VARResult.coefficient_covariance

property VARResult.coefficient_covariance: _CoefficientInference

Covariance of the coefficient matrix, sampling or posterior.

Returns whichever object describes how the numbers were produced. A shrunk fit carries its posterior covariance from estimation, because it cannot be reconstructed here: under a prior the joint covariance stops factoring as kron(sigma_u, inv(X'X)) and there is nothing to rebuild it from. Both objects answer the same questions, which is why everything downstream – bse, tvalues, conf_int, the coefficient table – reads one interface and never asks which it has.