Principal component error in high-dimensional factor models
Abstract
In a statistical factor model, principal components (or eigenvectors) of a sample covariance matrix serve as estimates of principal directions, the true drivers of co-movement of a collection of observed variables. We write the often substantial error in these estimates as a sum of two interpretable terms, which we show have almost sure asymptotic limits as the number of variables grows with sample size bounded. This scenario is commonplace in financial economics, genomics, machine learning and signal processing. Out-of-subspace error measures the distance from an estimate to the subspace spanned by population factor exposures. It can be expressed in terms of data, providing an estimable floor for asymptotic error. In-subspace error is driven by a rotation error within that subspace, which arises from the fixed sample size of the latent factor returns and cannot be estimated from data alone. We illustrate our error analysis with a three-factor simulation of the US public equity market, showing the dependence of the magnitude of the error and its components on dimension and sample size. In that simulation, out-of-subspace error dominates. Researchers who rely on principal component analysis to estimate factor models can use our results to quantify errors in model-based predictions and attributions.
Related
- High-dimensional Four-Level Structured Functional Principal Component Analysis for Unbalanced Level Data with Application to Concussion Data
- Two-sample tests for principal eigenvalues and eigenvectors in high-dimensional elliptical factor models
- Geometric Fluctuations of the sinΘ Distance in High-Dimensional Principal Subspace Estimation
- Nonparametric shrinkage estimation of high-dimensional variance component matrices
- Principal component analysis for a mix of stationary and nonstationary variables
BibTeX
@misc{bernstein2026principal,
title = {{Principal component error in high-dimensional factor models}},
author = {Alex Bernstein and Lisa R. Goldberg and Nicholas Gunther and Alec N. Kercheval and Tian Lan and Yian Lin and Dayi Yao},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2609.20550},
eprint = {2609.20550},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.20550},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Principal component error in high-dimensional factor models
AU - Alex Bernstein
AU - Lisa R. Goldberg
AU - Nicholas Gunther
AU - Alec N. Kercheval
AU - Tian Lan
AU - Yian Lin
AU - Dayi Yao
PY - 2026
DA - 2026/10/15
PB - arXiv
DO - 10.48550/arxiv.2609.20550
UR - https://arxiv.org/abs/2609.20550
AB - In a statistical factor model, principal components (or eigenvectors) of a sample covariance matrix serve as estimates of principal directions, the true drivers of co-movement of a collection of observed variables. We write the often substantial error in these estimates as a sum of two interpretable terms, which we show have almost sure asymptotic limits as the number of variables grows with sample size bounded. This scenario is commonplace in financial economics, genomics, machine learning and signal processing. Out-of-subspace error measures the distance from an estimate to the subspace spanned by population factor exposures. It can be expressed in terms of data, providing an estimable floor for asymptotic error. In-subspace error is driven by a rotation error within that subspace, which arises from the fixed sample size of the latent factor returns and cannot be estimated from data alone. We illustrate our error analysis with a three-factor simulation of the US public equity market, showing the dependence of the magnitude of the error and its components on dimension and sample size. In that simulation, out-of-subspace error dominates. Researchers who rely on principal component analysis to estimate factor models can use our results to quantify errors in model-based predictions and attributions.
ER -