st: makeCns

[ACavallo@lexecon.com]

Dear Statalisters,

I am trying to use the constraint command to define constraints for linear
combinations of coefficients in a linear regression model.  I want to do
some matrix algebra to compute tests of equality on these combinations.  I
am having trouble with makeCns.  I wrote a program that imposes my
constraints using constraint define, but when I issue the makeCns command,
I get the message:  r(301) Last estimates not found. [check this]

Here's my code:
#delimit ;
capture program drop hsddiff;
program define hsddiff;
  args prefix yst yend;

  local count=0;
  foreach y of numlist `yst' / `yend' {;
    local count = `count' + 1;

    if `y' == 1984 {;
      constraint define `count' `prefix'__`y'  ;
    };
    else {;
      constraint define `count' `prefix'__`y' - hsd__`y' ;
    };
  };
  constraint li;
  matrix makeCns 1-`count';
  matrix r=get(Cns);
  scalar cols=colsof(r)-1;
  scalar rows=rowsof(r);
  matrix R=r[.,1..cols];
  scalar cols=cols+1;
  matrix q=r[1..rows,cols];
  matrix v=e(V);
  matrix b=e(b)';
  matrix coefdiff=R*b;
  matrix coefvar=R*v*R';
  matrix matrix vv=vecdiag(coefvar);
end;

reg lnw black__* other__* female__* blackfem__* otherfem__*
      age__* agesq__*
      divsepwid__* married__* fnumper__* foklt18__*
      inmsa__* hmetrodum__*
      dropout__*
      $hsd
      somecoll__* bach__* grad__*
      [pweight=wgt] ,  robust;
parmest, saving(wtall,replace);

hsddiff bach 1980 1989;


here's my log:

. reg lnw black__* other__* female__* blackfem__* otherfem__*
>       age__* agesq__*
>       divsepwid__* married__* fnumper__* foklt18__*
>       inmsa__* hmetrodum__*
>       dropout__*
>       $hsd
>       somecoll__* bach__* grad__*
>       [pweight=wgt] ,  robust;
(sum of wgt is   1.0516e+09)

Regression with robust standard errors                 Number of obs =
697670
                                                       F(178,697491) =
1756.36
                                                       Prob > F      =
0.0000
                                                       R-squared     =
0.3450
                                                       Root MSE      =
.63906

------------------------------------------------------------------------------
             |               Robust
         lnw |      Coef.   Std. Err.      t    P>|t|     [95% Conf.
Interval]
-------------+----------------------------------------------------------------
 black__1980 |  -.2187249   .0124757   -17.53   0.000     -.243177
-.1942729

output omitted

  grad__1989 |    .385969   .0492801     7.83   0.000     .2893815
.4825564
       _cons |   3.805371   .0322174   118.12   0.000     3.742226
3.868516
------------------------------------------------------------------------------

. parmest, saving(wtall,replace);
file wtall.dta saved

. hsddiff bach 1980 1989;
     1:  bach__1980 - hsd__1980
     2:  bach__1981 - hsd__1981
     3:  bach__1982 - hsd__1982
     4:  bach__1983 - hsd__1983
     5:  bach__1984
     6:  bach__1985 - hsd__1985
     7:  bach__1986 - hsd__1986
     8:  bach__1987 - hsd__1987
     9:  bach__1988 - hsd__1988
    10:  bach__1989 - hsd__1989
last estimates not found
r(301);

end of do-file
r(301);

Thanks in advance!!


--Alex Cavallo
Lexecon
(312) 322-0208  voice
(312) 322-0218  fax

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/