[金融与经济学术论坛2017年第五期]
讲座题目:Validating Market Risk Factors and Forecasting Bond Risk Premia using Novel Factor Models
主讲人:Jianqing Fan, Princeton University
主持人:李平
时间:2017年3月20日下午2点
地点:北航新主楼A座928
摘要:
We provide econometrics methods and analysis for validating market risk factors andforecasting bond risk premium via factor models in which the latent factors canbe partially explained by several observed explanatory proxies. In financialfactor models for instance, the unknown factors can be reasonably wellpredicted by a few observable proxies, such as the Fama-French factors. Indiffusion index forecasts, identified factors are strongly related to severaldirectly measurable economic variables such as consumption-wealth variable,financial ratios, and term spread. To incorporate the explanatory power ofthese observed characteristics, we propose a new two-step estimation procedure:
(i) Regress the data onto the observables, and (ii) take the principalcomponents of the fitted data to estimate the loadings and factors. Theproposed estimator is robust to possibly heavy-tailed distributions, which areencountered by many macroeconomic and financial time series. With thoseproxies, the factors can be estimated accurately even if the cross-sectionaldimension is mild. Empirically, we apply the model to forecast US bond riskpremia, and find that the observed macroeconomic characteristics contain strongexplanatory powers of the factors. The gain of forecast is more substantialwhen these characteristics are incorporated to estimate the common factors thandirectly used for forecasts.
主讲人简介:
Jianqing Fan is Frederick L. Moore '18 Professor of Finance, Professor of Statistics, andformer Chairman of Department of Operations Research and Financial Engineeringat the Princeton University. He previously held professorships at CUHK,UNC-Chapel Hill, and UCLA. He has authored or co-authored over 200 articles onfinancial econometrics, computational biology, and various aspects oftheoretical and methodological statistics. His finance work focuses on theanalysis of high-frequency data, portfolio allocation, risk management, timeseries, high-dimensional data, and non-parametric modelling. His published workhas been recognised by the 2000 COPSS Presidents’ Award, the 2007 MorningsideGold Medal of Applied Mathematics, and a Guggenheim Fellowship in 2009,Academian of Academia Sinica 2012, Guy Medal in Silver, 2014. He is an ElectedFellow of the American Association for Advancement of Science, the Institute ofMathematical Statistics, and the American Statistical Association, and a pastPresident of the Institute of Mathematical Statistics. He is the co-editor ofJournal of Econometrics and associate editor of Journal of American StatisticalAssociation, and past co-editor of Annals of Statistics and Probability Theoryand Related Fields and has served as as co-editor of Econometrics Journal andas associate editor of Econometrica, and Journal of Financial Econometrics.
经管学院科研办
2017-03-13