Mode‐adaptive factor models

Tao Wang

Scandinavian Journal of Statistics, 2025, vol. 52, issue 3, 1206-1238

Abstract: Factor models are indispensable tools in economics and finance, providing valuable insights into the latent structures underlying complex datasets. Nevertheless, the prevalence of heavy‐tailed macroeconomic and financial data, often characterized by extreme values and greater skewness than that found in a normal distribution, presents significant analytical challenges. This article introduces mode‐adaptive factor models (MAFM) for robust factor analysis in high‐dimensional panel data, inspired by the equivalence between conventional principal component analysis and the constrained least squares method in factor models. Unlike traditional factor models that concentrate on mean estimation, MAFM leverage the mode to capture central tendencies more effectively, particularly in the presence of skewed and heavy‐tailed distributions. To facilitate MAFM for factor analysis, we develop an iterative mode regression algorithm that integrates the expectation‐maximization procedure, ensuring convergence to a stationary solution. We establish the theoretical properties of the MAFM estimators without imposing moment constraints on idiosyncratic errors and propose a mode information criterion for consistent factor number selection. We also suggest a data‐dependent bandwidth selection procedure to enhance the flexibility of MAFM. The simulation studies demonstrate the effectiveness of MAFM across diverse distributional settings. An empirical application to macroeconomic forecasting further highlights the practical advantages of MAFM, showcasing their robustness and efficacy in real‐world analyses.

Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
https://doi.org/10.1111/sjos.12785

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:bla:scjsta:v:52:y:2025:i:3:p:1206-1238

Ordering information: This journal article can be ordered from
http://www.blackwell ... bs.asp?ref=0303-6898

Access Statistics for this article

Scandinavian Journal of Statistics is currently edited by ÿrnulf Borgan and Bo Lindqvist

More articles in Scandinavian Journal of Statistics from Danish Society for Theoretical Statistics, Finnish Statistical Society, Norwegian Statistical Association, Swedish Statistical Association
Bibliographic data for series maintained by Wiley Content Delivery ().