Suppose we want to estimate how household expenditures on alcoholdepend on average household income and changes in average household income. Wetake a random sample of households for two periods. Assume the true populationmodel is:

:y_it = β_0 + δ_0 *d2t + β_1*(x_i)-bar + β_2*(x_it - (x_i)-bar) + a_i + u_it,

where the dependent variable is expenditures on alcohol and x represents household income. Suggestan estimation strategy to find unbiased estimators of and. [Assume the assumptions about the idiosyncratic error termhold.]