A public-benefit research institute · in formation
The Thesis Institute builds calibrated, open forecasts of the statistics and policies that shape public life — every prediction published with its full chain of reasoning, and scored against the record when the official numbers arrive. The track record is the product.
In formation · building in the openA public observatory of forecasts on government data — tax and benefit statistics, poverty, costs, and the outcomes of policy. Every forecast is reasoned in the open, carries a calibrated interval, and is scored against the official record once it resolves. The track record is the product.
Every forecast publishes its full chain of reasoning — the priors, the sources, and the drivers behind the number — not just a point estimate.
When the official number lands, every forecast is graded. Calibration is public, per forecast and per agent — so the record speaks for itself.
The whole stack is a public good — nothing about the forecasts, the data, or the models is hidden from the people they describe.
The Thesis Institute builds and stewards a connected stack of open forecasting infrastructure.
The open-predictions platform — a live docket of open forecasts and a ledger of resolved actuals, every prediction scored against reality, written by Brier-1.
Open-source microsimulation of tax and benefit policy for the US, UK, and Canada — and custom analysis for governments and researchers.
Calibrated synthetic populations — the substrate that powers policy simulation and grounds population-scale forecasting.