Dear all,
I am analysing the determinants of international currency adoption. I am interested in how their relative importance might be changing through time.
One of my hypotheses is that post-9/11, network effects and inertia weakened. To test this out, I've run panel data regressions on the pre and post-9/11 samples. I've also run a test on the full sample where I included a post-9/11 dummy variable and interaction terms. Just looking at the coefficients, my hypothesis looks valid, but I would like to know if these differences are robust or just a fluke. I would like to know if something did indeed change around 2001, if there was a structural break.
I have tried using the 'chowtest' function, which I have coded like this (post911 being the aforementioned dummy):
"chowtest varlist, group(post911)"
I assume this is not testing the panel regression (xtreg) I am actually interested in, right? How would I do that?
The above is also taking the potential break year as given. Besides 2001, I would be interested in learning about other years that might better qualify as constituting a structural break. Some in the literature have run so-called "rolling Chow tests," where they then plot a graph of the chi squared values of the chow test across time. I was wondering how to do this.
I am new to this forum and would much appreciate your help! If further details are helpful, I will gladly provide them!
I am analysing the determinants of international currency adoption. I am interested in how their relative importance might be changing through time.
One of my hypotheses is that post-9/11, network effects and inertia weakened. To test this out, I've run panel data regressions on the pre and post-9/11 samples. I've also run a test on the full sample where I included a post-9/11 dummy variable and interaction terms. Just looking at the coefficients, my hypothesis looks valid, but I would like to know if these differences are robust or just a fluke. I would like to know if something did indeed change around 2001, if there was a structural break.
I have tried using the 'chowtest' function, which I have coded like this (post911 being the aforementioned dummy):
"chowtest varlist, group(post911)"
I assume this is not testing the panel regression (xtreg) I am actually interested in, right? How would I do that?
The above is also taking the potential break year as given. Besides 2001, I would be interested in learning about other years that might better qualify as constituting a structural break. Some in the literature have run so-called "rolling Chow tests," where they then plot a graph of the chi squared values of the chow test across time. I was wondering how to do this.
I am new to this forum and would much appreciate your help! If further details are helpful, I will gladly provide them!
