The Bayesian Route, Orthogonality, and Probability Matching in the Gaussian AR(1) Model

This preprint studies probability matching priors in the stationary Gaussian AR(1) model. Using a Bayesian-route approach based on higher-order asymptotics, we connect Edgeworth expansions, posterior Laplace expansions, orthogonal nuisance parameters, and the Cox-Reid adjustment to characterize a matching-prior class for the autoregressive parameter. The record includes the manuscript, LaTeX source, simulation code, and simulation results.