Statement
U.S. monetary policy during the pre-Volcker period (approximately 1960-1979) is best characterized as “passive” — the Federal Reserve’s inflation response coefficient was below unity, violating the Taylor principle. Through the lens of standard New Keynesian DSGE models, this passive policy implies equilibrium indeterminacy: the propagation of fundamental shocks is not uniquely determined, and self-fulfilling sunspot fluctuations can influence equilibrium allocations. Bayesian estimation that covers both the determinacy and indeterminacy regions of the parameter space assigns ~80-97% posterior probability to indeterminacy for the pre-1979 sub-sample, while the post-1982 sample is overwhelmingly consistent with determinacy.
Evidence summary
- Lubik-Schorfheide (2004): direct Bayesian estimation allowing for indeterminacy on a three-equation NK model. Pre-Volcker: P(indeterminacy) = 0.80-0.97. Post-1982: P(determinacy) = 0.95-1.00. This is the primary evidence.
- Bianchi (2013): medium-scale MS-DSGE model estimates the Dove regime (inflation response ~0.95) prevailing in the 1970s, consistent with the passive policy finding. The MS framework shows this is a recurrent phenomenon, not a one-time structural break.
- Clarida-Gali-Gertler (2000): single-equation GMM estimation finds pre-Volcker inflation response < 1, the original finding that motivated the subsequent structural literature.
Conditions and scope
- Model-dependent: the association of psi_1 < 1 with indeterminacy requires the standard NK model assumptions (forward-looking IS, Calvo pricing, no endogenous investment in the simplest version).
- Sub-sample-dependent for the Lubik-Schorfheide approach (not for Bianchi’s regime-switching approach).
- The post-1982 determinacy finding is equally model-dependent.
Counter-evidence
- Dupor (2001): in a continuous-time model with endogenous investment, passive monetary policy can be consistent with determinacy.
- Sims-Zha (2006): find no evidence of changing monetary policy parameters in a Bayesian VAR; attribute the Great Moderation to changing shock volatilities rather than policy changes.
- Orphanides (2002): argues that when real-time data (rather than revised data) are used, the pre-Volcker Fed’s policy appears more active than suggested by retrospective estimates.
- Model misspecification concern: richer endogenous dynamics under indeterminacy can capture omitted propagation mechanisms, potentially biasing posteriors toward indeterminacy.
Linked ideas
- (none yet)
Open questions
- Does the indeterminacy finding survive in larger-scale DSGE models with financial frictions, wage rigidity, and investment?
- Can the pre-Volcker indeterminacy be reinterpreted through a regime- switching lens where agents anticipated the eventual Volcker disinflation?
- What are the welfare costs of the pre-Volcker indeterminacy?