The Price of Knowing You
What happens to consumers, firms and total welfare when a seller prices each person by what it knows about them — and which rules soften the harm without destroying the gains?
A synthetic market with heterogeneous consumers, four information regimes, six pricing algorithms, six behavioural consumer types, ten precisely-defined policy interventions, a competition extension and an econometric-recovery study that asks whether standard empirical methods would even detect these effects from observational data.
canonical runs Every number on this page comes from persisted Monte Carlo runs with manifests. demo marks browser-only demonstrations.
Headline (individualized boosting seller vs uniform price)
The six research questions and what we found
One chart: where the surplus goes
As sellers learn more, who keeps the surplus?
Steady-state surplus per period for 1,000 consumers, averaged over 20 seeds, no intervention. First best is the dashed line; the gap is deadweight loss.
Pre-registered hypotheses
Registered in hypotheses.json before any canonical run; evaluated mechanically by spo/analysis.py. A hypothesis is marked not supported when any of its conjuncts fails — the evaluation detail says which.
Explore
Market lab
Set population, information, algorithm, competition, behaviour and policy; run it in the browser; compare with the canonical run.
Consumer perspective
Who wins, who loses, by income, valuation, urgency and behaviour type. Lorenz curves of surplus.
Firm perspective
Information-vs-profit curves, learning speed, algorithm comparison, poaching under competition.
Policy comparison
Ten interventions, defined precisely, on the welfare/fairness/privacy frontier.
Welfare analysis
Surplus decomposition, frontier, behavioural sensitivity, robustness, competition.
Econometric recovery
Would OLS, fixed effects, DiD, IV or an RCT recover the true effect from the data an empiricist actually sees?