Learn from the past and plan for the future: a perspective on monetary policy

Good afternoon everyone,

Thank you very much for inviting me to join this distinguished panel here at the Barclays-CEPR Monetary Policy Forum. It’s a pleasure to be here. I feel particularly inspired knowing that we are assembled here just a few blocks away from Gordon Square, where John Maynard Keynes – one of the greatest economic minds of modern times – lived.

Preparing for this event, I learned that this beautiful venue is sometimes also called the ‘Titanic hotel’. The main reason being that the restaurant is said to be almost identical to the Titanic's dining room. Both were designed by Charles Fitzroy Doll, the hotel’s architect, and also the man who was commissioned to do the first-class parts of the ocean liner.

Let me say that although I feel the weight of the historical context of this beautiful place, thankfully I do not see a connection between the euro area economy and the doomed ship. In fact, the economy has sailed a steady course lately despite high global waves and tight monetary policy.  

Starting with inflation, I am confident to say that the inflation peak lies behind us. Chart 1 on the left-hand side shows the evolution of euro area headline inflation, decomposed into its main components since 2019. We see a clear disinflation since late 2022 when inflation stood at more than 10%. This largely reflects the unwinding of earlier energy and food prices spikes, as well as the easing of supply-side bottlenecks. The reduction in goods inflation from 6.8% to 0.9% has been very strong, too. Only services inflation still stands at an elevated 3.7%. The next phase of disinflation is likely to be more volatile, because the large swings in energy prices last year together with the staggered reversal of fiscal support measures have an impact on this year’s inflation readings via base effects.

Chart 2 provides a more structural decomposition of inflation, based on recent DNB research (Ascari et al. (2024). Among the set of models we use at DNB, this one focuses particularly on capturing supply-side constraints. While the model supports the notion that supply-side shocks were dominant drivers of the rise in euro area inflation, it still attributes more than one third of the inflation dynamics to demand shocks. For the disinflation period, the decline in demand is particularly important. The contribution of demand has narrowed to less than one fifth relative to the peak. Monetary policy – which is captured here largely as affecting demand – has thus supported the disinflationary process.  

But the reaction of monetary policy has also been important in the face of these supply-side shocks. Even though monetary policy has less effects on supply bottlenecks or on food and energy prices, the public is particularly sensitive to these salient aspects (see e.g. Coibion et al. (2020) and D'Acunto et al. (2021)). As a result, an extended period of above-target inflation and possible second round effects create the risk that high inflation will become entrenched and, at some point, that inflation expectations will drift away from the target. A timely return of inflation to target is therefore crucial. That’s why it is important for monetary policy to react forcefully to underline our commitment to price stability.

Taking a step back, many have argued that the ECB was too late in hiking rates. While I, with the benefit of hindsight, understand where that interpretation is coming from, I would like to make three points in defence of the Governing Council’s response.

First, it is important not to reduce the monetary stance to the policy rates alone. The Governing Council adjusted its forward guidance long before and tapered its bond purchases under the APP as early as December 2021.

Second, it is important to note that the energy shocks were a direct consequence of the unjustified Russian invasion of Ukraine. As such, these shocks were not only unprecedented, but with a war at the European borders, the uncertainty went beyond energy prices. In this situation, my colleagues and I on the Governing Council were wary to choke off the economy in light of possible contractionary effects of the war.

Finally, even if we had initially fallen somewhat behind the curve, the eventual response has been very forceful and united, thus underlining the Governing Council’s commitment to its 2% inflation target over the medium term. This interpretation is also supported by a simple – so-called thick – modelling approach of policy rate paths implied by a large set of Taylor rules shown here in Chart 3. Clearly, the ECB’s policy rate implied by these simple reaction functions lifts off later but catches up quite soon.

In my view, one key element of the success of monetary policy over the last few years has been in ensuring that the large inflationary shocks did not cause a de-anchoring of expectations. One way to assess the risk of de-anchoring of expectations is this model-based analysis shown here on Slide 4. In this euro area application, DNB staff assess the path of inflation that is consistent with anchoring long-run inflation expectations going forward. As you can see on the left-hand side, at the September 2023 Governing Council meeting, the projected path of inflation depicted by the green dots was slightly above the anchoring-compatible path of inflation, the blue-dashed line. Therefore, further policy tightening was justified as this clearly posed an upward risk to inflation. The right-hand side of Chart 4 shows the same analysis conducted at the March 2024 Governing Council meeting. As you can see, the latest projected path of inflation is consistent with anchoring long-run inflation expectations. So, our policy of running a tight ship seems to have paid off.

