I want to analyze the relationship between three capital markets. For doing so I got the respective market index on a daily basis for several years (CAC, DAX, FTSE). If I analyze those series, they are I(1) but they also present a volatility component. If I adjust a VAR over those series in first difference for identifying the order of a VAR to use for cointegration testing purposes, obviously, residuals also present a volatility component (it was expected) and I cannot proceed with the cointegration and causality tests. How should I deal with that component? Should I adjust a GARCH first for each series? If I work with monthly data, volatility "disappears" and I get and adequate VAR and cointegration test with one cointegration vector. Would it be wrong If I work with monthly data for a long run analysis?
it is better to use GARCH model for volatility. you can do co-integration and also error correction in garch model.
see Wong and lee (2009)
- Badges
- Science topic
- Similar topics
- Biological Science
- Histology
- Fixatives