Accompanying the economic recovery

Introductory speech by Mario Draghi, President of the ECB, at the ECB Forum on Central Banking, Sintra, 27 June 2017

For many years after the financial crisis, economic performance was lacklustre across advanced economies. Now, the global recovery is firming and broadening. A key issue facing policymakers is ensuring that this nascent growth becomes sustainable.

Dynamic investment that drives stronger productivity growth is crucial for that – and in turn for the eventual normalisation of monetary policy.

Investment and productivity growth together can unleash a virtuous circle, so that strong growth becomes durable and self-sustaining and, ultimately, is no longer dependent on a sizeable monetary policy stimulus.

The discussions we will have over the next two days – in particular understanding the puzzles of slowing productivity growth and persistently low investment – are therefore highly relevant for the path of the economy and of our monetary policy.

Yet as we anticipate the problems of tomorrow, we must also work on the issues of today. For central banks, this means addressing an unusual situation. We see growth above trend and well distributed across the euro area, but inflation dynamics remain more muted than one would expect on the basis of output gap estimates and historical patterns.

An accurate diagnosis of this apparent contradiction is crucial to delivering the appropriate policy response. And the diagnosis, by and large, is this: monetary policy is working to build up reflationary pressures, but this process is being slowed by a combination of external price shocks, more slack in the labour market and a changing relationship between slack and inflation. The past period of low inflation is also perpetuating these dynamics.

These effects, however, are on the whole temporary and should not cause inflation to deviate from its trend over the medium term, so long as monetary policy continues to maintain the solid anchoring of inflation expectations.

Hence we can be confident that our policy is working and its full effects on inflation will gradually materialise. But for that, our policy needs to be persistent, and we need to be prudent in how we adjust its parameters to improving economic conditions.

Monetary policy is effective in raising demand

Understanding inflation dynamics requires us to divide up the inflation process into two legs: the effect of monetary policy on aggregate demand; and the effect of aggregate demand on inflation. All the evidence suggests that the first leg is working well.

Though the euro area recovery started later than those in other advanced economies, we have now enjoyed 16 straight quarters of growth, with the dispersion of GDP and employment growth rates among countries falling to record low levels. If one looks at the percentage of all sectors in all euro area countries that currently have positive growth, the figure stood at 84% in the first quarter of 2017, well above its historical average of 74%. Around 6.4 million jobs have been created in the euro area since the recovery began.

The role of monetary policy in this growth story is clear. Since mid-2014, our monetary policy stance has been determined by the combination of three instruments: first, low policy interest rates; second, asset purchases in financial markets and targeted long-term refinancing operations for banks; reinforced by, third, forward guidance on both.

This has created strong downward pressure on financing costs, with rates falling steeply across asset classes, maturities and countries, as well as across different categories of borrowers. Converging financing conditions have in turn fed into rising domestic demand.

According to our Bank Lending Survey, our latest easing phase has coincided with a strong rebound in demand for consumer credit to purchase durable goods, while demand for loans for fixed investment has gradually firmed. At the same time, falling borrowing costs have reduced interest payment burdens and facilitated deleveraging, which is one reason why, for virtually the first time since 1999, spending has been rising while indebtedness has been falling. This is a sign that the recovery may be becoming more sustainable.

Just to put our measures into context, since January 2015 – that is, following the announcement of the expanded asset purchase programme (APP) – GDP has grown by 3.6% in the euro area. That is a higher growth rate than in same period following QE1 or QE2 in the United States, and a percentage point lower than the period after QE3. Employment in the euro area has also risen by more than four million since we announced the expanded APP, comparable with both QE2 and QE3 in the US, and considerably higher than QE1.

For the monetary transmission process to work, however, stronger growth and employment ought to translate into higher capacity utilisation, scarcity in production factors and – in time – upward pressure on wages and prices. And this is what we see.

The unemployment gap – the difference between actual unemployment and the non-accelerating inflation rate of unemployment (NAIRU) – is narrowing and is forecast by the Commission to close within the next two years. Surveys of business capacity utilisation are now above their long-term average levels. And inflation is recovering. Between 2016 and 2019 we estimate that our monetary policy will have lifted inflation by 1.7 percentage points, cumulatively.

