Abstract
In finding X-heterogeneous effects of an endogenous treatment D on a response Y where X is covariates and an instrument δ is available, we overcome the X-dimension problem, by conditioning on the “control propensity score” E(D|X,δ=0). We propose two estimators allowing for any form of Y.
Original language | English |
---|---|
Article number | 109747 |
Journal | Statistics and Probability Letters |
Volume | 196 |
DOIs | |
Publication status | Published - May 2023 |
Externally published | Yes |
Keywords
- Complier effect
- Endogenous treatment
- Instrumental variable estimator
- Propensity score
- Randomization
ASJC Scopus subject areas
- Statistics and Probability
- Statistics, Probability and Uncertainty