GIC Central Banking Series Conference

GIC Central Banking Series Conference
Paris, 14 May 2018

Speech by François Villeroy de Galhau,

Governor of the Banque de France

A perspective on the ECB’s Asset Purchase Programme

Ladies and Gentlemen,

It is a great pleasure to welcome you to Paris for this GIC Central Banking Series conference. This tenth GIC conference in partnership with the Banque de France is, as the previous ones, a very promising opportunity for professionals and monetary policy makers to exchange views, especially in the current euro area context.

It has been almost four years now since the Eurosystem incorporated several new instruments to combat the deflation risks that threatened the euro area in the wake of the crisis. One of our important tools was our Asset Purchase Programme (also referred to as APP – the euro-area version of Quantitative Easing) that was announced in its full blown version in January 2015. Looking ahead, our net asset purchases will stop when we see a sustained adjustment in the path of inflation (“SAPI”). We are not there yet, but we are more and more confident that the three criteria of this “SAPI”, namely convergence, confidence, and resilience[i], will be met. So, the time when our net asset purchases will end is approaching – and as I already said, whether it will be in September or in December is not a deep existential question.

Today is thus a good time to reflect on our experience with the non-standard measures. I would like to share with you some views on the following three questions. First, how effective were these non-standard instruments as a whole? Second, how much did the Asset Purchase Programme contribute to our exceptionally accommodative stance? Third and finally, how can central banks manage the risks they face when making their way to the “new normal” and exiting from such policy packages?

Needless to say, in all these issues, we benefit from the US experience, in particular of the Federal Reserve which has achieved, in recent years, a smooth path to normalisation. We are fortunate to have Loretta Mester (President and CEO of the Federal Reserve Bank of Cleveland) with us this morning.

**

I. Where do we stand on our unconventional tools as a whole?

Since 2014, we have been playing what I like to call a quartet of non-standard instruments. [Slide 2]

Our first non-standard instrument is the APP. I will be more specific in a short while on the channels through which this programme operates through stocks and flows. The size of our asset holdings related to APP is now approaching EUR 2.37 trillion and should reach EUR 2.55 trillion by the end of September 2018.

Our second instrument is the negative deposit facility rate: its purpose is to relax the zero lower bound constraint – therefore lowering the level of the yield curve.

Our third non-standard instrument consists in providing forward guidance on the path of future interest rates to make future monetary policy decisions predictable and to signal that rates will stay low for an extended period of time. This has contributed to flattening the short to medium term of the yield curve.

The fourth instrument is the provision of liquidity and credit to banks. Most recently, the second series of targeted longer-term refinancing operations (TLTROs) was launched in 2016. We are currently lending about EUR 760 billion through this instrument which eases credit conditions for firms and households.

These four instruments are therefore complementary and self-reinforcing. As a consequence, monetary policy measures have been supporting aggregate demand as well as employment, and the recent economic recovery in the euro area. The Banque de France just released this morning its first preliminary forecast for French GDP growth for Q2, at 0.3%. But this somewhat softer pace could be influenced by the exceptional holiday schedule in May. At this stage, new orders remain positively oriented and do not point to a trend change. Compared with the beginning of 2014, 7 million jobs have been created in the euro area. We estimate that, without monetary policy measures taken between 2014 and 2018, economic growth in the euro area would have been 1.9 percentage points lower in cumulative terms over the 2016-2020 period.^[ii] [Slide 3]

On the inflation front: annual HICP growth was 1.5% in 2017 and is expected to rise to 1.7% in 2020. Underlying inflation is set to strengthen, irrespective of short-run fluctuations in energy inflation. We clearly see the current slowdown [the 1.2% Flash HICP estimate for April, with a particularly low figure in Italy] as temporary, and we expect inflation to resume its progress in the coming months.

II. A comprehensive view on the Asset Purchase Programme

In theory, the effect of APP on yields is not straightforward: as Ben Bernanke said in a famous statement in 2014, “the problem with Quantitative Easing is that it works in practice, but it doesn’t work in theory”.^[iii] Today we probably better understand the channels through which this policy works, now that it has been implemented by all major central banks. We can distinguish three main channels through which our APP operates: a portfolio rebalancing effect through “duration extraction”, a signalling effect, and a reanchoring of inflation expectations.

Let me start with the portfolio rebalancing channel through “duration extraction”. By purchasing long-term assets, central banks replace long-term maturity and riskier assets with short-term maturity and safer assets for investors. Investors consider duration risks as a source of possible losses and they require an additional compensation – the term premium – for bearing such risks. By reducing duration risks for assets held by investors, the APP weighs on term premia and consequently on long-term yields. According to our estimations, holding a stock of assets equivalent to 10% of GDP lowers 10-year bond yields by about 45 basis points in the euro area. [Slide 4] This estimation implies that our APP has contributed to lower long-term yields by about 100 basis points. There is a broad consensus around these estimates for the euro area.^[iv] The magnitude of these effects is also quite similar to the ones observed in the United States or in the United Kingdom. Beyond extracting duration risk, the portfolio channel of APP also works through the exchange rate: rebalancing flows might result in increased demand for foreign assets by domestic residents and/or a repatriation of funds by non-residents.