Over recent years, policy decisions have been made in an environment of high uncertainty and large shocks. We knew that icebergs were looming in the dark, but it was hard to predict where. This uncertainty not only included the inflation outlook and the balance of risks, but was also related to the transmission of the extraordinary tightening of monetary policy. The Governing Council has therefore opted to take a data-dependent approach. Underlying this approach are three criteria. First, the inflation outlook in light of the incoming data. Second, the dynamics of underlying inflation, based on indicators which remove some of the high-frequency movements in inflation. And third, evidence on the strength of monetary policy transmission.

The focus on incoming data was particularly important, because making projections in an environment of high uncertainty is inherently difficult. Not surprisingly then, the forecast errors of the Eurosystem staff as well as those of other institutions have been sizeable. You can see this in Chart 5, which shows the inflation forecast errors from the Eurosystem staff projections since 2021.

Yet, what’s more important than the forecast errors themselves is understanding the underlying cause of these errors. And as you can see in Chart 5, the largest part of these forecast errors can be explained by assumptions about energy prices and their direct and indirect effects on inflation, indicated by the yellow and red bars. This is not surprising given the large shocks to energy prices we observed in this sample. Other non-energy related errors also play an important role in the latter part of the sample.

I’m convinced that this multi-pillar approach with a stronger focus on actual data realisations has served us well recently and that we need to continue to be data-dependent going forward. At the same time, as the decline in forecast errors in Chart 5 shows, there are also reasons to regain some confidence in our projections. This, together with the fact that a lot of the relevant and comprehensive data such as that on wages, productivity and profit margins are only published on a quarterly basis, suggests to me that projection round meetings of the Governing Council will be the key meetings for our interest rate decisions.

Currently, the monetary stance remains historically tight. This is not only the case for the level of policy rates themselves, but also when comparing real interest rates relative to the natural rate of interest, r-star. Chart 6 shows this gap between the ex-ante real interest rate and r-star for the euro area and the US, based on DNB research (Brand et al. (2021)). Of course, these estimates always have to be taken with a grain of salt given the sizeable model and estimation uncertainty as highlighted by the huge confidence bands. But the picture looks similar across a wide array of Eurosystem models. According to this model, the degree of monetary tightness is comparable across the two regions.

However, as the disinflationary path in the euro area continues and the medium-term inflation outlook further improves, we become increasingly confident that inflation will return to target in a timely manner. Consequently, it can soon be appropriate to ease the currently restrictive monetary policy stance, and gradually take our foot off the brake. As a result, policy rates will slowly but gradually move to less restrictive levels.  The precise timing, speed and scale of easing will also follow a data-dependent approach, with our projections being a key ingredient.

To give you a perspective on the path of policy normalisation based on the March Eurosystem staff projections, Chart 7 shows optimal policy scenarios. Optimal here means that monetary policy tries to minimise the deviations of inflation from target and output from its potential (for detailed methodology see Barnichon and Mesters (2023), McKay and Wolf (2023) and de Groot et al. (2021)). The range, in turn, is based on both a set of models and varying weights that the central bank attaches to inflation relative to growth. The upper end of that range thus corresponds to stricter inflation targeting. This means achieving a faster return of inflation to target, albeit at higher costs in terms of output. According to the chart, the optimal rate path would prescribe a reduction in the policy rate over the next quarters. Looking at the horizon until the end of the year, based on the March projections, optimal policy would have been broadly in line with 3-4 rate cuts, similar to the market pricing that was incorporated into these projections.

It’s important to note that these optimal rate paths rely on the inflation projections from March that include a quite rapid fall of inflation and slightly below-target inflation from the second half of 2025 onwards. This reflects, amongst others, a moderation of wage growth, a recovery of productivity growth and firms’ profit margins that absorb the higher wages. While the incoming data on new wage agreements does indicate some moderation since late 2023, wage growth has remained elevated and – according to forward-looking indicators – the path for 2024 is still expected to be quite bumpy. Productivity growth has remained low and is yet to pick up. Hence, we will have to await our next projections in June that will provide an updated assessment of the inflation outlook and the accompanying balance of risks.

I am aware I have been throwing a lot of technical stuff at you for the last few minutes, and I hope you are still with me. In any case, the important takeaway is that, although these interest rate scenarios can provide useful guidance, given the current environment we still have to avoid any commitments on a specific future rate path. Waters may still be rough in the time ahead, so it’s too early to declare that we have made a safe crossing. Our data-dependent approach allows us to assess the incoming data and inflation outlook moving from one meeting to the next and adjust our policy stance accordingly.   

Let me stop here. I am looking forward to an insightful discussion. Thank you for your attention!