Title: A test for the number of factors in dynamic factor models
Authors: Daniel Pena - Universidad Carlos III de Madrid (Spain) [presenting]
Abstract: An eigenvalue ratio test for the number of dynamic factors is presented. The test combines the advantages of those proposed previously for the eigenvalues of the covariance matrix, and for the cumulative sum of lagged covariance matrices. A pooled correlation matrix is defined as a weighted combination of the main observed correlation matrices and the proposed test is based on the ratio of consecutive eigenvalues. Some theoretical results are given to justify the good expected properties of the test and a Monte Carlo study is presented showing its good finite sample performance compared to previous approaches. The usefulness of the test is also illustrated in an example with real macroeconomic data.