 Thank you very much. Good morning, ladies and gentlemen. First of all, thank you to Professor Hopper and to Dr. Bilschan Hopper for inviting me and hosting this conference in these beautiful surroundings. I'm very honoured to be here at the Property and Freedom Society, which I've followed for many years, ever since I was an undergraduate student in economics in Berlin. And to be here personally for the first time really has in many ways exceeded my expectations, especially the food. But not exclusively the food. When I picked my topic for this talk, I came up with the provisional title, which was different. It was the Menace of Inflation, Admission and Denial. And I picked that title at a time when the price inflation rates in the eurozone went up sharply and the authorities at the ECB and elsewhere rather desperately tried to reassure the public that this was only transitory. Nothing to worry about. This is transitory. It is because of problems related to the lockdown recession and then the war in Ukraine, it certainly has nothing to do with monetary policy. By now this has changed. The narrative has changed. Basically nobody claims anymore that this is a transitory phenomenon. Everybody admits that the higher inflation rates are here for a while at least. And in fact it is now the official goal of the ECB not to return back to 2% price inflation as soon as possible, but rather to do it smoothly in a gradual process year by year, step by step. We had 8% this year. Next year we might have something like 6% or 7% and 5% and eventually we go back to 2%. Whether this will work out is highly questionable. Only recently the Kiel Institute for the World Economy has published the inflation forecasts for next year and they estimate that inflation will actually be even higher next year than it was this year. So given all that I thought it is not really necessary to talk too much about these public discussions of inflation because we all know that there are many problems related to this. Instead I wanted to look at some more fundamental problems with the official inflation measures and hence the new title, A Critique of Inflation Measurement. And I would like to start this providing some motivation. In fact having listened to Torsten Pollard who criticizes the role of the empirical economist this is very fitting because I want to do a little empirical exercise. So this is the money stock M1 in the Euro area plotted as an index. So the final quarter of 1998 is the base quarter where the index takes the value of 100 and then you can see over time how the money stock M1 has evolved throughout the existence of the Euro. Of course it was introduced in 1999 as an accounting currency and M1 by the way is the cash money in circulation as well as the money on regular bank accounts. So for the lay person that's what we consider to be money or that's the money that is readily available to be spent on goods and services. So it makes sense to look at M1 when we think about inflation. And you can see over a bit more than two decades the money stock M1 has increased by a factor of larger than 6 or compounded by more than 500%. You can calculate the average annual growth rate which would be about 8.5%. In a traditional sense, in a classical sense of the word this would be a measure of inflation. That's the expansion of the money stock at on average 8.5% per year. Of course we, when we talk about inflation in the modern context mean price inflation and it is very instructive to look at how price inflation compares to the evolution of the money stock M1 and here we have the harmonized index of consumer prices which is the conventional measure of inflation in the Euro area. It has grown by comparison on average only by 1.7% per year. So now this in and of itself doesn't mean that the index is wrong or anything although it is wrong. But there are of course other factors that come into play that could explain away some of the excess money production that we see here in the graph. For example, economic growth. To the extent that you have real economic growth the additional money that is created can be absorbed without unit prices of goods and services increasing. However, if you look at economic growth it has been rather weak. I take out here the lockdown recession and if we take out the lockdown recession then per year real GDP in the Euro area has increased by 1.6% per year. So now if you make a back of the envelope calculation you can see that price inflation as measured by the HICP and real economic growth do not correspond at all to the rate at which the money stock has expanded. So the money stock on average increases about 5% points faster than the HICP and real GDP put together. And this is really an explanatory gap where we have to think about where does the excess money go. If it doesn't push up prices if it is not absorbed by real economic growth where does it go? And there are different explanations for that. One typical explanation that economists like to give is, well, there should be a correspondence between the growth rate of the money stock and the growth rate of the HICP and the real GDP only in so far as other things are constant under the catalyst-paribus consumption. But catalyst is almost never paribus and so one thing that could have changed over the time is the demand for money. To the extent that the additional money is simply held in cash balances rather than spent on goods and services, it won't push up prices. So the HICP is not necessarily downward biased if indeed the demand for money has increased to such a large extent, year by year. If you think about it, this is very implausible. You would, if anything, expect the reverse to happen. In an inflationary environment where prices go up and even if only at 2% at a relatively moderate rate you would expect people to reduce their cash balance demand for money and rather spend the money faster. So mainstream economists would call this the velocity. They would argue partly this can be explained by a