why proc varmax output and proc reg outputs are same?

[swapnilsoni]

how coefficients are estimated in VAR model in sas? why proc varmax output and proc reg outputs are same?

Please check the following programs:

/*Program-1: VARMAX PROCEDURE*/

/*Estimate VAR(1) model with all 3 variables (Trivariate)*/

proc varmax data=sashelp.class ;

model Age Weight Height/p=1 ;

run;

/*PROGRAM2: REG PROCEDURE*/

/*Let's replicate this using PROC REG with newly created dataset with manual lags*/

/*New dataset*/

data lag_data;

set sashelp.class;

Age_l=lag(Age);

Weight_l=lag(Weight);

Height_l=lag(Height);

run;

/*Use PROC REG: gives same result as above*/

proc reg data=lag_data;

model Age=Age_l Weight_l Height_l;

run;

Concern: Both are giving same output. This should not happen since VAR uses multivariate OLS not conventional OLS used in Regression. If not so the what is use of VAR?

[gergely_batho]

Parameter estimates should be the same, standard errors and p-values should be different.
Please someone correct me if I’m wrong.