I have an issue with ARDL right now. I have done unit root test, cointegration test and an ARDL and got my values. Three of my variables are for different emission types and my dependent variable is healthcosts. The result showed that one of my emission types would have a negative relation with my dependent variable, which makes little sense. I´ve seen people using DOLS and FMOLS to check whether the ARDL is robust, so I tried this. The results from DOLS and FMOLS gave totally different significance for my emission variables, but not the coefficients which I've understod is good. The rest of my variables had also almost the same coefficients and significance. I am wondering how i should interpret this? Looking through the similar articles they barely comment on this, even though ive seen the same phenomena in those as well. Ive done some tests and cant really see what is wrong with the model either.
Ideally, I would hope that this means that the emission variables are inconclusive since they are shifting, but not the rest. But since I cant find any information on this, I'd like to hear what you guys who knows this have to say.
ARDL, DOLS and FMOLS
Moderators: EViews Gareth, EViews Moderator
-
- Posts: 17
- Joined: Wed Jun 23, 2021 1:48 pm
Re: ARDL, DOLS and FMOLS
Honestly, it's going to be extremely difficult to measure something like emissions' impact on health care costs, irrespective of the estimation technique.
I'd look for a natural experiment here, rather than a time series estimation.
But anyway, I think your conclusion that the results are inconclusive sounds sensible to me.
I'd look for a natural experiment here, rather than a time series estimation.
But anyway, I think your conclusion that the results are inconclusive sounds sensible to me.
-
- Posts: 2
- Joined: Wed Apr 19, 2023 10:34 am
Re: ARDL, DOLS and FMOLS
random_user wrote:Honestly, it's going to be extremely difficult to measure something like emissions' impact on health care costs, irrespective of the estimation technique.
I'd look for a natural experiment here, rather than a time series estimation.
But anyway, I think your conclusion that the results are inconclusive sounds sensible to me.
I know its going to be difficult, but ive seen a lot of studies doing the same thing and I will also add that the study is going to need added on research. My other variables are not differeing, so I will at least have some ideas of the effect that can be added. Thanks any way for the answer!
Return to “Econometric Discussions”
Who is online
Users browsing this forum: No registered users and 11 guests