1. / Home
  2. / Blog
Blog post

Including home-ownership costs in the inflation indicator is not just a technical issue

The ECB’s preferred method to include owner-occupied housing services in the inflation indicator would involve an asset price.

Publishing date
18 November 2021

This blog post is based on a study prepared for the European Parliament’s Committee on Economic and Monetary Affairs (ECON). The study is available on the European Parliament’s website.

 

Sixty-six percent of euro-area households live in the home they own. This comes with running costs, such as the cost of energy and water, maintenance and property taxes. Owners of course do not have to pay rent, but it does not mean that having the right to live in their own home is free in an economic sense. The home could be rented out and thereby owner-occupiers forgo rental revenues. Own home occupation can be viewed as the provision of a service (providing a place to live) by the owner to herself/himself.

The estimated cost of such a service is included in the national consumer price inflation indicators used in the United States and some European countries, such as Germany, the Netherlands and Czechia. But the European harmonised index of consumer prices (HICP) does not include this service cost. The HICP is used by the European Central Bank (ECB) to measure inflationary developments. As early as 2003, the ECB monetary policy strategy review said the “inclusion of owner occupied housing services in the HICP is desirable”, a call reinforced in the ECB’s 2021 monetary policy strategy review. This proposal is welcome.

The big question is how to do this. The 2021 review overview highlights the so-called “net acquisition” approach as the preferred way to include costs of owner-occupied housing in the HICP, possibly informed by a paper prepared by ECB staff for the strategy review. This is also the approach favoured by Eurostat. In contrast, in the United States and some EU countries like Germany, the ‘rental equivalence’ approach is used for including the costs of home ownership services in consumer prices. Eurostat and all EU countries also use the rental equivalence approach in national accounts to measure private consumption resulting from home ownership services and their costs.

However, the use of the net acquisition approach for HICP would involve an asset price component (ie the price of certain dwellings) in the consumer price indicator, which is clearly unwarranted (as we have flagged up in a study for the European Parliament's Economic and Monetary Affairs Committee). The consumer price index should measure the cost of living, not the cost of investing in assets. It is no surprise that other investment assets, including shares, bonds and investment fund shares, are not included in consumer prices. This blog post briefly explains the issue and quantifies the possible impacts of including owner-occupied housing in the HICP.

 

The net acquisition approach

According to the net acquisition approach, the purchase of a dwelling is recorded as consumption at the time the transaction takes place, as is done with other durable goods. This approach disregards the fact that the consumption of the good takes place over time. ‘Net acquisition’ in this context means purchases minus sales of dwellings of the household sector from/to other sectors, for example, netting out the sales by companies to households of newly-built houses, and the sales of old houses by households to companies (for redevelopment, for example). Transactions between households are excluded.

This approach is in line with the conceptual basis of the HICP, which considers transaction prices. The same treatment is applied to other durable goods, like cars and TV sets.

However, the market price for used cars and TV sets gradually declines and even approaches zero, in around 10 years for TV sets and 20 years for cars. By contrast, the market price of ‘used’ homes tends to increase and property can store value for decades, if not for centuries. That’s why in the national accounts dataset, the purchase of cars and TV sets is considered consumption, while the purchase of houses is considered as investment.

Including a house price index in HICP according to the net acquisition approach would thus include the price of a major investment asset. On average across 21 EU countries, main residences account for about 60% of households’ wealth assets (see Figure 7 here).

The net acquisition approach implicitly assumes that housing-related consumer costs follow the fluctuations of house prices, an assumption that is not justified. On the contrary, house price changes determine the capital gain (or loss) resulting from investing in a house and thus the house price index reflects the results of the investment motive for homeownership.

In fact, ECB President Christine Lagarde has highlighted that only the consumption element should be included in the inflation indicator:

"What was decided by the Governing Council was to account for the consumer cost of the owner-occupied house. So, it has nothing to do with the investment cost that an owner incurs; it has to do with the consumer cost that the owner of a house actually incurs."

Unfortunately, the separation of the consumption and investment components of a home purchase is an unresolved issue. The ECB has recognised this problem and called for further investigation of how to treat the consumption and investment components inherent in house purchasing. But there is no way to separate the two. In some countries, the cost of the land is considered the investment component and the cost of the building is considered the consumption component, but this is ad hoc. The building itself is also a store of value and its secondary market price can increase for many decades, thus behaving like an investment asset and not like a consumption good.

Since 2014, Eurostat has published an owner-occupied housing price index using the net acquisition approach, which has two parts: an acquisition component (accounting for around 80% for the euro area) and an ownership component (around 20%). The first contains the acquisition of dwellings from the non-household sector and self-building of dwellings, plus related costs, while the second includes expenditures related to owning and maintaining the dwelling. Due to the dominant (80%) role of house prices in this index, it is not surprising that this index bears similarities to the house price index (which considers the prices of both new and used houses) and the dwellings price index from the national accounts dataset (Figure 1).

