Programming the statistical procedures from SAS

Heteroskedasticity tests when have heteroskedasticity consistent standard errors

Reply
Super Contributor
Posts: 412

Heteroskedasticity tests when have heteroskedasticity consistent standard errors

Hi,

I am doing a regression which contains heteroskedasticity according to the White and Breusch-Pagan tests.

 

 

 proc model data=data;
      parms a1 b1-b21;
      car3 = a1 + b1 * concentration + b2 * instown_hhi + b3 * ln_mkt + b4 * lt_mkt + b5 * deal_mkt + b6 * ceo_centrality
+ b7 * ceo_tenure + b8 * ceo_tenure2 + b9 * total_boards_to_date + b10 * total_current_boards
+ b11 *  total_current_boards2 + b12 * ind_past_cfo_fraction + b13 * ind_past_cfo_dummy + b14 * Total_Directors_on_the_Board
+ b15 * intense_monitor + b16 * ind_past_cfo + b17 * sic_3 + b18 * govt + b19 * jv + b20 * sub + b21 * priv;
      fit car3 / white pagan=(21 concentration instown_hhi ln_mkt lt_mkt deal_mkt ceo_centrality
ceo_tenure ceo_tenure2 total_boards_to_date total_current_boards total_current_boards2 ind_past_cfo_fraction
ind_past_cfo_dummy Total_Directors_on_the_Board intense_monitor ind_past_cfo sic_3 govt jv sub priv ) 
      out=resid1 outresid;
   run;
   quit;

 

 

Then I try to mitigate for this heteroskedasticity with heteroskedasticity consistent standard errors using the hcc option. 

proc reg data= data ;
      model CAR3= concentration  INSTOWN_HHI  ln_mkt lt_mkt  deal_mkt ceo_centrality
 ceo_tenure ceo_tenure2  total_boards_to_date total_current_boards total_current_boards2 ind_past_cfo_fraction ind_past_cfo_dummy
Total_Directors_on_the_Board directors2  intense_monitor ind_past_cfo    sic_3 govt jv sub priv / hcc hccmethod=0   ;
output out=resids3 r=residual; 
   run;

What I would like to do is to redo the heteroskedasticity tests but this time for the model containing hs consistent standard errors.

 

Thank you!

Ask a Question
Discussion stats
  • 0 replies
  • 137 views
  • 0 likes
  • 1 in conversation