When Support Works—and When It Only Looks Like It Does

The entrepreneurship industry makes many claims about value, but the strongest test is not how much support is delivered. It is whether a clearly stated intervention changes the outcomes it was meant to change, compared with what would otherwise have happened. This chapter reads the OECD’s selected evaluation evidence to show how support can be positive, mixed, or unproven—and why those labels cannot be understood without objectives, comparators, metrics, survival, implementation, and cost.

The point is not to turn evaluation into a technical ritual. It is to protect a practical distinction established earlier: activity is visible, while additional impact has to be demonstrated. A programme can be busy, popular, and sincerely helpful without producing the result used to justify its budget. It can also produce a real benefit that is hidden when decision-makers look only at a different outcome.

A mixed result is a distribution, not a shrug

The OECD Framework reviews 50 selected evaluations from 28 OECD countries. The evaluations were chosen because their methods met the Framework’s reliability threshold, which means the set is more informative than a collection of participant testimonials or simple before-and-after counts. It is still not a census of all entrepreneurship support. The sample is selective, and programmes can cover more than one policy area.

Within that selected group, 23 evaluations were classified as showing positive impact, 21 as mixed impact, and 6 as showing no or negative impact. The classification is based on the pattern across the outcomes examined. “Mixed” does not mean that the study failed to reach any conclusion. It means that a programme improved some measures while leaving others unchanged or producing a different result. A programme might raise sales without improving profitability, or employment without improving survival.

That distribution resists two easy conclusions. The first is that entrepreneurship policy works whenever it is well intentioned. The second is that mixed evidence proves evaluation is too complicated to be useful. The more defensible conclusion is conditional: some interventions produce clear benefits, many produce benefits that depend on the metric, and a smaller group produces no positive result or a negative one. A policy portfolio therefore cannot be judged by a single success story or by a blanket claim about support as a category.

The selection boundary matters. Because the OECD included evaluations that met its reliability requirements, the 23 positive results are not casual evidence. But the framework also notes that reliable evaluations are unevenly available across policy areas, and that the reviewed programmes are not identical. A result from a finance intervention cannot be treated as a result from a training intervention simply because both are described as entrepreneurship support. The evidence is strong enough to reject a universal slogan and not strong enough to produce a universal ranking.

The mixed category is especially important for the entrepreneurship industry because its actors often promote intermediate results. More applications can be presented as demand. More pitches can be presented as ambition. More ventures can be presented as dynamism. The evaluation question is whether those changes connect to the programme’s intended outcome and survive the relevant time horizon.