Abstract:This article proposes a new approach to modeling high-dimensional time series by treating a $p$-dimensional time series as a nonsingular linear transformation of certain common factors and idiosyncratic components. Unlike the approximate factor models, we assume that the factors capture all the non-trivial dynamics of the data, but the cross-sectional dependence may be explained by both the factors and the idiosyncratic components. Under the proposed model, (a) the factor process is dynamically dependent and the idiosyncratic component is a white noise process, and (b) the largest eigenvalues of the covariance matrix of the idiosyncratic components may diverge to infinity as the dimension $p$ increases. We propose a white noise testing procedure for high-dimensional time series to determine the number of white noise components and, hence, the number of common factors, and introduce a projected Principal Component Analysis (PCA) to eliminate the diverging effect of the idiosyncratic noises. Asymptotic properties of the proposed method are established for both fixed $p$ and diverging $p$ as the sample size $n$ increases to infinity. We use both simulated data and real examples to assess the performance of the proposed method. We also compare our method with two commonly used methods in the literature concerning the forecastability of the extracted factors and find that the proposed approach not only provides interpretable results, but also performs well in out-of-sample forecasting. Supplementary materials of the article are available online.

Noise dressing of the correlation matrix of factor models

Multi-Wavelength Surveys for Galaxies Hidden by the Milky Way

EIGENVALUE DISTRIBUTIONS OF VARIANCE COMPONENTS ESTIMATORS IN HIGH-DIMENSIONAL RANDOM EFFECTS MODELS

Asymptotic locations of bounded and unbounded eigenvalues of sample correlation matrices of certain factor models -- application to a components retention rule

The bulk of the stock market correlation matrix is not pure noise

The Application of Spectral Distribution of Product of Two Random Matrices in the Factor Analysis

The Application of Spectral Distribution of Product Matrices of Large Dimensional Random Matrices in the Factor Analysis

Time-Varying Matrix Factor Models

Factor Models for Matrix-Valued High-Dimensional Time Series

Asymmetric correlation matrices: an analysis of financial data

Inferring financial stock returns correlation from complex network analysis

A new method to estimate the noise in financial correlation matrices

Factor Strength Estimation in Vector and Matrix Time Series Factor Models

Factor correlation and the cross section of asset returns: A correlation-robust machine learning approach

Factor Models for Portfolio Selection in Large Dimensions: The Good, the Better and the Ugly

Knowing Factors or Factor Loadings, or Neither? Evaluating Estimators of Large Covariance Matrices with Noisy and Asynchronous Data

The Dynamic, the Static, and the Weak factor models and the analysis of high-dimensional time series

Modeling High-Dimensional Time Series: A Factor Model with Dynamically Dependent Factors and Diverging Eigenvalues

A Score-Driven Conditional Correlation Model for Noisy and Asynchronous Data: An Application to High-Frequency Covariance Dynamics

Approximate Factor Models with Strongly Correlated Idiosyncratic Errors

Dynamic partial correlation models