Hi all,
I am trying to estimate a hedonic price model. The first step used is estimating the Box Cox and predicting: lhsonly, rhsonly, lamda and theta. Then, I want to estimate the Vuong test to compare and identify the best transformation. Hence, manually to compare e.g., log-lin vs. lin-log, I use the following:
gen lhsonlysq = lhsonly^2
gen rhsonlysq = rhsonly^2
gen diff_log_lin_vs_lin_log = lhsonlysq - rhsonlysq
sum diff_log_lin_vs_lin_log
*mean=-4.29*
gen diff_log_lin_vs_lin_log_B=diff_log_lin_vs_lin_log-(-4.29)
gen diff_log_lin_vs_lin_log_B2=diff_log_lin_vs_lin_log _B^2
sum diff_log_lin_vs_lin_log_B2
*mean=13912.15*
then, I continue with some more estimations in Microsoft Excel to find the Vuong test for these two transformations and the rest of the combinations.
QUESTION: Is there an easier way to calculate the Vuong test directly from STATA 14, and without using excel?
Best regards,
Petjon
0 Response to Vuong test for hedonic price model - Regression
Post a Comment