About

A research firm with a public track record

Kestrel Economics builds calibrated forecasts of mining permit decisions in North America, and publishes them. Every forecast. Every score against the decision the regulator actually rendered. Every miss.

Every forecast is checkable against the decision the regulator eventually renders. Hits and misses both stay in the record. Trust gets built that way, over time and in public, rather than from anything we claim about ourselves.

Kestrel publishes a calibrated forecast for every major mining permit in North America. The weekly report puts that forecast in front of you on a recurring schedule. The analyst dashboard, in development, lets your team work with the forecasts and the underlying decisions directly.

How we do the work

Prior knowledge counts

Start from what is already known

We do not start each forecast from a blank slate. Base rates for how permits of this type, in this jurisdiction, under this regime have moved through review go in up front. So does the accumulated knowledge of which factors have historically mattered. The model begins informed.

Forecasts move when the facts move

Update as new evidence arrives

A forecast is not a one-time call. When a court ruling lands, a hearing concludes, a minister changes, or a new technical report is filed, the forecast moves. The size of the move depends on how much the event actually tells you about the outcome, and we are explicit about that calculation.

Process-fit, not one-size-fits-all

Models shaped to how permitting actually runs

Permitting is not a single yes-or-no event. It is a sequence of stages, holds, referrals, hearings, and reviews, and different permits move through different ones. The models follow that structure rather than collapsing it into a single probability, and different permit types and jurisdictions get treatments fit to their process.

Uncertainty in the open

Show the range, not just the number

Every forecast comes with a range. Where the historical record supports a tight estimate, the range narrows. Where it does not, the range widens, and the factors widening it are named. A confident-sounding number with no range is the easiest way to mislead someone making a real decision.