When can OLS predict parameters
When linear in paramters
I.e it is fine if not linear in variables
How to turn non-linear parameters into ways OLS applicable.
Logs
What is β₀ in log form
β₀ = e to the α
Or lnβ₀= a
Log linear model vs semi log model
Semi log is when only the dependent variable has been log transformed.
Feature of the log linear model
Slope parameters are interpreted as elasticities
E.g in lnY= a + β₁lnX + lnε ,
If X is income, β₁ is the income elasticity!!