Andrews-Quandt test for unknown breakpoint.
Carries out the Andrews-Quandt test for parameter stability at some unknown breakpoint.
Syntax
eq_name.ubreak(options) trimlevel @ x1 x2 x3
You must provide the level of trimming of the data. The level must be one of the following: 49, 48, 47, 45, 40, 35, 30, 25, 20, 15, 10, or 5. If the equation is specified by list and contains no nonlinear terms, you may specify a subset of the regressors to be tested for a breakpoint after an “@” sign.
Options
wfname = series_name | Store the individual Wald F-statistics into the series series_name. |
lfname = series_name | Store the individual likelihood ratio F-statistics into the series series_name. |
prompt | Force the dialog to appear from within a program. |
p | Print the result of the test. |
Examples
equation ppp.ls log(spot) c log(p_us) log(p_uk)
ppp.ubreak 15
regresses the log of SPOT on a constant, the log of P_US, and the log of P_UK, and then carries out the Andrews-Quandt test, trimming 15% of the data from each end.
To test whether only the constant term and the coefficient on the log of P_US are subject to a structural break, use:
ppp.ubreak @ c log(p_us)
Cross-references