decrease in velocity. The money circulates at a lower pace in the economy. And at the inception of the euro, the euro system estimated that there would be such a decline in the velocity of about 0.5% per year. If you take that, okay, let's run with the estimate of the euro system, 0.5% does not at all explain away the gap of 5 percentage points that we have seen in the previous slide. And even if you increase that to 1% or 2%, which is again rather implausible, it doesn't explain away the gap. And this brings me to the second possibility and this is of course the hypothesis that I want to stress here in the talk. The HICP is underestimating general price inflation. It is downward biased. And another piece of suggestive evidence can be found in some survey data that is compiled by the European Commission since 2004. They ask people in the euro area about their perceived price inflation, about the inflation perceptions. And if you calculate the median of this perceived inflation, which is plotted here in red, you can see that it is persistently above the official inflation measure. Interestingly, it is on average about 5 percentage points above the officially measured price inflation. Now you might say this is a coincidence. Or is it? But we here at the Property and Freedom Society are very fond of the wisdom of the crowds and so we take this very seriously. If the public says the inflation is really 5 percentage points higher, there might be something to it. Now, and this brings me to the important theoretical problems that are related to the HICP. If we accept that second explanation to play a significant role, then we have to ask, so why is it then that the HICP is underestimating general price inflation? And then if you think about it, there are basically two possibilities. So it might be the case that there is over proportionate price inflation outside of the HICP basket. So the HICP tries to measure consumer prices, prices of consumer goods and services, and it doesn't look at other markets. So it leaves things out in the economy and maybe in these areas in these markets that are left out in the HICP, there is an over proportionate amount of inflation going on. The second possibility is that the HICP is indeed downward biased in what it attempts to measure. It measures consumer prices and it has an in-built downward bias in the construction of the index. And I would argue that there is a very potent mix of both of these factors at play that can show or explain why the HICP is indeed too low, is underestimating general price inflation. So let us look at the first point here. Potentially over proportionate inflation rates outside of the HICP basket. The HICP as I mentioned focuses on consumption goods and services and more precisely it focuses on private consumption and even more precisely on private present consumption. So the only thing that matters for the construction of the HICP are consumption goods and services that are bought by private individuals in the present. So two things are systematically excluded from the calculation of the HICP. The first thing are so-called public goods. We could call that state consumption. Expenditure by the government is excluded. All of the things that the government buys, infrastructure, expenses for education, for health, security and so on, all of that is not included in the HICP. It is true that the HICP has, for example, a sub-index for education and for health. They have a weight of 1% and 5% respectively in the overall index. This is of course from a perspective of the economy as a whole too small a weight because as a society we spend much more on these things because of the government spending that goes on. What the HICP considers is only the private part of expenditure on education and health. So public goods are excluded and there is potentially an area where we have over-proportioned inflation. Surprise, surprise. The second element are future goods. So that means saving is excluded and saving is nothing else but an additional vision for future consumption. As the HICP focuses only on present consumption, any future goods are excluded in the index. So all assets that you can imagine, stocks, real estate, land and so on, gold, bitcoin, are excluded in the index. And this is problematic if you think about the HICP as being a general inflation measure. I believe that it takes also account of some of the future goods that are important for the general or average household. The households are not only consumers in the present but they also save us for the future and they try to provide for future consumption at least to some extent. And so if you want to analyze the general standard of living or things like that, you might think that future goods are something that we should take account of. Now it is indeed the case that in all of these areas, if you take the available data, you find over proportionate rates of price inflation. So here this is data for Germany for the same time period that we looked at before. The average HICP price inflation in Germany was 1.6%. So a bit lower than the Eurozone average. And if we compare that to, for example, stock price inflation as measured by the German stock market index, the DAX, we see that there is a big gap on average. On average stock price inflation was 5.2% per year. And it is of course very volatile. That's a particularity of stocks. But on average, and that's important when we consider this rather long-term perspective, on average the price inflation in stocks is much higher. So now you might argue, okay, this is because of economic growth and the businesses make productivity gains or the expected profits go up and that's why the stocks go up. That's rather a sign of health. And I would hold against that just the observation that for example during the lockdown recession, the DAX has increased by 5% per year. Of course it has gone down initially, but the monetary expansion was so strong that stock prices were pushed up again very quickly. And so it's hard