Bayesian Autoregressive Modeling

Specification and estimation of Bayesian univariate autoregressive models.

Description

The priors may be: Litterman random walk plus drift prior; Raynauld-Simonato seasonal random walk plus drift prior or Canova prior: seasonal and non-seasonal unit roots via stochastic constraints + Litterman prior. Combination is achieved using Theil-Goldberger mixed estimation.
Estimation uses Theil-Goldberger mixed procedure or OLS.
Calibration may be performed using axial search.
Analytics based on the AR filter.

• MATLAB release: MATLAB 7.8 (R2009a)