[ ] Find citation: I'm treating beaver presence as an absorbing treatment, because lodging lasts so long and control efforts have only started recently.
[x] Controls. Some potential are: annual precip/temp averages, nat'l or global prices of major cereal commodities, population, flood vulnerability, any-neighbor-treated
[x] Spatial spillovers. This hinges on whether you believe a) that farmers reallocate within their property (and these properties extend beyond the 1km grid cells I've created), or b) that there is reorganization along the extensive margin if farms go out of business in cell $i$ and new farms enter the market in adjacent cell $j$ (or if farmers migrate to adjacent, not-yet-treated cells. Though this assumes farmers have full information about beaver presence in nearby grid cells).
Run event studies of land use on beaver presence the following form using Sun and Abraham (2021)'s staggered event DiD estimator.