• