Working Paper: NBER ID: w16714
Authors: Ulrich K. Müller; James H. Stock
Abstract: We propose a Bayesian procedure for exploiting small, possibly long-lag linear predictability in the innovations of a finite order autoregression. We model the innovations as having a log-spectral density that is a continuous mean-zero Gaussian process of order 1/√T. This local embedding makes the problem asymptotically a normal-normal Bayes problem, resulting in closed-form solutions for the best forecast. When applied to data on 132 U.S. monthly macroeconomic time series, the method is found to improve upon autoregressive forecasts by an amount consistent with the theoretical and Monte Carlo calculations.
Keywords: Bayesian forecasting; VAR models; time series; residual predictability
JEL Codes: C11; C22; C32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| slight misspecifications in finite order VAR models (C32) | improvements in forecast accuracy (C53) |
| Bayesian approach (C11) | incorporation of residual predictability (C59) |
| incorporation of residual predictability (C59) | better forecast compared to standard autoregressive forecasts (C53) |
| model specifications (using AIC and BIC) (C52) | resulting forecast accuracy (C53) |
| Bayesian method (C11) | improvements upon traditional methods (AIC and BIC) (C52) |