Kong Xin-Bing，Liu Cheng（刘成）

Publication:

Journal of Econometrics

Abstract:

In this paper, we introduce a nonparametric test against the constancy of the factor loading matrix of a high-dimensional continuous-time factor model using high-frequency data. The central limit theorems on the test statistics with and without perturbation are established under the null hypothesis that the factor loading matrix is constant as time evolves. The tests perform well in size and power. Interestingly, the test statistic without perturbation converges at a rate that depends not only on the sample size but also on the dimension through the cross-sectional dependence of the residual process, which is a distinctive feature that contrasts with the low-dimensional setting. Extensive numerical studies, including Monte Carlo simulations and real data analysis, validate the performance of our test. (C) 2018 Elsevier B.V. All rights reserved.

【Keywords】Continuous-time factor model;Factor loading matrix;High dimensional itô process

链接地址：https://www.sciencedirect.com/science/article/pii/S0304407618300393?via%3Dihub