*
* Example 6.3.2 on pp 197-198
*
open data oshorts.dat
data(format=free,org=columns) 1 57 overs
*
boxjenk(ma=1,maxl,demean) overs
compute unrlogl=%logl
*
disp "Davis-Dunsmuir test" %beta(1) "vs .05 critical value" -1.0+6.80/%nobs
*
* To do the likelihood ratio version of the test, we need to evaluate the
* likelihood at the hypothesized value (theta=-1). This is done by using
* METHOD=EVAL (which evaluates, rather than estimating) and feeding in the value
* with the INITIAL option.
*
boxjenk(ma=1,maxl,demean,initial=||-1.0||,method=eval) overs
compute rlogl=%logl
disp "LR version of test" -2.0*(rlogl-unrlogl)