help ghansen
-------------------------------------------------------------------------------

Title

ghansen -- Gregory-Hansen test for cointegration with regime shifts

Syntax

ghansen depvar indepvars [if] [in] , break(string) lagmethod(string) [maxlags(integer) trim(real) level(real)]

options Description ------------------------------------------------------------------------- Main break type of structural break of cointegration vector lagmethod Method for lag selection in adf test maxlags Maximum lags to search in adf test trim fraction of data range to skip at either end when examining possible break points level confidence level for downward t statistic lag selection in adf test ------------------------------------------------------------------------- by is allowed; see [D] by. You must tsset your data before using ghansen; see [TS] tsset. depvar and indepvars may contain time-series operators; see tsvarlist.

Description

ghansen performs the Gregory-Hansen test for cointegration with regime shifts proposed in Gregory and Hansen (1996) on a dependent variable depvar and up to four independent variables in indepvars . The test's null hypothesis is no cointegration against the alternative of cointegration with a single shift at an unknown point in time. ghansen does not allow gaps within the observations of the time series.

Options

+------+ ----+ Main +-------------------------------------------------------------

break(string) is required and specifies the type of break in the cointegration vector. Details are provided in Gregory and Hansen (1996) and Gregory and Hansen (1996a).

break(level) specifies a break in the constant term.

break(trend) specifies a break in the constant and the trend.

break(regime) specifies a break in the constant and the slope.

break(regimetrend) specifies a break in the constant, the slope and the trend (Gregory and Hansen 1996a)

lagmethod(string) is required and specifies the algorithm to choose the number of lags in the Augmented Dickey Fuller tests.

lagmethod(fixed) sets the number of lags equal to the maximum lags specified in maxlag(integer)

lagmethod(aic) chooses the number of lags that minimize the Akaike information criterion.

lagmethod(bic) chooses the number of lags that minimize the Bayesian information criterion.

lagmethod(downt) chooses the number of lags such that the last lag is significant according to its t-statistic, starting from a maximum number of lags.

maxlags(integer) is optional and specifies the maximum number of lags to include in the Augmented Dickey Fuller test, or the number of lags to include if lagmethod is fixed. maxlags is required if lagmethod is fixed. Default is T^0.25, where T is the sample size.

trim(real) is optional and specifies the fraction of the data range to skip at either end when examining possible break points. Default is 0.15. trim must be a positive real number between 0 and 0.25.

level(real) is optional and specifies the cutoff confidence level for the downt lag selection method. Default is 0.95.

Examples

. webuse balance2 ,clear

Break in level, lags chosen by Akaike criterion, maximum of 5 lags.

. ghansen y i c, break(level) lagmethod(aic) maxlags(5)

Change in regime , 5 lags chosen by user.

. ghansen y i c, break(regime) lagmethod(fixed) maxlags(5)

Change in regime and trend , downward t lag selection at the 99% confidence level, trim 10% of sample.

. ghansen y i c, break(regimetrend) lagmethod(downt) level(0.99) trim(0.1)

Saved results

ghansen saves the following in r():

Scalars r(tstat) t-statistic for adf test r(lag) lags for adf test r(breakptadf) estimated breakpoint for adf test r(za) za statistic r(breakptza) estimated breakpoint for za test r(zt) zt statistic r(breakptzt) estimated breakpoint for zt test

Macros r(break) Type of break r(badfdate) estimated break date for adf test r(bzadate) estimated break date for za test r(bztdate) estimated break date for zt test

Author

Jorge Pérez, Inter-American Development Bank jperezperez@iadb.org

References

Gregory, Allan W. and Bruce E. Hansen (1996). "Residual-Based Tests for Cointegration in Models with Regime Shifts", Journal of Econometrics 70(1), pp. 99-126.

Gregory, Allan W. and Bruce E. Hansen (1996a). "Tests for Cointegration in Models with Regime and Trend Shifts", Oxford Bulletin of Economics and Statistics 58 (3), pp. 555-60.

Hansen, Bruce E. (n.d.) Residual-based tests for cointegration in models with regime shifts. Programs -- Cointegration. Retrieved August 1, 2011 from http://www.ssc.wisc.edu/~bhansen/progs/joe_96.html

Acknowledgements

This command is based on R code provided in Hansen (n.d.) I am grateful to Muhammad Anees and Nicholas Cox for help in identifying a bug in a previous version of the program.

Disclaimer

This program is provided without warranty of any kind. The author is not responsible for any cost derived by the usage of this program.

Also see

Help: [TS] vec, [TS] vecrank, zandrews (if installed),