Hi,
i have a question about my model for the thesis.
I have identified the periods that go back to the bursting of bubbles within the series of oil and natural gas prices through a methodology. I would like to include the variables =1 in the presence of a bubble and =0 otherwise in a VAR model. I would like to estimate a model that allows me to understand how the aforementioned bubbles influence the price of the US stock market index.
Yes, you can use dummy variables in a VAT model in the same way that you would in an OLS regression.
Yes, as an explanatory variable, of course, not among the variables of the VAR. See e.g. Lütkepohl's book, chapter 7. He used them as seasonal dummies.
Obviously, you try to avoid a clear specification of your model by using a VAR estimation method. I think you should proceed in two steps. First, explaining the movements of oil and gas prices (with the bubble dummies as explaining variables; how do you determine, if there is a bubble or not). Here you can try to find out, if the stock market index has an effect on energy prices by including it as explaining variables. The second step would be to test the effect of oil and gas price on this index. I think, that a VAR estimation would mix-up the effects in both directions and, therefore, hide more than it shows.
https://stats.stackexchange.com/questions/181109/building-var-vector-autoregression-model-with-dummy-variables-in-r
- Badges
- Science topic