This question already has an answer here:
- Interpretation of log transformed predictor and/or response 3 answers
- How to interpret transformed variables in multiple regression? 1 answer
- Interpreting Log-Transformed Percentages in OLS 1 answer
- Interpreting the change in two logs in a regression 1 answer
- Interpreting coefficients of a dlog-dlog regression 2 answers
Can anyone suss out the interpretation of beta here:
D.log(gdp) = alpha + beta(log(military expenditure %)
Where D.log(gdp) = growth of GDP (first difference of log(GDP)
Would it be:
A 1% point increase in military expenditure is associated with a beta% point change in the growth of GDP?
A 1% increase in military expenditure is associated with a beta % change in the growth of GDP?
N.B. Also what implication does the underlying variable military expenditure have being a %, wouldn't that make the interpretation a percentage point?