Asset price inflation in the context of inflation measurement and monetary policy

From Felix Großmann  and  Gerd Kommer  

Many private investors are currently concerned about the risk of rising inflation. This is happening against the background of the very expansionary monetary policy of most Western central banks - including the ECB - since around 2008 and inflation, which is still low but has recently increased. In this context, critics of monetary expansion and low interest rate policy often put forward the argument of so-called asset inflation or asset price inflation (API).

The argument essentially says that the officially reported inflation measure, consumer price inflation (CPI), under-represents actual inflation. This is because the CPI only expresses the increase in the price level of consumer goods and services, but not the long-term, stronger increase in the price level of investment goods (hereinafter “investment assets”). In any case, the money supply expanded by the central bank is partly reflected in increased investment asset prices.

The fact that the CPI is actually too low is being used as an argument by the central banks to prove that their expansionary monetary policy has not yet triggered inflation. If the API were included in an expanded inflation measure, the actual inflation rate would have been higher for years.

API also contributes to economic inequality among the population because only the wealthier half of society benefits from it. The poorer half do not own any significant assets and therefore do not benefit from increased real estate and stock prices.

A few figures to illustrate: The price (price) of the “Global Stock” asset class, measured for example by the MSCI World (Standard) Index, rose by an average nominal 11.4% per annum (in euros, Total Return Index) in the ten years from 2011 to 2020. German house prices increased by 6.5% p.a. in the same period. If one were to add the net rental yields of an estimated 2.5%, we would be at 9% p.a. This compares to the German CPI of just 1.2% p.a.

In this blog post we want to look at the main practical arguments for or against the concept of adding an asset inflation component (API) to the CPI.

For the sake of brevity, we will hereinafter call the expansion or broadening of the conventional CPI indicator to include API into “expanded price inflation” “EPI”.

Clarification on housing costs in the context of consumer price inflation

Before we get to the arguments for and against the introduction of an EPI, we would like to note that “housing costs” in the sense of rents, i.e. a real estate component, are already included in the general CPI today. The statement of this fact is often “forgotten” by the media, book authors and internet bloggers in the context of the API topic.

Depending on the national CPI definition, warm rents have a weight of up to 35% in the general CPI shopping basket. Here are four concrete examples of the weighting of rents, i.e. housing costs, in the conventional CPI definition.

In western countries, the long-term average “warm rent cost ratio” (the share of the average warm rent in the average net household income) is between 25% and 35%. This suggests that the magnitudes of these weights are realistic. (We do not know why the housing cost weight in Eurostat's Harmonized Index of Consumer Prices/HICP is so low in comparison.)

However, the CPI does not take into account housing costs in relation to owner-occupied residential property. We will discuss the implications of this particular aspect below. As already mentioned, the costs or price changes of other investment assets are not included in the CPI.

The history of the discussion of an API-enhanced inflation measure

The question of adding an API component to the CPI is old hat in economics. As early as 1953, almost 70 years ago, the two US economists Alchian and Klein published an article on this subject, which has since been quoted in literally every academic publication on the question of the conceptual adequacy of the CPI (Alchian/Klein 1953). Another EPI essay classic – Goodhart 2001 – is also rather old. Both Alchian/Klein and Goodhart were mainstream economists and supported the introduction of an EPI, but were also aware of the many theoretical and practical hurdles to such a measure.

Since the EPI issue is actually simple in theory, all of the important arguments have been on the table for decades, easily understandable and freely accessible to everyone.

However, to date, ongoing collection and publication of an EPI has not been introduced by national statistical authorities in any Western country. The central banks also do not formally base their monetary policy on an PPI, although a sharp increase in asset prices, especially for stocks and real estate, is almost constantly discussed in the publications and announcements of the central banks, at least if it occurs at a given point in time - there can be no question of ignoring it or keeping it quiet.

A single example out of probably hundreds of possible examples over the past three decades is the famous speech by then US Central Bank chief Alan Greenspan in December 1996, in which he warned of “irrational exuberance” in relation to the sharp rise in US stock prices. (Greenspan's warning was ultimately wrong because it was far too early. US stock prices then climbed by a cumulative 112% over three and a quarter years before the dot-com crash began in April 2000. Anyone who sold their shares because of Greenspan's warning in December 1996 probably earned less over the following years than a strict buy-and-hold investor who simply ignored the central bank boss.)

