Cookies en Macro by Mark.Usamos cookies esenciales para inicio de sesión, idioma y seguridad. Con tu permiso, también usamos cookies opcionales de preferencias, para analítica y para medición publicitaria con el píxel de Meta (Facebook). Lee nuestra política de cookies.
Bayesian shrinkage on a vector autoregression - Minnesota-style priors stabilize forecasts for medium-sized macro systems.
Follow the same equations and assumptions one section at a time or in a continuous reading.
Hover or focus a symbol defined in this model to find its exact matches. Select it to keep the highlight; Escape clears the selection.
Write the -variable in stacked regression form. Let be the matrix of observations, the matrix of lagged regressors ( including the intercept), and the coefficient matrix. The model is , where is with rows .
Vectorize: is , is , and the likelihood is . The prior on given under the natural conjugate (normal-inverse-Wishart) setup is , where is the prior mean and is a single prior covariance that the Kronecker structure forces on every equation, up to the scaling. This is symmetric cross-shrinkage: the asymmetric Minnesota variance with cannot be written in the form.
The Minnesota prior sets specific entries in and . For variable 's equation, the prior mean on its own -th lag is - on the first lag only, zero on own lags ( for a random-walk variable, 0 for a stationary variable) - and the prior mean on variable 's -th lag is 0. The prior variance for the coefficient is , where if and if , and is the residual standard error from a univariate for variable .
Standard reduced-form VAR in matrix regression form.
Vectorized form enabling standard Bayesian normal-linear-model machinery.
Natural conjugate prior. Conditioning on makes the posterior tractable.
The Minnesota prior variance formula. Own lags get more prior variance (less shrinkage) than cross lags.
Why conjugacy matters
The normal-inverse-Wishart prior is the only prior family for which the multivariate regression posterior is available in closed form. Any other prior - horseshoe, spike-and-slab, non-parametric - requires MCMC or variational approximation.
In a linear state-space representation, B maps the innovation into the contemporaneous state. A delayed variable has a zero row in B and moves only after A propagates the shock. A jump is therefore a timing restriction, not a charting artifact.
A nonzero point at h=0 means the shock changes the plotted variable on impact. It is not the end of a smooth pre-shock segment.
Do not smooth through h=0 when the model says the shock enters contemporaneously.
Stable eigenvalues shrink repeated applications of A. If a variable is reached through another state first, the response rises before the stable decay dominates, producing a hump-shaped path.
Post-impact movement follows the model recursion. A hump means delayed propagation, not interpolation by the chart.
Do not treat a hump as evidence of a linear ramp unless the model includes that ramp.
The shock is dated at horizon zero. A response at h=0 is an impact effect, not the end of a hidden ramp.
A state can move on impact only through the relevant row of B; otherwise it moves after A propagates the shock.
The plotted variable can jump directly through D or indirectly through the state vector.