A Dynamic Stochastic General Equilibrium model is a macroeconomic forecasting framework built from optimizing households and firms, nominal rigidities, and stochastic shocks, estimated to project output, inflation, and policy rates — and the probabilities of outcomes like recession — under internally consistent equilibrium dynamics.
How it works
DSGE models derive aggregate behavior from microfounded agents who maximize utility and profit subject to budget constraints, with sticky prices and wages creating a role for monetary policy. The system is hit by stochastic shocks (technology, demand, policy), then estimated using Bayesian methods to generate forecasts and full predictive distributions for key macro variables.
Why it matters now
In the 2025-2026 late-cycle debate, structural models like the New York Fed's DSGE anchor the dovish-growth end of the forecast distribution; its recession-probability and disinflation projections feed directly into how desks read the Fed's reaction function against survey- and market-based alternatives.
Example
The New York Fed's DSGE model is published quarterly alongside FOMC cycles; in our briefing it sat at the dovish-growth end of the dossier, putting the probability of a recession at 46 percent — a structural read that diverged from market-implied and survey-based estimates and pulled the consensus forecast distribution toward softer growth and faster disinflation.
Frequently asked
- What is a DSGE model?
- A DSGE model is a macroeconomic framework that derives forecasts from optimizing households and firms, nominal rigidities, and random shocks within a general-equilibrium system. It produces internally consistent projections for output, inflation, and interest rates, plus full probability distributions for outcomes like recession, and is estimated with Bayesian methods. Central banks including the New York Fed publish them quarterly.
- How does a DSGE model differ from a Bayesian VAR?
- A DSGE model imposes explicit economic theory — optimizing agents, budget constraints, sticky prices — that restricts how variables interact, while a Bayesian VAR
Glossary · Bayesian VAR
A Bayesian vector autoregression is a multivariate time-series model that estimates the joint dynamics of several macro variables while imposing prior beliefs to discipline the large number of parameters. The priors shrink coefficients toward parsimonious benchmarks, taming overfitting and producing more stable forecasts and impulse responses than unrestricted VARs.
Read more →
- Why does the New York Fed's DSGE model matter for rate forecasts?
- The New York Fed's DSGE model matters because it provides a structural, internally consistent benchmark for output, inflation, and recession probabilities that desks weigh against market- and survey-based forecasts. Published each quarter, it often sits at the dovish-growth end of the forecast distribution, shaping how analysts read the Fed's reaction function and disinflation path in the 2025-2026 cycle.
- What are the main criticisms of DSGE models?
- DSGE models are criticized for poor forecasting around turning points, reliance on representative-agent assumptions that miss heterogeneity and financial frictions, and an empirical record that failed to anticipate the 2008 crisis. Post-2008 work added financial sectors, the zero lower bound, and heterogeneous-agent (HANK) extensions, but skeptics argue the theoretical restrictions still distort real-world dynamics.