In recent years, the Bundesbank has also commented several times and often in detail about the sharp increases in the price of residential real estate and their implications for financial market stability. The ECB, in coordination with the national statistical authorities of the 19 eurozone countries, is currently examining the extent to which the inclusion of residential property prices in an inflation indicator makes sense and is feasible (Siedenbiedel 2020).

The federal government also apparently sees an EPI as fundamentally positive. It reported at the end of 2020: "Germany supports the initiative to include owner-occupied residential property in the harmonized consumer price index calculated for European purposes, even if uniform methodological implementation across Europe would be a very big challenge."

In our opinion, anyone who looks more closely and impartially at the issue of introducing an EPI from an official perspective must come to the conclusion that its non-introduction so far has failed not because of the ideological reluctance of monetary politicians or governments, but because of some less spectacular, namely technical and conceptual, issues that are outlined in this article.

Before we get to the technical arguments, we want to clarify two things.

(a) The definition of inflation indicators (the technical determination of how inflation is measured) as well as the procedural implementation of inflation measurement lies in the competence of the 26 national statistical authorities in the EU, as well as that of the EU statistics authority Eurostat. The inflation indicators are neither defined by the national central banks nor by the ECB, nor is the ECB responsible for measuring inflation. The ECB has no authority to issue instructions to Eurostat and therefore cannot change anything in this area on its own initiative. Only Eurostat can do this together with the national statistical authorities. These are part of the government apparatus, not part of the national central banks or the ECB.

(b) The influence of any expansion of the money supply on investment assets in general and individual asset classes in particular cannot be quantified in concrete numerical values ​​using any method in the world. At best, one can say purely qualitatively that an expansion of the money supply tends to contribute to the nominal increase in the price of investment assets.

The pro argument for an EPI

Investment assets are acquired by private households for wealth creation and retirement provision - primarily in the form of real estate and shares or financial products based on them. Creating wealth or investing in this sense is actually making provisions for future consumption. This logic seems to suggest considering API for the purposes of measuring “general inflation.” This is all the more true since investment assets such as real estate and stocks must have a positive real return in the long term - measured by the “classic” CPI - and have had them in the last 200 years.

Investment assets such as stocks, bonds or real estate are priced by the capital market in such a way that they have a positive, risk-adequate “expected return” above consumer goods inflation, i.e. a positive, real (CPI-adjusted) return. (In the very long run, the expected return is approximately the same as the average return.)

For example, the real, CPI-adjusted return of the equity asset class in the long term, starting from a medium, normal valuation level, is around 5% to 5.5% per annum. This is the average value over the past 120 years for 23 countries for which data of sufficient quality is available. If you only take the last 60 years or so, the result is a slightly higher average.

Bonds, real estate and other investment assets each have their own expected real return, which is lower than that of stocks, but also well above zero.

The six most important technical problems of an EPI

(a) The time lag problem

If you wanted to take the price changes of owner-occupied residential properties into account in an EPI key figure, you would first of all be confronted with the practical problem of time lag in data collection.

Based on the “data collection infrastructure” that exists today in most countries, including Germany, the price changes of residential real estate would only be available to the statistical authorities much later than the price changes of normal consumer goods. This is tricky because inflation needs to be measured as up-to-date and as quickly as possible - especially for the purposes of the central banks' monetary policy. If one were to try to take real estate prices into account in a normal inflation indicator today, it would probably drastically slow down the speed of publication, i.e. the timeliness of the inflation indicator.

In order to overcome this practical calamity, a government initiative would have to be launched to greatly improve the quality and speed of data collection for real estate prices. We assume that this would be possible.

This problem naturally does not exist with other investment assets such as stocks.

