Estimation and forecasting with time-varying parameters models and sequential method
Sun, Zhendong
This item's files can only be accessed by the System Administrators group.
Permalink
https://hdl.handle.net/2142/124632
Description
Title
Estimation and forecasting with time-varying parameters models and sequential method
Author(s)
Sun, Zhendong
Issue Date
2024-03-29
Director of Research (if dissertation) or Advisor (if thesis)
Amir-Ahmadi, Pooyan
Doctoral Committee Chair(s)
Amir-Ahmadi, Pooyan
Committee Member(s)
Bernhardt, Dan
Xie, Shihan
Chen, Yuguo
Department of Study
Economics
Discipline
Economics
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
Economic Forecasting
Time-Varying Parameters Model
Sequential Monte Carlo
Abstract
In this research, we examine the use of time-varying parameters (TVP) models for out-of-sample forecasting within the realms of macroeconomics and finance. From a methodological perspective, the efficacy of the Sequential Monte Carlo (SMC) method in estimating TVP models is emphasized. Notably, SMC provides a distinct computational edge, requiring substantially less processing time relative to the traditional Markov Chain Monte Carlo (MCMC) method, all the while preserving predictive accuracy. Furthermore, we augment a generic SMC approach by incorporating the variational Bayes method, thereby enabling it to estimate large TVP models with an integrated variable selection prior. Empirically, we embark on a detailed exploration of three out-of-sample predictive applications in the fields of macroeconomics and finance: 1) the estimation of US GDP and inflation via a trivariate VAR model; 2) the forecasting of monthly returns of the S$\&$P500 index, which integrates a comprehensive set of 143 predictors; and 3) the nowcasting of US GDP using a TVP VAR model enriched with mixed-frequency variables. Consistently, across these analytical domains, findings suggest that TVP models bolster predictive capabilities, surpassing both their fixed-parameter counterparts and other advanced methodologies.
Use this login method if you
don't
have an
@illinois.edu
email address.
(Oops, I do have one)
IDEALS migrated to a new platform on June 23, 2022. If you created
your account prior to this date, you will have to reset your password
using the forgot-password link below.
