Committee on Economic and Monetary Affairs Working group Monetary Dialogue prep meeting (EGOV)
Legend
Transcript markers
- text Sentence being played
- text Search result
- text Intervention cited by the link you opened
Click a sentence to play the video from that moment. Select text to quote it with the speaker, timestamp and link.
More options
ECON European Parliament
Follow its sittingsSpeaker 3
We see something similar for inflation expectations. If we ask households, which inflation rate do you expect for the next 12 months? The median respondent in the CES survey responded with a rate close to 3%, somewhat below 3%. So, this is the upward bias, or the gap between perceptions and reality, that we are talking about today. Is this a problem for the ECB's mandate to achieve price stability?
The first argument I want to make is that I want to give you further context. So, next slide, please. If we look at the longer history of data that we have, you see that. Please, next slide. You will see that this upward bias in inflation perceptions is a general phenomenon that is not only present today but was also present in the 2000s, the 2010s, and before the inflation peak. The data is not yet visible to you, so we need the next slide.
Yes. Back to one slide. Okay. Very good. So you see that, with the exception of the recent inflation peak, inflation perceptions and expectations have already been elevated in the 2020s, in 2020, in the 2010s, and before that.
So this is really a general phenomenon, and the data doesn't suggest that it's inconsistent or a problem for the ECB's manner to reach price stability. Okay. Now, let me get to the full argument on the next slide. Why do we think that the upward bias itself is harmless from the perspective of price stability? First of all, it's really only a household-level phenomenon.
Today, we see that the expectations of financial market participants are well calibrated, and the expectations of professional forecasters are close to 2%. Of course, their actions will transmit monetary policy to households. For firms, the phenomenon is less attenuated; it is much smaller in the data. It's really a household-level phenomenon that we're talking about. That's the first point.
The second point is that it is a general phenomenon. We've seen it even when inflation is low. So, per se, it doesn't seem to be inconsistent with reaching price stability. Now, why do we observe this bias in the data? There are plausible methodological and psychological reasons for this.
One is that households tend to focus on large and unrepresentative price increases for goods that they most frequently purchase. And this leads to an upward bias in their perceptions. Second, in general, there is very little attention to inflation by households, in particular when inflation is low. Third, there is also a measurement problem. In these surveys, households have to come up with a number.
Reasonable forecasts would typically range from 1 to 3%, but the scale they have access to ranges from negative numbers to very large positive numbers. It's very easy to be misled by this range and provide large numbers, which frequently happens in surveys. And for all of these reasons, there's really no evidence that we have that the upward bias itself changes behavior. The research that we have argues that inflation expectations are extremely important for behavior, but it focuses on changes in expectations, not on the level per se, and the upward bias is a phenomenon that is specific to the level of inflation perceptions and expectations. So, there's a general consensus in the academic literature that, rather than focusing on the level of expectations, which we know has always been, and probably always will be, upward bias, we should focus on differences in expectations over time.
For this reason, the upward bias in the level, we think, is mostly harmless from the perspective of price stability. Next slide, please. Of course, there is also another side to the argument. The bias reflects very limited knowledge about inflation on the part of households, and this can be problematic. Two reasons and two examples.
First, it means that not only in the Euro area, but also outside, say, the United States, households persistently perceive inflation to be higher than the central bank's target, which, of course, can create a credibility challenge. Second, households' limited understanding of inflation is limiting their economic decision-making in a broader sense. They don't understand what inflation means, why it's important, for investment decisions, what the debt-eroding effect of inflation can be, where it comes from, when inflation peaks, and so on. This limited understanding also limits their possibility to optimally respond to inflation episodes or to take inflation into account when it is low. And this is a reason for concern.
Okay. So, our general recommendation is that high inflation literacy would be desirable. It's probably difficult to achieve, and it's clearly outside the sole scope of the ECB's policy. So, this is a broader objective for policy institutions. And in the short and mid-term, we don't expect huge changes here.
So the ECB should really take into account that when they communicate with households, they communicate with people who imperfectly understand inflation and have very different views on the phenomenon, and this needs to be considered in communication. Finally, the ECB should continue its path of tracking and researching household expectations. The data that I've shown you today, the research behind that data, is partly possible because the ECB, but also other central banks across the world, are taking expectations very seriously, tracking them and measuring them. And we think that's the right path and it should be continued by the ECB in the future. That's all I have.
