work and not quite disabled enough to be counted. The Great Resignation was not a burst of confidence. It was a workforce that had shrunk, leaving employers bidding for the workers who remained.

In that situation unemployment is low for the same reason it is low in a town after a flood: the people are not there. The demand the Fed proposed to reduce was the demand of those who were still consumers but no longer producers, and reducing it by raising rates would reduce production first, because the workers who could be laid off were the ones producing. The labour signal the Fed was steering by was not a measure of strength. It was a measure of damage, and it was not transitory.

Why it mattered

If the diagnosis was right, the Fed was about to raise rates hard into a supply-constrained economy, which brings a recession without bringing prices down proportionately, the stagflation combination described earlier in this section. The people singing the recessional would have been right that this was not a recession like any other, and wrong about what that meant. The Fed met that week and raised by three-quarters of a point, its third such rise in a row, and its chair said the labour market remained out of balance.

Later note. Unemployment stayed below 4 per cent through 2023 while inflation fell by two thirds, which is not what the Fed’s model predicted and not what this page predicted either. Prime-age participation recovered fully; participation among older workers did not. The best available reading in 2026 is that the labour market was both things at once: short of the workers who had left, and able to absorb the demand that remained once supply chains and energy prices normalised. The Fed’s diagnosis and this page’s were each half right, which in economics is above average.