statsmodels.tsa.vector_ar.var_model.VARResults.mse

VARResults.mse(steps)

Compute theoretical forecast error variance matrices

Parameters:

stepsint

Number of steps ahead

Returns:

forc_covsndarray (steps x neqs x neqs)

Notes

$$\mathrm{MSE}(h)=\sum _{i=0}^{h-1}\mathrm{\Phi }{\mathrm{\Sigma }}_u{\mathrm{\Phi }}^T$$