Delivering a symmetric mandate with asymmetric tools: monetary policy in a context of low interest rates

Speech by Mario Draghi, President of the ECB, at the ceremony to mark the 200th anniversary of the Oesterreichische Nationalbank, Vienna, 2 June 2016

Exactly 200 years ago today, Emperor Francis I issued two imperial decrees assigning the Oesterreichische Nationalbank the exclusive right to issue banknotes and to stabilise the finances of the empire. This was part of a growing realisation, across Europe, of the key role that central banks could play in securing monetary stability.

In 1816, monetary stability in Austria meant bringing the currency under control in the wake of the ruinous Napoleonic wars.

In the two centuries since then, the cost of not delivering stability has been painfully displayed: in the devastating effects of excessive inflation, such as the 1400% annual increase in prices recorded here in the early 1920s; but also in the terrible consequences of deflation, as in the 1930s.

This is why most modern central banks have price stability mandates. And it is why legislators have made those mandates symmetric – central banks are expected to fight persistent inflation undershooting as vigorously as they fight persistent overshooting.

In the ECB’s case, our aim is to keep inflation below but close to 2% over the medium term. Today, this means raising inflation back towards 2%. And the series of measures we have adopted in recent years – bringing policy rates into negative territory, engaging in large-scale asset purchases, and providing banks with long-term refinancing on conditional terms – are geared exactly to achieving that.

Why 2% inflation is the right objective for monetary policy

But if our aim is to avoid both excessive inflation and deflation, why do we not set a 0% inflation objective? The answer is that the rate of inflation we aim for and our ability to stabilise prices are intertwined. Or put another way, a steady state inflation rate of 2% inflation is itself a shock absorber which allows us to deliver stability. There are several reasons for this, but let me highlight two in particular.

First, a moderately positive level of inflation facilitates the adjustment of relative prices, which helps prevent short-term shocks from morphing into longer-lasting disturbances. This stems from the fact that, even in the most flexible economies, nominal wages and prices are “sticky” and slow to adjust downwards.

In that context, when demand falls, 2% inflation allows real wages to adjust downwards even if nominal wages do not. That in turn helps keep unemployment lower than would otherwise be the case, and prevents the consequences of the downturn from lasting longer than they need to – for instance, by eroding the human capital of the jobless, leading to permanently higher structural unemployment.

What is true within economies is also true across regions – and this is particularly so in a multi-country monetary union like the euro area. A 2% inflation objective means that less competitive countries can lower costs and prices relative to the area average, which allows them to recover competitiveness without destabilising consequences.

The need for such a buffer was explicitly acknowledged by the Governing Council in 2003 when we clarified our definition of price stability.^[1] It has since helped countries to adjust competitiveness when required. Initially, aiming at 2% allowed some core economies to lower relative prices in a fairly painless way, as other “catching up” economies had higher inflation rates and were pulling up the area average. Today, the positions of those groups have reversed. But the rationale remains exactly the same.

The second reason why a 2% objective helps absorb shocks is that it supports the implementation of monetary policy in adverse conditions. A small positive buffer creates more scope to support the economy through cutting nominal interest rates and reduces the likelihood of running up against the effective lower bound.

This is because, for a given equilibrium real interest rate, a higher inflation objective implies higher nominal rates over the cycle. In line with the results from a larger research literature, ECB simulations made in 2003 suggested that a 2% objective would substantially decrease the probability of nominal rates reaching zero.^[2] This was also recognised by the Governing Council as a reason for aiming closer to 2%.^[3]

What was true then has become even more relevant now. The studies in 2003 assumed an equilibrium real rate of around 2%, so with a 2% inflation objective the equilibrium nominal rate would be around 4%. Evidence suggests, however, that demographics-induced high savings and low productivity growth have led equilibrium real rates to fall.^[4] Aiming for 2% inflation is hence even more crucial today to get nominal interest rates safely away from the lower bound.

Yet it is also important to stress that the fall in real rates is by no means pre-determined. It can be reversed at least in part by structural reforms that raise productivity and participation rates. By increasing the potential for profitable investment opportunities, and by reducing the need for precautionary savings, such measures would raise the equilibrium real rate, all things being equal.

Why the objective should be symmetric

All this justifies the level of our inflation objective in the steady state. But it is equally important that we pursue our objective symmetrically.

