* PS10.1, using DATA10-1, for Example 10.1
* The preliminary results for this example are replicated.  However, we could
* not replicate the results from Table 10.1 in Section 10.1.  The estimated
* R-squared and slope coefficients in SHAZAM are close to those in the textbook.
* However, the estimated constants are quite different.  SHAZAM uses a
* nonlinear optimization method which is different from that used by ESL.
* The constants might be different because of division by 1-rho1-rho2 etc.
*
* The TIME command used below specifies that the beginning year is 1964 and
* it is quarterly data.  The SAMPLE command indicates that the data begins
* in the first quarter of 1964 and ends at the second quarter of 1991.
*
TIME 1964 4
SAMPLE 1964.1 1991.2
READ(data10-1) PERIOD R M D 
*
* The GENR command and LAG function is used to generate the lagged values for
* the money supply (variable M) and budget deficits (D).  To lag the money
* supply back one period the command is GENR M1=LAG(M) and to lag the money
* supply back two periods the command is GENR M2=LAG(M,2).  To lag three and
* four periods back refer to the commands below.
*
GENR M1=LAG(M)
GENR M2=LAG(M,2)
GENR M3=LAG(M,3)
GENR M4=LAG(M,4)
GENR D1=LAG(D)
GENR D2=LAG(D,2)
GENR D3=LAG(D,3)
GENR D4=LAG(D,4)
*
* Model A
*
SAMPLE 1965.1 1991.2
*
* Estimate Kitchen Sink Model by OLS.
*
OLS R M M1 M2 M3 M4 D D1 D2 D3 D4 / RESID=UHAT RSTAT
*
GENR UT1=LAG(UHAT)
GENR UT2=LAG(UHAT,2)
GENR UT3=LAG(UHAT,3)
GENR UT4=LAG(UHAT,4)
*
SAMPLE 1966.1 1991.2
*
* Auxilliary Regression for LM test.
*
OLS UHAT M M1 M2 M3 M4 D D1 D2 D3 D4 UT1 UT2 UT3 UT4 
GEN1 LM=$N*$R2
distrib LM / type=chi df=4
*
SAMPLE 1965.1 1991.2
*
* The ORDER=4 option estimates the model using the least squares procedure.
*
AUTO R M M1 M2 M3 M4 D D1 D2 D3 D4 / ORDER=4
AUTO R M M1 D D4 / ORDER=4
AUTO R M M1 D / ORDER=4
AUTO R M M1 / ORDER=4
DELETE / ALL
STOP