Hi again and thanks for your reply. My initial datasets have this shape: MONTH AA_CDS_FPC AA_E_FPC (name dataset AA_CDS_EQUITY)

MONTH DOW_CDS_FPC DOW_E_FPC (name dataset DOW_CDS_EQUITY)

MONTH CPB_CDS_FPC CPB_E_FPC (name dataset CPB_CDS_EQUITY)

Below I report the two codes I wrote:

1)Code for Cross-sectional regression: I run the same regression:

%scan(&C,&ic,@)_cds_fpc = A + B * %scan(&C,&ic,@)_e_fpc + error

for three different companies (tickers: AA, DOW,CPB). For each of them I create a file with the Adjusted R2 as only variable included:

%macro regression; ods output Reg.MODEL1.Fit.%scan(&C,&ic,@)_cds_fpc.FitStatistics=%scan (&C,&ic,@)_fit;

PROC REG data=%scan(&C,&ic,@)_cds_equity; MODEL %scan(&C,&ic,@)_cds_fpc=%scan(&C,&ic,@)_e_fpc; run; ods output close; %mend regression;

/* In the file “=%scan(&C,&ic,@)_fit” the R2 and Adjusted-R2 are respectively the first and second observations of the variable cValue2, while the third observation is null. */

%macro fit; data %scan(&C,&ic,@)_cValue2; set %scan(&C,&ic,@)_fit; keep cValue2; run; %mend fit;

%macro R2; data %scan(&C,&ic,@)_R2; set %scan(&C,&ic,@)_cValue2; i = _n_ ; rename cValue2=R2_%scan(&C,&ic,@); run; %mend R2;

%macro adj_R2; data %scan(&C,&ic,@)_adj_R2; set %scan(&C,&ic,@)_R2; if i=1 then delete; if i=3 then delete; drop i; rename R2_%scan(&C,&ic,@)=adj_R2_%scan(&C,&ic,@); run; %mend adj_R2;

%macro mainloop; %do ic=1 %to 5; %merge; %regression; %fit; %R2; %adj_R2; %end; %run; %mend mainloop; %let C=AA@DOW@CPB; %mainloop;

2)Code for Time-series regression:

I run routinely the same regression for each company (for example here AA): Aa_cds_fpc = A + B * aa_e_fpc + error and I include at each step one more observation (observation=month; total observations: 85, start from 10th month-observation)

data aa_cds_equity; merge results.aa_cds_pc results.aa_e_pc; run;

%MACRO REG(N); %DO I = 10 %TO 85; ods output Reg.MODEL1.Fit.aa_cds_fpc.FitStatistics=aa_fit; PROC REG data=aa_cds_equity(OBS = &I); MODEL aa_cds_fpc=aa_e_fpc; RUN; ods output close;

data aa_cValue2; set aa_fit; keep cValue2; run;

/* In the file aa_fit” the R2 and Adjusted-R2 are respectively the first and second observations of the variable cValue2, while the third observation is null. */

data aa_R2; set aa_cValue2; i = _n_ ; rename cValue2=R2_aa; run;

data aa_adj_R2; set aa_R2; if i=1 then delete; if i=3 then delete; drop i; rename R2_aa=adj_R2_aa; run;

%END; %MEND; %reg;

First of all, I suspect that the procedures I’m following are quite clumsy. Secondly, I would need a mix of these two codes. In fact, what I really want is to have for each company the series of the 75 adjusted-R2s obtained from the time-series regressions. My ideal output would be:

MONTH AA_adj_R2 DOW_adj_R2 WMT_adj_R2 T_adj_R2 CPB_adj_R2

Jun2004 Feb2004 March 2004 … Jul2010

Please, help if you can and thank you in advance. MM