Aurore Lalucq
Thank you. I don't know if there are any questions. So, okay. So next will be Frank Smith. Welcome.
And you have the floor for, let's say, six minutes.
Speaker 6
Thank you very much, Madam Chet, for the invitation. Let me also start by saying that the views I express are my own and not necessarily those of the BIS. I very much enjoyed reading the note prepared by Peter, and it very much captures the state of knowledge about what drives household inflation perceptions and expectations. Rather than comment on the note itself, I thought I would present some additional international survey evidence based on two recent BIS publications. They very much confirmed the findings and conclusions of Peter.
Speaker 7
It shows that the findings are quite universal. So, next slide, please. The survey covers 31 countries and was last run in March and April. We are currently doing a new round, but these results are not yet available. Next slide, please.
Speaker 6
Okay. Your first finding, from this international evidence covering 31 countries, is that indeed household inflation expectations are upward biased relative to actual inflation and to inflation forecasts that you see on the left panel here. So the average or the median is more than 6% in advanced economies, whereas actual inflation early last year was around 2%, and also inflation forecast by professionals was around 2%. As Peter mentioned, when you ask households where they get their information on inflation, it's mostly based on food prices, grocery prices, oil, and energy prices, and that's across the world. It's also important to know that there's quite a bit of heterogeneity in this bias in expectations and perceptions across households.
Typically, you will get lower biases and lower inflation expectations for male respondents, for older respondents, and for more educated and higher-income respondents. Next slide, please. As mentioned, I think we can skip this. Next slide. So one factor driving high inflation expectations is the perception of an increase in the cost of living post-COVID, which was associated with a loss in real income, as you see on the right-hand side of this graph.
So, as the middle graph shows, those households that perceived a large increase in inflation post-COVID had inflation expectations that were twice as high in advanced economies as those that perceived a low inflation perception. So perceptions are very important for the formation of inflation expectations, and these expectations are very persistent. Next slide, please. I mean, one important driver of the bias in household inflation expectations is the knowledge and the credibility of the central bank. In our sample, and you see that on the left panel, about half of households recognize the central bank's name, and less than half know that price stability is the central bank's goal.
So the knowledge about what the central bank is and what it does is quite limited. And that, again, is an important factor in driving differences in the bias of inflation expectations that you see in the middle panel. So informed households have a much lower bias than those that are not informed, and often they are relatively close to the inflation forecasts of professionals. When we ask them where they get their information, I mean, in the vast economy, it's still very much TV and radio, but, of course, social media is also becoming more and more important. And that raises one of the questions: to what extent do central banks need to invest more in reaching out through social media to explain what they do and what they try to achieve.
Next slide, please. So finally, how relevant is the bias for central banks? I mean, here, this is a little bit of a complicated slide, but basically, the bottom line of this slide is to say that household inflation perceptions do matter for consumer confidence, sentiment, and therefore may have an impact on behavior such as savings and precautionary savings behavior. But I agree with Peter that when you look through the empirical academic evidence, the results of to what extent household biases affect behavior are mixed. And I would also agree that, in the end, it's not an obstacle for the central banks to achieve their targets, as we also see in the current context in the Euro area.
The next slide. Yes. These are just the key takeaways, but I don't think I need to go through them again. So let me thank you for the attention.
Okay.
Speaker 11
Curiosity. Concerning the fact that the upper bias is influenced by food and energy prices, together with the fact that there's a lot of heterogeneity, as you said. Can you elaborate a little more on the sources of heterogeneity? Because you mentioned education, you also mentioned income groups. But I don't know if you have some evidence on this, because I think it would be very interesting as a source of explanation.
Speaker 12
So, education or income groups—if you have some more evidence on that. Thank you.
Yes, please.
Speaker 6
Okay. Yes. I mean, these findings, and Peter can also confirm that it's lower income households that typically have higher inflation expectations. There are a number of factors that can explain that, but, of course, one of them is, for example, that to the extent that lower-income households spend more on energy and food than, say, richer households, their perceptions may also be more affected by those prices. So some of that heterogeneity that you see in the data may have a common cause.
