Quasi-Maximum Likelihood Estimation of Dynamic Factor Models with Time-Varying Loadings

• Grant number: 25/25711-4
• Support Opportunities: Scholarships abroad - Research Internship - Doctorate (Direct)
• Start date: February 02, 2026
• End date: August 01, 2026
• Field of knowledge: Physical Sciences and Mathematics - Probability and Statistics - Statistics
• Principal Investigator: Pedro Alberto Morettin
• Grantee: Davi Oliveira Chaves
• Supervisor: Esther Ruiz
• Host Institution: Instituto de Matemática e Estatística (IME). Universidade de São Paulo (USP). São Paulo , SP, Brazil
• Institution abroad: Universidad Carlos Iii De Madrid, Campus De Getafe, Spain
• Associated to the scholarship: 23/11547-2 - Factor models with vector autoregressive dynamics, BP.DD

Abstract

Given the increasing availability of data in recent years, Dynamic Factor Models (DFMs) have become particularly advantageous for modeling multivariate time series, as they are able to represent the dynamic evolution using a small number of parameters. However, in fields such as econometrics, the conventional DFM may be overly restrictive, as it assumes constant factor loadings, an unrealistic assumption when long periods of time are considered. Since quasi-maximum likelihood (QML) estimation methods for DFMs offer some analytical advantages over principal component analysis methods, such as jointly modeling the serial dependence in the factors, this project aims to develop an appropriate EM-type algorithm to perform QML estimation of DFMs with time-varying loadings.