Interpretation of coefficients: d(lnY)=c+b*(lnX)+e*d(lnZ)
Posted: Sat Jul 04, 2015 6:48 am
Dear all,
what is the proper interpretation of coefficients in time series regression when using first differencing on logs of the DV and in certain IVs. The log transformation was chosen to aim at elasticities. Differencing was chosen to achieve stationarity.
As d(lnY) [first difference formulation of log in Eviews] can be understood as e.g "growth rate", I now wonder if the elasticity notation holds in this case: e.g. that a 1% change of X leads to a b% change in d(lnY)?
Does this further apply to coefficients of differenced logs of IVs, e.g. given by e*d(lnZ) term, which could be "inflation"? Then, a 1% change in Z leads to a e% change of d(lnY)?
Thank you so much for clarification!
Best,
Sebastian
what is the proper interpretation of coefficients in time series regression when using first differencing on logs of the DV and in certain IVs. The log transformation was chosen to aim at elasticities. Differencing was chosen to achieve stationarity.
As d(lnY) [first difference formulation of log in Eviews] can be understood as e.g "growth rate", I now wonder if the elasticity notation holds in this case: e.g. that a 1% change of X leads to a b% change in d(lnY)?
Does this further apply to coefficients of differenced logs of IVs, e.g. given by e*d(lnZ) term, which could be "inflation"? Then, a 1% change in Z leads to a e% change of d(lnY)?
Thank you so much for clarification!
Best,
Sebastian