I just have one important question that I have not been able to answer in the past few days.
The question is about my analysis on Eviews (Eviews10) in order to identify the day of the week effect in volatility. In Eviews I do not have the option to exclude the constant in the variance equation, hence I only include 4 dummies in order to avoid multicollinearity. This means that the constant represents the Wednesday effect on volatility (the dummy I did not include).
Given the fact that I am using a modified version of Garch with dummies I do not think that the standard properties of Garch are still valid.
Given the fact that in the variance equation the constant represents the effect on one day of the week it is not possible to demand for that constant to be positive and significant. For example here I exclude the Wednesday dummy to avoid multicollinearity. The constant represents the Wednesday effect on volatility, therefore it can be negative if the shift in volatility is negative on Wednesdays and if there is no effect at all it could be insignificant. My point is that the sum of the Garch coefficients should be less than one, positive and statistically significant which is the other fundamental restriction of the Garch model, but I cannot expect that the constant will always be positive and significant. Can you please comment on this?
I have attached a screenshot of the model so you can understand better what I am talking about.
For econometric discussions not necessarily related to EViews.
1 post • Page 1 of 1
Who is online
Users browsing this forum: Google [Bot] and 5 guests