Local Average Treatment Effect
Also known as · LATE · LOCAL effect · treatment-effects
The Local Average Treatment Effect (LATE) is the average treatment effect for the subpopulation whose treatment status is changed by the instrument (or by crossing the RDD cutoff). IV does not estimate the average effect for everyone in the sample — it estimates the effect for compliers, the subset of units who would take the treatment if and only if the instrument pushed them to. In RDD, LATE is the effect for units close to the cutoff.
When to use
Always interpret IV and RDD estimates as local effects, not population averages. The Angrist draft-lottery estimate is the effect of military service for men whose service decision was changed by their draft number — not for someone who volunteered regardless of the draft. PS_5's RDD estimates the union → recall effect for firms with near-50% vote shares, not the average effect across all unionised firms. This is the internal-validity vs external-validity trade-off: LATE is highly credible for the compliers but may not generalise to the whole population.