Re-discovering the Phillips curve

László Andor

01 October 2014



As a contribution to the EU’s institutional transition, a conference entitled “Labour Economics after the Crisis” took place in Brussels on 18-19 September 2014. The aim was to draw lessons from Europe’s protracted jobs crisis and from the experience of the Barroso II Commission as regards macroeconomic and employment policies.1

My key ‘handover’ message was that Europe needs to re-discover the Phillips Curve and pursue higher inflation if the EU is to make progress towards the 75% employment rate target of the Europe 2020 Strategy.2

The European labour market is currently adversely affected by three key macroeconomic developments:

  • First, there is a persistent gap between effective aggregate demand and potential output in most Member States – combined with high unemployment, a large overhang of private debt, low inflation and nominal interest rates close to their lower bound.

The shortage of demand (for both consumption and investment) is related to demographic trends, but also to rising inequalities and an increase in savings which are not channelled to the real economy but parked in financial instruments or metropolitan real estate.

  • Second, there is an unprecedented polarisation in economic and employment outcomes across the Eurozone.

This is linked to the incomplete character of the monetary union, notably the lack of aggregate demand management and absence of a shared fiscal capacity. The Eurozone’s design has, up till now, forced macroeconomic adjustment to unfold predominantly through internal devaluation.

  • Third, Europe struggles to reap the full job potential of structural changes under way, such as technological progress and further globalisation.

The reason is that our labour market institutions, but also product markets, financial sector and public investment agencies are not capable of reallocating labour and capital in a flexible but secure way towards activities with a strong job-creation potential.

The bottom line is that secular stagnation, permanently depressed employment and rising inequalities may well become reality if Europe’s economy continues to be characterised by a large overhang of private debts from the past and at the same time low inflation (see Teulings and Baldwin 2014).

How does inflation matter for monetary, fiscal and structural policies?

Many years ago the Phillips curve used to encourage macroeconomic policy-makers to pursue full employment by using monetary and fiscal policies at their disposal. Today, the inverse relationship between unemployment and (wage or price) inflation may appear as controversial in Europe. After all, we live in an era of financial capitalism and two decades ago we established a monetary union geared towards price stability more than full employment.

In 1958, William Phillips highlighted the inverse relation between unemployment and nominal wage inflation in the UK in the previous century. A lot of similar empirical research followed and Samuelson and Solow soon re-stated the “Phillips curve” as an inverse relationship between unemployment and inflation.

The policy implication of the Phillips curve is that increasing aggregate demand through monetary and/or fiscal policies is considered sufficient to increase labour demand and thereby bring unemployment down, provided that we accept the higher inflation that goes with it.

In the late 1960s, however, Milton Friedman claimed on the basis of empirical evidence concerning the USA that “inflation is always and everywhere a monetary phenomenon”, i.e. that inflation rates are proportional to the growth of the money stock.

Furthermore, in the early 1970s, economists like Robert Lucas argued that only unanticipated inflation developments would create temporary deviations from equilibrium, given that rational economic agents take all available information into account when setting prices and wages. According to this school of thought, output and unemployment are essentially independent of inflation and expansionary fiscal policy is useless since economic agents will respond to it by reduced consumption in anticipation of higher future taxation.

It was also concluded as part of this critique that in the long term, the Phillips curve becomes vertical: the economy reaches a so-called natural rate of unemployment which depends only on structural factors. Accordingly, policy-makers’ focus shifted to structural reforms which would bring down structural unemployment.

Figure 1. The textbook Phillips Curve and its outward shift due to the oil price shocks and resulting stagflation of the 1970s

Source: (

The critique of the Phillips curve was warranted by developments in the 1970s and 80s, namely the supply-side shocks of the oil crises, which understandably drove up inflation (including wages), without reducing unemployment.

But let us have a look at what actually happened since then.

The Phillips curve has become more horizontal, not vertical

We can start with the example of France because it is so close to the European average in many ways.

Figure 2. The (original) Phillips curve for France, 1970-2013

Source: European Commission, DG EMPL.

Throughout most of the period from 1970 to 1990 in France, increases in unemployment were associated with falls in nominal wage growth, just like the traditional Phillips curve would suggest.

  • The years after the 1973 and 1979 energy crises, when both unemployment and wage inflation increased, were notable exceptions.
  • Starting in the mid-1990s and especially more recently, the relationship between unemployment and wage inflation has flattened out.
  • In other words, while unemployment rose further, there was only a slight decrease in wage inflation.

In the case of Spain we can observe a similar but more clear-cut pattern.

  • The relationship between unemployment and wage growth since has basically flattened out since the onset of the crisis in 2008.
  • Wage inflation has been low and unemployment grew.

Figure 3. The (original) Phillips curve for Spain, 1970 - 2013

Source: European Commission, DG EMPL.

