讲座题目：Daily Tracking of Economic Conditions via Mixed-frequency Dynamic Factor Models with Serial Correlations
主讲人：王 霞 中国人民大学 副教授
主持人：尚玉皇 西南财经大学 教授
讲座时间：2021年6月3日星期四 中午 12:00-13:00
We introduce a daily dynamic factor model for mixed-frequency growth rate data to track the high-frequency evolution of economic conditions. The model not only allows for serial correlations and conditional heteroscedasticity in idiosyncratic components but also incorporates both flow and stock variables measured at various frequencies. We derive the proper temporal aggregations for low-frequency variables that allow the observation intervals for variables collected at low-frequency to vary with calendar time. By representing the proposed model to a state-space form, we estimate unknown parameters and the common factor via the maximum likelihood method based on the Kalman filter. Monte Carlo studies demonstrate the excellent performance of our estimation and clarify the importance of proper temporal aggregation and allowing serial correlations in idiosyncratic components. In an empirical application to the U.S. economic data, we extract the daily real activity indicator, which captures the recession more accurately and timely than FRED's coincident economic activity index and Aruoba et al.'s (2009) ADS index. We also aggregate the daily common factor to get the monthly and quarterly common factors, which are similar to Mariano and Murasawa's (2003) monthly result and the quarterly GDP growth rate respectively, verifying the reasonability of our daily activity indicator.
王霞，经济学博士，中国人民大学经济学院，副教授，主要从事计量经济学理论及其在经济金融中的应用等研究工作，研究领域包括非线性时间序列分析、宏观经济实时监测预警等。多篇文章发表在International Economic Review、Journal of Econometrics、Journal of Business & Economic Statistics、Econometric Theory等国际主流期刊和《经济研究》等国内顶级期刊上，曾主持国家自然科学基金青年项目、面上项目、教育部人文社科研究青年项目等多项国家和省部级课题。