Gaussian Transforms Modeling and the Estimation of Distributional Regression Functions (with Richard Spady).
Forthcoming Econometrica.
   We propose flexible Gaussian representations for conditional cumulative distribution functions and give a concave likelihood criterion for their estimation. Optimal representations satisfy the monotonicity property of conditional cumulative distribution functions, including in finite samples and under general misspecification. We use these representations to provide a unified framework for the flexible Maximum Likelihood estimation of conditional density, cumulative distribution, and quantile functions at parametric rate. Our formulation yields substantial simplifications and finite sample improvements over related methods. An empirical application to the gender wage gap in the United States illustrates our framework.
Econometric Society World Congress 2020 presentation.

Identification of Treatment Effects under Limited Exogenous Variation (2025, with Whitney Newey). arXiv.
   Nonparametric identification of treatment effects requires strong support conditions. To alleviate this requirement, we consider varying coefficients specifications for the conditional expectation function of the outcome given a treatment and control variables. This function is expressed as a linear combination of either known functions of the treatment, with unknown coefficients varying with the controls, or known functions of the controls, with unknown coefficients varying with the treatment. We use this modeling approach to give necessary and sufficient conditions for identification of average treatment effects. We extend our analysis to flexible models of increasing dimension and relate conditional nonsingularity to the full support condition of Imbens and Newey (2009), thereby embedding semi- and non-parametric identification into a common framework.

Heterogeneous Coefficients, Control Variables, and Identification of Multiple Treatment Effects (2022, with Whitney Newey).
Biometrika, 109(3), pp. 865–872. First arXiv version September 2022.
   We use control variables to give necessary and sufficient conditions for identification of average treatment effects. With mutually exclusive treatments we find that, provided the heterogeneous coefficients are mean independent from treatments given the controls, a simple identification condition is that the generalized propensity scores (Imbens, 2000) be bounded away from zero and that their sum be bounded away from one, with probability one. Our analysis extends to distributional and quantile treatment effects, as well as corresponding treatment effects on the treated. These results generalize the classical identification result of Rosenbaum and Rubin (1983) for binary treatments.

Control Variables, Discrete Instruments, and Identification of Structural Functions (2021, with Whitney Newey).
Journal of Econometrics, 222(1), Part A, pp. 73-88. First arXiv version September 2018.
   Control variables provide an important means of controlling for endogeneity in econometric models with nonseparable and/or multidimensional heterogeneity. We allow for discrete instruments, giving identification results under a variety of restrictions on the way the endogenous variable and the control variables affect the outcome. We consider many structural objects of interest, such as average or quantile treatment effects.

Semiparametric Estimation of Structural Functions in Nonseparable Triangular Models (2020, with Victor Chernozhukov, Iván Fernández-Val, Whitney Newey, and Francis Vella).
Quantitative Economics, 11(2), pp. 503–533. First arXiv version November 2017.
   We introduce two classes of semiparametric nonseparable triangular models that are based on distribution and quantile regression modelling of the reduced form conditional distributions of the endogenous variables. We show that average, distribution and quantile structural functions are identified through a control function approach that does not require a large support condition. We propose a computationally attractive three-stage procedure to estimate the structural functions. We provide asymptotic theory and uniform inference methods for each stage.

Simultaneous Mean-Variance Regression (2018, with Richard Spady). arXiv. Supplementary material.
   We propose simultaneous mean-variance regression for the linear estimation and approximation of conditional mean functions. In the presence of heteroskedasticity of unknown form, our method accounts for varying dispersion in the regression outcome across the support of conditioning variables. Simultaneity generates outcome predictions that are guaranteed to improve over OLS prediction error, with corresponding parameter standard errors that are automatically valid. Under misspecification, the corresponding mean-variance regression model weakly dominates the OLS location model under a Kullback-Leibler measure of divergence, with strict improvement in the presence of heteroskedasticity. We obtain large improvements over OLS and WLS in terms of estimation and inference in finite samples.

Dual Regression (2018, with Richard Spady).
Biometrika, 105(1), pp. 1–18 (lead article). First arXiv version October 2012.
   We propose dual regression as an alternative to the quantile regression process for the global estimation of conditional distribution functions under minimal assumptions. Dual regression provides all the interpretational power of the quantile regression process while avoiding the need for repairing the intersecting conditional quantile surfaces. Our approach introduces a mathematical programming characterization of conditional distribution functions. We apply our general characterization to the specification and estimation of a flexible class of conditional distribution functions, and present asymptotic theory for the corresponding empirical dual regression process.

Cross-Validation Selection of Regularization Parameter(s) for Semiparametric Transformation Models (2017, with Senay Sokullu).
Annals of Economics and Statistics, 128, pp. 67–108.
   We propose cross-validation criteria for the selection of regularisation parameter(s) in the semiparametric instrumental variable transformation model proposed in Florens and Sokullu (2016). This model is characterized by the need to choose two regularisation parameters, one for the structural function and one for the transformation of the outcome. We consider two-step and simultaneous criteria, and analyze the finite-sample performance of the estimator using the corresponding regularisation parameters. We find that simultaneous selection of regularisation parameters provides significant improvements in the performance of the estimator.

Research Grants

Principal Investigator, UKRI Frontier Research Guarantee (ERC Starting Grant) (2023-2028):
“Policy and Treatment Evaluation under Model Uncertainty (PoTEMU)”

Principal Investigator, ESRC New Investigator Grant (2019-2022):
“Policy Evaluation Beyond Averages: Distributional Impact Analysis”