The second important channel is that APP has had a clear signalling effect on the future path of short-term interest rates. APP has served as a credible commitment to keeping interest rates low for a long time, reinforcing our forward guidance policy. Recent Eurosystem research has shown that communication surrounding the APP has been key for the predictability of our monetary policy. As a result, markets were able to better understand our policy reaction function, leading to a reduction in uncertainty and unintended volatility and consequently risk premia on long-term yields. This effect is even much more significant when the size of monthly purchases has been clearly indicated.^[v] APP has finally provided state-contingence to our forward guidance.

The third transmission channel of APP comes from the re-anchoring of long-term inflation expectations [Slide 5]. When deflationary risks are rising – in particular when the central bank is thought to be close to its effective lower bound (ELB) – monetary policy actions based only on interest rates become ineffective. In that case, implementing APP reassures private agents on the central bank’s ability to fulfil its inflation mandate. Long-term inflation expectations from survey data bounced back after the APP announcement in January 2015, putting an end to the downward trend observed since 2012.^[vi]

A positive by-product of the APP has been to contribute to a further reduction in fragmentation risks in the euro area after 2014. By purchasing long-term sovereign bonds, the APP has helped to reduce risks of speculative attacks on sovereign bonds that were not justified by economic fundamentals. Doing so, the APP has improved the financial soundness of credit institutions – especially in periphery countries. Overall, by reducing fragmentation risks on bonds yields, the APP has improved monetary policy transmission.^[vii] [Slide 6] But we should be crystal-clear: reducing unwarranted fragmentation does not mean that the ECB is subject to any kind of fiscal dominance. Nobody should expect us to delay warranted monetary policy normalisation in order to accommodate debt problems of any Member State.

While the APP had clear benefits, some have argued that our prolonged accommodative monetary policy may have had detrimental side effects on financial stability. The main potential risk comes from high levels of debts (private and public) which would become more difficult to sustain with higher interest rates. I partly share these concerns and I think we need a better understanding of the side effects of QE. However, up to now, we have not seen any general sign of build-ups of excessive leverage or asset price bubbles in the euro area although vigilance is in order in some particular segments, such as high-yield corporate bonds or some prime commercial real estate.^[viii] Since the crisis, we have made a lot of progress in addressing vulnerabilities in financial stability – in particular through stronger capital and liquidity requirements on banks – and designing better macroprudential policies. Still, we also need broader liquidity stress-tests and macroprudential tools for non-bank finance.^[ix]

That being said, I think that an important question ahead is how to better take into account our new financial stability mandate, along with our core price stability mandate, in our monetary policy decisions.

III. Managing the exit

I would like to conclude these remarks with a short discussion of how we should handle a few important challenges on our path of gradual normalisation. Let me start with two clarifications and then touch upon one broader open question.

The first clarification is that APP will continue to shape our monetary stance even if we stop our net asset purchases. As we have learnt, stocks do matter, so that the end of our net asset purchases will not mean the end of the APP. The Governing Council is committed to reinvesting principal payments from maturing securities for an extended period of time after the end of our net asset purchases, and in any case for as long as necessary. The recurrent and sizeable purchases induced by the policy of reinvesting maturing assets will continue to extract duration risk from the market, to signal relatively low interest rates, and to favour the anchoring of long-term inflation expectations.

The second clarification is that, with the end of our net asset purchases approaching, our stance will rely more on our forward guidance on interest rates. Without a doubt, our communication will be adjusted given that the current guidance on policy rates is explicitly conditioned on the end of net asset purchases. As far as the first rate hike is concerned, we could give additional guidance on its timing –“well past” meaning at least some quarters but not years – and its contingency on the inflation outlook. Beyond that, the updated formulation of our communication will have to continue providing guidance on future interest rate movements over a relevant horizon, in order to avoid unintended volatility of the yield curve, especially during the transition period.

Now looking broadly over a longer horizon, I believe we will also need a better understanding of the key features defining “normal times” in the post-crisis economic environment. This need was forcefully recalled by Vítor Constâncio in a recent speech.^[x]

If the real natural rate of interest were to be lower for a long period – due to structural factors like the savings glut –, one should expect the nominal interest rates to hit their effective lower bound more often in the future. Active balance sheet policies could then be de facto included in our “normal-time” monetary policy toolkit. This could imply that we keep operating under a large balance sheet and excess liquidity. While the size of the balance sheet will definitely decline as a share of nominal GDP, we still do not know for sure what its new normal level will be. Alternatively or additionally, there is currently an academic debate about whether or not the quantitative definition of price stability should be revised. Loretta Mester has recently contributed to this debate in the United States.^[xi] For the euro area I would stick to the greatest caution: changing the inflation objectives either by increasing or lowering them, would blur the credibility of our commitment to price stability.

**

But let me come back to my core message about the APP. In the euro area, our current operational framework and policy strategy have proven to be flexible enough to deliver on price stability even in extreme conditions like the ones we have experienced over recent years. And this can give us confidence – notwithstanding patience and prudence – about the exit and the “new normal”, whenever they come.