Hello,
I've use xtbond2 to estimate a growth model for the Italian NUTS-3 provinces, where growth of Value Added is my dependent variable and log of VA is my independent with other explanatory variables. Number of obs. are 700 and numeber of groups are 7 (100 provinces).
Among the explanatory variables is per-capita expenditure for social security services (ss) expressed in Euro per thousand of VA, whose etimated coefficient is equal to .018748.
I've tried to calculate elasticity with margin after the regression, at mean values as in:
margins, eyex(L.ss) atmeans
This gives me 21.83529
(given that average value for L.ss is equal to .0999557 and average value for my dep. var. (growth of VA) is equal to .00000858).
This means that as per-capita expenditure for social security services increases by 1% then growth of VA increases by 21.83%.
I don't think this make sense, in economic terms. I think I miss something on the command, or else.
Any suggestion on this?
Thanks a lot.
Giorgioca
I've use xtbond2 to estimate a growth model for the Italian NUTS-3 provinces, where growth of Value Added is my dependent variable and log of VA is my independent with other explanatory variables. Number of obs. are 700 and numeber of groups are 7 (100 provinces).
Among the explanatory variables is per-capita expenditure for social security services (ss) expressed in Euro per thousand of VA, whose etimated coefficient is equal to .018748.
I've tried to calculate elasticity with margin after the regression, at mean values as in:
margins, eyex(L.ss) atmeans
This gives me 21.83529
(given that average value for L.ss is equal to .0999557 and average value for my dep. var. (growth of VA) is equal to .00000858).
This means that as per-capita expenditure for social security services increases by 1% then growth of VA increases by 21.83%.
I don't think this make sense, in economic terms. I think I miss something on the command, or else.
Any suggestion on this?
Thanks a lot.
Giorgioca