Proc PANEL/TSCSREG: Error: regarding missing values, weird

Arsenio_Staer — Mon, 23 May 2011 22:46:54 GMT

Hey guys,

So I have a quick question. The same dataset and model fit perfectly well in stata using the whole gamut of specs but in SAS 9.22 and EG4.3 it givs an error

Error:

ERROR: Each observation for time point date=19940103 has a missing value for at least one
variable in MODEL statement .
NOTE: The SAS System stopped processing this step because of errors.

Code:

------------------------
ROC TSCSREG DATA = WORK.SORTTempTableSorted

OUTEST=WORK.TSCSPanelDataAnalysisXSH_EQUIV_F(LABEL="Panel Data Analysis of WORK.XSH_EQUIV_FIN_FULL_REG_NOMISS") CORROUT;
MODEL RET= curr_fund_dret equiv_pct_dvol pdvol_inter VWRETD /
FIXONE
FIXTWO
RANONE
RANTWO
CORRB
;
ID ticker date
;
/* -------------------------------------------------------------------
End of task code.
------------------------------------------------------------------- */
RUN; QUIT;
-----------------------

I tried Proc Panel, still the same, tried all different fixed random effects still the same thing. Apparently if for the first observation its where the series start one of the variables has missing values for all groups. But this shouldn't matter at all! The valid data will just be limited to the ones where the observations have values for all variables in the model, right? Maybe it tried to fit the model day by day not that it would make any sense? some sort of missing option or something?

I may post some data but it's big and with no attachments and no table markup available, it's gonna be a mess.

Would greatly appreciate any help!

Thanks!

Re: Proc PANEL/TSCSREG: Error: regarding missing values, weird

Arsenio_Staer — Tue, 24 May 2011 00:56:46 GMT

More data,

Same model as below, here is first lines of the dataset

ticker date RET VWRETD curr_fund_dret equiv_pct_dvol pdvol_inter
ADI 19940103 -0.040609 -0.003734
ADI 19940104 0.015873 0.002820
the last 3 variables are missing and it's fine cause later in the series

ADI 20011231 -0.016397 -0.009728 -0.030577249
ADI 20020102 0.021626 0.003861 0.04181677 0.038828028 0.001623663
ADI 20020103 0.050937 0.009825 0.081652147 0.059559576 0.004863167
ADI 20020104 0.006085 0.006973 -0.00063584 0.049063525 -3.11966E-05

and so on for all the groups. The same dataset and same specification works ok in Stata. Anybody please? Proc Panel should just discard uncomplete observations and proceed to populate the x'x with complete values.

Thanks!

Arsenio

topic Re: Proc PANEL/TSCSREG: Error: regarding missing values, weird in Statistical Procedures

Proc PANEL/TSCSREG: Error: regarding missing values, weird

Re: Proc PANEL/TSCSREG: Error: regarding missing values, weird