TSAYTEST Procedure |
@TSAYTEST performs an arranged regression test for threshold autoregression (though it can be used in more general situations). This is based upon the recursive estimation of a regression with data points added in order based upon a desired threshold series. From Tsay(1989). An alternative procedure for testing for the same type of break is @THRESHTEST.
@TSAYTEST(threshold=threshold series,other options) depvar start end
# list of regressors(in regression format).
Parameters
|
depvar |
dependent variable |
|
start, end |
range for regression. By default, the maximum range permitted by all variables involved in the regression. |
Options
THRESHOLD=series to use for threshold values
D=delay of the threshold series [0]
TITLE="title for output" ["Tsay Arranged Autoregression Test"]
[PRINT]/NOPRINT
GRAPH/[NOGRAPH]
If GRAPH, a SCATTER plot of the threshold series against the recursive residuals is created.
Variables Defined
|
%CDSTAT |
the computed test statistic (REAL) |
|
%SIGNIF |
the marginal significance level (REAL) |
Example
This does tests for threshold effects using different lags of the dependent variable, in both direct and reverse order (reverse order by doing - the series).
*
* Replication of Balke and Fomby(1997), "Threshold Cointegration,"
* International Economic Review, vol 38, no 3, 627-45.
*
cal(m) 1954:7
open data irates.xls
data(format=xls,org=columns) 1955:01 1990:12 fedfunds mdiscrt
*
set spread = fedfunds-mdiscrt
@dfunit(lags=12) fedfunds
@dfunit(lags=12) mdiscrt
@dfunit(lags=12) spread
*
* Pick lags
*
@arautolags(crit=hq) spread
*
* Tsay threshold tests with direct ordering
*
do d=1,4
@tsaytest(title="Threshold Test with Delay="+d,thresh=spread,d=d) spread
# constant spread{1 2}
end do d
*
* And with reversed ordering
*
set reverse = -spread
do d=1,4
@tsaytest(title="Threshold Test Reversed with Delay="+d,thresh=reverse,d=d) spread
# constant spread{1 2}
end do d
Sample Output
This is for the reversed order test in the example. It appears there may be a threshold break, with lag 1 as the threshold variable showing the greatest effect.
Threshold Test Reversed with Delay=1
F(3,424) 26.979 Signif 0.000
Threshold Test Reversed with Delay=2
F(3,424) 18.219 Signif 0.000
Threshold Test Reversed with Delay=3
F(3,423) 10.567 Signif 0.000
Threshold Test Reversed with Delay=4
F(3,422) 5.425 Signif 0.001
Copyright © 2026 Thomas A. Doan