I understand that the long run multiplier formula will depend on whether m estimating my equation in level or difference...
I am estimating equation in difference...I calculated the long run multiplier by -(coeff of independent variable divided by coeff of lagged dep. variable ). But how should I calculate its significance or it will remain same as in long run results of ARDL ? Any clarity on this would be appreciated.
You analyze each variable separately, i.e. the b coefficients are on lags of the same variable, I don't know if you are using monthly data but there should not be interaction terms.
I am using quarterly data and there are no interaction terms. But i did not understand. For instance, I have a model like this:-
yt= a +b yt-1+ cx1t + d x2t....
Now for lung run multiplier of X1t on yt...i calculated it like = -c/b ..for significance.. if coefficient on x1t is significant ..for long run mulitplier also it will be significant..? ?Please tell me is it right ?
@Nip : Can u please elaborate more on delta method...or give me some reference for paper that have used it ?
Green's "Econometrics" text explains the delta method. There is a brief summary at https://www.stata.com/support/faqs/statistics/delta-method/
NOTE: In your model (yt= a +b yt-1+ cx1t + d x2t....) the long run multuplier for x1 is m=c/(1-b), not "-c/b" as you have suggested.
The delta method gives the asymptotic variance of m as w'.V.w, where w is a column vector containing the derivatives dm/dc, dm/db and V is the Variance matrix for c, b.
@Vince : thanks for reference..But i read that if my dependent variable is difference and not level..in that case long run multiplier is -c/b whereas for level it is c/(1-b)...
Yes, if your model is actually yt - yt-1= a +b yt-1+ cx1t + d x2t...., then the long-run mu;tiplier for the level is -b/c.
My very brief definition, "The delta method ... " remains valid.
You are welcome. I hope my error did not confuse you - I wrote "-b/c" instead of your (correct) "-c/b".
Yeah..it's typo error...no probs...thanks once again for ur help Vince....
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running