 

The rental equivalence approach

The other main method for incorporating the costs of owner-occupied housing services is the so-called ‘rental equivalence’ method. This means estimating what the market rent would be for an equivalent dwelling (size and quality) in the same location (also called owners’ equivalent rent or imputed rent). This method is used in the United States, Germany and some other countries for including home ownership services in the national consumer price index and in the national accounts of all EU countries to calculate private consumption expenditure and its price.

The intuition is straightforward: by living in a house they own, the owner doesn’t have to pay to rent. So, the consumption cost can be approximated using the estimated rental value of the property, considering its size, quality, location and other characteristics.

But the rental equivalence approach is difficult when rental markets are thin, due to the lack of comparability between the rental and owner-occupied housing markets. Moreover, when rental market regulation imposes controls on rental price changes for sitting tenants, the rental price index might not reflect market developments. These are significant drawbacks, but national statistical offices are able to use this method for national accounts.

Figure 1 shows that actual rent prices for housing (both from the HICP and the national accounts dataset) and the imputed rentals (from the national accounts dataset) develop more smoothly than housing prices. This difference is relevant because the inclusion of price developments of one or the other in the HICP would have different implications for the ECB’s inflation target.

Implications for the inflation indicator

Adding alternative measures of costs of owner-occupied housing to the HICP has different impacts depending on the approach:

  • Using the net acquisition approach, data on the owner-occupied housing price index inflation is available starting in 2011 for the euro area and most EU countries. In the euro area, this inflation rate was very similar to the dwellings inflation rate from national accounts (Figure 1). For Belgium, Denmark, Lithuania and Sweden, data on owner-occupied housing price index inflation is available from 2006 or 2007, showing a similar pattern as the dwellings price index. Thus, for the net acquisition approach, we use data on the owner-occupied housing price index when available and the dwellings price index for earlier years.
  • Using the rental equivalence approach, data on the price index of imputed rents for housing from the national accounts is available from 1995.

For simplicity, we assume that the weight of owner-occupied housing in an augmented HICP would be the same as the weight of imputed rentals in the national account dataset of each country and each year, as well as for the euro area (using country-specific and time-varying weights). On average, this weight was 11.6% from 1999 to 2019 in the euro area.

In the euro area, on average from 1999 to 2019, the difference between the actual HICP and the augmented indicators would have been relatively small, 0.07% when using the net acquisition approach and 0.05% when using the rental equivalence approach (Figure 2).

These numbers do not suggest dramatic differences on average over 1999-2019, yet in some years the differences would have been larger, for example, at a time of a housing boom or a housing bust. In 2006, euro-area inflation would have been 2.5% with owner-occupied housing (when using the net acquisition approach) instead of the reported 2.2%, deviating even more from the ECB’s definition of price stability. For countries that experienced rapid house price increases and price declines before and after the global financial crisis, such as Estonia, Ireland, Latvia, Lithuania and Spain, the difference between HICP and its augmented version (including the net acquisition approach) would have been particularly large in some years.

Conceptually, the way owner occupation is included in the HICP will have implications for the scope of the ECB’s inflation target. Will this target continue to refer to costs associated with consumption (if the rental equivalence approach is used)? Or will the indicator partly involve an asset price component, which is more related to financial stability concerns (if the net acquisition approach is used)? This is a crucial conceptual question, not just a technical issue about the inclusion of the costs of living of homeowners in their own homes.

 

Recommended citation:

Darvas, Z. and C. Martins (2021) ‘Including home-ownership costs in the inflation indicator is not just a technical issue’, Bruegel Blog, 18 November

About the authors

  • Zsolt Darvas

    Zsolt Darvas, a Hungarian citizen, joined Bruegel as a Visiting Fellow in September 2008 and continued his work at Bruegel as a Research Fellow from January 2009, before being appointed Senior Fellow from September 2013. He is also a Senior Research Fellow at the Corvinus University of Budapest.

    From 2005 to 2008, he was the Research Advisor of the Argenta Financial Research Group in Budapest. Before that, he worked at the research unit of the Central Bank of Hungary (1994-2005) where he served as Deputy Head.

    Zsolt holds a Ph.D. in Economics from Corvinus University of Budapest where he teaches courses in Econometrics but also at other institutions since 1994. His research interests include macroeconomics, international economics, central banking and time series analysis.

  • Catarina Martins

    Catarina works at Bruegel as a Research analyst. She studied her BSc in Economics at the University of Porto. She then pursued an international quantitative MSc in Economics via the QTEM Network, studying the first semester at the University of Porto, the second at HEC Montréal and the third semester at Solvay Brussels School of Economics and Management (SBS-EM).

    Before joining Bruegel, Catarina worked at the European Central Bank in the Directorate General Market Operations, the department responsible for the implementation of Monetary Policy. As part of DG-M, Catarina joined the Money Market and Liquidity division, where she worked very closely with topics related to financial markets, benchmark reforms and liquidity developments. She had previously done a quantitative internship in Banco de Portugal at BPLim Microdata Research Laboratory of the Economics and Research department.

    Catarina is interested in various areas of economics and financial topics and has developed over time a fascination for data-related work. She is fluent in Portuguese and English and has a good command of Spanish and French.

Related content