Andrews and Ploberger brekapoint tests
resout = AndPlob(resin,t1,t2,arg1,...,argn)
* resin = an ols result tlist
* t1 = starting breakpoint index or percentage, may be integer in [k+1,T-k-1] or percentage in (0,1).
* t2 = ending breakpoint index or percentage, may be integer in [k+1,T-k-1] or percentage in (0,1), must equal or exceed t1.
* arg1,..., argn = optional arguments that can be:
- 'noprint' if the user does not want to print the results
- 'nboost=xx' where xx is the number of bootstrap replications (default=1000)
* resout = a result tlist, with
. resout('meth') = 'AndPlob'
. resout('res all') = the original ols tlist
. resout('res 1') = the estimation results over the period before the break
. resout('res 2') = the estimation results over the period after the break
. resout('res 2') = the estimated break point
. resout('SupF') = Andrews Sup F statistics
. resout('ExpF') = Andrews and Ploberger Exponential F statistics
. resout('AveF') = Andrews and Ploberger Average F statistics
. resout('SupF pvalues') = Andrews Sup F p-values:
- asymptotic
- boostrapped under homoskedasticity
- boostrapped under heteroskedasticity
. resout('ExpF') = Andrews and Ploberger Exponential F p-values
- asymptotic
- boostrapped under homoskedasticity
- boostrapped under heteroskedasticity
. resout('AveF') = Andrews and Ploberger Average F p-values
- asymptotic
- boostrapped under homoskedasticity
- boostrapped under heteroskedasticity
. resout('nboost') = # of boostrap draws