to believe that this is due to productivity gains. So it is primarily driven by monetary inflation and that's why it should be taken account in the measure of price inflation as well. The second element that I would like to show you are real estate prices or residential property prices. In Germany, over the entire period they have increased by 2.6% per year. In the first half of the period, the inflation rate when it comes to real estate was low in Germany. That's a particularity of Germany. If you look into other countries, take France, Spain, Italy, Portugal, of course they have massive amounts of price inflation in real estate markets. In Germany this has only started after the great recession and with the beginning of unconventional monetary policy measures. Since then, if you take the second half of the period, the average rate of inflation in real estate markets is about 4.5% in Germany. So also over proportionately high. And the last element that I want to show to you are public goods. Now the problem with public goods is of course that there are no market prices for public goods so it's very difficult to directly calculate an index, a price index for public goods. But what you can do instead, what you can use as a proxy, are of course the costs of provision. What does it cost to provide public goods? Well it's tax money. The tax revenue of the government is the price the public pays for all the public goods that the government spends the money on. And if we look at the tax revenue in Germany we can see that well here too we have an over proportionate development. The tax revenue has increased at a faster rate than the general price inflation rate. So the tax burden of the average household in Germany has increased over time. Public goods have become more expensive over time. That's in relative terms when you compare that to the general rate of price inflation. And this has not taken account of the quality of public goods. Now so this, all of these empirical facts show that indeed there are blind spots of the HICP that when we would take account of those we would end up with a much higher assessment or estimation of the general rate of price inflation. So part of this gap that I've shown you can be explained away if you were willing to take a broader index of inflation. Now this is independent so far of an inbuilt bias in the HICP. Now that's the part that I want to talk about now. There might be an inbuilt bias in the HICP and that it underestimates the rate of price inflation for consumer goods and services of course. And there are good reasons to believe that. There's an extensive literature on that topic on the potential biases of price indices. It goes back more than 100 years. But an important contribution in the modern context was the work of the Boscon Commission that published a series of papers in the mid-1990s and they found in upward bias in the consumer price index for the United States. Very similar studies were published for other countries for Germany for example and for France and they found the same results. There's an inbuilt upward bias in the consumer price index as they were used at the time. So these findings led to changes in the methods of calculation which I want to explain now. The first source of this supposed upward bias are quality changes in the goods and services included in the basket. So the Boscon Commission argued that the statistics do not sufficiently take account of quality improvements in the product. So if you have a good that becomes more expensive but at the same time for whatever reason there is an increase in the quality of the good then this price increases not necessarily a sign of inflation but rather you should reduce the observed price by the equivalent of the quality improvement measured in money. So you reduce the observed prices and you end up with a lower inflation measure. And this argument, the Boscon Commission argued that about half a percentage point is due to quality changes. This led to a more explicit quality adjustments in the goods included in the HICP. And when you take a look at the website of the German statistical office, the Federal Statistical Office of Germany, they list nine types of quality adjustments that they use. The most famous one are the hedonic quality adjustments. Those are quantitative regressions where they estimate the money price or the price equivalent of a quality change and then they deduct that from the observable price. But my favorite are the judgmental quality adjustments. And I found that absolutely outrageous that they list that on their website. They say, and I quote loosely or paraphrase loosely from the website, they say that in cases where the quality change cannot be objectively quantified, it might be up to the expert to just make a judgment of what the money equivalent of the quality improvement is. And so they engage in discretionary price adjustments and do not take the raw data, the raw price data as the observant. The big problem here, or one of the big problems, is that there is no transparency in terms of the data. You cannot find the raw data before quality adjustments. You only have the data after quality adjustments. And even there, you don't find all the data. It's not publicly available. We had a meeting, a colleague of mine and I, we had a meeting with employees of the German Statistical Office and we asked them, okay, can you give us an example of where you correct for a quality improvement? And they were like, sure, yeah, here, computers, mobile phones, all of that, constant improvements, we correct the prices, of course. Okay, wonderful. Can you also give us an example of where you take account of a potential quality deterioration? That silence. They couldn't give us an example. I mean, this is scandalous, right? But when you think about it, this is also not all that surprising. Of course, when you have quality improvements, any business would openly advertise that. Well, yeah, the computer has gotten faster and so on. If there