I'm running a dynamic panel model. I've read through the PROC PANEL documentation here.
I'm relatively familiar with the Arellano/Bond/Bover and Blundell/Bond System GMM procedures for dynamic panels, but I want to make sure I understand exactly what is being in SAS. In particular, the relationship between the INSTRUMENT and MODEL statements is unclear. Could someone please confirm or correct my intiution outlined below? Thanks in advance!
*********
the INSTRUMENT statement (p. 1386-7) gives the following example:
proc panel data=test;
/* Call this model (1) */
INSTRUMENTS depvar pred=(x1 x2) exog=(x3 x4 x5) correlated=(x3 x4 x5);
model y = y_1 x1 x2 / gmm1 maxband=6 nolevels ginv=g4 artest=5;
/* And this model (2) */
INSTRUMENTS pred=(x2 x4) exog=(x3 x5) correlated=(x3 x4);
model y = y_1 x2 / gmm1 maxband=6 nolevels ginv=g4 artest=5;
id cs ts;
run;
Here's what I understand is happening:
The dynamic panel options seem really fantastic in PROC PANEL but I gotta say I've found the documentation to a little murky in places, especially with respect to the specification of instrumental variables.
(NOTE: I've cross-posted this question on StackOverflow: http://stackoverflow.com/questions/40752195/understanding-instrument-statement-for-dynamic-panel-mod...)
To answer your questions,
1. This is correct. By default (NOLEVELS not specified), both difference and level equations are used.
2. This is correct.
3. This is not correct. PROC PANEL assumes an AR(1) model in developing the instruments. Making this work with higher-order AR models would require some coercion of the INSTRUMENTS statement. In my experience, it would be difficult to get the full instrument matrix just right because PROC PANEL won't innately recognize the higher orders. This is an area where could improve PROC PANEL in the future.
4. This is not correct. See 3.
5. Correct. When you specify NOLEVELS, the CORRELATED option plays no role for the reason you described.
To answer your questions,
1. This is correct. By default (NOLEVELS not specified), both difference and level equations are used.
2. This is correct.
3. This is not correct. PROC PANEL assumes an AR(1) model in developing the instruments. Making this work with higher-order AR models would require some coercion of the INSTRUMENTS statement. In my experience, it would be difficult to get the full instrument matrix just right because PROC PANEL won't innately recognize the higher orders. This is an area where could improve PROC PANEL in the future.
4. This is not correct. See 3.
5. Correct. When you specify NOLEVELS, the CORRELATED option plays no role for the reason you described.
Thank you bobby_sas!!!
I feel much better having these points confirmed.
Regarding the current (practical) limitation of PROC PANEL to AR(1) models, that's an important point to know! I wish it were noted in the document (or, if I missed it, noted more prominently). In any case I'll probably be working with an AR(1) anyhow, so it's fine for my present purposes 🙂
SAS Innovate 2025 is scheduled for May 6-9 in Orlando, FL. Sign up to be first to learn about the agenda and registration!
Learn how to run multiple linear regression models with and without interactions, presented by SAS user Alex Chaplin.
Find more tutorials on the SAS Users YouTube channel.