NCGP reform: what it does to the research workforce.

How to read it. The model is calibrated to ARC and government data - the rising fellowship share, the near-flat-to-declining real budget, the observed career-stage shares, the growth in industry collaboration and Indigenous participation. Its behavioural relationships are the main uncertainty, and the most consequential of them are exposed as levers rather than buried in the code. Read shapes and turning points, not point values. A forward run is a scenario, not a forecast: it is an instrument for seeing how a conclusion depends on the reforms and on the assumptions you are willing to make. Where a number matters to a decision, treat it as directional.