statsmodels.tsa.vector_ar.irf.IRAnalysis.err_band_sz3#

IRAnalysis.err_band_sz3(orth=False, svar=False, repl=1000, signif=0.05, rng=None, burn=100, component=None)[source]#

IRF Sims-Zha error band method 3. Does not assume symmetric error bands around mean.

Parameters:

orthbool, optional: Compute orthogonalized impulse responses. The default is False.
svarbool, optional: Use structural IRFs. The default is False.
replint, optional: Number of MC replications. The default is 1000.
signiffloat, optional: Significance level for the confidence interval, between 0 and 1. The default is 0.05, giving a 95% confidence interval.
rngint, array_like of int, numpy.random.Generator, or numpy.random.RandomState, optional: Source of random numbers used for the Monte Carlo replications. If rng is None, a new Generator is created using fresh entropy from the operating system. If rng is an int, a new RandomState instance is created, seeded with rng; this integer-seeding behavior is deprecated and will change to creating a Generator in a future release. If rng is already a Generator or RandomState instance, that instance is used.
seedint, array_like of int, numpy.random.Generator, or numpy.random.RandomState, optional: Deprecated since version 0.15: seed has been deprecated. In-line with SPEC-007, use rng for passing a random number generator or seed.
burnint, optional: Number of initial simulated obs to discard. The default is 100.
componentarray_like of int, optional: Sequence of length neqs giving the index of the column of eigenvector/value to use for each error band. Note: the period of impulse (t=0) is not included when computing the principal component. If None, the column of the largest eigenvalue is used for each element.

Returns:

ErrorBand: A result object with fields lower and upper.

References

Sims, Christopher A., and Tao Zha. 1999. “Error Bands for Impulse Response”. Econometrica 67: 1113-1155.