Given the “gap” between inflation projections and our policy objective, we decided to launch an expanded asset purchase programme. We had to act because the inflation outlook had deteriorated, while our previous measures had not worked as effectively as we expected. In this environment our decision was fully consistent with our reaction function, which we had communicated last year, and it was only natural for markets to anticipate our move. Indeed faced with an increasing risk of a too prolonged period of low inflation, we intensified our stimulus to a degree necessary to bring inflation closer to 2%. We did it in line with our mandate and price stability objective. And I have no doubt that the measures we took will bring the intended effects.
Ladies and gentlemen ,
Thank you very much to the Financial Times for inviting me to open this conference today.
As you know, the ECB Governing Council last month decided to launch an expanded asset purchase programme to bring inflation back to levels closer to our aim of below but close to 2% over the medium term. What I would like to talk to you about this morning is the context in which we took that decision and how we are calibrating our policy response.
The context for our recent decisions
A stylised way of thinking about why we introduced our recent measures is to use the concept of the “inflation gap”. This gap is the difference between our aim of an inflation rate below but close to 2% over the medium-term and our projection for inflation over that same horizon, based on our current monetary policy stance and the information on the state and evolution of the economy that we have access to today. If the inflation gap widens, and the monetary policy stance remains unchanged, it follows that there is a “missing stimulus” to which we may need to respond.
At its meeting on 22 January, the Governing Council assessed that two factors had come together which had led to a widening of this “inflation gap” and had hence rendered our stimulus insufficient.
The first was the fact that the credit easing package we decided in June-October 2014 had delivered less stimulus and thus less support to inflation than initially expected. And the second was the fact that, in parallel and independent of the monetary policy stance, the inflation outlook for the euro area had further deteriorated towards the end of the year. This contributed to a further widening the inflation gap that our measures needed to close.
Let me explain each point in some more detail.
The initial impact of credit easing
For some time inflation dynamics in the euro area have been surprising on the downside, and by June last year our inflation projections were suggesting that the risk of us not achieving our mandate over the medium term was increasing. At that time we foresaw annual HICP inflation at 0.7% in 2014, 1.1% in 2015 and 1.4% in 2016.
While our policy stance was undeniably very accommodative, our assessment was that it was not being transmitted to credit conditions faced by the real economy in large parts of the euro area, and that this in turn was hindering the economic recovery in the euro area as a whole. What we observed was a feedback loop between weak macroeconomic developments and weak demand for and supply of credit. The effect was downward pressure on our medium-term inflation outlook.
Our response was the credit easing package we introduced between June and October last year. This package was designed to achieve two types of transmission effects. The first was a qualitative effect – to strengthen the pass-through per euro of liquidity injected to private sector borrowing costs. The second was a quantitative effect – to generate volumes of liquidity sufficient to activate portfolio rebalancing channels.
While the effects of such a package of course take time to materialise, our view was that by the beginning of this year we should be in a position to make an initial assessment of whether it was working as intended. And when the Governing Council evaluated the effectiveness of the credit easing measures, our conclusion was that the pass-through effects of the package were satisfactory, but the quantity side was below expectations.
In terms of pass-through, our measures now seem to be working their way through markets and the economy. Bank lending rates to non-financial corporations (NFCs) have been trending downwards since the third quarter of last year. By the end of last year, the composite cost of borrowing indicators in Italy and Spain had declined by approximately 70 basis points and 50 basis points, respectively, when compared to the middle of the year. This should be seen in a context where lending rates in these countries had not been very responsive to monetary policy actions for at least three years.
These developments coincided with the first targeted longer-term refinancing operation (TLTRO) and the announcement of the asset-backed securities purchase programme (ABSPP) and covered bond purchase programme (CBPP3). I think it is therefore fair to say that these measures have had a positive effect on compressing spreads.
In addition, on average net redemptions of loans to NFCs have moderated, with net lending turning slightly positive in November. Results from the January 2015 Bank Lending Survey have also been encouraging and indicate a further net easing of credit standards across all loan categories in the final quarter of last year. And we are seeing that banks that participated in last year’s two TLTROs have been reducing loans by less than in the previous year, and by less than non-participating banks, which have not altered their lending behaviour to the same extent.
