The numbers everyone repeats · 6.1

Most jobs in 2030 do not exist yet

Most jobs in 2030 do not exist yet. 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 work hours tracker.

A figure in wide circulation holds that most children starting school today will work in jobs that do not yet exist. It has a traceable origin, the origin is not a study, and the person credited with it has said so.

Where it came from

The number entered general circulation through a major economic forum's report in 2016, which footnoted it to a popular book. The author of that book has since written publicly that the figure was never the product of research, that she had offered it as an illustration, and that she was surprised to find it cited as a finding.

It continues to appear in government strategies, in education technology marketing and in conference keynotes, generally without the footnote and always without the retraction.

What the evidence actually shows

New occupations do appear continuously. Statistical agencies revise occupational classifications precisely because of it, and studies of new job titles find them emerging steadily across the twentieth century and into this one.

They also account for a modest share of employment. The large occupations of today are nursing, driving, teaching, retail, care and cleaning, and there is no serious basis for expecting them to be replaced by categories that do not currently exist within a decade.

The more accurate statement

Most people working in 2030 will be doing jobs that exist now, and many of those jobs will contain different tasks than they do today. Task composition changes fast; occupational categories change slowly.

That version is less dramatic, is supported, and has a different policy implication: the case for continued training within occupations is strong and the case for abandoning occupational preparation is not.

Why it spread

Because it is useful. It supports urgency, it supports spending on new provision, and it makes a case that existing institutions are obsolete. A figure that supports what somebody was going to argue anyway travels further than one that complicates it.

The citation chain

Following this figure back is a short exercise and a useful one. The report cites the book, the book offers it as an estimate, and the chain ends. There is no dataset and no method.

Doing the same to any striking statistic takes fifteen minutes and succeeds surprisingly often, which is the practical skill this whole part of the site is about.

What did happen to occupations

Some disappeared: typists, switchboard operators, film processors, much of typesetting. That is real and it was concentrated in specific occupations rather than spread across employment.

Far more changed internally. A bank clerk in 1980 and one today share a title and share few tasks. The category persisted, the content did not, and the workers who managed the transition mostly did it inside the same occupation and frequently inside the same employer.

The prediction record

Forecasts of imminent transformation of the occupational structure have been made in every decade since the nineteen-thirties. The transformations that occurred were real, slower than forecast, and different in shape from what was described.

That record does not prove the current forecasts wrong. It does argue for reading them as scenarios rather than as findings, and for asking what would count as evidence either way.

What to do with the claim

Ask what the underlying study measured, and expect that there was no underlying study. Then ask the more useful question, which is which tasks within existing jobs are changing, because that one has answers.

What actually should be taught

The claim is normally deployed to argue that specific skills are pointless and general adaptability is everything. The evidence does not support the first half: occupational skills continue to command a return, and the fastest-growing occupations have specific requirements.

The defensible version is that people will need to learn new tasks within their occupation repeatedly, which is an argument for continued training rather than against initial preparation.

The version that would be worth measuring

How much the task content of existing occupations changes per decade, measured directly from occupational descriptions. Some work of this kind exists and it finds substantial change within stable categories, which is the interesting result and has no slogan attached.

The retraction that did not travel

The author's public correction of the figure is findable in a few minutes and has been online for years. The figure continues to appear in documents published since.

Corrections do not propagate along the same channels as the claims they correct, which is a general property of this kind of statistic and not a failure of any particular publication.

For broader context, consult Future of Jobs Report.