VARXResult._trailing_blocks#

VARXResult._trailing_blocks()[source]#

The distributed-lag coefficients, in design order.

One (m, k) block per exogenous lag, transposed to the row layout of deterministic so the mixin can stack them under the endogenous lags.

Return type:

tuple[NDArray[double], ...]

Returns:

s + 1 blocks, contemporaneous term first.

Example

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> y, x = rng.standard_normal((120, 2)), rng.standard_normal((120, 3))
>>> res = VARX(y, x, order=1, exog_order=1).fit()
>>> [block.shape for block in res._trailing_blocks()]
[(3, 2), (3, 2)]