Consider the VAR() model
|
or
|
When the parameter vector has a prior multivariate normal distribution with known mean and covariance matrix , the prior density is written as
|
The likelihood function for the Gaussian process becomes
|
|
|
|
|
|
Therefore, the posterior density is derived as
|
where the posterior mean is
|
and the posterior covariance matrix is
|
In practice, the prior mean and the prior variance need to be specified. If all the parameters are considered to shrink toward zero, the null prior mean should be specified. According to Litterman (1986), the prior variance can be given by
|
where is the prior variance of the th element of , is the prior standard deviation of the diagonal elements of , is a constant in the interval , and is the th diagonal element of . The deterministic terms have diffused prior variance. In practice, you replace the by the diagonal element of the ML estimator of in the nonconstrained model.
For example, for a bivariate BVAR(2) model,
|
|
|
|
|
|
with the prior covariance matrix
|
|
|
|
|
|
For the Bayesian estimation of integrated systems, the prior mean is set to the first lag of each variable equal to one in its own equation and all other coefficients at zero. For example, for a bivariate BVAR(2) model,
|
|
|
|
|
|
The mean squared error (MSE) is used to measure forecast accuracy (Litterman 1986). The MSE of the -step-ahead forecast is
|
where is the number specified by NREP= option, is the time index of the observation to be forecasted in repetition , is the actual value at time , and is the forecast made periods earlier.
The Bayesian vector autoregressive model with exogenous variables is called the BVARX(,) model. The form of the BVARX(,) model can be written as
|
The parameter estimates can be obtained by representing the general form of the multivariate linear model,
|
The prior means for the AR coefficients are the same as those specified in BVAR(). The prior means for the exogenous coefficients are set to zero.
Some examples of the Bayesian VARX model are as follows:
model y1 y2 = x1 / p=1 xlag=1 prior;
model y1 y2 = x1 / p=(1 3) xlag=1 nocurrentx prior=(lambda=0.9 theta=0.1);