(b) The volatility problem

There is no doubt that an EPI indicator would be more volatile than the current CPI. High fluctuations from quarter to quarter are already a technical problem in the conventional CPI, which is why core inflation (core CPI or core inflation) has always been calculated in addition to the “normal” CPI. The Core CPI is the normal CPI excluding the notoriously volatile energy product group (oil, natural gas).

Since there are strong short-term price movements up and down in energy prices (and these world market-related movements largely have nothing to do with monetary policy), it makes sense to exclude energy from the measurement of inflation with regard to questions of monetary policy control.

However, in an EPI, the sometimes strong short and medium-term fluctuations of investment assets up and down would in principle further intensify the volatility problem.

The result would almost inevitably be the use of three variables: the conventional CPI, the core CPI (excluding energy prices) and the EPI.

(c) The problem of mixing different time horizons

When a household purchases an investment asset, such as a property or stocks, it benefits from it over many years - just as when a company buys a machine or a building, it benefits from it over a period of several years. This is a fundamental difference to a consumer good, where the benefit only exists in the short term until it is used up.

For this reason, corporate accounting rules require companies to “capitalize” capital goods (as distinct from expenditures on inputs that are immediately fully consumed) on the company’s balance sheet and then “depreciate” them (approximately) in line with annual consumption. This depreciation approximately represents the proportionate annual consumption in the company's profit and loss statement.

This means that from a statistical perspective, only a small portion of the purchase price of an investment asset in a private household represents actual consumption in the current year. However, the aim of measuring inflation is to measure short-term, more current  There could be price changes and accordingly one would have to use something like the depreciation principle for companies.

On the one hand, the depreciation principle would greatly mitigate the impact of a price change for an investment asset in an EPI, but on the other hand, it would greatly complicate the technical consideration of the price change of investment assets in the EPI.

(d) The problem of paradoxical implications

In the Eurozone, the proportion of households with owner-occupied homes, the homeownership ratio (HOR), is 68%. This means that a clear majority of all households in the Eurozone own their own residential property; only a minority of households are renters.

At this point, too, a note on the interpretation of the HOR is appropriate: The national HOR in the EU fluctuates between 95% for Romania at the upper end and 52% for Germany at the lower end. It is a common misconception that a high national HOR in the 190+ countries of the world is somehow a reliable sign of national prosperity in the sense of a high gross domestic product per capita. In fact, “poor” countries tend to have higher HORs than richer countries. In Europe, the richest country, Switzerland, has the lowest HOR at 43%.

Back to the actual topic: If the prices of residential real estate rise more than the goods in the CPI basket in a certain period of time, then this price increase creates a increase the purchasing power of the owner households concerned. You could and can now buy more CPI basket goods, including more housing utility, because rents regularly rise noticeably less than purchase prices in a broad, strong upward movement in property purchase prices.

This effect is paradoxical in that the basic character of the inflation indicator - as it is rightly perceived and understood by the population today - is reversed and in a certain way distorted. Inflation is perceived by the population as something negative because it worsens the purchasing power of existing financial assets and the future salary income of the population as a whole. House price inflation ultimately does the opposite. In the Eurozone, house price inflation increases the purchasing power of 68% of households and reduces it for only the remaining 32%. This means an improvement at the level of the entire population. Since the HOR in the Eurozone is almost 70% and is likely to be in this order of magnitude or higher worldwide, and since most governments in the world are striving to further increase the HOR (albeit often unsuccessfully), one can conclude that the paradox in question applies practically everywhere: property price inflation positively influences the purchasing power of the entire population on balance.

It is true that house price inflation contributes to wealth inequality in the population as a whole, all other things being equal, but in our view it cannot be the purpose of a central bank's monetary policy to combat inequality in the distribution of wealth. This is the sole responsibility of governments, not central banks. Furthermore, the state has sufficient means and means at its disposal to regulate economic inequality, even without monetary policy. For reasons of space, these will not be discussed further here.

(e) The problem of the lack of significance of pure asset price changes

Measuring the value or affordability of investment assets such as real estate and stocks by their absolute purchase price level (or index level for stock indices) or by the change in the absolute purchase price over time is - unlike with consumer goods - fundamentally wrong.

