Testing for Common Autocorrelation in Data Rich Environments
Cubadda GianlucaHecq Alain
CEIS Research Paper
This paper proposes a strategy to detect the presence of common serial correlation in high-dimensional systems. We show by simulations that univariate autocorrelation tests on the factors obtained by partial least squares outperform traditional tests based on canonical correlations.
Keywords: Serial correlation common feature; high-dimensional systems; partial least squares. JEL code: C32;
JEL codes: C32
Date: Friday, December 4, 2009
Revision Date: Friday, December 4, 2009