报告题目：Addressing Parameter Instability in Latent Factor Models: A Combined Discrete Fourier Transform Based Test
报告嘉宾： 王霞 中国人民大学
报告时间：2024年11月25日(周一) 16:30-18:00
报告地点：中国科学院大学中关村校区教学楼S406
腾讯会议：917-837-482
内容摘要
The existence of parameter instability causes the inconsistent estimation of factor models when apply principal component analysis due to the violation of constant factor loadings assumption. Since the signal of such structural changes is usually contained both in the residuals and factors, we propose a combined test to detect the structural changes in the factor loadings using discrete Fourier transformation (DFT) approach. Our combined DFT-based test considers the information of estimated residuals and factors, while allowing for self-determined weights to balance their relative significance. Simulation results show that our combined test have reasonable size and better power in either abrupt breaks or smooth changes especially when breaks or changes are in large magnitudes. We further apply our test statistics to the U.S. macroeconomic data and Chinese stock market data, where our test demonstrates robustness and has better performance especially in the financial data analysis.
主讲人简介
王霞，中国人民大学吴玉章特聘教授，博士生导师，入选国家级青年人才。2013年于厦门大学王亚南经济研究院获得经济学博士学位。目前从事理论计量经济学以及宏观经济监测与预测等研究工作，研究成果发表在 International Economic Review, Journal of Econometrics, JBES， 经济研究等高水平期刊上。主持多项国家自然科学面上、青年基金。
