The current model rules out an expensive national design under weak effects. It also shows why a small, persistent improvement across an entire cohort could matter over a forty-year working life. Both belong on the same page. Neither gets to impersonate a forecast.
The current 1% affected-population sensitivity applies only to the modelled effective affected employed population. It is equivalent to approximately 0.162% averaged across the full annual cohort. Under that narrow case, the mature full design produces:
Interpretation: this version should not be approved for national rollout. Better to discover that in a workbook than after commissioning a billion-dollar machine.
| Stage/design | Estimate | Status |
|---|---|---|
| Stage 0 feasibility, co-design and full business case | $5.3m | Immediate request |
| Three-year targeted pilot | $97.6m | Conditional |
| Universal core + targeted accelerator | $875.0m/year | Lower-cost comparator |
| Mature full design | $1.526b/year | Planning comparator—not requested |
The defensible question is: What if stronger skills, qualifications, employment stability and financial capability increase the cohort’s average real lifetime earnings by 1%, 5% or 10%?
It does not mean every student receives an identical pay rise. Some people may experience no measurable earnings change; others may experience much more. It means the cohort’s total inflation-adjusted lifetime earnings are, on average, 1% higher because capability and employment pathways improved.
The website shows 1% as a plausible lower sensitivity, 5% as ambitious and evidence-dependent, and 10% as an upside stress test—not a prediction.
At a genuine 1% average cohort uplift, the current sensitivity produces a social BCR of approximately 1.82x at 2%, 1.39x at 4% and 0.96x at 8%.
That is far more promising than the mature full design. But it currently assumes the lower-cost structure retains the full modelled high-risk effect. That assumption has not been demonstrated and must be attacked vigorously.
At the same 1% whole-cohort-average sensitivity, the mature full design is approximately 1.05x at 2%, 0.80x at 4% and 0.55x at 8%.
The lesson is not “assume 5% and declare victory.” The lesson is to test whether a leaner design can produce a broad, persistent effect at a defensible cost.
The long-term fiscal problem explains why the question matters. It does not automatically make an expensive programme good value.
Follow the whole financial discussion, reply to analysts, or download the source data and workbook.