is a quality improvement, you are open about it. You make it public, you advertise it. So it's very visible. It's easier to take account of that. And there are quality deteriorations, which also take place, arguably. But those are hiding. There is an incentive for businessmen to hide the quality deterioration. So they are not as easily observed. And so this whole enterprise of trying to correct for quality adjustments comes with an inbuilt bias towards only taking account of the quality improvements, which reduce the inflation measurement, but leaving out the quality deteriorations in the product, which would increase the inflation measure. The second point are substitution effects. So the economists in the Boskin Commission observed, surprisingly, that people change their consumption behavior over time. There are changes in the consumption pattern. And it so happens that the goods that are replaced have a relatively high rate of price inflation. And the goods that are come in to the actual consumption basket have a relatively low price inflation rate. And so the economists and the statisticians argued, well, we have to adjust the basket over time, right? We have to adjust the weights. If people stop buying the goods with a relatively high price inflation rate and instead buy the goods with a relatively low price inflation rate, then we should adjust the weights. We have to give a higher weight to those goods that they now actually consume, which just so happens to be the goods with a relatively low price inflation rate. And of course, this is exactly what has been done in the U.S. and in Europe as well. The HICP is normally a less-Paris index. On their website, they say it's a less-Paris type index, which is actually more accurate because it's not really a less-Paris index anymore. The less-Paris index works as follows. You take a base year, you fix a basket, and you do not change the basket, and then you observe how the price of that basket evolves over time. That would be your inflation measure. What the statisticians involved in the calculation of the HICP do now is to update the basket every year following the arguments of the Boschian Commission. Now, this is problematic, and this has partly been pointed out almost a hundred years ago by an Austrian economist, Gottfried von Haberler, in his habilitation thesis. This is problematic from a welfare economic point of view because it might not just be a genuine change in preferences that is the cause of the change in consumption pattern, but it might just reflect, these changes in the consumption pattern might just reflect a strategy of avoiding higher costs. People buy those products with lower price inflation rates not because they intrinsically value them more now that preferences have changed, but just because they are cheaper in relative terms. They are more expensive to those other goods that are more expensive. And if that's true, then of course the quality of the overall basket deteriorates because of that. Now, in that way it is linked to quality changes. But the quality changes in the statistics are not made on the level of the overall basket, but only on the level of individual goods. So if the quality of the basket changes, no adjustments are made. So to take an example, instead of buying at the organic fruit store, you go to the cheap chain supermarket where prices have remained relatively low. And in the statistics that reduces the price inflation rate, but of course it also reduces arguably the quality of the products you get. And if you were to correct that, price inflation would be higher. So what I want to argue here is then that the reactions to these initial upward biases in the HICP index, if they existed at all, are very likely to have turned these biases into downward biases and make us now underestimate the general rate of price inflation with the HICP. Okay, so the final question of course might be, well, do we need measures of price inflation? My answer would be not for monetary policy. There's actually a big problem that monetary policy has focused on a nebulous construct like a price inflation index in order to manage their monetary policies and their monetary expansion. In fact, of course, I don't have to tell you that I'm talking to a libertarian audience here. The ideal would be not to have monetary policy at all. But if you have monetary policy, then you would want to, and you want to have a rule so that policy is not discretionary, then you want to have a rule which is pretty much objective that uses variables that are measurable, clearly measurable without much doubt. And the price inflation index certainly does not work, and especially the one that is used today. And it is responsible for, in my opinion, an over expansion of the money stock because the consumer price inflation as measured by the HICP has not reacted much to the tremendous expansion of the money stock that has been going on. Remember, a factor of six over two decades. The actual inflation that was going on outside of the HICP was not actually taken into account, and the official measure gave a pretext for continuing the monetary expansion. And all of the adverse effects of that, of course, amplified. Nonetheless, I would say it is important to think about how we could potentially improve an inflation index. We don't need it for monetary policy, but it still is very informative when you engage in economic history in the Missassian sense, for example, in order to assess the development of real wages, standard of living, and so on. That is still informative, although there are many problems and we cannot objectively measure it. I believe that if we were not only to get rid of monetary policy, but altogether get rid of public statistics and instead privatize the measurement of inflation, we would get much better inflation measures that are much more informative and would give us a much clearer view on what is going on in the economy. And with that, I close, and thank you very much for your attention.