Working Paper: NBER ID: w18501
Authors: Kenneth L. Judd; Lilia Maliar; Serguei Maliar
Abstract: We introduce an algorithm for solving dynamic economic models that merges stochastic simulation and projection approaches: we use simulation to approximate the ergodic measure of the solution, we construct a fixed grid covering the support of the constructed ergodic measure, and we use projection techniques to accurately solve the model on that grid. The grid construction is the key novel piece of our analysis: we select an ε-distinguishable subset of simulated points that covers the support of the ergodic measure roughly uniformly. The proposed algorithm is tractable in problems with high dimensionality (hundreds of state variables) on a desktop computer. As an illustration, we solve one- and multicountry neoclassical growth models and a large-scale new Keynesian model with a zero lower bound on nominal interest rates.
Keywords: Dynamic Economic Models; Stochastic Simulation; Projection Methods; High Dimensionality
JEL Codes: C61; C63
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| merging of stochastic simulation and projection techniques (C53) | accurate solutions for high-dimensional dynamic economic models (E13) |
| proposed algorithm (C69) | accurately compute solutions for models with up to 80 state variables (C51) |
| EDS method (C51) | solutions with maximum unit-free approximation errors consistently smaller than 0.01 (C60) |
| EDS method (C51) | significantly lower error rate compared to perturbation methods (C52) |
| proposed algorithm (C69) | effective when dealing with the zero lower bound on nominal interest rates in New Keynesian models (E12) |