Speaker 3
I can add another example, echoing what Frank just said. We often see in the data that women have higher inflation expectations than men, but that is easy to account for statistically by taking into account who takes care of the groceries, which is often women. And once you control for that, you don't see that difference anymore. So again, you see that exposure and experiences that people have in their everyday lives shape their inflation expectations. And then, heterogeneity in such experiences and exposure will also lead to heterogeneity in expectations and perceptions.
But I think today, given that we talk about the upward bias, that's probably of second order. First order is really that we do see this upward bias across large parts of the population, yet we do see it even when inflation is low. In academic research, we focus on differences across time rather than on this level of upward bias in perceptions and expectations.
Aurore Lalucq
Thank you. So, it's now time to move to the presentation of Ms. Laura Bottazzi is a full professor of economics at the University of Bologna. She's also a fellow at the Institute for Economic Research at Bocconi University, Edinburg Center for Economic Policy Research in London. So you will also have six minutes.
Speaker 17
So, thank you so much for the invitation. And let me start by acknowledging, of course, that the report has been written by my colleagues at Buchen University and Chicago University. So, the report is divided mainly into two parts. So I would concentrate, hopefully, for three minutes on one part and three minutes on the other part. So, in the first three minutes, I want to concentrate on the determinants of inflation in the Euro Area in the period 2024 to 2025.
Speaker 14
So, inflation—the core of our analysis—shows that inflation is close to the target. We already said that, in the sense that we have non-energy industrial goods prices and inflation around 2% now. And the decrease in headline inflation was mainly due to, and you can also look at the graph if you want. So, can I put some arrows? No, go back. I put some arrows to go back.
Speaker 13
I put some arrows, yes, because you can directly go to the graph. So, thanks to the energy inflation, the impact was mainly negative. Service inflation went down slowly, but we have a general decrease. So, what is behind this euro area? So if you were to look at our first report last year, you can go back, thanks. We were somewhere scared of two possible shocks.
One shock is coming from the tariff for the U.S. S. And the other one is coming from China. Let me go to the index of inflation view. So click on the arrow, then on OK.
Speaker 16
The Citigroup Index of Surprise Inflation tells us a lot. What does "a lot" mean? It does not measure inflation expectations, but measures the difference between realized inflation and expected inflation, okay? Expected inflation comes from consensus and various surveys, waiting for their volatility. And you mainly see three blocks.
Speaker 13
One is U. S, one is China, and the other one is Europe. If you look at the U. S. Immediately, you see, when the index is positive, it means that realized inflation is higher than the expected one.
Speaker 15
When it's negative, expected inflation was higher than realized inflation. So mainly started from the mid-2025 expectations. So we were expecting higher inflation than realized in the U.S. Yes, okay? So, we were expecting higher inflation due to tariffs, of course.
Speaker 14
And then, we were also expecting lower demand. Demand in the U.S. S. It has been highly resilient. And there's been a lot of volatility, as you can also see in the index, due mainly to the volatility of the tariffs, so trade uncertainty.
Speaker 13
If we look instead, demand was resilient in the U.S. S. And that's why, when we also look at the amount of exports that Europe had towards the U.S. S. Didn't decline, really.
We're expecting a decline in exports. We don't see any kind of decline in exports from the Euro Area towards the U.S. S. When you look at Europe instead of China, you can see that there is some kind of volatility. There was an upward trend in some ways, so expected inflation was higher than before.
But mainly, it has become quite stable, okay? When we look at China instead, we have a negative. So, the line is negative. So, realizing inflation is always below the expected level. And in China, in fact, what happened is that China has exported deflation in general.
And so, we see that there was not a shock of U. S. As we expected last year, at the same moment and the same time, there was another report. But mainly, we have seen a shock. You can go back to the slide. So the shock that you had was mainly on China.
So, the external shock to China was relevant. Export, we said, import from China has instead increased, and you can see it directly by clicking there. So, we have an increase. Can you click on the arrow there? So imports on China have increased, and then you click. Thank you.
