for the model with a constant and
Wfk-r
= Wk-r
Wk-r
12a
aWk-r
for the model with a time trend. To estimate μr and σ2 used in Theorem
2 for the model with a constant term and a linear time trend, the Brown-
ian motions Wk-r are replaced by a (k - r) dimensional vector of Gaussian
random walks with T = 500. The mean and variances are computed from
20,000 replications of the stochastic expressions.
Table B.1: Asymptotic values of μr and σ2
|
constant |
linear trend | |
|
sig. lev. |
2 222 μr σr |
2 222 μr σr |
|
k — r = 1 |
3.051 6.826 |
5.301 10.94 |
|
k — r = 2 |
9.990 18.46 |
14.35 26.02 |
|
k — r = 3 |
20.88 35.03 |
27.31 45.79 |
|
k — r = 4 |
35.67 57.49 |
44.13 70.82 |
|
k — r = 5 |
54.33 86.00 |
64.71 101.9 |
|
k — r = 6 |
76.94 119.7 |
89.16 136.9 |
18