Yet the second leg of the inflation process – the transmission from rising demand to rising inflation – has been more subdued than in the past. How can this be explained?

The link between output and inflation is determined by three main factors: external shocks to prices; the size of the output gap and its impact on inflation; and the extent to which current inflation feeds into price and wage setting. In different ways, each of these factors has been relevant for the delayed reaction of inflation to the recovery.

Temporary external shocks

Starting with external factors, one explanation for the slow improvement in inflation dynamics is that we are still suffering the after-effects of price shocks in global energy and commodity markets, which have led output and inflation to move in different directions.

Inflation has indeed been subject to such shocks over recent years, most notably the oil and commodity price collapse in 2014-15. This not only depressed the cost of imported energy, but also lowered global producer prices more generally. ECB analysis suggests that the global component in the underperformance of euro area inflation in recent years was considerable. In 2015-16, around two thirds of the deviation of euro area headline inflation from a model-based mean can be accounted for by global shocks linked to oil prices.

Even though the downward pressure on inflation from past oil price falls is now waning, oil and commodity prices are still having a dragging effect – if only because they continue to lack a clear upward trend. In fact, lower oil and food prices than those assumed in the March 2017 projections are an important factor behind the downward revision of our latest inflation forecasts.

Oil-related base effects are also the main driver of the considerable volatility in headline inflation that we have seen, and will be seeing, in the euro area.

Falling import prices partly explain the subdued performance of core inflation, too. This is because imported consumer products account for around 15% of industrial goods in the euro area. Likewise, changes in commodity prices feed through into some services items and into industrial goods produced with high energy intensity. As a result, in the first quarter of 2017, oil-sensitive items were still holding back core inflation. This illustrates that core inflation does not always give us a clear reading of underlying inflation dynamics.

Global factors therefore do appear to be weighing on the path of inflation today. But our analysis suggests that the drivers of low oil prices at present are mainly supply factors, which a central bank can typically look through.[1] And even if supply factors affect the path of inflation for some time, with inflation expectations secure, they should not ultimately affect the inflation trend.

Uncertainty over slack and its impact on inflation

A second explanation for the discrepancy between real developments and inflation is uncertainty surrounding the size of the output gap and its impact on inflation. This might be happening for a variety of reasons.

One possible reason is that we are currently experiencing positive supply developments. In particular, we do observe that, as the recovery strengthens, the supply of labour is rising too. Labour force participation has been growing consistently over the last few years, buoyed especially by increases in participation rates of older workers. We also see some evidence that labour supply has become more elastic due to immigration, particularly in strongly growing economies such as Germany.[2] Since 2007, the euro area participation rate has risen by around 1.5 percentage points, whereas in the United States it has fallen by more than 3 percentage points in the same period.[3]

Structural reforms in labour markets have been a factor in this labour supply boost.

Similarly, past reforms in product markets may also have had a positive effect on the supply side by reducing price mark-ups, increasing productivity and raising growth potential.

Another reason why there is some uncertainty over slack is the correct notion of unemployment – that is, there may be residual slack in the labour market that is not being fully captured in the headline unemployment measures. Unemployment in the euro area has risen during the crisis, but so too has the number of workers who are underemployed (meaning that they would like to work more hours) or who have temporary jobs and want permanent ones. This has implications for inflation dynamics, since these people might prioritise more hours or job security over higher wages in employment negotiations.

If one uses a broader measure of labour market slack including the unemployed, underemployed and those marginally attached to the labour force – the so-called “U6” – that measure currently covers 18% of the euro area labour force.[4] Phillips curve models that employ this measure appear to be more successful in predicting inflation.[5]

A third reason why slack might be larger is the effect of so-called “global slack”. This is the notion that globalisation has made labour supply characteristics more uniform across the globe and labour markets more contestable. Conditions in foreign labour markets could therefore have a dampening effect on domestic inflation even as domestic slack is shrinking.

The evidence, however, is not clear-cut. For example, new ECB analysis finds only limited support for the thesis that global slack is weighing on euro area inflation today, over and above the impact it has on global prices.[6]

Alongside the question of the level of slack is the impact of slack on inflation. This is the well-known debate on the slope of the Phillips curve. There are indeed reasons to believe that wage and price setting behaviour in the euro area might have changed during the crisis in ways that slow the responsiveness of inflation.