Okay. So, total Euro Area imports from China have increased almost. So, we have data from 2024 and 2025, showing almost a 7.2% increase in imports from China. Okay, that is relevant when we want to highlight the effect of inflation. So, if you want to, you can also go back.
So, that is one element. So, inflation in the Euro Area was determined by an external shock, but we also have some kind of internal shock. E, great heterogeneity. There is great heterogeneity in inflation among the European countries. That's why it's also relevant for the European Central Bank, especially for communication to highlight.
If you can click there, it shows regenerative inflation in member countries. The red one is not weighted by GDP. The other line, instead, the blue one, is weighted by GDP. And you see the declines because the smallest country shows a high divergence with respect to the percent. Go down and then look at the core, and in particular, there is a difference between core-periphery inflation differential that you can see there. Okay, now go back to the core-periphery inflation.
You're clicking the wrong one. Okay. Okay, in any case, even if you don't, we have no time to go through it. The core-periphery inflation differential is really high. So, the main, big final message here is that this inflation has been successful; external forces play a big role, especially China. Why?
We were also able to estimate that imports from China have increased by 7.2%, which is well determined by a decline in coinflation headline inflation; over two years' time, we estimated around 0.24%. So, China has really exported this inflation. Now let's look at the communication of the central bank instead, because we don't have much time. So, what we did, of course, is that communication is highly relevant. So, what we did was apply a new methodology in order to understand communication and to stay away from possible subjective biases.
We use a large language methodology, and it's very nice, by the way, so I have to acknowledge our younger colleague who wrote a thesis dissertation this year. And it's nice because instead of evaluating using distortion and meter subject methodology, you can click on the communication on the first one, and I can show you the methodology. Okay. Methodology, in reality, looks very simple. It's complicated when you have to apply, but it's very simple in itself.
So every time we know that there is a meeting, there are four kinds of documents that are prepared and published by the European Central Bank: monetary policy statement, press conference transcript, economic bulletin, and meeting accounts. So, we have an extractor. So, we have a specialized extractor that takes part of this document, okay, citation without changing any words, that is highly relevant and identifies rate decision, inflation references, growth references, forward guidance, country mentions, and central bank spillover. So, they take exactly these mentions. And then we have two different kinds of agents: one DAB agent and one HOC.
So the HOC agents tend to argue, say, look at the documents and the stance, and argue the stance is okay. And it shows why that is the case, highlighting and emphasizing inflation risk, wage pressure, vigilance, and also acknowledging the opposite vision by highlighting and explaining why it is a weak position. The DAB agent does exactly the same, okay? So it takes exactly the same kind of documents and then it tries to justify, mainly emphasizing growth risk and disinflation, why the statements are mainly dovish. And we must also acknowledge the position of the hawkish one and explain why it is the case that they are weak.
Once we do that, there is an extra third actor that enters, who is neutral, looks at the position of dove and hawk, and also looks at the same statement they've been going through, highlights and identifies the exaggeration and cherry-picking argumentation of one or the other, evaluates the general position, and picks a winner with a sort of confidence. So, 50% confidence means they are equal; 70% confidence in one or the other means they have a higher position in one or the other. And then it produces a final stance, whether or not, a confidence level, argument strength, key uncertainty, and communication. If you go to the other slide, then go back and go to the other slide again, you can see the number. You always have to click.
You always have to click. Yeah. So, what are the main arguments they control? So, it's rather clear. So they pin down, they look at different argumentations.
So for sure, the door looks more like growth. The other one looks more at inflation in services, for example. So, they do have different arguments. What is nice? Okay, what is nice?
Actually, because there is this kind of debate, it means the debate also exists at the level of the European Central Bank, inside. Then, when we look at what realizes this, this kind of analysis captures exactly the realization of the choice of the European Central Bank. So if you look at the dovish period when you have a 200 basis points decrease in the main interest rate, okay, which is between 2024 and 2025, it captured exactly this change and became the dovish period. Okay, it's very nice from that point of view. So, it captured the debate.
The debate helps because when you look at the language model, first of all, large language models are not perfect; we know they cannot be replicated exactly, and they also have hallucinations. So the nice debate creates the ability to better evaluate the test of communication. So they also give you the option to go back, okay, and also look at the distance and lowest. Okay, ECB communication also gives clarity over time. So we see the clarity is increased and continues to increase.
