*   R. Carter Hill, William E. Griffiths and Guay C. Lim,  
*   Principles of Econometrics, Fourth Edition, Wiley, 2011.

* Chapter 9.2 Finite Distributed Lags

READ (okun.dat) / names
GEN1 nobs=$N

* Generate change in employment
GENR du = u - LAG(u)

* Generate lagged variables
DO #=1,3
GENR g#=LAG(g,#)
ENDO 

* Estimation results Table 9.2, page 346.
* Adjust the sample period for undefined observations
*   Lag length 3
SAMPLE 4 nobs
OLS du g g1 g2 g3
*   Lag length 2
SAMPLE 3 nobs
OLS du g g1 g2

* Chapter 9.3 Serial Correlation

SAMPLE 1 nobs
* Use the ARIMA command to calculate the sample autocorrelation function,
* results in Equation (9.16), page 349 and Figure 9.6, page 350.
ARIMA g / GRAPHAC NLAG=12 ACF=r

* Calculate the test statistics at bottom of page 349.
SAMPLE 1 4
GENR z=SQRT(nobs)*r
DISTRIB z / TYPE=NORMAL MEAN=0 VAR=1 CDF=cdf
* Get a p-value for a two-sided test.
GENR pvalue=2*(1-cdf)
PRINT r z pvalue
STOP