BBCTest reference

[long...@gmail.com]

Hi Matthieu,

Could you please refer me to the paper(s) that underlie the code of the function "BBCTest" (that is used to test of unit root against a stationary three regime SETAR alternative)? I can't find any references in either the help file or the source code.

Many thanks,

Long 

[Matthieu S]

Hi Long

Oh good point, I never added the reference. It is:  Frédérique Bec; Mélika Ben Salem; Marine Carrasco (2004) Tests for Unit-Root versus Threshold Specification With an Application to the,  Journal of Business & Economic Statistics; Oct 2004; 22, 4; ABI/INFORM Global

I would recommend running some tests/simulations before using that function.

Best,

Matthieu

--
You received this message because you are subscribed to the Google Groups "tsdyn" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tsdyn+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tsdyn/6e62e824-324a-4f87-aefc-7df70227c67c%40googlegroups.com.

[LV]

Many thanks for the reference, Matthieu.

To follow that up, how should we interpret the test results? 

+ The function gives us test statistics and critical values at 0.9, 0.95 and 0.99. Am I correct in saying that if the statistic is smaller than the value in the middle (for 0.95 level), that means we cannot reject the null of unit root (for the process as a whole)?

+ And I take it that this test is not designed to test for unit root in each regime, only for the whole process?

Regards,

Long Vo

You received this message because you are subscribed to a topic in the Google Groups "tsdyn" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/tsdyn/-FQpfRtyDzY/unsubscribe.
To unsubscribe from this group and all its topics, send an email to tsdyn+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tsdyn/CAEYvigLFkkm79kaKNvgN1%2BLtq5ik9rs3RLwqs%2B_SmsVD-4_%3DwQ%40mail.gmail.com.

[Matthieu S]

HI Long

1) Well it depends if you want to test at 5% or 10%, but a value lower than the quantile indicates non-rejection at that probability

2) Correct, it tests for all regimes being unit root, see p. 385.

Please note that there is a warning "his function should be considered as experimental. No consistency checks with original paper could be made". I would highly recommend you trying the function against the Monte Carlo in the paper to check if the implementation behaves as expected.

Best,

Matthieu

To view this discussion on the web visit https://groups.google.com/d/msgid/tsdyn/CAFoC1mQV3CUEj0C5DeTz6wGE9ytk%2BYnqdQs6QzNQuL-ukgC8BA%40mail.gmail.com.