The semi-parametric midas models and some of their applications: the impact of news on the stock volatility Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 21, 2019
  • Chen, Xilong
    • Affiliation: College of Arts and Sciences, Department of Economics
  • In the first essay, I examine whether the sign and magnitude of discretely sampled high frequency returns have impact on expected volatility over some future horizon. Technically speaking, I introduce semi-parametric MIxed DAta Sampling (henceforth MIDAS) regressions. I show that the asymptotic distribution of semi-parametric MIDAS regressions depends on mixture of sampling frequencies. Also novel is the parametric specification I consider to deal with (intra-daily) seasonality. In the empirical work, I find that moderately good (intra-daily) news reduces volatility (the next day), while both very good news (unusual high intra-daily positive returns) and bad news (negative returns) increase volatility, with the latter having a more severe impact. The asymmetries disappear over longer horizons. I also introduce a new class of parametric models with close ties to ARCH-type models, albeit applicable to high frequency data. In the second essay, I extend the semi-parametric MIDAS model to multivariate case and find that besides the asymmetric effect, the market-wide news and firm-specific news interactively affect the individual firm's future volatility and using both of them can increase the out-of-sample forecast performance. In the third essay, I propose a new type of semi-parametric MIDAS index model, which potentially applies in a variety of fields, and investigate its estimation and asymptotics.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Ghysels, Eric
Degree granting institution
  • University of North Carolina at Chapel Hill
  • Open access

This work has no parents.