This is easy to show. A property that costs 500,000 euros can be “cheaper” than one that costs 300,000 euros. This logic also applies to stocks and other investment assets that produce cash flow.

The question of whether an investment asset is expensive or cheap, has become more expensive or cheaper in a certain period of time, does not depend on its absolute price level or its changes, but on its valuation. Price and valuation are two different things. If a share price increases by 1% in a calendar year (in a naive interpretation, asset inflation took place here), but the company's profit increased by 5%, then this share is cheaper become, not more expensive. The same applies to real estate relative to the net rents, the “profit” in real estate.

Inflation in the usual sense of the word is a synonym for “becoming more expensive”. There is nothing wrong with this interpretation - it is correct. However, an investment asset (unlike a consumer good) can only actually become more expensive if its valuation goes up. However, rising purchase prices for assets are not valuation indicators (we have the details on this using the stock market here shown).

Anyone who follows this logic cannot measure API in terms of rising asset prices, but that is exactly what proponents of EPI usually suggest. However, if one were to use valuation metrics within an EPI instead of mere prices, this would probably have enormous complexity consequences in practice.

(f) The problem of historical and international comparability

An EPI would probably be difficult to compare across countries, as there is little guarantee that all of the world's 190 countries would measure the API in a sufficiently similar way and incorporate it into the EPI. Similarly, an EPI would probably not allow long-term historical comparisons because it could only be determined a few decades backwards.

However, since the CPI would still be available for such cross-country or historical comparisons, this argument is likely to have only limited relevance.

All in all, we can conclude that the introduction of any form of EPI would entail a number of technical difficulties and conceptual ambivalences, but with good will these could be overcome.

Would an PPI indicator change anything at all in the practice of ECB monetary policy?

In our opinion, this question cannot be answered from the current perspective. Only practical experience after the introduction of an EPI could show this.

It can be assumed that central banks, as well as politicians and the population, (a) would quickly get used to a higher and more fluctuating PPI and (b) the parallel use of several inflation indicators would become commonplace in monetary policy and public discussion, e.g. B. the old CPI, the core CPI and the new EPI.

Furthermore, since the introduction of the welfare state at the beginning of the 20th century, one of the main goals of economic and tax policy has been economic redistribution. Against this background, one could argue that it cannot be the task of the central bank and thus of monetary policy to pursue social equality goals because this is a core task of the elected government. The government generally has a variety of instruments to pursue and achieve these goals. Anyone who demands that the central bank be involved in social policy could overload and endanger the independence and quality of monetary policy.

Conclusion

The development of an EPI key figure, however structured, would be possible in principle, but would require overcoming a considerable number of practical hurdles and conceptual question marks. In our opinion, these are underestimated and sometimes swept under the table by the proponents of an EPI.

In our opinion, the fact that an EPI has not yet been introduced in over 190 countries around the world, despite 70 years of scientific discussion, is not due to the ill will of governments and central banks.

Currently it looks as if the EU is seriously considering the introduction of a PPI that includes house prices.

Rents with heating are already included in the national CPI figures with a weight of around 30% (less in the EU “HICP”).

It is unclear whether an EPI would result in significant changes to the monetary policy of the relevant central banks.

In our opinion, it is unclear to what extent a social policy mandate (e.g. “contributing to preventing the increase in economic inequality”) should be imposed on the monetary policy of central banks. Social policy should, as in the past, be a responsibility of governments, not central banks.

literature

Alchian, Armen/Benjamin Klein (1953): “On a Correct Measure of Inflation”; In: Journal of Money, Credit and Banking, 1973, Vol. 5, issue 1, pp. 173-91

Goodhart, Charles (2001): “What Weight Should be Given to Asset Prices in the Measurement of Inflation?” In: The Economic Journal; Vol. 111, No. 472, June, 2001, pp. F335-F356

Israel, Karl-Friedrich/Günther Schnabl (2020): “Alternative Measures of Price Inflation and the Perception of Real Income in Germany”; CESifo Working Papers, No. 8583; September 2020

Siedenbiedel, Christian (2020): “Inflation lacks real estate prices”; March 8, 2020; In: FAZ.net (Frankfurter Allgemeine Zeitung)