Professor at the University of Chicago-Department of Economics
I did my PhD at CREST and Université Paris I, under the supervision of Jean-Marc Robin.
You can see my CV HERE.
Labour economics, Microeconometrics, and econometric theory.
With a special interest in latent variable modelling and panel data.
I teach Econometrics at the graduate level.
You can see some class notes at the bottom of this page.
Keywords: Unobserved heterogeneity, dimension reduction, panel data, structural estimation.
Summary: We develop two-step and iterative estimators based on a discretization of unobserved individual heterogeneity. We view discrete estimators as approximations, and study their properties in environments where population heterogeneity is unrestricted.
[FORTHCOMING] Earnings and Consumption Dynamics: A Nonlinear Panel Data Framework (with Manuel Arellano and Richard Blundell)
To appear in Econometrica
Keywords: Earnings Dynamics, Consumption, Latent Variables.
Summary: We develop a flexible framework to study the nonlinear relationship between shocks to household earnings and consumption over the life cycle.
[NEW] Nonlinear Panel Data Methods for Dynamic Heterogeneous Agent Models (an overview prepared for the Annual Review of Economics), with Manuel Arellano.
[UNDER REVISION] A Distributional Framework for Matched Employer Employee Data (with Thibaut Lamadon and Elena Manresa)
Keywords: Matched employer employee data, sorting, job mobility, models of heterogeneity.
Summary: We develop methods to estimate the contributions of firms and workers to earnings dispersion that allow for complementarities and dynamics.
[FORTHCOMING] A survey on Sample Selection in Quantile Regression (prepared for the Handbook of Quantile Regression), with Manuel Arellano.
To appear in Economic Journal
Keywords: Earnings Inequality, Social Security data, Unemployment, Business cycle.
Summary: We use Spanish Social Security data to document the evolution of earnings inequality since the end of the 1980's. Male inequality is strongly countercyclical, and partly reflects the effects of the housing boom and bust on the construction sector.
A complement of the paper using tax data is:
Applied Economics, 45, 2013, 4212-4225.
Econometrica, 85(1), 1-28, January 2017.
Keywords: Quantile regression, sample selection, gender wage gap, copula.
Summary: We propose a method to correct quantile regression estimates for sample selection, and apply it to study wage and employment in the UK.
Code of the quantile selection estimator (in matlab), to be used on any data set.
Econometrics Journal, 19(3), C61-C94, October 2016.
Keywords: Panel data, quantile regression, Expectation-Maximization.
Summary: We introduce a class of quantile regression estimators for short panels. Our correlated random-effects approach uses quantile regressions as flexible tools to model conditional distributions.
Economic Journal, 126(594), 1342-1371, August 2016.
Keywords: labor turnover, compensating differentials, teacher labor markets, sample selection.
Summary: We propose a method to estimate workers' preferences for job amenities using data on job changes. We apply it to administrative data on Dutch primary school teachers.
Annals of Statistics, 44(2), 540-563, April 2016.
Keywords: finite-mixture models, hidden Markov models, simultaneous matrix diagonalization.
Summary: We develop an approach to identify and estimate latent structure models, such as finite mixture models or hidden Markov models.
A companion paper is Nonparametric Estimation of Non-Exchangeable Latent Variable Models (with Koen Jochmans and Jean-Marc Robin)
To appear in Journal of Econometrics
Journal of Royal Statistical Society, Series B, 78(1), 211-229, January 2016.
Keywords: finite-mixture models, repeated-measurement data, re-weighting, two-step estimation.
Summary: We develop a practical two-step procedure to nonparametrically estimate finite mixture models from data on repeated measurements.
Econometrica, 83(3), 1147-1184, May 2015.
Keywords: Discrete heterogeneity, panel data, fixed effects, democracy.
Summary: We propose an alternative to fixed-effects estimation in linear panel data regression that allows for group-level time-varying unobservables. We use this approach to document the evolution of income and democracy in the last part of the XXth century.
Econometrica, 80(4), 1337-1385, July 2012.
Keywords: Panel data, incidental parameters, inverse problems.
Summary: We propose a general method (a ``nonlinear within transformation'') to difference out the individual fixed effects in likelihood-based panel data models.
Review of Economic Studies, 79(3), 987-1020, July 2012.
Keywords: Panel data, random coefficients, multiple effects, nonparametric identification.
Summary: We provide identification results and construct estimators for variances, and more generally densities, of individual fixed effects in linear panel data models.
Discussion of papers prepared for the World Congress of the Econometric Society (Shanghai, 2010), 338-352, September 2011.
Keywords: Latent variables models, lasso, penalization.
Summary: A discussion of papers by Susanne Schennach and Victor Chernozhukov. We apply a penalized least squares (``lasso'') density estimator to a simple measurement error model.
Annual Review of Economics, 3, 395-424, 2012.
Keywords: Panel data, incidental parameters, partial identification.
Summary: A survey of recent advances in panel data, including a discussion of the computation of random-effects estimators, and partial identification.
Review of Economics and Statistics, 93(2), 479–494, May 2011.
Keywords: Selective education, difference-in-differences, treatment effects, quantiles.
Summary: We introduce a method to estimate distributions of potential outcomes in difference-in-differences models. We apply it to study the effect of selective secondary education in the UK.
Review of Economic Studies, 77(2), 491-533, April 2010.
Keywords: Factor models, nonparametric deconvolution, earnings dynamics.
Summary: We propose a nonparametric estimator of factor distributions in linear factor independent factor models. We use it to estimate the distribution of earnings shocks in a simple permanent/transitory model, on PSID data.
Journal of Applied Econometrics, 24(5), 763-795, June 2009.
Keywords: Compensating wage differentials, job mobility, amenities.
Summary: We build and estimate a structural job search model of wages, non-wage amenities, and job mobility on European data. We find strong preferences for some amenities, which are not reflected in wage/amenity correlations.
Journal of Econometrics, 149(1), 12-25, April 2009.
Keywords: Factor analysis, Independent Component Analysis, higher-order moments.
Summary: We propose a method to estimate linear independent factor models using higher-order moments. We apply it to schooling data on test scores, and to the Fama-French data on stock returns.
Gauss codes for the ``quasi-JADE'' estimator
Econometrica, 77(2), 489-536, March 2009.
Keywords: Panel data, integrated likelihood, bias reduction
Summary: We show under which condition panel data integrated likelihood estimators reduce first-order bias on common parameters and average marginal effects. Our framework covers fixed-effects, random-effects, and Bayesian approaches as special cases.
Review of Economic Studies, 76(1), 63-92, January 2009.
Keywords: Inequality, mobility, earnings dynamics, copula.
Summary: We build a model of earnings dynamics that combines a flexible modelling of the marginal distributions (inequality) with a tight parameterization of the dynamics (mobility). We use it to compute inequality in Present Values. We estimate the model on the French Labor Force Survey.
Keywords: EM algorithm, standard errors.
Summary: A simple method to estimate asymptotic standard errors of sequential EM estimators, using the generalized information identity.
Statistical Methods in Econometrics
A gentle introduction to quantile regression