The Unassuming Economist

Global Housing Watch Newsletter: February 2020

A remarkable new data set constructed by Jean-Charles Bricongne, Alessandro Turrini and Peter Pontuch allows direct comparison of house prices across countries—illustrating big differences in how many years of income it takes to buy a house—and provides suggestive evidence of when house prices may be at risk of correction. Prakash Loungani summarizes the data set and poses a few queries to its creators.

How expensive on average are houses in the United States relative to Australia? And how many years of income would it take the average person to buy the average house in each country?

It would seem that the answers to these questions should not be too difficult to provide. After all there are a number of data sets that give cross-country data on house prices in various countries. But most of the data sets provide price indices, not the actual price of houses. They can be used to compare appreciation in house prices across countries over some time period, but don’t necessarily tell us whether houses in the United States are more affordable than in Australia and how the relative affordability has changed over time.

Figure 1: House price per square meter in 2016 (in euros)

A new data set called HouseLev put together by Jean-Charles Bricongne, Alessandro Turrini and Peter Pontuch fills the gap. It provides the average price per square meter of housing in 40 countries, from as far back in 1970 for most countries to the most recent available, generally 2017 or 2018.

The authors use two methods, top-down and bottom-up, with the latter intended to provide a check on the former or a back-up estimate when the top down method is not feasible due to missing data.

The top-down method computes the average price as the ratio of the total value of dwellings and associated land to the total floor area of dwellings. The total value of dwellings and land (‘the numerator’) is generally taken from the national income accounts of countries; the total floor area (‘the denominator’) is from the census data of countries. Divide one by the other and, presto, you have the average price of a house.

So now we have an answer to the first question posed earlier. As shown in Figure 1, it turns out that the cost of the average house price in the United States is under 1500 euros per square meter and over 5000 euros per square meter in Australia. (For metrically-challenged U.S.-centric readers, that’s about $150 per square foot in the United States.)

Of the countries in the data set, houses in Bulgaria are the cheapest at under 300 euros per square meter and the most expensive houses are in Hong Kong (SAR)—prices there are literally off the charts—nearly 30,000 euros per square meter!—and hence not shown in Figure 1.

How reliable are these estimates? Here’s where the bottom-up approach comes in. The authors have painstakingly gone through the websites of real estate agents and collected the data on sales offers in different locations. These data are then aggregated up to give a country-level average. Reassuringly, for most countries, the top-down approach and the bottom-up approach give similar estimates of house prices—the median difference in only 7 percent and the biggest difference is 12 percent.

Figure 2: Years of income required to purchase a house in 2016

Figure 2 shows that it takes under 4 years of average income to buy a house of average size in the United States and over 16 years in Australia. The median across countries is 10 years.

The creators of the data set also looked into whether high ratios of house prices to incomes can signal a correction in house prices. They find that if the ratio is greater than 10, there is concrete risk of a significant downward correction of house prices in the following three years (see their paper, Bricongne et al. (2019), for the details).

Q&A with the creators:

Q. Is the data set publicly available?

A. The data set that covers forty countries is publicly available with detailed annexes of prices in levels in national currencies, in euros, in PPP (purchasing power parity) and with price to income ratios, in the annex of the European Commission Discussion Paper available at the following link: https://ec.europa.eu/info/sites/info/files/economy-finance/dp101_en_houselev.pdf

Q. You mentioned plans to extend the dataset to the regional level? What is the status of this work?

A. Work is ongoing in the Commission, with first results covering all EU countries expected for Autumn 2020. The Commission, Banque de France and the OECD are cooperating and interacting on these efforts. Initial raw results are available already at a basic regional level for the EU27 countries (the so-called NUTS2 level), with time series broadly spanning the 2010s.

Since national accountants quite rarely publish data at a local (NUTS3) level, the principal source to produce regional house prices in levels is the one based on web scraping, as well as aggregates from administrative data. The main issues that we encountered with the construction of sub-national price levels is comparability, which is why we focus mainly on improving stratification algorithms and refining our data on housing stocks, incorporating also satellite data.

Q. Will the data set be kept updated on a regular basis? By whom?

A. The dataset will be updated at least annually by the European Commission with Banque de France also contributing to possibly expanding the country coverage. The intent is to give access to the dataset to all potential users. The dataset will initially be accessible at the ‘browsable’ website of the REFINE network (real estate finance and economics network: https://www.institutlouisbachelier.org/en/programme/refine-real-estate-finance-and-economics-network-2/) but other solutions are also being considered.

Q. Has there been any reaction to your estimates (e.g. some validation from country authorities or realtors; anyone challenging your estimates)?

A. HouseLev is a database of estimates: these are not official statistics and they have not been validated neither by national statistical authorities or Eurostat nor by realtors. The estimates have been constructed with the objective of reliability, transparency and comparability, keeping in mind that full comparability is a tremendous challenge for non-homogenous objects such as dwellings. To get estimates as comparable as possible, a common concept of floor area is adopted (“useful floor area”). When possible, comparisons are made on the basis of more than one methodology for computing average prices per square meter.

The data have been shared with EU Member States and used by different institutions: central banks, ministries of finance, academics. A number of comments have been received by experts in the fora where preliminary versions of the dataset have been presented (European Commission, ECB, Banque de France, Paris Dauphine, OECD ACPR—the French Supervisor for banks and insurance companies). On a few occasions, the feedback has helped improving the estimates for some countries or to better qualify the results. Such feedback is being continuously used to incrementally improve the data methodology country by country.

