The authors estimate a sticky-price dynamic stochastic general-equilibrium model with a financial accelerator, à la Bernanke, Gertler, and Gilchrist (1999), to assess the importance of financial frictions in the amplification and propagation of the effects of transitory shocks. Structural parameters of two models, one with and one without a financial accelerator, are estimated using a maximum-likelihood procedure and post-1979 U.S. data. The estimation and simulation results provide some quantitative evidence in favour of the financial-accelerator model. The financial accelerator appears to play an important role in investment fluctuations, but its importance for output depends on the nature of the initial shock.