Developments in Germany since unification also suggest that the Phillips curve holds. Decreases in unemployment in Germany in recent years have been associated with (slightly) rising wage levels.

Figure 4. The (original) Phillips curve for Germany, 1992 - 2013

Source: European Commission, DG EMPL.

All three cases display an inverse relationship between unemployment and wage growth, like the original Phillips curve. However, we nowadays observe relatively major differences in unemployment rates, while inflation remains rather low and stable.

  • What all three cases show is that the Phillips curve has not become vertical as the monetarists had predicted; it is much closer to being horizontal in recent years.

Why has the Phillips curve flattened?

One obvious explanation is that inflation in Europe has simply been very low in the last 20 years. This can be associated with the establishment of the euro where the ECB has price stability (not full employment) as its primary objective and where convergence towards low inflation rates represents one of the key accession criteria.

Adjustment to economic shocks in the Eurozone tends to occur not through expansionary fiscal or monetary policies that would drive up inflation and reduce unemployment, but through internal devaluation which leads to low inflation or outright deflation, accompanied by high levels of unemployment.

Moreover, wage-inflationary pressures are much less likely nowadays than in the 1970s or 1980s because labour-union density has decreased considerably and collective bargaining has become more decentralised and easy to opt out from.

In many countries, such as Spain or Germany, the nominal unit labour cost has also been rising less than overall prices, due to deliberate wage constraint or inefficient product markets. The result has been a further compression in aggregate demand and a more pronounced impact on unemployment.

Figure 5. Unit labour cost and GDP price deflator (net of indirect taxes and subsidies) in France, Spain and Germany since 2000

Source: European Commission, DG EMPL.

What should we learn from the Phillips curve?

In short, empirical evidence suggests that the Phillips curve continues to be relevant, only it has been perhaps neglected in macroeconomic policy. We rightly pursue structural reforms aiming at reducing the long-term structural rate of unemployment, such as the Youth Guarantee3 or investments in education. But we may have forgotten that the way to the long term leads via several episodes of the short term.

Since 2010, we have allowed aggregate demand in Europe to drop and unemployment to shoot up. Today we see that this has negative impact on medium-term growth prospects.

  • A protracted period of low or negative growth causes hysteresis (decline in human and fixed capital), in turn undermining the growth potential.
  • Unemployment caused by a cyclical downturn becomes structural in its consequences if the downturn is not tackled.


In conclusion, assuming that we take the EU’s objective of full employment seriously even with low projected GDP growth, we need to embrace negative real interest rates.

Nominal interest rates should be kept at very low levels and inflation expectations need to rise.

Economic agents (and especially the savers among us) need to accept the idea that very low or even negative returns on their capital are necessary for the sake of the whole economy’s growth. Conditions for borrowing need to become more favourable, especially as regards young people and SMEs.

In practice, a negative real interest rate means some form of redistribution from savers to debtors. It can take many forms, such as debt relief, ‘helicopter money’ or simply increased taxation of capital gains or wealth and greater fiscal support to lower-income groups. Greater redistribution would benefit the whole economy at the current juncture as it would strengthen aggregate demand, including demand for investment.

In any event, higher inflation expectations and lower unemployment cannot be achieved on the basis of only fiscal or only monetary policies, nor only through structural changes like wage increases in surplus countries. Concerted action through fiscal, monetary and structural policies is needed, as in the concept of “Abenomics” and as recently argued also by the President of the European Central Bank.[4]


Andor, László (2013). "Can we move beyond the Maastricht orthodoxy?", VoxEU column, 16 December 2013.

Council of the European Union (2013), "A recommendation on establishing a Youth Guarantee".

Teulings, C and R Baldwin (2014), Secular Stagnation: Facts, Causes and Cures, VoxEU e-book, 15 August 2014, .


1 The conference papers are available at

2 This article elaborates on part of a speech presented at the EU conference on "Labour Economics after the Crisis" under the title "Towards a European Labour Model" (18 September 2014,

3 The Youth Guarantee is a structural reform aiming to ensure that all people under the age of 25 receive a good-quality offer of a job, continued education, apprenticeship or traineeship within four months of becoming unemployed or leaving school. A recommendation on establishing a Youth Guarantee was adopted by the Council of the European Union in April 2013; cf.

4 The recent speech of Mario Draghi in Jackson Hole emphasised the need for more accommodative monetary policies and a more expansionary aggregate fiscal stance of the Eurozone, so that Europe can implement structural reforms without risking further short-term contraction in GDP and further deflationary pressure. Draghi, M., "Unemployment in the euro area", speech at the Annual central bank symposium in Jackson Hole, United States, 22 August 2014,



Topics:  Labour markets

EU Commissioner for Employment, Social Affairs and Inclusion