The standard story about changing industry is that you just need to reinvent yourself. What that story never reports is how much it costs, and for how long.
Louis Jacobson, Robert LaLonde and Daniel Sullivan measured it in work published in 1993 in the American Economic Review, using Pennsylvania administrative data: they followed high-tenure workers who lost their jobs in mass layoffs, comparing them with similar workers who stayed employed.
The result is harsh and surprisingly stable:
- At the moment of job loss, the earnings drop is around 40%.
- Five years later, those who changed jobs still earn about 25% less than the trajectory they would have had.
- The loss is larger for those who change industry than for those who find a similar job in the same industry.
It isn’t a sentence. It’s a price, and like all prices it can be negotiated if you know it in advance.
Why the loss is industry-specific
Labour economics has a name for the thing you lose: specific human capital. It isn’t your intelligence, it isn’t your work ethic. It’s the knowledge that only holds value inside a given industry or firm: who to call, how a quote is put together there, what the unwritten rules are.
Robert Topel documented back in 1991, in the Journal of Political Economy, that wages rise markedly with tenure in the same firm and the same industry. That growth is precisely what resets when you switch sector, which is why the literature draws a sharp line between mobility within an industry and mobility across industries.
Gueorgui Kambourov and Iourii Manovskii, in 2009 work in the International Economic Review, pushed the distinction further: they argue the relevant human capital is tied to the occupation more than to the firm. An accountant who changes company loses little. An accountant who becomes a salesperson loses a lot.
| Type of change | Loss of specific capital |
|---|---|
| Change firm, same occupation | Low |
| Change industry, same occupation | Medium |
| Change occupation, same industry | Medium |
| Change occupation and industry | High |
The practical reading is immediate: changing one variable at a time costs far less than changing both at once.
Do retraining programmes work? It depends on the type
Here is the most useful and least cited evidence. David Card, Jochen Kluve and Andrea Weber published in 2018 in the Journal of the European Economic Association a meta-analysis of over 200 evaluations of active labour market policies worldwide. The picture:
- In the short run (up to a year) average effects are close to zero.
- In the medium to long run (two to three years) effects become positive and more substantial.
- Programmes with human capital content and job-search assistance perform better than generic subsidies.
- Programmes work better for women and for the long-term unemployed, and in recessions.
There’s one effect worth knowing because it’s counterintuitive: the lock-in effect. While you’re on a full-time course you search less for work, so in the first months your outcome is worse than that of someone who didn’t take the course. The advantage comes later. Evaluate these programmes at six months and you conclude they don’t work; evaluate them at three years and you conclude the opposite.
The right question isn’t “does this course work”. It’s “how long does this course take to work, and can I afford to wait that long”.
The most studied American case
The US Trade Adjustment Assistance programme, designed for people who lose work to foreign competition, was the subject in 2012 of a large-scale evaluation commissioned by the Department of Labor, comparing participants with a matched comparison group.
The result, put bluntly: in the first years participants did worse than the comparison group, due to lock-in; by four years the difference had largely closed; and the biggest gains were concentrated among those who completed training connected to a job actually found in that field. Training that doesn’t lead to a job in the field trained for produces no return.
This is where the economic literature and personal experience converge: the value isn’t in the course, it’s in the transition the course makes possible.
How to lower the cost of switching
- Change one variable at a time. First the occupation inside the industry you know, or the industry while keeping the occupation. The double jump is the most expensive configuration in the data.
- Inventory what’s transferable. Not “I’m good with people”, but “I managed a 200,000 euro budget” or “I trained 40 colleagues”. The difference is verifiability, and it applies to writing your CV too.
- Work out your horizon before the course, not after. If the positive effect arrives at two to three years, you need savings or partial income to cross that period. That’s the deciding variable, not motivation.
- Choose training with an identified destination. In the data the return concentrates on those who end up in a job consistent with the path. Before enrolling, check where previous graduates ended up.
