Survey_COW_2003

Abstract

In preparation for the stochastic simulations, we first computed the structural shocks of the model based on U.S. data from 1980 to 1999.13 Since the non-negativity constraint for nominal interest rates was never binding during this period and our model is otherwise linear, we obtained the structural shocks by solving the model analytically for the reduced form using the AIM implementation (Anderson and Moore, 1985, and Anderson, 1997) of the Blanchard and Kahn (1980) method for solving linear rational expectations models. The structural shocks also provide a good indication of the historical fit of our model. A further indication of the good empirical fit of our model is obtained from a comparison of the implied autocorrelation functions of inflation and output with the empirical autocorrelation functions implied by an unconstrained bivariate VAR. Based on the covariance matrix of structural historical shocks, we generated 100 sets of artificial normally-distributed shocks with 100 quarters of shocks in each set from which the first 20 twenty quarters of shocks were discarded in order to guarantee that the effect of the initial values die out. We then used the sets of retained shocks to conduct stochastic simulations under alternative inflation targets, while imposing the non-negativity constraint on nominal interest rates.