We describe methods for assessing estimated dynamic stochastic general equilibrium (DSGE) models. One involves the computation of alternative impulse responses from models constrained to have an identical likelihood and the same contemporaneous signs as responses in the DSGE model. Others ask how well the model matches the data-generating process; whether there is weak identification; the consequences of including measurement error with growth rates of non-stationary variables; and whether the model can reproduce features of the data that involve combinations of moments. The methods are applied to a large-scale small-open economy DSGE model, typical of those used at policy institutions.
|Publication status||Published - 2018|