SAS：如何为300多个外生变量修改Proc模型

Question

I have to run a Newey-West regression with over 300 exogenous variables (due to multiple lags and time dummy variables). This is the typical way to run a SAS Newey-West regression:

proc model data=two;
     endo r_invest;
     exog r_int r_gnp;
     instruments _exog_;
     parms b0 b1 b2;
     r_invest=b0 + b1*r_int + b2*r_gnp;
     fit r_invest / gmm kernel=(bart,5,0) vardef=n;
     run;
     quit; 

Say that I have 390 time dummy called tdum1 to tdum390 , how can I avoid to write r_invest=b0 + b1*r_int + b2*r_gnp +b3*tdum1....b390*tdum390; ?

Answer 1

I'm not sure of the exact syntax of this procedure, but this macro will write out the variables you have listed in the comments below your code. You can edit it to fit the exact syntax you need.

%macro writeOutMyVars();

    %do i=1 %to 390;
        + b%eval(&i+2)*tdum&i  
    %end;

%mend;

proc model data=two;
     endo r_invest;
     exog r_int r_gnp;
     instruments _exog_;
     parms b0 b1 b2;
     r_invest=b0 + b1*r_int + b2*r_gnp %writeOutMyVars(); ;
     fit r_invest / gmm kernel=(bart,5,0) vardef=n;
run;
quit;

If you add "options mprint;" to the top of your code then the macro output will be written to the log so you can see what it is doing.