
Bayesian Parallel Computation using Griddy-Gibbs Sampler 629
TABLE 30.4
Estimation res ults of EGARCH(1,1) model using S&P 500.
ˆ
θ Mean S.D. 2.5% 97.5% SIF
ˆµ
y
0.0292 0.0085 0.0128 0.0451 2.5 8
ˆα
1
0.9839 0.0022 0.9789 0.9873 2.7 6
ˆ
β
1
0.1201 0.0101 0.1013 0.1414 2.3 7
ˆγ
1
−0.0959 0.0080 −0.1 117 −0 .0804 2.28
log-marginal likelihood = 20,511
Notes: SIF = Simulation Inefficiency Factor and Computing Time = 14.5470 mins.
Table 30.4 presents the posterior mean, poster ior standard deviation and
the credible interval with coverage probability 0.95 for each parameter of the
EGARCH(1,1) model. The favo rable results can be seen from the low stan-
dard deviations