*
* Example 10.3.1 from pp 351-352
*
all 1000
compute a0=1.0,a1=0.5
*
* This draws the first value from the unconditional distribution (which is mean
* 0, variance a0/(1-a1)); the remaining ones are generated recursively using the
* previous value. Because this uses random numbers, you'll get somewhat different
* results each time you run this (and will also get somewhat different results
* from the ones in the book).
*
set(first=%ran(sqrt(a0/(1-a1)))) arch = %ran(sqrt(a0+a1*arch{1}**2))
*
graph(footer="Figure 10-7 A Realization of an ARCH process")
# arch
*
* Do autocorrelations of the series and its square
*
@acf(number=40) arch
set asq = arch**2
@acf(number=40) asq
*
* Do independence tests
*
@bdindtests arch