Stochastic Volatility in DSGE models

A number of recent papers have concluded that stochastic volatility plays a prominent role in describing the business cycle, particularly for the characterization of monetary policy. The impact of including stochastic volatility in DSGE models remains, however, unexplored. This paper therefore deals with the estimation of DSGE models when structural innovations have volatilities that are allowed to vary over time. In particular, we develop an efficient algorithm for estimating DSGE models subject to stochastic volatility that allows for jointly inferring the model’s parameters, underlying shocks and time varying volatilities. We apply our algorithm to the estimation of a model of the US business cycle and show the implications of the inclusion of stochastic volatility for the shape of impulse responses and variance decomposition