Approximation Errors of Perturbation Methods in Solving a Class of Dynamic Stochastic General Equilibrium Models

We establish the standard procedures in solving a class of dynamic stochastic general equilibrium models with a second-order approximation to the policy function. The innovation is that the means of exogenous shocks vary with the variances. The numerical results show that second-order perturbation methods generate large approximation errors regarding to the responses of policy functions to shocks.