Source Themes

How Much Overlap Failure Can Doubly-Robust T-Statistics Handle?

We show that the Augmented IPW estimator's simple t-statistics can remain well-calibrated even when strict overlap fails and there are many propensities near zero.

Dynamic Bargaining between Hospitals and Insurers (Job Market Paper)

The large literature on vertical market bargaining assumes contracts last for one period, but actual hospital-insurer contracts last for multiple years and are negotiated as a multiple of a benchmark price that changes over time. I study dynamic regulations to those benchmark prices by extending the existing single-period approach to vertical market bargaining to enable contracts that are formed at different times.

Sensitivity Analysis for Linear Estimators

We extend Dorn and Guo (2022)'s characterization of bounds for Tan's marginal sensitivity model to considerably more general assumptions and estimands.

B-Learner: Quasi-Oracle Bounds on Heterogeneous Causal Effects Under Hidden Confounding

We propose a meta-learner for conditional average treatment effect (CATE) bounds that can efficiently estimate sharp bounds.

Doubly-Valid/Doubly-Sharp Sensitivity Analysis for Causal Inference with Unmeasured Confounding

*Accept with minor revisions, Journal of the American Statistical Association.* We propose semiparametric average treatment effect (ATE) bounds estimators with novel robustness properties: double sharpness and double validity.

Sharp Sensitivity Analysis for Inverse Propensity Weighting via Quantile Balancing

We provide sharp bounds for an existing sensitivity analysis that are valid under minimal conditions and sharp given a consistent quantile regression estimate.