1.Local Whittle Memory Parameter Estimation of Long Memory Process in the Presence of Low Frequency Contaminations (Job Market Paper) (Jie Hou and Pierre Perron)
We propose local-Whittle estimator of the memory parameter of a long memory time series process which has good properties under an almost complete collection contamination process that have already been identified in the literature. Those contaminations include: 1. processes whose spectral density function dominates at low frequency level in the frequency domain (also called low frequency contamination, denoted by LFC), examples of which are Random Level Shifts (denoted by RLS thereafter), deterministic level shifts and deterministic trends; 2. short memory dynamics; 3. additive noise. We give the asymptotic property of the proposed estimator. Through extensive simulation study, we show that the proposed estimator has substantial efficiency gains compared to existing semi parametric estimators and still remain valid on cases where the current estimators fail to apply. We analyze the effects of bandwidth selection on the performance of the estimators under different contaminations and the tradeoffs between bias and RMSE among those estimators. We recommend general rules to select the bandwidth and the exact form of the estimator in order to minimize the bias and RMSE.
2.A Pivotal Test Statistic against Structural Changes in Multi-Equation Regression Systems (Jie Hou and Pierre Perron)
We introduce an interesting relationship between test statistics against differnt alternative hypothesis in which the covariance that is allowed to change varies. We utilize this relationship to construct a pivotal likelihood ratio test statistic which is invariant under the different alternative hypothesis on structural changes in multivariate regression models. It is a multi-dimensional expansion of Zhou and Perron (2008). We then use this test statistic to expand their result of testing problems TP-1 to TP-4 to multi-dimensional case.
Work in Progress
1.“Memory Parameter Estimation of S$P 500 Data Robust to Low Frequency Contaminations”
2.“Using Interactive Effects Factor Analysis to Improve the BLP model”
Hou, Jie and Yang, Lijun (2006) “Triangularization of a class of C1 unipotent maps,” the Fifth International Conference on Information and Management Sciences, Chengdu, China (ISTP Indexed).