The numbers everyone repeats · 6.3

The great resignation, in the series

The great resignation, in the series. What is actually the case, and how it compares with what is repeated.

A practical software reference for this part of the discussion is Monitask's overview of monitoring software.

A rise in resignations occurred, it was real, and the name attached to it did most of the interpretive work.

What the series shows

The American quits rate, the share of employees voluntarily leaving each month, rose to about three per cent in late 2021 from a pre-pandemic level near two point three.

That is a substantial increase. It is also a continuation: the quits rate had been rising steadily since around 2010 as the labour market tightened, and the peak sat perhaps thirty per cent above the existing trend rather than departing from it entirely.

By 2023 and 2024 the rate had returned to approximately where the trend line would have put it.

Where people went

Mostly to other jobs. Job-to-job transitions accounted for the great majority of the movement, and the flow out of the labour force altogether, while it rose, was a much smaller component and was concentrated among older workers taking early retirement.

A reshuffle is a different phenomenon from an exodus, and the available data supports the first.

Who was moving

Disproportionately workers in leisure, hospitality, retail and other low-paid in-person sectors. Quits rates in those industries rose most and from the highest base.

The coverage overwhelmingly described knowledge workers reassessing their priorities during lockdown, which describes a different population from the one the data identifies. This site's argument, again.

Why they were moving

The most economically legible explanation is that wages at the bottom rose quickly and workers moved to capture the increases. Wage growth was strongest in exactly the sectors with the highest quits, which is what a tight labour market looks like.

Reassessment of life priorities may also have occurred. It is harder to measure and it was the explanation that received the coverage.

What the name did

Converted a labour market statistic into a cultural event. Once named, the phenomenon acquired an explanation, a moral, and a large volume of commentary that did not consult the series.

It also acquired a second life as an argument in unrelated disputes, including about remote work, which the data does not speak to either way.

The employer response

Wage increases at the bottom, signing bonuses, and in some sectors improved scheduling. Those responses are informative: employers behaved as though the cause was compensation and conditions, whatever the commentary said.

Revealed preference is generally the better guide when stated explanations and behaviour diverge.

What was genuinely new

The speed of the reallocation and its concentration in the lowest-paid sectors. Also a measurable rise in business formation, which was real and sustained longer than the quits spike.

Both are more interesting than the narrative and neither had a name.

The reading rule

When a phenomenon acquires a name before it acquires a series, find the series. In this case it is published monthly, it is free, and it covers thirty years, which makes the trend line easy to draw and easy to compare with whatever is being claimed.

What the same period looked like elsewhere

Quits data of comparable quality exists for few countries, and where it does the pattern was weaker. The phenomenon was substantially an American one and was reported internationally as though it were general.

Which is a further instance of the same reading error: a statistic about one population, applied to all of them.

The retirement component

Exits from the labour force rose most among older workers, and a substantial part of that was early retirement enabled by asset price rises rather than by any reassessment of working life.

It is the part of the phenomenon with the clearest financial explanation and it received the least attention, because it does not support any argument about workplace culture.

What to compare it with

The quits rate in the late nineteen-nineties reached comparable levels during a similarly tight labour market, without acquiring a name or a literature.

That comparison is available in the same free series and it takes about two minutes, which is roughly the effort that separated the coverage from the data.

The business formation figure

Applications to form new businesses rose sharply in the same period and, unlike the quits rate, stayed elevated for years afterwards. It is the most durable change in the data from that period and it received a small fraction of the attention.

For broader context, consult Job Openings and Labor Turnover Survey.