FLUX—General Nyblom fluctuations test

[TomDoan]

@FLUX is a general procedure for implementing the fluctuations test proposed by Nyblom(1989), "Testing for Constancy of Parameters Over Time", JASA, vol 84, pp 223-230. The inputs are either the set of scores from instructions like MAXIMIZE or GARCH or the derivatives of u from instructions like NLLS, which you can obtain with the DERIVES option on those instructions. It computes both individual and joint test statistics, gives an approximate significance level for each and an approximate date/entry at which the fluctuation statistic is farthest from zero.

Detailed description

[fadimohamed]

Hi Tom,

i have a multivariate asymmetric garch model, i want to test the stability of the model, is the flux test applicable to the multivariate specification?

Fadi

[TomDoan]

Yes. GARCH and MAXIMIZE both define the DERIVES VECTOR[SERIES] for either univariate or multivariate models. The biggest problem is that when the parameter set gets too large, you don't get good power against a break in one part of it, but it's hard to test for structural breaks in a GARCH model using anything else because of the recursive calculation of the likelihood.

[c912]

I don't understand the interpretation of the test. If the p-value is <5% ther is stability or if the p-value is >5%?

[TomDoan]

Null is stability, so <5% would reject.