This repository contains MATLAB code accompanying the paper
"Efficient Estimation of Nonlinear DSGE Models"
by
Each subdirectory implements a specific model with its own estimation or filtering routine. The general structure of the filtering procedure alternates between:
- Solution Step: Solves the model locally at the current state forecast.
- Filtering Step: Constructs state-space matrices consistent with the current local solution and performs a single step of the time-varying Kalman filter.
Each subdirectory contains a dedicated README.md with details on the model, data, and estimation or filtering demonstration.
MATLAB code for full-information Bayesian estimation of the New Keynesian Diamond-Mortensen-Pissarides (NKDMP) model using a Random-Walk Metropolis-Hastings sampler.
- Three policy variables: market tightness, inflation, marginal cost
- Five structural shocks: productivity, discount factor, monetary policy, matching efficiency, separation rate
- Estimation on U.S. data from 1966:Q1 to 2019:Q4
- Newton solver with symbolic equilibrium conditions
MATLAB code for estimating the Diamond-Mortensen-Pissarides (DMP) model.
- Uses data simulated from the global solution as the data-generating process
- Benchmarks the computational performance of existing filters
MATLAB code for estimating the standard three-equation New Keynesian model.
- Solves the textbook three-equation New Keynesian model
- Full-information Bayesian estimation using data from 1966:Q1 to 2007:Q4
MATLAB code for a stochastic growth model with AR(1) productivity and stochastic volatility.
- Demonstrates how to implement the filtering procedure when the volatility of productivity follows an AR(1) process
MATLAB code for a stochastic growth model with regime-switching volatility.
- Demonstrates the filtering procedure when productivity dynamics are determined by a discrete Markov process
- MATLAB R2024a or later
- Symbolic Math Toolbox
- Statistics and Machine Learning Toolbox
This is a work in progress. Code will be updated periodically.
For questions, suggestions, or concerns, feel free to reach out: mccrary.65 at osu.edu