StudentBVARResult.innovation_covariance

property StudentBVARResult.innovation_covariance: NDArray[float64]

The innovation covariance scale * df / (df - 2).

Reported alongside the scale rather than instead of it, because the scale is what the structural factorization and the predictive consume. Requires df > 2, which the estimator guarantees.

Example

>>> import numpy as np
>>> y = np.random.default_rng(0).standard_normal((120, 2))
>>> res = StudentBVAR(y, order=1).fit(df=5.0, n_draws=30, n_burn=10, seed=0)
>>> bool(np.allclose(res.innovation_covariance, res.sigma_u * 5.0 / 3.0))
True