For example, structural reforms that have increased firm-level wage bargaining may have made wages more flexible downwards but not necessarily upwards.

Likewise, we see today that firms are absorbing input costs through lower margins due to uncertainty over future demand, which would also tend to temper price pressures. Indeed, ECB estimates show that, if we take into account the unusually large and persistent shocks of the past years, the Phillips curve for core inflation may well be somewhat flatter recently. However, insofar as the slope of the Phillips curve depends nonlinearly on the cyclical position, it may steepen again when the economy reaches and surpasses full potential.

While these various reasons might delay the transmission of our monetary policy to prices, they will not prevent it. As the business cycle matures, the higher demand resulting from positive supply developments will accelerate price pressures, while firms’ pricing power will increase and the broader measures of slack will converge towards the headline measures.

As shown in the United States, the gap between the headline unemployment rate and those broader measures typically opens in recessions and shrinks in expansions. Currently, it is converging to the minimum levels recorded before the 2001 and 2007 recessions.

So just as for oil and commodity price shocks, we can be reasonably confident that the forces we see weighing on inflation are temporary – so long as they do not feed more lastingly into inflation dynamics.

Low inflation feeding into price and wage setting

This brings us to the third possible explanation for why growth might be diverging from inflation: the hypothesis that a persistent period of low inflation is in fact feeding into price and wage setting in a more persistent way.

What is clear is that our monetary policy measures have been successful in avoiding a deflationary spiral and securing the anchoring of inflation expectations. In the past, as interest rates approached zero, some did question our ability to add sufficient accommodation to combat a prolonged period of too-low inflation. We answered those doubts by demonstrating that we would take any measures necessary within our mandate to deliver our mandate, and that those measures were effective in further easing financial conditions. Deflation risk premia, which had been growing in 2014 and 2015, have now been more or less priced out of market-based inflation expectations.

That being said, a prolonged period of low inflation is always likely to be exacerbated by backward-lookingness in wage and price formation that occurs due to institutional factors, such as wage indexation.

This has plainly happened in the euro area. ECB analysis finds that, compared with long-term averages, low past inflation dragged down wage growth by around 0.25 percentage points each year between 2014 and 2016.

The evidence as to whether backward-lookingness has increased recently is mixed. There were signs that indexation had fallen in the early part of the crisis, and ECB empirical estimates suggest that the weight of past inflation in current inflation has decreased. Yet there is also evidence that indexation has returned in some large euro area countries. In Italy, for example, backward-looking indexation of wages now covers around one-third of private sector employees.[7]

Still, even if indexation rose, it would only create inertia in price formation: it would not obstruct the transmission process. As economic slack shrinks, upward pressure on prices will materialise and gradually enter the indexation ratchet.

So once again we see temporary forces at work that should not affect medium-term price stability. And this assessment is broadly what we see in market-based inflation expectations today. Interpreting with some caution, they are now consistent with the picture that our policy is effective, but that its full effects will take time to materialise.

Accompanying the recovery

So what do these various explanations imply for our monetary policy stance?

The first point to make is that we face a very different situation today from the one we encountered three years ago. Then, we also faced global shocks and significant labour market slack. But the recovery was still in its infancy. Global growth was slowing, depressing both import prices and net exports. Fiscal policy was less supportive than it is now. And headline inflation was much lower than today and inflation expectations more fragile, creating a higher risk of low inflation becoming entrenched.

In this context we faced another risk, too: of permanent damage to the economy through so-called “hysteresis effects”. Given the large degree of slack at the time, the risk was that if output remained below potential for too long, we would see a permanent destruction of productive capacity. The output gap would close the “wrong way” making the losses permanent.

When we said we wanted our policy to have effects without undue delay, we meant we wanted the output gap to close “upwards” – and before such hysteresis effects could develop. This is why we had to act forcefully.[8]

Now, we can be confident that our policy is working and that those risks have abated. The threat of deflation is gone and reflationary forces are at play. And since one of the drivers of inflation today is positive supply developments, this should feed back positively into potential output rather than produce hysteresis. In these conditions, we can be more assured about the return of inflation to our objective than we were a few years ago.

