Forecasting and conditional projection using realistic prior distributions,10.1080/07474938408800053,Econometric Reviews,Thomas Doan,Robert Litterman, and is shown to improve out-of-sample forecasts relative to univariate equations. Although cross-variable responses are damped by the prior, considerable interaction among the variables is shown to be captured by the estimates We provide unconditional forecasts as of 1982:12 and 1983:3. We also describe how a model such as this can be used to make conditional projections and to analyze policy alternatives. As an example, we analyze a forecast made in 1982: 12 Although no automatic causal interpretations arise from models like ours, they provide a detailed characterization of the dynamic statistical interdependence of a set of economic variables, information that may help in evaluating causal hypotheses without containing any such hypotheses. The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search. , Todd (1984), Litterman (1986), and Spencer (1993), use a Bayesian VAR (BVAR) model to overcome the over-parameterization problem… propose a formula to generate standard deviations that depend on a small numbers of hyper-parameters: w, d, and a weighting matrix f(i, j )t o reduce the over-parameterization in the VAR and VEC models… ) for Bayesian modification is more suited for restricting and reducing the parameters in a VAR representation in first differences as the basic idea underlying the prior being set for the means and variances of the variable coefficients is that the variables are normally unrelated random walks rather than a cointegrated system encountered in a VEC model… suggested a formula to generate standard deviations as a function of small numbers of hyperparameters: w, d, and a weigting matrix f(i, j). This approach allows the forecaster to specify individual prior variances for a large number of coefficients based on only a few hyperparameters… suggested a formula to generate standard deviations as a function of small number of hyperparameters w, d, and a weighting matrix f(i, j). This approach allows the forecaster to specify individual prior variances for a large number of coefficients based on only a few hyperparameters… …Banbura et al. (2009) apply this result to a large Bayesian VAR (BVAR) with Litterman (1986) and sums of coefficients priors ( ).Banburaetal.(2009)findthattheforecastingperformanceandtheimpulse responses to a monetary policy shock from their large model, which contains 108 US variables, compare favourably to those of smaller VARs… …Another modification of the Minnesota prior, motivated by the frequent practice of specifying a VAR in first differences, is the sums of coefficients prior of Source.