On the quantity side, the less favourable assessment is a reflection of the lower than expected outturns for the TLTRO and a slower contribution from the outright purchases. This has meant weaker portfolio rebalancing effects than we foresaw in the original calibration of the package. For example, TLTRO borrowings enable banks to pay down market debt, thus passing on cash to bank bond investors and encouraging them to rebalance their portfolios and find alternative investment opportunities. A lower uptake in the TLTRO therefore naturally weakens the strength of these effects.
The inference was thus clear: while our measures had been effective in certain ways, relative to what we envisaged when we adopted our June-October package, they had not been sufficient. They had not provided the degree of policy support to the inflation trajectory that we saw as necessary to close the inflation gap.
The deteriorating inflation situation
In parallel, the inflation gap had widened as the inflation outlook had materially worsened towards the end of last year. When we took the decision in January most indicators of realised and expected inflation in the euro area stood at, or close to, their historical lows.
While the main driver of inflation developments in the months before was the sharp fall of global oil prices, the conditions we faced made this shock different from a classic positive supply shock, and we could not afford to “look through” the effects on inflation. Indeed, several factors suggested that the present situation was more worrying than past episodes of oil-induced disinflation, particularly the most recent case in 2009 following the collapse of Lehman brothers.
First, the persistence of low inflation across a range of statistical metrics was higher than in 2009. This reflects the fact that, at that time, disinflation was very abrupt with inflation rates expected to quickly revert back into positive territory. The current episode is characterised by a prolonged downward inflation trend, and we expect inflation to remain in negative territory until latter stages of this year.
Second, since last summer market measures of inflation expectations have become more sensitive to realised inflation, even at maturities and at horizons that we would normally expect to be more resilient. This is in stark contrast with 2009, where inflation expectations hardly moved, even over the shorter end of the inflation forward curve.
Finally, measures of core inflation are less sticky today compared to 2009, which implies a higher risk of second round effects materialising from low realised inflation and falling inflation expectations. This in turn creates a higher likelihood that, absent policy action, low inflation will become entrenched and will not bounce back when oil prices normalise.
The importance of fulfilling the mandate
In sum, what we saw was a real risk of not achieving our aim of an inflation rate of below but close to 2% over the medium-term. And let me stress that there is no room for complacency when inflation stays too low for too long. There are real costs to falling short of our objective. Those costs are both institutional and economic.
First, in institutional terms, it is important to remember the very unique environment in which we operate in the euro area. It involves multiple actors in multiple countries, and optimal economic outcomes can only be achieved if each of those actors follows the rules in its domain, thereby ensuring a consistent set of policies at the aggregate level. This is why, for example, we constantly stress the importance of countries respecting the Stability and Growth Pact rules.
However, if we were to miss our mandate as soon as it became difficult to achieve, how would it affect the incentives of others in the euro area to fulfil their own mandates? Could we still expect others to stick to their commitments – such as implementing structural reforms and fiscal consolidation – in difficult times? And perhaps even more importantly, would our credibility be preserved if we are given a narrow mandate but cannot deliver it?
Second, in economic terms, a too prolonged period of low inflation can negatively affect both consumption and investment. Certainly, an oil price shock improves the terms of trade, increases households’ real disposable income and reduces the input costs for firms, thus boosting aggregate demand. But if inflation expectations become de-anchored and second round effects set in, then the sign of the effect can change. What is initially a positive supply shock can morph into a negative demand shock.
This is first of all because at the zero lower bound falling inflation implies rising real interest rates, causing households to postpone consumption and firms to delay investment, which can then feedback into still lower inflation. In fact, when we took our decision expected real rates had risen by almost half a percentage point in the previous few months alone.
These effects are then amplified in the presence of high levels of debt – as we see in the euro area today. The reason is that if second round effects take hold, the debt burden in real terms increases. The result is that a higher proportion of firms’ and households’ income has to be diverted to debt service, eroding the positive term-of-trade gain arising from lower oil prices. And this may have negative implications for aggregate demand due to the different propensities to consume and invest of borrowers and lenders.