This more favourable balance of risks has been already reflected in our monetary policy stance, via the adjustments we have made to our forward guidance.

Another considerable change from three years ago is the clarification of the political outlook in the euro area. For years, the euro area has lived under a cloud of uncertainty about whether the necessary reforms would be implemented at both the domestic and Union levels. This acted as a brake on confidence and investment, which is tantamount to an implicit tightening of economic conditions. Today, things have changed. Political winds are becoming tailwinds. There is newfound confidence in the reform process, and newfound support for European cohesion, which could help unleash pent-up demand and investment.

Nevertheless, we are still in a situation of continuing slack, and where a long period of subpar inflation translates into a slower return of inflation to our objective. Inflation dynamics are not yet durable and self-sustaining. So our monetary policy needs to be persistent.

This is why the Governing Council has repeatedly emphasised that a very substantial degree of monetary accommodation is still needed for underlying inflation pressures to build up, and to support headline inflation in the medium term. This is reflected in our forward guidance on net asset purchases and interest rates, as well as our decision to reinvest the principal payments received under the APP for as long as necessary.

With reflationary dynamics slowly taking hold, we now need to ensure that overall financing conditions continue to support that reflationary process, until they are more durable and self-sustaining.

As the economy continues to recover, a constant policy stance will become more accommodative, and the central bank can accompany the recovery by adjusting the parameters of its policy instruments – not in order to tighten the policy stance, but to keep it broadly unchanged.

But there is an important caveat that we need to consider. Financing conditions are not only determined by the calibration of central bank instruments, but also by other market prices, some of which are significantly affected by global developments.

In the past, especially in times of global uncertainty, volatility in financial market prices has at times caused an unwarranted tightening of financial conditions, which has necessitated a monetary policy response.

So in the current context where global uncertainties remain elevated, there are strong grounds for prudence in the adjustment of monetary policy parameters, even when accompanying the recovery. Any adjustments to our stance have to be made gradually, and only when the improving dynamics that justify them appear sufficiently secure.


Let me conclude. Our assessment of the outlook for inflation and for monetary policy can be summed up in three messages.

The first is confidence that monetary policy is effective and the transmission process will work. All the signs now point to a strengthening and broadening recovery in the euro area. Deflationary forces have been replaced by reflationary ones.

While there are still factors that are weighing on the path of inflation, at present they are mainly temporary factors that typically the central bank can look through. However, a considerable degree of monetary accommodation is still needed for inflation dynamics to become durable and self-sustaining. So for us to be assured about the return of inflation to our objective, we need persistence in our monetary policy.

And, finally, we need prudence. As the economy picks up we will need to be gradual when adjusting our policy parameters, so as to ensure that our stimulus accompanies the recovery amid the lingering uncertainties.

[1] For a fuller account of the response of monetary policy to demand and supply shocks in the oil market see Draghi M. (2015), “Global and Domestic Inflation”, speech at Economic Club of New York, 4 December.

[2] Deutsche Bundesbank (2016), “The Phillips curve as an instrument for analysing prices and forecasting inflation in Germany”, Monthly Report, April.

[3] Source: AMECO database (workers aged 15-74 for the euro area and 16-74 for the US).

[4] ECB (2017), “Assessing labour market slack”, Box 3, Economic Bulletin, Issue 3.

[5] See Cœuré, B. (2017), “Scars or scratches? Hysteresis in the euro area”, speech at the International Center for Monetary and Banking Studies, Geneva, 19 May.

[6] ECB (2017), “Domestic and global drivers of inflation in the euro area”, Economic Bulletin, Issue 4.

[7] Banca d’Italia (2017), box on “Private sector contract renewals in 2016”, Economic Bulletin, No. 1, January.

[8] For a fuller explanation of this point see Draghi, M. (2016), “On the importance of policy alignment to fulfil our economic potential”, 5th Annual Tommaso Padoa-Schioppa Lecture at the Brussels Economic Forum 2016, Brussels, 9 June.

ECB Forum on Central Banking

Media contacts