Editing Openai/69640cff-e5e0-8003-a8d1-a78b118ca18b (section)

==== 3) “Low-skilled at entry upgrades to high productivity? Deliveroo rider → high income sounds fantastical.” ====

Your skepticism here is warranted. The literature does not justify the rosy inference “entered low-skilled ⇒ later high-skilled in large numbers.”

What the evidence actually supports is heterogeneity:
* There is strong evidence of occupational downgrading on arrival for immigrants (doing jobs below formal qualifications).
* And there is UK evidence that for at least some cohorts, downgrading does not quickly unwind. For example, a UK government-commissioned analysis using ASHE–Census linked data finds substantial migrant pay gaps at entry and that gaps narrow only partially over the lifecycle; it also reports “very little evidence of substantial labour market improvement” for some downgraded groups studied.

On the other hand, newer administrative/cohort work often finds that earnings rise with years since arrival, and some cohorts catch up relatively quickly—but that can reflect visa selection (more “skilled-worker-route” entrants), and compositional attrition (the lowest earners are more likely to leave the payroll data / leave the country / move into self-employment). A UK administrative-data discussion of recent cohorts reports that median earnings for the 2022 cohort rose substantially and “caught up” to the median, and that earnings tend to rise with length of stay.

So: it’s not “fantastical” that some people transition upward (especially students → graduates; skilled-worker dependents → workers; etc.). But it is unjustified to treat that as the modal story for low-paid gig workers. The research support is closer to:
“Some upgrade; many partially upgrade; a non-trivial share remain low-paid for long periods.”