uroot (version 1.3)

CH.test: Canova-Hansen test

Description

This function computes the Canova-Hansen statistic for testing the null hypothesis of stationary seasonal cycles against the alternative of a unit root at either a single or a set of seasonal frequencies.

Usage

CH.test (label, frec, f0, DetTr, showcat)

Arguments

label

a list object with information about the series. The list consist of the following elements. vari, a ts object with the data of the series, s, the periodicity of the series (4, 12, or 1 for quarterly, monthly, or anual data),

frec

a vector indicating the frequencies to analyse.

f0

a 0-1 (No-Yes) vector of length one indicating wether a lag of the dependent variable is included or not in the auxiliar regression.

DetTr

a logical argument. If TRUE a linear trend is included in the auxiliar regression.

showcat

how the results are showed and stored. If TRUE a brief explanation reports the results. If FALSE a list object stores the results.

Value

• CH test statistic.

Details

Elements of frec must be set equal to 0 if the season assigned to this element is not considered and equals to 1 for the frequencies to analyse. The position of each frequency in the vector is as follows: c(pi/2, pi) for quarterly series and c(pi/6, pi/3, pi/2, 2pi/3, 5pi/6, pi) for monthly series.

The lag truncation parameter for the covariance matrix is $round(s*(N/100)^{0.25})$, where $s$ is the periodicity of the series and $N$ the length of the series.

References

F. Canova and B.E. Hansen (1995), Are seasonal patterns constant over time? A test for seasonal stability. Journal of Business and Economic Statistics, 13, 237-252.

See Also

CHseas.test.

Examples

## Analyse frequency pi, including a first order lag and but not a linear trend.
     data(AirPassengers)
     AirP <- list(vari=AirPassengers, s=12, t0=c(1949, 1), N=length(AirPassengers))
     CH.test(label=AirP, frec=c(0,0,0,0,0,1), f0=1, DetTr=FALSE, showcat=TRUE)