Proc Syslin

Scottcsumb — Tue, 08 Mar 2022 18:23:35 GMT

In the SAS SAS/ETS® 13.2 User’s Guide: ..The SYSLIN Procedure page 2112 it defines predetermined variables as both exogenous and lagged endogenous variables. Next it defines instrument variables as predetermined variables. Is that statement indicating you have to include all exogenous variables as instrument variables when using 3SLS? thanks for the clarification in advance

Re: Proc Syslin

pink_poodle — Wed, 09 Mar 2022 02:48:53 GMT

Could you please attach the part of the documentation that you are referring to?

Re: Proc Syslin

Scottcsumb — Wed, 09 Mar 2022 22:11:45 GMT

Hello pink_poodle. Thank you for your email. I provided a copy of the page
out of the SAS guide. Below that is my 3SLS code. scott
*Below is the copied text: *

*Page. 2112 F Chapter 29: The SYSLIN Procedure*
In this system, quantity demanded depends on price, income, and the price
of substitutes. Consumers
normally purchase more of a product when prices are lower and when income
and the price of substitute
goods are higher. Quantity supplied depends on price and the unit cost of
production. Producers supply more
when price is high and when unit cost is low. The actual price and quantity
sold are determined jointly by the
values that equate demand and supply.
Since price and quantity are jointly endogenous variables, both structural
equations are necessary to adequately
describe the observed values. A critical assumption of OLS is that the
regressors are uncorrelated with
the residual. When current endogenous variables appear as regressors in
other equations (endogenous
variables depend on each other), this assumption is violated and the OLS
parameter estimates are biased and
inconsistent. The bias caused by the violated assumptions is called
simultaneous equation bias. Neither the
demand nor supply equation can be estimated consistently by OLS.
*Variables in a System of Equations*
Before explaining how to use the SYSLIN procedure, it is useful to define
some terms. The variables in a
system of equations can be classified as follows:
• *Endogenous variables, *which are also called jointly dependent or
response variables, are the variables
determined by the system. Endogenous variables can also appear on the
right-hand side of equations.
•* Exogenous variables* are independent variables that do not depend on any
of the endogenous variables
in the system.
• *Predetermined variables *include both the exogenous variables and lagged
endogenous variables, which
are past values of endogenous variables determined at previous time
periods. PROC SYSLIN does not
compute lagged values; any lagged endogenous variables must be computed in
a preceding DATA step.
• *Instrumental variables* are predetermined variables used in obtaining
predicted values for the current
period endogenous variables by a first-stage regression. The use of
instrumental variables characterizes
estimation methods such as two-stage least squares and three-stage least
squares. Instrumental variables
estimation methods substitute these first-stage predicted values for
endogenous variables when they
appear as regressors in model equations.
Using PROC SYSLIN
First specify the input data set and estimation method in the PROC SYSLIN
statement. If any model uses
dependent regressors, and you are using an instrumental variables
regression method, declare the dependent
regressors with an ENDOGENOUS statement and declare the instruments with an
INSTRUMENTS statement.
Next, use MODEL statements to specify the structural equations of the
system.

The data is from an experimental auction: The endogenous variables are
winning prices. The instrument variables are lagged winning prices and
structural characteristics of the auction mechanism. Ten independent
auctions were held. So for instance, *Size* is the number of subjects in a
particular auction. Number of subjects participating varied across the ten
auctions held. The model statements represent the prices of the winning
bids for the 3 meat products being auctioned. Note: endogenous variables
are included on the right hand side of the model startment (bolded). The
other right hand side variables are exogenous variables representing
information from a demographic survey completed by subjects just prior to
the auction.

proc syslin data=a 3sls ;
*endogenous* prBIS prBF93 PrBF80;

*instruments * lagprbis lagpr80 lagpr93
dum80 dum93 *SIZE* dumround2 dumround3 dumround4 dumround5 ;

Beef80: model PrBF80= *prBIS prBF93* biscorrect B93correct BNI
b80text b80juice b80flav Q10 Q15 Q16 female /overid;

Beef93: model prBF93 = *prBIS PrBF80* biscorrect B80correct BNI
b93text b93juice b93flav Q10 Q15 Q16 female / overid;

Bison: model prBIS = *prBF93 PrBF80* b80correct b93correct BNI
bistext bisjuice bisflav Q10 Q15 Q16 female Dumbybis /overid;

run;

Sorry for taking so long to reply. Work has been very busy.

Re: Proc Syslin

pink_poodle — Thu, 10 Mar 2022 03:01:56 GMT

I do not have a deep understanding of the subject. To me what it’s saying is when exogenous variables are used to predict endogenous variables they are called instrumental. It seems to imply that exogenous variable should be listed as instrumental when used for predicting endogenous variables.

topic Re: Proc Syslin in SAS Forecasting and Econometrics

Proc Syslin

Re: Proc Syslin

Re: Proc Syslin

Re: Proc Syslin