Proc Panel Warning message

[mspak]

Dear all,

I ran a program using proc panel:

proc panel data=rd5;
instruments depvar;
model RD = LAGRD RDRD MB SGWTH CASH_AT1 CHG_WC STKISSUES CHG_CHOLD DBRISSUES YR01 YR02 YR03 YR04
YR05 YR06 YR07 YR08 YR09 YR10-YR15 IND01 IND04 IND05 IND06 IND07 IND08 IND09 IND10
/ gmm nolevels twostep maxband=10 artest = 2;
id gvkey fyear;
run;

I would like to know the AR test result, however, the log showed:

WARNING: The ARTEST test cannot be calculated for one or more lags due to missing values in the response and/or explanatory variable.

Can anyone provide a soultion to fix the problem? 

My another question is the MAXBAND option in proc panel:

MAXBAND=integer

specifies the maximum number of time periods (per instrumental variable) that are allowed into the moment condition.

If I specify as 10, does it mean that estimation is by system GMM with lagged levels dated t-10 used as instruments for equation in differences and lagged differences dated t-10 used as instruments for equation in levels?

Thank you. 

Hope to find a solution soon. 

[bobby_sas]

mspak asks:

I would like to know the AR test result, however, the log showed:

WARNING: The ARTEST test cannot be calculated for one or more lags due to missing values in the response and/or explanatory variable.

Can anyone provide a soultion to fix the problem? 

If the missing values are due to calculations performed by PROC PANEL when generating lagged variables, then this can be alleviated by using the CLAG statement (instead of the LAG statement).  This will replace missing values with cross-sectional means.  Otherwise, you would have to manually exclude the "missing" observations from the analysis by using the WHERE= option. 

My another question is the MAXBAND option in proc panel:

MAXBAND=integer

specifies the maximum number of time periods (per instrumental variable) that are allowed into the moment condition.

If I specify as 10, does it mean that estimation is by system GMM with lagged levels dated t-10 used as instruments for equation in differences and lagged differences dated t-10 used as instruments for equation in levels?

If you specify MAXBAND=10, then lagged levels UP TO AND INCLUDING t-10 are used as instruments in the difference equations, and lagged differences UP TO AND INCLUDING t-10 are used as instruments in the level equations.

--Bobby

[mspak]

Dear Bobby,

Thank you for your reply.

I have deleted missing variables due to proc panel lags generation before running GMM using proc panel. If the missing variables are available in the dataset, it is impossible for me to get the results from proc panel GMM. I am able to view all the results except for the AR test.

Since I set the Maxband in the proc panel, I am not sure whether it is due to the lagged variables generated by the proc panel GMM model?

Thank you.

[bobby_sas]

It is certainly possible that the missings are caused by internally-generated variables resulting from setting MAXBAND.  Does the problem persist if you reduce the value of MAXBAND?  

I'd be glad to take a closer look if you can email me the code and data that reproduces the problem.

--Bobby