What does this research mean for the field?

Two-stage least squares (TSLS) estimators that include covariates parametrically do not have a Local Average Treatment Effect (LATE) interpretation, meaning standard empirical practices often misinterpret these estimates. Novelty: ClaimNovelty.CONTRADICTORY. Consensus alignment: ConsensusAlignment.CHALLENGES_CONSENSUS.

What question did this study set out to answer?

This research investigates the conditions under which two-stage least squares can accurately estimate local average treatment effects.

May 16, 2026The Review of Economic Studies

When is TSLS Actually LATE?

View Full Paper

Ask AI

Bookmark

Authors

CBChristine BlandholPrinceton University JBJohn BonneyStanford University MMMagne MogstadStatistics Norway

Discussion

Member takes

Implication

Analysis of TSLS specifications reveals limitations in estimating local average treatment effects.

Key Points

This research investigates the conditions under which two-stage least squares can accurately estimate local average treatment effects.
Examined various two-stage least squares specifications used in empirical studies.
Analyzed the impact of including covariates in TSLS estimations.
Evaluated thirteen empirical studies to assess LATE applicability.
LATE interpretation of TSLS does not generally apply when covariates are included in the specification.
Most TSLS specifications reflect treatment effects for both compliers and always/never-takers.
Only saturated specifications that control for covariates nonparametrically can accurately yield a LATE interpretation.

Cite This Study

Blandhol et al. (2026) studied this question.

synapsesocial.com/papers/6a080a29a487c87a6a40c0a3 https://doi.org/10.1093/restud/rdag029

View Full Paper

Ask AI

Bookmark

When is TSLS Actually LATE?

Discussion

Member takes

Implication

Key Points

Cite This Study

When is TSLS Actually LATE?

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider