Page 1 of 1

Does mixed-frequency Vector Autoregression help to improve forecasting performance?

Posted: December 17th, 2015, 8:54 am
by EdisonCruise
I am now using VAR to forecast asset return. I am trying to combine daily asset data with monthly or weekly macroeconomic data in order to improve the forecasting performance of my model. I indeed find some literature on mixed-frequency VAR. Typically the research work by Eric Ghysels. However, I find Eric Ghysels?s works are mainly limited in two-factor model, one low-frequency data time series and one high-frequency data time series. Is there similar study on multi-factor model?Eric Ghysels? works also focus on investigating the impact between low-frequency and high-frequency data. Is there similar study on asset return forecasting?Actually is there any literature on using mixed-frequency data and multifactor VAR model to forecast asset return? Any potential to work?Any comment is appreciated.