Released on 05/04/2019:
• | Added sequential estimation of the marginal likelihood for DSGE-VAR models. Model sequences can be selected and the tools run via items on the Model sequence menu-item on the File menu. |
• | Computing observation weights on the observed variables of the projected estimates of state variables and structural shocks is now supported for diffuse initialization based on the initial parameter values and the posterior mode estimates. The results can be displayed in terms of either decompositions of the projected variables or as weights on observed variables from a particular time period. The tools are available from the Observation weights menu-item on the Decompositions item on the Tools menu. |
• | Stationary predictive variances and the decomposition into state, shock, measurement error and parameter uncertainty can be performed from the prior or the posterior draws for DSGE models. This tool is available from the Predictive Distributions menu-item on the Tools menu. The same type of computations can be performed for DSGE-VAR models from the Predictive Distributions item on the Tools menu-item on the DSGE-VAR menu, but where the total stationary predictive variance is decomposed into shock and parameter uncertainty. |
• | Parameter scenarios can now also deal with new values for the calibrated parameters. |
Page url: http://www.texlips.net/yada/help/index.html?4_70.htm