Measuring monetary policy: rules versus discretion

In this paper, we propose a novel method to measure the strength of commitment versus discretion in monetary policy. We estimate a Taylor-type monetary policy rule with time-varying heteroskedasticity, decomposing the policy into rule-based and discretionary components. Deviations from the committed rule, in form of volatility of a policy shock, are linked to macroeconomic variables, disclosing the nature and strength of discretion of a central banker. We estimate our model for the period 1967–2005 focusing on the determinants of volatility of policy shocks in the USA. The proposed heteroskedastic model provides a better fit to the data than the standard Taylor rule. Inflation has positive significant effects on volatility of shocks in the full sample, pre-Volcker, and Volcker periods, while during the Greenspan era, the degree of discretion is decreasing in headline inflation, but increasing in core inflation. We also find significant positive association between stock market volatility and policy discretion, but no relationship between the shock volatility and output gap.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic €32.70 /Month

Buy Now

Price includes VAT (France)

Instant access to the full article PDF.

Rent this article via DeepDyve

Similar content being viewed by others

Fiscal rules and volatility: the role of stabilising properties and compliance

Article 17 August 2021

The Lucas hypothesis on monetary shocks: evidence from a GARCH-in-mean model

Article 29 June 2017

The Macroeconomic Effects of the Federal Reserve’s Conventional and Unconventional Monetary Policies

Article 10 June 2024

Notes

For analytical derivation of policy rules under discretion and commitment, see Clarida et al. (1999) and Clarida et al. (2000).

In the literature, “monetary policy shock” is often referred to as an unanticipated monetary policy action. In this paper, we refer to monetary policy shocks as deviations from the policy rule.

For the early literature, see Clarida et al. (2000), Orphanides (2004) and the references therein.

Friedman (2006) concludes that the era may stand as the modern day pinnacle of “discretion” rather than “rule.”

Orphanides (2003) shows that the systematic component of the monetary policy in the Greenspan era was not much different from earlier times. Kahn (2012) analyzed the transcripts of Federal Open Market Committee (FOMC) during this period and found a large number of references by committee members to monetary policy rules. On the contrary, Blinder and Reis (2005) conclude that “Federal Reserve policy under his chairmanship has been characterized by the exercise of pure, period-by-period discretion, with minimal strategic constraints of any kind, maximal tactical flexibility at all times, and not much in the way of explanation.” (p.14).

In the pre-Volcker period 1967Q1–1979Q2, the Fed-Reserve had been chaired by William Martin, 1951–1970; Arthur Burns, 1970–1978; and William Miller, 1978–1979.

Although Greenbook projections are available for a longer period, we use data up to 2005Q4 in order to exclude effects of possible structural changes in the rule resulting from Ben Bernanke’s appointment as the chairman of the Federal Reserve, as well as the effects of the Great Recession and unconventional monetary policy measures implemented after 2008.

The core measure of inflation excludes food and energy prices from the price index. We obtain data on VIX index from the FRED database of the Federal Reserve Bank of St. Louis. Choice of timing of variables is motivated in the next section.

The null hypothesis that the mean difference is equal to zero is rejected at 1% significance level with P-value = 0.003.

The correlation coefficients are 0.81 between \( \pi _^c \) and \( \pi _^h \) , 0.80 between \( \pi _^h \) and \( \pi _^d \) , 0.89 between \( \pi _^c \) and \( \pi _^d \) .

The correlation between the measures of output gap \( x_ \) and \( x_^ \) is 0.43. See Enders (2015), p.70.

We discuss alternative specifications with additional regressors in the mean equation in the Robustness section.

References

Acknowledgements

We thank Jordi Galí, Christian Brownlees, the Editor of the Journal, two anonymous Referees, and participants of Armenian Economic Association’s 2016 Annual Meeting for helpful comments and suggestions.

Author information

Authors and Affiliations

  1. Department of Economics and Business, Universitat Pompeu Fabra, 25-27, Ramon Trias Fargas, 08005, Barcelona, Spain Narek Ohanyan
  2. American University of Armenia, 40, Baghramyan Ave., 0019, Yerevan, Armenia Aleksandr Grigoryan
  1. Narek Ohanyan