Maximum a Posteriori Estimation of Dynamic Factor Models with Incomplete Data

TL;DR

This paper introduces a Bayesian approach for estimating dynamic factor models with incomplete data, incorporating shrinkage priors and adaptable algorithms, improving estimation robustness and flexibility.

Contribution

It extends existing EM-based methods by integrating penalized estimation with shrinkage priors, handling arbitrary missing data patterns in dynamic factor models.

Findings

Method performs favorably in Monte Carlo simulations.

Algorithm handles arbitrary missing data patterns.

Comparable or better than maximum likelihood estimation.

Abstract

In this paper, we present a method of maximum a posteriori estimation of parameters in dynamic factor models with incomplete data. We extend maximum likelihood expectation maximization iterations by Ba\'nbura & Modugno (2014) to penalized counterparts by applying parameter shrinkage in a Minnesota prior style fashion, also considering factors loading onto variables dynamically. A heuristic and adapting shrinkage scheme is considered. The algorithm is applicable to any arbitrary pattern of missing data, including different publication dates, sample lengths and frequencies. The method is evaluated in a Monte Carlo study, generally performing favourably, and at least comparably, to maximum likelihood.