From the IMF’s latest report on Croatia:

“Housing prices have begun to accelerate, mainly in the capital and coastal areas. Average housing prices grew 8.0 percent, but 12.2 percent in Zagreb (yoy, September 2019). This increase should be seen in context of higher real wages, better employment prospects, growing consumer confidence, as well as declining interest rates. Tourism is the main driver of real estate price developments in Zagreb and the coast. Investment properties for short-term rentals have grown rapidly. This is facilitated by a favorable flat-tax on short-term rentals compared to higher taxation on long-term rentals. Market observers note that some of these purchases are not loan-financed, but they still assume that the majority is financed by bank loans. The market has also been supported by the government’s housing loan subsidy program for young first-time house buyers introduced in 2017 and the reduction of the real estate transfer tax since 2019. According to the CNB’s housing price index, real estate prices are now beginning to reach pre-crisis levels. Staff recommended that housing prices should be monitored with a holistic approach taking into account mortgage lending, general purpose loans that might be diverted to real estate, as well as government housing subsidies on the demand side. Also, the impacts that the current tourism boom and tourism rental taxation policies have on the supply of housing for purchase need to be taken into consideration. The mission welcomed current research efforts of the CNB to better gauge housing affordability.”

 

From BIS:

“Although residential and commercial real estate prices are increasingly moving in sync and the role of international investors is growing, this does not mean that there is a global real estate market, a report by the Committee on the Global Financial System finds.

Property price dynamics: domestic and international drivers documents recent trends in residential and commercial property prices in over 20 countries, gives an overview of key drivers of price developments and describes policy initiatives used to manage associated risks to the economy and financial stability.

Property prices have been rising, reaching record highs in many countries. As prices appear high in comparison to simple rule-of-thumb valuation benchmarks, such as rents and incomes, some central banks are concerned about the consequences of a potential correction. In many cases, however, current price developments can be largely explained by fundamental drivers such as interest rates and income, the report finds.

“A key takeaway is that even if prices (both residential and commercial) have become more synchronised over the past decade, this doesn’t imply that we now have a global real estate market,” said Study Group Chair Paul Hilbers, Director of Financial Stability at the Netherlands Bank.

“Significant differences in cross country price dynamics reflect the strength of local drivers. Some drivers are more important in some countries than in others.”

A third highlight is evidence of the growing role of international investors in many markets. Policymakers have found they need alternative tools to deal with foreign buyers. These investors do not fund their purchases through local banks, so fiscal tools like higher stamp studies may be more effective than macroprudential policy.

The CGFS is a central bank forum for the monitoring and analysis of broad financial system issues. It supports central banks in the fulfilment of their responsibilities for monetary and financial stability by contributing appropriate policy recommendations.”

From Eurofund paper by Tadas Leončikas and Sevinç Rende:

• “In Europe, people living in the capital city generally have a better quality of life than people living in other parts of a country. On this basis, it seems that capital cities are indeed the best places to live.
• For most countries, residents of the capital city score higher on life satisfaction on average than people living outside the capital.
• Life satisfaction in a capital city for the most part is closer to the national average than to the averages of other capital cities. This finding suggests that, despite concerns that capitals increasingly operate independently of their nations, national-level factors are still important in shaping differences in well-being both between countries and between capital cities.
• Capital cities have, by and large, larger proportions of people who report feeling resilient – able to cope during times of hardship – compared to other urban centres and rural regions in the same country. Some characteristics of city populations – such as a younger age profile and higher educational attainment – contribute to resilience, while others, such as housing insecurity, erode it. The findings suggest that some other latent factor, possibly related to opportunities for economic advancement and improving one’s living standards, could underlie the extra resilience that capital cities provide.
• Capital city residents tend to be more satisfied with how democracy works in their country compared to the population outside the capital. This difference is largely associated with socioeconomic background, which on average is more advantageous in capital cities. Capital city residents also tend to have greater trust in national institutions and be more critical of local or municipal authorities than people in the rest of the country.”

On cross-country:

• ECB’s house price headache too big to solve – Reuters
• Guidelines for the Implementation of the Right to Adequate Housing – UN
• Are current measures of housing affordability fit for purpose? – Housing Europe
• Smashed avocado on toast and the housing market – ING
• What are macroprudential tools? – Brookings

On the US:

• S. Homeowners Four Times As Likely To Be Equity-Rich Than Seriously Underwater – ATTOM
• How to Make Housing Markets Fairer for All – E21
• The Fed Faces a Housing Conundrum – Bloomberg
• San Jose’s Severe Housing Shortage Is Easy to Fathom – Wall Street Journal
• Are you waiting for house prices to drop during the next recession to buy a home? Why you could have a very a long wait – MarketWatch
• History of Housing Policy in the United States – Portland State University
• Housing Sentiment Nears Survey High as More Consumers Expect Mortgage Rates to Remain Favorable – Fannie Mae
• The Geography of Housing Market Liquidity During the Great Recession – Federal Reserve Bank of St. Louis
• The GSEs and the Economic Cycle: Realistic Expectations – Harvard Joint Center for Housing Studies

On other countries:

• [Rwanda] Housing policies in Rwanda – International Growth Centre
• [United Kingdom] Where are we in the property cycle? – Financial Times
• [United Kingdom] Britons Priced Out of Housing Market Face Renting Into Old Age – Bloomberg
• [United Kingdom] K. Housing Market Sees Renewed Optimism in January, RICS Says – Bloomberg