A0652
Title: Identification and estimation of parameter instability in high dimensional approximate factor models
Authors: Yiru Wang - University of Pittsburgh (United States) [presenting]
Ruiqi Liu - Texas Tech University (United States)
Abstract: A novel approach is introduced for estimating structural break ratios in the factor loadings of high-dimensional approximate factor models, where the breaks occur at unknown common dates and the number of factors is unknown. The method is based on the observation that the sum of the numbers of pseudo factors in the pre- and post-split subsamples is minimized when the sample is split at the structural break. By appropriately transforming these criteria using the eigenvalue ratios of the covariance matrices of the pre- and post-split subsamples, consistent estimators are derived for the structural break ratios. Notably, the framework exhibits remarkable flexibility in accommodating weak factors and can be easily extended to handle multiple breaks. A data-driven process is also introduced to determine the number of breaks. Monte Carlo simulations demonstrate the good performance of the proposed estimators. Furthermore, in an empirical analysis of the FRED-MD dataset, two structural breaks are identified around January 1983 and March 2009.