Monetary policy operates to a large extent by guiding expectations, a process which sets in motion a series of automatic stabilisers in the economy. For instance, if markets expect central banks to react to negative shocks by increasing monetary accommodation, when a shock actually happens, it will immediately lead to lower real interest rates as the central bank’s reaction function is taken into account. That will in turn result in higher consumption and investment, which helps offset the initial shock.

This mechanism hinges crucially on the central bank being credible on both sides – being just as committed to fighting too low inflation as too high inflation.

And while this is true generally, it is all the more so in the special circumstances we face coming out of the crisis – namely, where private debt stocks are high and policy interest rates are close to the lower bound. In those conditions, any perception that the central bank might tolerate persistent downward inflation misses would be especially costly.

It would lead first of all to a dis-anchoring of inflation expectations, which would cause real yields to rise mechanically. This would be contractionary and could not be offset by lowering policy rates even further. And with fixed nominal debts, lower inflation would trigger redistribution from borrowers to creditors, which would prolong the debt overhang and exacerbate the contraction due to the different propensities to consume and invest of those two groups. This is not an argument for raising inflation targets, as that would only create redistribution in the other direction. But it is an argument for central banks to fulfil their objectives.

However, while our mandate is symmetric, and our commitment to our mandate is symmetric, there is an asymmetry in the tools we can use to achieve it, which stems from the existence of a lower bound for interest rates.

When inflation is too high, we can always raise interest rates to a level that will rein in demand and eventually prices. And as this is widely understood from historical experience, our credibility relies only on one parameter: our willingness to fight excessive inflation. Our ability to do so is taken for granted.

When inflation is too low, however, there are limits to how far interest rates can be cut, because of the existence of a non-interest bearing substitute for bank deposits in the form of cash. And since we have no intention to do away with cash, central banks fighting too low inflation may have to resort to tools other than policy rates – what are often referred to as “unconventional” tools. This is necessary to deliver price stability, but it also comes with additional complications.

First, unconventional policies require us to operate in a broader range of markets, which means that the risk of unintended distortionary effects is inevitably larger than when using conventional tools. Does this imply we should refrain from using them when they are needed to restore price stability, as is the case today? The answer is clearly no, because we operate under a framework of monetary dominance. Our task is not to maximise the chances of price stability, under the constraint of not creating side effects. There are always side effects to monetary policy. And we are not at liberty to choose to fail our mandate.

But in pursuing our mandate, we ought to try and minimise distortionary effects as much as possible. And this is exactly what we have done through the design of our measures. This is one reason, for instance, why we concentrated our asset purchases in the most “commoditised” markets, such as government bonds. It is also why we have shifted the relative weight from rates towards other tools, so as to avoid as much as possible unintended adverse consequences for the banking sector.

A second complication associated with unconventional policies is that the public inevitably knows less about their transmission channels and effects. This is understandable because there is only limited historical experience of using such policies. But the body of publicly available research on their effects is rapidly growing. Like other central banks we have carried out considerable empirical work to calibrate our monetary policy stimulus.

And I am confident that, over time, our experience with unconventional tools will fill the remaining knowledge gap. One should not forget that the ability of central banks to bring high inflation under control was also doubted in the past, most famously in the 1970s, until empirical evidence put the debate to rest. But in the meantime, central banks have to demonstrate that there is no discontinuity when interest rates reach zero – unconventional measures can work as well as conventional ones.

How unconventional monetary policy works

Why is this the case? Conventional monetary policy operates by steering real money market rates below the prevailing equilibrium real rate, which in turn stimulates demand and inflation. But when the equilibrium rate is so low that the central bank cannot bring its policy rate sufficiently below it, the capacity to increase the degree of stimulus through moving short-term rates becomes limited. Unconventional tools can, however, still be effective in these conditions.

This is because, in reality, there is not just one interest rate which determines saving and investment in the whole economy. There is in fact a constellation of rates, which apply to different maturities, to different types of financial instruments, to different borrowers and lenders. So even without policy rates moving much, it is still possible for the central bank to stimulate the economy by lowering the level of all those interest rates. This can be effective in any circumstances, but particularly so when risk premia have risen due to market fragmentation or unwarranted uncertainties.

Reflecting this broader channel of monetary transmission, we have deployed a three-pronged strategy to inject additional stimulus into the euro area economy.

First, our forward guidance allows us to lower longer-term rates by steering expectations of future short-term rates. Having opened up the possibility that policy rates can turn negative has also contributed to flattening the whole yield curve, by removing the upward bias to yields that came from the perception that rates could only go up, not down. Indeed, we have stated unambiguously that policy rates would remain at current levels or lower for an extended period of time.

