Bayesian Sensitivity Analysis for Set-identified Models
Holder： Yizhou Kuang（University of Manchester）
Time：2024-05-09 15:10-17:00
Location：Room 217, Guanghua Building 2
Abstract:
This paper proposes a new algorithm to conduct sensitivity analysis for set-identified structural models. It combines standard Bayesian methods with a characterization of observationally equivalent parameters. The algorithm finds the range of posterior means and the Bayesian credible region of both the structural parameters and any parameters of interest, ensuring robustness against the selection of priors within a class that produces identical marginal likelihoods. I provide theoretical support for this algorithm and apply the method to the model in Cochrane (2011), as well as to DSGE models in An and Schorfheide (2007) and Smets and Wouters (2007), to show its relevance in policy analysis. I find that, in set-identified models, parameters of primary interest like impulse responses can be have very different implications based on the prior, even within the same class. Additionally, optimal monetary policy rules varies with the choice of prior in that class. It is also shown that standard estimation procedure may unexpectedly alter the analysis significantly in highly structural models even when the set-identified parameters are calibrated at the `true' values.
About the Speaker:
Dr. Kuang is a Lecturer (Assistant Professor) at the Department of Economics, University of Manchester. His research focuses on Bayesian econometrics, partial identification, and information theory, with a general interest in both the theoretical and applied dimensions of econometrics and macroeconomics.
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
