I'm not sure exactly how you want the results. It seems as though you want data sets, so I went ahead and did this. The solution is in the ODS TRACE statement. It prints all the HTML results and their respective names to the log. Then you trace down exactly what you need. Note: I have no experience with the AUTOREG procedure, but I went ahead and created a fake data set.   data have; input value time; datalines; 1 1 3 2 5 3 7 8 9 7 2 9 ; run; ods trace on; /* Trace the HTML results in the log. */ proc autoreg data = have; model value = time; run; ods trace off; /* Turning off for good practice. */ /* Extract the part of the report you want. I name the output data set here 'want'. I also mimicked your call for the parameter estimates */ ods output parameterestimates = capmparam fitsummary = want (where = (label2 = "Total R-Square")); proc autoreg data = have; model value = time; run; /* Just showing the 'want' data set via PROC PRINT */ proc print data = want; run; Obs Model Label1 cValue1 nValue1 Label2 cValue2 nValue2 1 Model1 Durbin-Watson 1.9201 1.920052 Total R-Square 0.2091 0.209074 ... View more

Hello,   Do you mean that you do not get parameter estimates for the explanatory variables smb and hml? In the picture of your dataset, the two variables (smb and hml) are constant. That's maybe / probably why. It would be better to include the LOG. That is more telling than just the code.   Cheers, Koen   ... View more

You have a time series of data, and your current program captures the last value in each month. (BTW, I see "fflib" and some familiar variable names, is this fama-french data?).   By "I wish to filter only December data.", are you really only trying to get the last value in each year?  If so, then instead of  if intck('month',date,nxt_date)>0; use if intck('year',date,nxt_date)>0; Also you can use a MERGE statement to simplify your program instead of two SET statements (one conditional).  It's a bit neater. data fflib.mnme_all: merge fflib.me_all fflib.me_all (firstobs=2 keep=date rename=(date=nxt_date)); if intck('year',date,next_date)>0; run; But finally, your data appears to be monthly.  So (1) your filter eliminated no data, and (2) you can use @ChrisNZ 's suggestion to use a WHERE statement with the MONTH function: data fflib.mnme_all: set fflib.me_all (where=(month(date)=12); run;   ... View more

My third footnote links to a macro that converts datasets to data step code.   A simple data step for your dataset will look like this: data mktcap6yrs; length reits $8 cy 8 dsname $41 me_group 8 ; input reits cy dsname me_group ; datalines; AAREIT 834849900 WORK.YR16 1 ; Add additional datalines as needed to illustrate your task. ... View more

You wouldn't change the data.  But you could change how the values print by using PROC FORMAT, then apply the format when printing to translate the five numbers into words. ... View more

Using option PROC REG + OUTEST= get these estimator, And use PROC APPEND to combine all these estimator. ... View more

Thank you for all the help! ... View more