Macro/Trade Seminar | Binary Models with Extreme Covariates: Estimation and Prediction by Laura Liu
School of Economics
Invites you to a
Macroeconomics and Trade seminar presented by
Laura Liu
(University of Pittsburg)
Binary Models with Extreme Covariates: Estimation and Prediction
Tuesday 14 November 2023
2:00pm – 3.30pm
Level 6 Seminar Room (650)
A02 Social Sciences Building
Camperdown Campus
The University of Sydney NSW 2006
This paper presents a novel semiparametric approach to studying the effects of extreme events on binary outcomes. With recent occurrences of extreme events, it is important to understand their impact on binary outcomes such as employment status, recession, and country default. Our approach is based on Bayes’ theorem and Pareto approximations at the tail, allowing for a flexible relationship between covariates and outcomes. We further extend the discussion to static and dynamic panel data models accounting for individual-specific tail thickness. We establish consistency and asymptotic normality of the proposed estimator. We evaluate its finite-sample properties in Monte Carlo simulations. In the empirical application, we use a panel of small banks to study whether small banks become riskier when local housing prices experience a significant decline, a channel crucial in the 2007-08 financial crisis.
For further information contact: Macroeconomics and Trade seminar series coordinator Dr James Graham (james.a.graham@sydney.edu.au)
For all upcoming seminars in School of Economics see Our events and Calendar