- Make the transition in parallel if you can. Working full-time and studying evenings is exhausting, but it removes the lock-in effect. It’s a trade-off the data makes defensible.
- Use weak ties, not strong ones. Opportunities in a different industry almost always come from peripheral contacts. I’ve written about it in professional networking.
- Look at where demand is heading, not where it’s high today. I’ve collected the data in most in-demand skills.
My take: the hard part is no longer being the expert
I think the most common mistake isn’t picking the wrong industry, it’s underestimating the identity part. Switching sector means going from competent person to person who asks questions, and that shift is socially uncomfortable more than it is technically hard. I suspect the people who succeed are above all those who can be at ease in the beginner’s position for a non-trivial stretch of time.
I also suspect age matters less than people say and life stage matters much more. It isn’t that you learn worse at 45; it’s that at 45 you have a mortgage. Those are two different constraints and they need different tools.
Here’s where I think this ends up, though I can’t prove it: the worst period isn’t the beginning, it’s the middle, when you’ve already lost the old status and haven’t yet built the new one. People who quit, quit there.
Frequently asked questions
Is 50 too late? The earnings-loss data shows the cost grows with the tenure lost, not with age as such. Which means the constraint is economic and plannable, not cognitive.
Long course or start straight away? Depends on your financial buffer. Card’s meta-analysis shows training programmes pay off in the medium run, but produce a lock-in effect in the short run. If you can’t cross that period, parallel training is the rational choice.
Do certifications help when changing industry? They help where they’re a formal entry requirement. As a generic signal the evidence is weak: I’ve written about it in which certifications are worth it.
How long does it really take? In the active labour market policy literature positive effects show up at two to three years. It’s an order of magnitude, not a promise, but it’s more honest than the six months you read in the ads.
Should I say in the interview that I come from another industry? Yes, and it’s better to say it first, explaining what you bring that transfers. I’ve collected the typical questions in job interview questions.
Sources
- Jacobson, L. S., LaLonde, R. J., & Sullivan, D. G. (1993). Earnings Losses of Displaced Workers. American Economic Review, 83(4), 685-709. The persistent earnings losses of high-tenure workers.
- Card, D., Kluve, J., & Weber, A. (2018). What Works? A Meta Analysis of Recent Active Labor Market Program Evaluations. Journal of the European Economic Association, 16(3), 894-931. Over 200 evaluations and the short-versus-medium-run difference.
- Kambourov, G., & Manovskii, I. (2009). Occupational Specificity of Human Capital. International Economic Review, 50(1), 63-115. Human capital tied to occupation rather than firm.
- Topel, R. (1991). Specific Capital, Mobility, and Wages: Wages Rise with Job Seniority. Journal of Political Economy, 99(1), 145-176. Wage growth tied to tenure.
- Mathematica Policy Research / Social Policy Research Associates (2012). Estimated Impacts for Participants in the Trade Adjustment Assistance Program. Report for the U.S. Department of Labor. The comparison-group evaluation and the lock-in effect.
- Heckman, J. J., LaLonde, R. J., & Smith, J. A. (1999). The Economics and Econometrics of Active Labor Market Programs. In Handbook of Labor Economics, Vol. 3A, 1865-2097. The methodological framework for evaluating these programmes.
- Autor, D. H., Dorn, D., & Hanson, G. H. (2013). The China Syndrome: Local Labor Market Effects of Import Competition in the United States. American Economic Review, 103(6), 2121-2168. The slowness of adjustment across industries.
- Cortes, G. M., & Gallipoli, G. (2018). The Costs of Occupational Mobility: An Aggregate Analysis. Journal of the European Economic Association, 16(2), 275-315. Estimating the cost of moving between occupations.
- OECD (2021). Training in Enterprises: New Evidence from 100 Case Studies. The comparative picture on adult training across OECD countries.
This article was written following the process described in how we write. If you find a figure reported incorrectly, write to me: I’ll correct it and flag it at the bottom of the article.



