Illustrative · Kestrel Pathways is a constructed operator, not a real provider. Its shape, scale and contract mix are drawn from how the UK employability sector actually works, and every figure is indicative unless it carries a source.
The argument
What an employment programme is actually for
Every instruction in this study rests on one claim: the purpose of the programme and the measure that funds it are not the same thing, and almost every pathology in the sector is a rational response to that gap.
The purpose is durable income and the stability that comes with it. A person who is working, earning enough to be worth the trouble, and not at risk of losing it by Christmas, is the thing the programme is for. Everything else is instrumentation.
The funded measure is a job start, followed by evidence of continued employment at four weeks, thirteen weeks and twenty six weeks. It is auditable, it is timely enough to pay against, and it correlates with the purpose. It is also not the purpose, and the gap between them has a shape that can be described precisely.
Chapter 1 · The claim
The gap, stated precisely
Four mechanisms, each with a name, the way it works, who carries the cost of it, and what it would take to measure. No scores, no risk matrix. A mechanism that has never been measured says so rather than being assigned a number that implies it has.
-
The proxy becomes the purpose
Unquantified- Mechanism
- What is paid for is measured, what is measured is managed, and what is managed becomes what the organisation believes it is doing. Within two contract cycles the sentence "we get people into work" quietly means "we evidence job starts".
- Who carries it
- The participant furthest from work, whose case takes eleven months and whose outcome may never be claimable.
- To measure it
- Effort per participant plotted against assessed distance from work. If the line slopes the wrong way, the proxy has won. Nobody in the sector publishes this.
-
The window ends and so does the interest
Unquantified- Mechanism
- Payment stops at twenty six weeks, so contact stops at twenty six weeks. What happens in month eight is invisible to the funding model and therefore to the operation, even though month eight is where durability is decided.
- Who carries it
- The participant who loses the job in month nine and returns to the start of a queue, and the taxpayer who pays for the second attempt.
- To measure it
- Employment status at twelve and twenty four months, which requires data the provider does not receive and the commissioner does not routinely share back.
-
The referral is not the beginning
Unquantified- Mechanism
- The provider clock starts at referral. The participant clock started when the job was lost, the illness began or the sentence ended, often years earlier. The programme optimises a window inside an episode it did not define and cannot see the start of.
- Who carries it
- The participant, who experiences the whole episode and is judged on the part inside the window.
- To measure it
- Time from the triggering event to referral, which is knowable for most cohorts and is almost never reported alongside programme performance.
-
Success is attributed to whoever is holding the file
Unquantified- Mechanism
- Many people would have found work without the programme. Counterfactual measurement is expensive, slow and politically uncomfortable, so it is rarely commissioned, and the whole sector reports gross outcomes as though they were additional.
- Who carries it
- The commissioner, who buys less additionality than the numbers suggest, and honest providers, who cannot prove they are better than the ones inflating.
- To measure it
- A randomised or quasi-experimental design with an untreated comparison group. It exists in the literature, rarely in live contracts, and nobody bidding will ask for one.
Chapter 2 · Fairness
Why the commissioner is not the villain
It is tempting, from inside a provider, to treat the funding model as an imposition by people who do not understand the work. That reading is comfortable and mostly wrong. A public body spending public money on a private provider needs evidence that would survive an audit and a select committee, and "our advisers worked very hard" is not that evidence. Outcome funding is what happens when a system that was defrauded by effort-based payment tries not to be defrauded again.
The honest position is that the model is a reasonable answer with known side effects, and that both parties know what the side effects are. What is missing is not good intentions. It is a measure of the thing itself, held by somebody, published, and used to check whether the proxy is still pointing at it.
Chapter 3 · The ask
What would close the gap
Three things, none of which requires abandoning outcome funding. Pay a differential rate by assessed distance from work, so the hardest cases are worth attempting rather than worth avoiding. Extend the measurement, though not necessarily the payment, to twelve months, so durability becomes visible. And share employment data back to providers at twelve and twenty four months, so the sector can see what it actually produced.
None of those is in the gift of a provider. All of them are in the gift of a commissioner, and a provider that argues for them publicly is arguing against its own short term interest, which is roughly the test of whether it meant the values on the previous page.