Counterfactual analysis in macroeconomic models
In this paper, James Hebden and Fabian Winkler present a novel, computationally efficient method for policy analysis that requires minimal structural model information and can handle various policy regimes and constraints. In particular, the authors propose a computational procedure to solve for policy counterfactuals in linear models with occasionally binding constraints using a sequence-space approach.
- The method requires only two inputs:
– Projections of variables relevant to the policy problem
– Impulse response functions of these variables to monetary policy instruments - The approach can compute solutions for:
– Instrument rules
– Optimal discretionary policies
– Optimal commitment policies
The methodology can handle occasionally binding constraints (e.g., effective lower bound on interest rates) efficiently. It does not require knowledge of structural model equations or filtering of structural shocks and it can accommodate stochastic changes in projections while honoring state-contingent commitments under optimal policy.
Main findings:
- The authors apply their method to compute counterfactuals for the U.S. economy between 2020-2023 using:
– FRB/US model
– Modified version of the del Negro, Giannoni, and Schorfheide (2015) model
– Baseline projections from FOMC Summary of Economic Projections - They compare outcomes under different policy regimes:
– Simple Taylor-type rules
– Optimal discretionary policy
– Optimal commitment policy - The method can efficiently handle revisions to economic projections and honor past policy commitments.
- The approach facilitates comparison of policy effects across different models, addressing model uncertainty concerns.
Key results:
- Optimal policy under commitment would have made strong promises to keep interest rates low in 2020.
- These promises would have been honored even as inflation rose subsequently.
- The optimal commitment policy path would have appeared to be “behind the curve” relative to simple Taylor-type rules.
- This apparent lag reflects the time-inconsistency of the optimal policy.
Computation of Policy Counterfactuals in Sequence Space
Authors: James Hebden, Fabian Winkler
From: Federal Reserve Board
Counterfactual analysis with minimal modeling assumptions
In this paper, Caravello et al. propose a novel approach to policy counterfactuals that relies less on full model specification, and demonstrates its usefulness for evaluating monetary policy questions. The authors propose a new “VAR-Plus” approach for evaluating policy counterfactuals that combines empirical time-series evidence with structural modeling.
- The approach has two main steps:
– VAR step: Estimate reduced-form projections and some policy causal effects using vector autoregressions and structural VAR methods.
– Plus step: Use structural models to match and extrapolate beyond the VAR evidence on policy effects. - The method leverages a theoretical identification result showing that policy counterfactuals are pinned down by just two sufficient statistics: reduced-form projections and policy causal effects.
- This approach requires weaker assumptions than full DSGE models and avoids having to specify the underlying shocks driving business cycles.
Main findings:
- The VAR evidence alone often largely determines the counterfactuals, with less role for model-based extrapolation.
- Standard RANK and HANK models extrapolate policy effects similarly, while behavioral models with less forward-looking pricing differ substantially.
- The results highlight the importance of empirical evidence on persistent policy changes and discriminating between rational and behavioral models.
Key results:
- Average business cycle: A counterfactual inflation targeting policy could have achieved lower output gap volatility and slightly more stable inflation.
- Great Recession: Without a lower bound constraint, interest rates would have been cut to around -4%, reducing the output gap at the cost of moderately higher inflation.
- Post-COVID inflation: There is large uncertainty about the optimal policy response, depending on assumptions about how persistent rate changes affect the economy.
Evaluating Policy Counterfactuals: A VAR-Plus Approach
Authors: Tomás E. Caravello, Alisdair McKay, Christian K. Wolf
From: MIT, FRB Minneapolis