What we are concerned about here is of course not the extreme type of debt-deflation scenario that Irving Fisher had in mind when studying the experience of the 1930s. We are rather concerned about attrition: that the difficult process of recovering from the crisis and working through the debt overhang in the euro area becomes even more complicated and protracted.
Finally, it is important to remember that a too prolonged period of low inflation is not only a cause of demand weakness, but also a symptom of it – as is evidenced in the decline in core measures of inflation. The latter reflects underutilised resources, which if persistent can produce hysteresis, scarring effects on the unemployed and underinvestment. This in turn can permanently lower potential growth, which is also inimical to working through a debt overhang.
Calibrating our response
In this context, with the risks of a too prolonged period of low inflation materialising, the Governing Council saw a clear case for scaling up our monetary stimulus sufficient to respond to it. In fact, nearly all Governing Council members agreed that in the situation we faced, if there were still a margin for further interest rate cuts, we would have used it.
The question was however more complex: how to act when rates reach the lower bound in a monetary union like the euro area. At this point, is moving from interest rates to asset purchases as the main instrument of monetary policy simply a linear progression? Or does the specific institutional makeup of the euro area mean that there are non-linearities when the lower bound is reached, which in turn make large scale asset purchases of private and sovereign paper an exceptional tool?
My own view is that when the central bank is in danger of missing its mandate, and it has a tool at its disposal that is both legal and effective, it has to use it. Indeed, the Governing Council is unanimous that the asset purchase programme is a monetary policy tool in a legal sense. Yet at the same time, it is clear that the design of an asset purchase programme including sovereign bonds has to reflect the specificities of the euro area institutional framework with a single monetary policy and many fiscal authorities. And the Governing Council decision reflects these considerations.
We have taken a significant monetary policy step, showing that we have the capacity and determination to deliver our price stability mandate, and that we act in full independence. But we have also acknowledged, through the loss sharing arrangement, the reality that the euro area is not a fiscal union and it is for politicians, not central bankers, to decide if and when it will become one.
Importantly, the expanded asset purchase programme was also in line with the communication on our reaction function that we introduced last year to help observers understand our strategy in an environment of heightened uncertainty. In a speech in April 2014 President Draghi laid out three contingencies that would warrant further policy action by the ECB. And as each of those contingencies materialised, we responded as we indicated we would.
In particular, we communicated explicitly that should we enter our “third contingency” – a worsening of the medium-term outlook for inflation and/or a loosening in the anchoring of inflation expectations – we would deploy a broad-based asset purchase programme. It is therefore no surprise that market participants, seeing much the same economic and financial data as we do, have largely anticipated this move.
It is important to keep this in mind as it helps clarify some misconceptions about what drives the Governing Council’s decisions. Long-term nominal interest rates have declined by around 100 bps since the middle of last year precisely because we specified our reaction function and markets understood it. Therefore, the argument that asset purchases are unnecessary because market interest rates have already come down, or that the ECB is “following” markets, misses the direction of causality.
The transmission of asset purchases to the real economy
How does the asset purchase programme add extra monetary stimulus to the euro area economy? It does so along five dimensions.
The first is the overall size of our interventions, which will be substantial. The second is the pace, which will accelerate to €60 billion per month. The third is the composition of purchases, which will include securities issued by euro area sovereigns and EU institutions. The fourth is the maturity, which extends from 2 to 30 years. And the fifth dimension is the horizon of our new measures, which will last until end-September 2016 or until we see a sustained adjustment in the path of inflation consistent with our aim of achieving inflation rates close to 2% over the medium term.
This combination of features will provide for full control over the scale and intensity of the monetary stimulus. And while it will work through several transmission channels, we expect it to be particularly powerful in terms of portfolio rebalancing effects. Put simply, the displacement of investors’ portfolios will be larger, faster and take place across more market segments. This should also help reduce market segmentation arising from excess risk aversion, improve confidence and lead to a generalised easing of financing conditions for euro area firms and households.
While such effects have been extensively studied in other jurisdictions where large-scale asset purchases were introduced, the way in which they work may differ in the euro area compared with other regions given our specific financial structure – namely the prevalence of bank-based over market-based intermediation. Indeed, I expect that an important transmission channel for our new measures will be via banks.
Banks may tend to follow a pecking order strategy in their response to large-scale asset purchases. In the first instance, those that sell their securities to the central bank, and thus have an excess reserve position, may react by paying down market debt. This would create scarcity in bond markets, thus inducing a portfolio rebalancing effect and leading in turn to a compression of yields.
When this liability management reaches its limit, we would then expect that those more favourable market funding conditions to translate into lower lending rates for borrowers in the real economy – which is a key aim of our measures. This should be aided by the effect of portfolio rebalancing on the returns on securities and its implication for banks’ decisions to grant loans.
Specifically, as expected returns on securities will be compressed, maintaining net interest income will require banks to shift their portfolios away from securities and towards loans to firms and households. Indeed, one reason that credit growth has been weak, especially in peripheral countries, is that the risk adjusted returns for banks on loans have been very subdued relative to returns on securities portfolios – especially domestic government bonds.
Moreover, as credit conditions further ease and the macroeconomic outlook improves, there should be an endogenous effect on borrowing conditions. Insofar as banks expect lower delinquencies thanks to improving macroeconomic conditions, they should require lower credit risk premia from firms and households, implying lower lending rates. And insofar as asset prices are boosted by our measures, collateral values will rise in tandem, implying higher lending volumes.
Finally, we expect all these effects to be reinforced by the way we have calibrated our programme, in particular the open-ended commitment to return inflation back close to 2%. This strengthens the signal emanating from our decisions, which is visible in price movements at the longer end of the yield curve. And it should also support the firm anchoring of inflation expectations, which is crucial to avoid the self-reinforcing dynamics associated with a too long period of low inflation that I mentioned above.
The pivotal role of governments
Still, it is clear that if our measures are to generate a faster and more sustained recovery then governments also have to play their part. In particular, at both the micro and macro levels, structural reforms are pivotal to encourage firms to capitalise on the easing of financial conditions and undertake new investment projects.
At the micro level, any investment project will happen if the return on investment sufficiently exceeds the cost of capital. Monetary policy can increase the wedge between the two on the downside, by lowering the cost of capital. But structural reforms can increase the wedge on the upside, by raising returns – for example by reducing costs arising from unnecessary red tape.
At the macro level, monetary policy can influence investment through lowering the real interest rate. But there also has to be a fundamental reason to invest, which comes down chiefly to developments in labour participation and total factor productivity. Here again structural reforms can complement monetary policy by, for instance, reactivating structurally unemployed workers and improving allocative efficiency across firms, thus raising potential growth.
Indeed, perhaps the biggest risk we face with our new measures is that they fall on barren ground because governments are not doing enough to raise confidence in the future. If firms are simultaneously revising down their growth expectations as we expand our monetary policy, we will not see the impact we expect and which is required.
Yet the opposite story is also possible. If governments are determined in addressing their structural challenges, all the conditions are there for growth to surprise on the upside. We are already seeing, in countries like Spain, how improved financing conditions together with targeted structural reforms can lead to stronger-than-expected growth outcomes. And if we take advantage of the window of opportunity that exists today, that possibility is there for the whole euro area.
Let me conclude.
The success of the euro area depends on every party doing its job, including the central bank. For us that means being fully symmetric in how we interpret our mandate: acting as decisively when the inflation outlook undershoots our aim as we would when it exceeds it. This is our responsibility in the Treaty and our duty to the firms and households of the euro area.
Remember that firms and households have entered contracts and entertain economic relationships based on the expectation that we will meet our inflation objective. If we fail, then we make their recovery from the crisis harder. And that loss of credibility would in turn make it harder for us to deliver our mandate in the future.
That is why, faced with an increasing risk of a too prolonged period of low inflation, we increased our stimulus to a degree necessary to bring inflation closer to 2%. And we will continue to monitor the progress towards a sustained adjustment in the path of inflation.
I would like to thank John Hutchinson for his contribution to this speech.