Discussion and Analysis by Charles Porter:
The Federal Reserve Bank of Philadelphia received media attention this week following their proclamation of the death, or at least the weakening, of the Phillips Curve. Commentary on this observation has regarded the European and United States macroeconomies alike. This three-part article challenges such a notion and, instead, demonstrates an uncertain and complex Phillips Curve within the Eurozone economy. To achieve this, we simultaneously guide you through the Phillips Curve.
Part One: Where has the Phillips Curve Gone?
In addition to the publication emanating from Philadelphia, this article is motivated by the annualised Eurozone Consumer Price Index (CPI) inflation and unemployment data that was released on Thursday 31st August. The CPI inflation rate was anticipated by the market to only increase from an annualised 1.3% to 1.4%. However, inflation was published at 1.5%; a trend predicted by our analysts. Meanwhile, Eurozone unemployment was recorded to hold at 9.1%. The release of these statistics attracted numerous comments about the defiance, and end, of a Eurozone Phillips curve.
Prior this release, Ardo Hanson, ECB Governing Council member, had hinted at an explanation for the apparent defiance of the Phillips Curve. Rhetoric surrounding the Jackson Hole symposium similarly stimulated commentary on the supposed relationship. As will be explained below, the inflation and unemployment statistics defy the macroeconomic relationship described by the Phillips curve; inflation increased by 0.2% without being induced by a decline in the unemployment level. This article offers an explanation of why this may have happened using economic theory.
Whilst this inflation rate revision may be unremarkable, it does demand further attention to the anticipation of monetary policy tightening from the European Central Bank (ECB) in the near future. Greater inflationary pressure, whilst still comfortably short of the ‘just below 2%’ target, does award weight to Mario Draghi’s, president of the ECB, assertion at the Jackson Hole symposium that “the global recovery is firming up”. In order to elucidate, and then challenge, the death of the Phillips Curve, we revisit the macroeconomic theory and apply it to the Eurozone.
The Phillips Curve:
One of the most popular and commonly cited macroeconomic theories is the Phillips Curve. It receives its name from LSE economist William Phillips who provided the impetus for the relationship in his 1958 publication. In this publication, Phillips investigated the relationship between unemployment and inflation (although under the guise of wage inflation). His article found, and subsequent qualifications endorsed, a strong inverse correlation along a curve.
Therefore, what was defended and extrapolated was a law of economics: lower unemployment, and therefore higher employment, generates a higher rate of inflation. The converse was therefore also discovered: lower rates of inflation coincide with higher rates of unemployment. Given the prevailing economic paradigm and the dominance of the-then mainstream economic thought, the causal incidence is provided by fiscal, government, expenditure.
Phillips, and the pioneers of what became the Phillips curve, examined the relationship over a vast time period and proposed a stable relationship. The 1958 publication even stretched across nearly a century, 1860-1957. The challenges that the Phillips curve now faces, and indeed those that it has already faced in the history of the theorised relationship, are evident from the first graph below. Figure 1 shows our simulation of the Phillips curve by taking quarterly (annualised) Eurozone inflation and unemployment data:
Testing the Phillips Curve
Figure 1: The Phillips Curve? The price inflation rate was plotted against the unemployment rate, both expressed as annualised percentages using data retried from the OECD database. A perfect Philips Curve would produce a downward sloping (inverse) relationship.
Two features of the above graph, one indicative and one analytical, support the assertion that the Phillips Curve no longer applies within modern economies, with little redeeming features. Over the medium-long run, 11 years, which this graph depicts, it is immediately discernible that the cluster of observations does not conform accurately to the line of best fit. In fact, if the polynomial line of best fit wasn’t mathematically fitted, it is unlikely that visual observation alone would lead an individual to draw a curve, or even linear line, of best fit with an appropriate gradient. Therefore, either a number of potential Phillips Curves would be admissible from the above graph, or none at all.
The second, empirical, feature of Figure 1 that supports the null-significance of the Phillips curve concerns the R2 value. This number, presented within each figure, measures how well the model (the, Phillips, line of best fit) describes the data from which it is drawn. Therefore, it describes whether the model is a good fit for the data or not. The closer an R2 value is to 1 (representing perfection), the better the model is. An R2 value of 0.3027 for a second order polynomial is essentially meaningless, thereby generating a similar conclusion for the existence of a Phillips Curve over this period. The R2 value quoted above therefore empirically validates the indicative analysis.
Part Two Addresses this limitation: