The following summarises a Bloomberg Opinion column by Justin Fox, former editorial director of the Harvard Business Review. It reflects the author’s personal views.
- A divergence has opened by age. Job-switching among workers 25 and older is back on the rise and at high levels relative to the past two decades, while those aged 16 to 24 are treading water near historic lows.
- That matters because of what job changes have historically done. A seminal 1992 study found early-career job changes accounted for more than a third of wage growth among a sample of 10,000 young American men tracked from the late 1950s through the early 1970s, describing it as the principal method by which workers improve their position on their own initiative.
- The rewards have returned. In early 2023 job switchers saw wage growth more than two percentage points above those who stayed. That premium disappeared during the weak labour market of 2025 and has been growing again lately.
- Fox offers two possible AI-related explanations for the youngest cohort’s stagnation: research showing employment of younger workers lagging in industries most exposed to generative AI, and the ease of mass-applying having gummed up the online job-matching infrastructure of LinkedIn, Indeed, Glassdoor and ZipRecruiter. He cautions it is too early to say with confidence.
What Happened?
Contrary to popular impression, early-career job switching has declined over the long run rather than increased, with the main drop coming in the 2000s. Fox notes that is a complicated picture: it forms part of a broader decrease in business dynamism since the 1980s that may be weighing on productivity, but a 2022 study found that very short employment spells have become much less common and that better matching accounts for about a third of the decline in turnover over two decades. The surge during the Great Resignation of 2021 and 2022 has reversed across all age groups.
Why It Matters?
The entry-level squeeze may be operating through two channels at once, which is the implication worth drawing out. If AI reduces the volume of routine work that justified junior roles, there are fewer positions at the bottom of the ladder, and if the youngest workers are also not moving between the positions that exist, they lose the mechanism that historically delivered a third of early-career wage growth. Citigroup has just cut its analyst programme from three years to two, framed as competition for talent, and the unstated reason is that the apprenticeship had less content once AI absorbed the modelling and document work. That logic applies well beyond banking. For clients with children entering the workforce, this is the practical consequence: the traditional route to early earnings growth is narrowing specifically for the people who need it. The platform observation is an investable one that will not appear elsewhere. If generative AI makes applications costless to send, the signal-to-noise ratio on job boards collapses and the matching function those businesses sell degrades. That is a direct threat to the value proposition of LinkedIn, Indeed, Glassdoor and ZipRecruiter, and it is a case of AI undermining a service rather than enhancing it. Worth watching for evidence in those companies’ engagement and conversion metrics. The switcher premium returning is the actionable fact for individuals. More than two percentage points of additional wage growth for moving, in an environment where overall incomes are rising about 3% while employer healthcare costs climb 8% and inflation runs at 3.7%, is a meaningful difference in real terms. Fox’s point, drawing on Allison Schrager’s argument that young Americans have become more risk-averse, is that the data supports taking the risk rather than avoiding it.
What Next?
Whether the 16 to 24 cohort’s job-switching recovers as the labour market does, or stays depressed while older workers’ rises, is the question that separates a cyclical story from a structural one. Watch entry-level hiring in AI-exposed industries, since that is where the research Fox cites points. For the job platforms, application volumes against successful placements are the metric that would show whether the matching infrastructure is genuinely degrading. The switcher wage premium published by the Atlanta Fed is the simplest ongoing measure of whether moving still pays.
Affected Tickers and Coins: ZIP, MSFT, 6098
Source: Bloomberg Opinion