So the main analysis at the end shows that clarity is increasing whenever the central bank has to take a position, like changing the interest rate, okay, or changing the interest rate up or down. There, you see that the clarity of the communication is very high. When it is the case that it is not so clear, there is a positive situation. So, when the central bank decided not to do anything, like in this period, you see that the clarity index goes down and there is less clarity in the documentation. Although that, for us, is the moment in which maybe the market needs more clarification in general. The other element, which I cannot really show you, is that there is also no mention of the originality among countries of the Euro Area.
So the countries mainly cited are the big countries, not the small ones. The small one is the one that also has higher inflation, and inflation is different from the rest of the other countries. Okay. So, I want to highlight that transparency should also increase during pauses. There is an explicit treatment of inflation.
We'd like to have a more explicit treatment of inflation heterogeneity, which becomes a problem, of course, when you have a common shock that might well have different effects on different countries. And we, therefore, hope to have a clearer articulation or trade-off during post periods, okay? But the analysis is very nice because it helps analyze all positions and all discussions within the European Central Bank. So, I don't have much time. We have plenty of things in our report.
Thank you so much.
Speaker 11
A clarification: the clarity score. I didn't really understand the definition of this score.
Speaker 13
When the third agent looks at the dovish and the hawkish positions, the other two agents try to identify how much the position is really linked. He tries to cancel in some way all the picking arguments and all exaggerations in the argumentation because the hawkish would like to read more the hawkish position, and the dovish read more in the document the dovish position. Instead, the neutral one looks at and compares directly the statements of the two other agents with the actual test. That's why they never changed. So, the first instructor never changed the wording. The word remains.
They all look the same. And confidence is mainly about how much, in reality, their statement aligns with the statement of the ECB, okay? So, it declares who wins, okay? Confidence means how much you are in one position or another. So a confidence index of fifty-fifty means that there are no winnings, so the communication is not very clear; one cannot win or lose.
When it's higher than 50, it means one direction or the other; the arguments of one position or the other are stronger in your lines. But that possibility of discussing eliminates any possible bias that the agents might have, because you start by looking at something with some distortion if you're in opposition or on the other side. So it eliminates, in some way, the distortion that you have when analyzing a test. It's very nice and very powerful. Because before, we were looking instead, you know, at a very individual level, so you let each individual look at the test and evaluate it directly.
And each of us has our own, not only a priority, but way of reading.
Speaker 19
Well, thank you. Thank you so much. No more requests for the floor? No. Well, thank you.
Aurore Lalucq
There are two things I want to say. First, everyone has a view in this committee on the monetary policy of the ECB. Everyone's got a view on the interest rate that they should adopt. But when it comes to preparing the monetary dialogue, our attendance is very low, which is not acceptable. We had, this morning, a hearing with the candidate for the vice chair of the ECB, and especially some political groups, I would say, asked some questions on the military policy.
But at the end of the day, there is only my dear Ludec Nida Maja in the room. So, this is not acceptable. This is my first point. And I will make sure that the coordinator has this message. The second point is that when it comes to the ECB, I think we also need to understand that the ECB, as you all know, is not a political institution.
And the ECB has to respect the treaties. And when it comes to communication, it is not an easy task for the ECB. So I think that at some point, it is also up to us as members of the European Parliament, or people who are elected or in the political field, to maybe say something about it. But I think that this is not totally the job of the ECB because the ECB is not free to say many things. That's my first point.
The second point is that when it comes to inflation, we also have to acknowledge that the inflation, especially the recent one caused by what happened with Russia, is very hard to find with such an interest rate. And this is also something that we politicians should explain to people. I mean, when we have this kind of shock, you can increase or decrease your interest rate. At the end of the day, the price of the gas will be what it is. Since the ECB is not in the room, but I know they listen to us sometimes, I just wanted to make this quite clear.
But thank you so much for your input. And yeah. But I have to go; I have a meeting with the minister.
Speaker 21
Yeah. Okay. Thank you so much for your time, and I hope that next time, we have more attendees in this room. Thank you.
Ask anything about this video — I analyse its transcript to answer.