Evaluation of the New Keynesian Phillips Curve: evidence from the Euro Area and United States

ABSTRACT

Empirical relevance of inflation expectations in the New Keynesian Phillips Curve (NKPC) is highly controversial in the macroeconomics literature. With this in mind, this article evaluates the purely forward-looking NKPC useful for policy analysis with respect to their abilities to account for the dynamic relationship between output and inflation. Our findings show that the NKPC heavily relying on firms’ forward-looking behaviour is hardly supported by the Euro Area and the US data. The failure of the NKPC in matching the data is consistently observed across the sub-samples divided before and after the early 1980s. For comparison, we also investigate the performance of the hybrid NKPC and the traditional backward-looking Phillips curve associated with ad hoc price indexation assumptions.

KEYWORDS:

• Phillips curve
• forward-looking behaviour
• backward-looking behaviour
• inflation expectations

JEL CLASSIFICATION:

• E31
• E32

Acknowledgments

We are grateful to Mark Taylor (the editor) and an anonymous referee whose comments were very helpful

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

¹ Smets and Wouters (2007) document evidence that the observed correlation in the US data can be explained by the hybrid New Keynesian Phillips Curve (NKPC). However, this model cannot be used for policy analysis. With this in mind, we study whether the microfounded NKPC that is useful for policy analysis is able to deliver a reasonable approximation of the observed correlation. This paper also differs from Smets and Wouters (2007) in that we investigate the role of expectations about future inflation and output as well as monetary policy in accounting for the observed correlations from the Euro Area and the United States.

² The AR(1) and ARMA(1,1) processes for the demand and supply shocks are empirically motivated due to Ireland (2007) and Smets and Wouters (2007). As shown in Table 1, the estimates of the AR(1) and ARMA(1,1) coefficients are highly significant.

³ The AWM data set takes the short-term interest rate from the ECB Monthly Bulletin. The data set is available from 1970:1 through 2014:4. The AWM data set can be found at http://www.eabcn.org/page/area-wide-model.

⁴ This reverse dynamic correlation is observed even after the trend component of inflation is removed from the US inflation rate (Kim and Yie 2016). See the companion paper of mine for detailed discussions about issues regarding trend inflation.

⁵ This work is motivated from the fact that the macroeconomics literature has found a substantial shift in monetary policy. For example, Clarida et al. (2000 Clarida, R., J. Galí, and M. Gertler. 2000. “Monetary Policy Rules and Macroeconomic Stability: Evidence and Some Theory” Quarterly Journal of Economics 115(1): 147–180.) document evidence that the parameters and are estimated to be higher in the post-1980 period than in the pre-1980 period.