| Library: | times |
| See also: | kpss |
| Quantlet: | adf | |
| Description: | Calculation of the two tabulated testvalues (the general one is tau) for the Augmented Dickey-Fuller Test of a unit root in an autoregressive process for with a constant. If the third argument is "trend" a linear trend is calculated. The third return value is a vector of critical values for tau. |
| Usage: | {rho,tau,critvalues}=adf(z,p{,trend}) | |
| Input: | ||
| y | series to be estimated | |
| p | number of included past differences in the regression | |
| trend | string (not needed), a linear trend is calculated if trend = "trend" | |
| Output: | ||
| rho | scalar, testvalue | |
| tau | scalar, testvalue | |
| critvalues | vector, critical values (1%, 5%, 10%, 90%, 95% and 99%); these values are calculated for sample sizes (rows of y) greater than 500 | |
library("times")
randomize(1234)
x=normal(1000)
x=cumsum(x)
{rho,tau,cv}=adf(x,7)
rho
tau
cv
; x is a random walk. thus we expect tau between -2.57 and -0.44
Contents of rho [1,] 4.8222 Contents of tau [1,] -1.8666 Contents of cv [1,] -3.43 [2,] -2.86 [3,] -2.57 [4,] -0.44 [5,] -0.07 [6,] 0.6
library("times")
randomize(1234)
x=normal(1000)
x=cumsum(x)
{rho,tau,cv}=adf(x,7,"trend")
rho
tau
cv
; x is a random walk. thus we expect tau between -3.12 and -1.25
Contents of rho [1,] 7.0198 Contents of tau [1,] -2.0605 Contents of cv [1,] -3.96 [2,] -3.41 [3,] -3.12 [4,] -1.25 [5,] -0.94 [6,] -0.33
| Library: | times |
| See also: | kpss |