Second, our asset purchases help us further lower yields across maturities and asset classes by compressing risk premia in the markets where we intervene. That in turn triggers portfolio rebalancing out of those markets and brings down borrowing costs across the whole constellation of rates, thereby producing broad stimulating effect. This is complemented by the negative rate on our deposit facility, which speeds up the process of asset reallocation and reinforces the downward pressure on financing costs.

Importantly, this stimulus reaches the economy independently of whether financing is dominated by banks or capital markets. Just like any investor, banks have to judge the risk-adjusted return on capital when they allocate assets, and the benchmark that is typically used is the return available on risk-free government bonds. So as we purchase government bonds, we tilt the calculation in favour of loans to the real economy: by compressing the return on the securities we buy, and by improving the economic outlook, hence reducing risks on loans.

This is also supported by the third prong – our targeted longer-term refinancing operations (TLTROs) – which is specifically aimed at galvanising bank lending to the private sector. While the initial goal of the TLTROs was to improve the monetary transmission process, further recalibrations have made it progressively more expansionary. Under the rules we introduced in March this year, banks that meet their lending benchmarks will be able to borrow ex post from the ECB at negative rates, which propagates the effects of our conventional policy more directly to the economy.

Monetary policy and the economic recovery

In the two years since our policy package was launched, we have seen the effects of these measures in practice.

Events studies conducted by ECB staff find that our measures have had a major impact on long-term sovereign bonds, and spillovers to yields of other asset classes have been significant, too, especially for euro area financial and non-financial corporate bonds.^[5] Our analysis also finds that our policy package has had a substantial direct effect on bank lending rates, as well as a substantial indirect effect on lending conditions through its marked impact on long-term government bond yields.^[6]

Improved financing conditions have led in turn to higher growth and inflation. Eurosystem modelling shows that, without our policy measures, inflation would actually have been negative since 2015. In 2016 it would have been at least half a percentage point lower than we forecast currently, and around half a percentage point lower in 2017. And the impact of our measures on euro area GDP is also estimated to be sizeable, helping raise output by around 1.5% in the period from 2015-18.^[7]

The latest data show that these positive effects are only gaining in strength as our measures work their way through the economy. The euro area continues to benefit from a domestic demand-led recovery, and we can see how our policy is supporting this by looking at the spending components that are especially sensitive to financing conditions, namely consumption of durables and investment.

Last year, after a multi-year contraction, consumption growth of durables in the euro area surged to a rate unseen since late 2006. The contribution of fixed capital formation to output growth, which had been extraordinarily sluggish since the start of recovery in 2013, also progressively strengthened. And in the last quarter of 2015 – the most recent observation for which we have a full composition breakdown – investment surpassed consumption as the main driver of growth.

The recovery has also withstood a decline in external demand linked to the major slowdown in world trade. Over the past 20 years, the growth rate of world trade has been lower than it was last year on only two occasions: in the aftermath of the dotcom bubble in the early 2000s, and again in 2009 following the collapse of Lehman Brothers. In both episodes there was a sharp fall in euro area growth to close to or below zero. In 2015, however, the steep drop in trade did not produce a slowdown in the euro area economy. In fact, year-on-year growth even picked up throughout 2015, despite the slump in world imports.

This was in part due to the increasing resilience of the domestic economy. But euro area exporters, after a long spell of losses, were also able to regain market shares in world demand. Monetary policy was again a key factor in explaining this atypical resilience of euro area exports.

Importantly, these positive effects of our measures have not been accompanied by significant distortions that might start to tip the cost-benefit analysis – for instance, by excessively harming bank profitability and so damaging the main transmission channel of our policy. This is in part because of the way we have designed our tools, as I mentioned above. But clearly, the best way to ensure that it remains the case is to get back to our objective soon.

This point was captured well by Federal Reserve Chairman Paul Volcker after he hiked rates steeply in 1980, when he noted that, “I am worried about those financial institutions, and the worst thing that can happen to them is [for us to] fail to do the job and get the interest rate turn fairly soon. But the way to get the interest rate turned is not by hastening it prematurely.”^[8]

That is our position and, as the Governing Council underlined today, the momentum of the euro area's economic recovery continues to be supported by our monetary policy. This fosters the return of inflation towards 2%.

We are on the right track, but we take nothing for granted either. The Governing Council will closely monitor the evolution of the outlook for price stability and, if warranted to achieve our objective, we will act – as we have always done – by using all the instruments available within our mandate.