n.V1-7.09 | VOCATIONAL TRAINING
01| Teach people a trade and they will get a job. It is the most intuitive solution in this book part, the most politically popular, and among the least supported by evidence.
02| McKenzie's critical review of active labour market policies in developing countries covers the family: vocational training, wage subsidies, job search assistance, migration assistance. The conclusion is that many of these policies are considerably less effective than policymakers assume, with a large share of evaluations finding no significant impact on either employment or earnings. Not small effects. No effects.
03| The usual explanation, and the one that keeps the training industry intact, is that the training was fine but everything around it was missing — no capital to start with, no employer nearby hiring, no transport to reach the job, no childcare, no certificate anyone recognises. On this reading the intervention is incomplete rather than wrong, and the fix is to add the missing pieces.
04| McKenzie's own reading points elsewhere. He suggests urban labour markets in many of these settings work reasonably well, with fewer market failures than commonly assumed — which, if true, means the constraint was never a skills gap. People are not unemployed because they lack the skill; they are in low-paid work because low-paid work is what the local economy has, and adding a certificate to a market with no vacancies produces a certified person in the same job.
05| Both explanations are consistent with the same nulls, and that is exactly why the nulls have not settled anything. Twenty years of disappointing training evaluations have been absorbed by the sector as evidence that training needs more wraparound, which is the reading that requires no institution to stop doing what it was doing.
06| There is a way to separate them, and it has been partly run. If the constraint is a missing complement, then supplying capital instead of training should do less than training, or at best the same. If the constraint was never skills, cash should do more. Rwanda tested that directly — a USAID workforce readiness programme against a cost-equivalent unconditional transfer — and the transfer moved more outcomes than the training did. That result gets its own chapter later; here it is enough to say that the fork has been probed once and the answer went against the skills-gap reading.
07| A second piece of the puzzle cuts the other way. Where training programmes do produce large and lasting earnings gains, the mechanism is not the training content but the destination: participants are moved into higher-paying industries and occupations, not merely into employment. That is the subject of the next chapter, and it suggests the useful variable in training is not what is taught but where the trained person is aimed.
08| What all of this implies for the standard programme is unflattering. A vocational course that trains people for the occupations already available around them is, on the evidence, likely to produce nothing measurable, because it is adding supply to the part of the labour market where supply is not the constraint. The programmes that work are the ones that change which market the person enters — and doing that requires knowing something about employers, which most training providers do not, and being able to place people, which most are not funded to do.
09| Open: nobody has run the test that would separate the two explanations cleanly — training held constant while the complementary inputs are varied. The bundled programmes come closest, and they bundle rather than decompose.
