Zivot-Andrews Unit Root test allowing for a single break in intercept and/or tr > end
zandrews [varname] [if exp] [in range], maxlags(#) trim(#) break() generate(varname) lagmethod() level(#) graph
zandrews is for use with time-series data. You must tsset your data before using these routines. Varname may contain time-series operators. zandrews may be applied to a single time series within a panel.
Description
zandrews computes the Zivot-Andrews (1992) unit root test for varname.
Options
maxlags(#) specifies the number of additional lags or the maximum number of lags to consider. It must be a positive integer. If not specified, T^0.25 is used. Maxlags must be given with lagmethod (input).
trim(#) specifies the fraction of data range to skip at either end when examining possible break points. It has a default value of fifteen percent (0.15). TRIM must be a positive real number greater than 0 and less than 0.25.
break() selects which elements of the process are allowed to have a break. BREAK can be INTERCEPT, TREND, or BOTH. The default selection is intercept.
lagmethod() Selects the method for deciding the number of additional lags. If INPUT, the number of lags given by the MAXLAGS option is used. If AIC, the AIC-minimizing value is used. If BIC, the BIC-minimizing value is used. If TTEST, the number of lags is that for which the last included lag has a marginal significance level less than the cutoff given by the LEVEL option.
level(#) specifies the cutoff significance level for LAGMETHOD=TTEST. It has a default value of 0.10. LEVEL must be a real number between 0 and 0.25.
graph requests a graph of the unit-root test statistics for the different break points. The default is NOGRAPH.
Examples
. webuse turksales
. dfuller sales . zandrews sales, graph . zandrews sales, break(trend) . zandrews sales, break(both) trim(0.10) . zandrews sales, lagmethod(BIC) . zandrews D.sales, graph . webuse grunfeld
. zandrews invest if company==3, break(trend) graph
References
Andrews, D., Zivot, E. 1992. Further evidence on the Great Crash, the oil price > shock, and the unit-root hypothesis. Journal of Business and Economic Statistics 10, 251-7 > 0.
Acknowledgements
This module was translated from RATS code. Much of the work of implementing and testing its features was done by Margaret Green, sponsored by an Undergraduate Research Assistantship from the College of Arts and Sciences, Boston College.
Author
Christopher F. Baum (baum@bc.edu) Boston College
Also see
dfuller, dfgls