School of Economics

  Invites you to a

Econometrics seminar presented by

Fu Ouyang

(Queensland University)

High-Dimensional Binary Choice Models with Unknown Heteroskedasticity or Instrumental Variables

 Wednesday 31 May 2023

2.00pm – 3.30pm

Level 6 Seminar Room (650)

A02 Social Sciences Building
Camperdown Campus

The University of Sydney NSW 2006 

Via Zoom: 844 7479 5992

Our paper proposes a practical special regressor approach for high-dimensional binary choice models. The model we consider places no distributional assumptions on the error term, allows the error to be heteroskedastic, and permits endogenous regressors. However, the existing method proposed by Lewbel (2000) requires nonparametric conditional density estimation and suffers from the curse of dimensionality, making it infeasible in high-dimensional settings. To address this issue, we introduce a data-driven dimension reduction method for nonparametric kernel estimators that combines distance correlation-based screening and cross-validation procedures. We then construct a GMM estimator with SCAD penalty to handle cases where there are many candidate instrumental variables, some of which may not be valid. We’ve fully developed the asymptotic properties of both our selection and estimation methods and have shown through Monte Carlo simulations that our proposed procedure performs well in finite sample sizes.

For further information contact: Econometrics seminar series coordinator Dr Ye Lu (ye.lu1@sydney.edu.au)

For all upcoming seminars in School of Economics see Our events and Calendar