The American Beer Association has asked you to determine how the demand for beer responds to changes in the tax on beer, using data for 2020 for each of the 50 states in the US. Consider each of the following regression models:
Model I: Bi = β0 + β1Ti + other variables + Ui
Model II: log Bi = b0 + b1 log Ti + other variables + Ui
where Bi = beer sold in state i per year (in millions of gallons); Ti = total sales tax, per gallon of beer, in state i; and Ui = error term. The other variables include the population and average household income in state i.
Which regression model would you estimate (Model I or Model II) to determine how the demand for beer responds to changes in the tax on beer, and why would you use it? (Ignore the possibility of missing-variables bias.)
Describe precisely how you would use the parameter estimates you would obtain from the model you chose in Part (A) to determine the elasticity of demand for beer with respect to the tax on beer.