Research

Working Papers

Estimating causal effects of discrete and continuous treatments with binary instruments (with V. Chernozhukov, I. Fernandez-Val, and Sukjin Han) [arXiv]

• We propose a framework for identifying average and quantile treatment effects of discrete and continuous treatments with binary instruments. Our identification results lead to tractable distribution regression estimators.

Selection and parallel trends (with D. Ghanem and P. Sant'Anna) [arXiv] [CESifo WP 9910]

• What selection mechanisms are compatible with parallel trends assumptions?

The power of tests for detecting p-hacking (with G. Elliott and N. Kudrin) [arXiv] 

• We study theoretically and in simulations the implications of likely forms of p-hacking on the distribution of p-values across studies and the power of tests for detecting it.

(When) should you adjust inferences for multiple hypothesis testing? (with P. Niehaus and D. Viviano) [arXiv] [Related comment on FDA guidance on multiple testing]

• We propose an economic model of the academic publication process to explain whether and why multiple testing procedures arise as solutions to economic decision making problems, and if they do, what features of the economic environment determine the right procedures to use.

Green governments (with N. Potrafke) [arXiv] [CESifo WP 8726] [Non-technical summary in German (ifo Schnelldienst)]

• We estimate the causal effect of a Green government on economic and environmental outcomes.

Protectionism and economic growth: Causal evidence from the first era of globalization (with N. Potrafke and F. Ruthardt) [arXiv] [CESifo WP 8759]

• We estimate the effect of protectionism on short-term economic growth in 19th century Sweden.

Bias correction for quantile regression estimators (with G. Franguridi and B. Gafarov), resubmitted to the Journal of Econometrics [arXiv] [CESifo WP 9046]

• We develop a higher-order theoretical framework for comparing different quantile estimators and propose novel bias correction methods.

A t-test for synthetic controls (with V. Chernozhukov and Y. Zhu), resubmitted to the Journal of Political Economy. [arXiv] [R-Package] 

• We propose a practical and robust t-test for making inferences on the ATT in synthetic control settings.

Pairwise valid instruments (with Z. Sun), conditionally accepted at the Journal of Econometrics [arXiv]

• We propose a method for estimating treatment effects when the instruments are partially invalid.

Publications

Factorial designs, model selection, and (incorrect) inference in randomized experiments (with K. Muralidharan and M. Romero), accepted at The Review of Economics and Statistics [Accepted Version] [NBER WP 26562]

• We study the use of factorial designs in experimental economics and discuss econometric methods for improving power while controlling size.

Toward personalized inference on individual treatment effects (with V. Chernozhukov and Y. Zhu), PNAS (Commentary), 2023. [Published Version]

• We discuss how to robustify conformal inference methods for predicting individual treatment effects.

Omitted variable bias of Lasso-based inference methods: A finite sample analysis (with Y. Zhu), The Review of Economics and Statistics, 2023. [Published Version] [Replication] [arXiv]

• We study the finite sample behavior of post double Lasso and debiased Lasso.

Detecting p-hacking (with G. Elliott and N. Kudrin), Econometrica, 2022. [Published Version and Replication] [arXiv] [Longer Version with more Results][Generic R-Code for the Proposed Tests] [R-Package by Sebastian Kranz]

• We provide theoretical foundations and statistical tests for detecting p-hacking based on distributions of p-values across studies.

An exact and robust conformal inference method for counterfactual and synthetic controls (with V. Chernozhukov and Y. Zhu),  Journal of the American Statistical Association, 2021.  [Published Version] [arXiv] [Video and slides (Chamberlain Online Seminar, May 2020)] [R-Package]

• We develop generic inference methods for synthetic control and related methods such as factor models, interactive FE models, matrix completion approaches, and (fused) time series panel data methods.

Distributional conformal prediction (with V. Chernozhukov and Y. Zhu), PNAS, 2021. [Published Version] [arXiv] [Replication]

• We propose transformation-based conformal prediction methods that are adaptive to heteroscedasticity and leverage powerful regression methods for estimating conditional distributions.

Decentralization estimators for instrumental variable quantile regression models (with H. Kaido), Quantitative Economics, 2021. [Published Version and Replication Code] [Generic R-code]

• We propose computationally efficient and easy-to-implement fixed point estimators for estimating linear IVQR models.

A comparison of two quantile models with endogeneity, Journal of Business and Economic Statistics, 2020. [Published Version]

• I establish a close connection between the two most popular quantile models with endogeneity and provide a characterization of IVQR when the underlying rank similarity assumption is violated.

Generic inference on quantile and quantile effect functions for discrete outcomes (with V. Chernozhukov, I. Fernandez-Val, and B. Melly).  Journal of the American Statistical Association, 2020. [Published Version] [arXiv] [R-Package]

• We propose practical methods for constructing simultaneous confidence intervals for quantile and quantile effect functions with discrete outcomes. 

A closed-form estimator for quantile treatment effects with endogeneity, Journal of Econometrics, 2019. [Published Version] 

• I develop semiparametric IVQR plug-in estimators for settings with binary treatments and binary instruments, building on and generalizing the closed-form solutions derived in Wuthrich (2020).

Local average and quantile treatment effects under endogeneity: A review (with M. Huber),  Journal of Econometric Methods, 2019. [Published Version] 

• We provide an extensive review of the literature on local average and quantile treatment effects with a focus on recent developments and extensions.

Financial incentives and physician prescription behavior, evidence form dispensing regulations (with D. Burkhard und C. Schmid), Health Economics, 2019. [Published Version] [Preprint]

• We exploit regional variation in dispensing regulations in Switzerland to estimate the causal effect of financial incentives on physician prescritption behavior.

Exact and robust conformal inference methods for predictive machine learning with dependent data (with V. Chernozhukov and Y. Zhu), Proceedings of COLT 2018 (Conference of Learning Theory). [Published Version] [arXiv] [Replication]

• We extend conformal prediction to time series settings.

Hedonic valuation of the perceived risks of nuclear power plants (with S. Boes and S. Nüesch), Economics Letters, 2015. [Published Version]

• We exploit quasi-experimental variation induced by the Fukushima disaster to estimate the effect of perceived risk on rental prices in Switzerland.

Handbook Chapters 

Instrumental variable quantile regression (with V. Chernozhukov and C. Hansen), Handbook of Quantile Regression.  [Published Version] [arXiv] [IVQR R-Package by Yu-Chang Chen] 

• We provide a review of the IVQR model with a focus on classical and modern estimation and inference procedures.

Local quantile treatment effects (with B. Melly), Handbook of Quantile Regression. [Published Version] [Preprint]

• We review the local quantile treatment effects framework with a focus on extensions and recent developments.