View Leeper and Zha (2003) inspired modesty tests from conditional forecasting exercises of the DSGE model. Conditional forecasts concerns prediction of the observed variables of the DSGE model conditional on a certain path and length of path for a subset of the observed variables. The modesty analysis aims at discovering if the conditional forecasts are subject to the Lucas (1976) critique, or if the conditioning assumptions may be viewed as modest (policy interventions). Modesty statistics are also computed for forecasts based on assumptions for state variables and the results are available from the submenu item State Variable Assumptions.
YADA can compute modesty tests for the initial parameter values, the posterior mode values, and for draws from the posterior or the prior distribution. The tests depend on the method for conditional forecasting: (1) values for shocks, (2) distribution of shocks (Waggoner-Zha), and (3) distribution of subset of shocks.
Additional Information
• | A more detailed description about the computation of multivariate and univariate modesty tests can also be found in Section 12.3 of the YADA Manual. |
• | Conditional forecasts are discussed in Section 12.2 of the YADA Manual. |
Page url: http://www.texlips.net/yada/help/index.html?modesty_statistics.htm