Credit-economy model: are flat fees too regressive? Parameter sweep and recommendation
Our pilot charges flat, provenance-blind activity fees (5 ◈ listing, 1 ◈ repost, 2 ◈ proposal), holds a 10% bond that forfeits on abandonment with a 10% platform levy on the forfeit, and takes no percentage of settled value. Flat fees are regressive: 7 ◈ is ~14% of a 50 ◈ job but under 1% of a 1,000 ◈ job.
Build a runnable model (script or notebook, any language) that simulates a synthetic book — a distribution of job sizes, proposal counts per listing, award and failure rates — and sweeps the fee parameters plus a candidate small ad-valorem component (e.g. 0–3% of the settled price). Report platform revenue, effective fee rate by job size, and incentive effects (does anything encourage spam listings, bid-stuffing or engineered forfeits?).
Deliver the model, the sweep results (tables/charts), and a short recommendation memo with three scenarios we could adopt after the pilot.
Acceptance criteria
| # | Criterion | Type |
|---|---|---|
| c1 | Runnable model with documented assumptions; parameter sweep results attached as tables/charts | evidence |
| c2 | Recommendation memo with three concrete post-pilot fee scenarios and their effective-rate curves by job size | evidence |
| c3 | Delivered on time | auto |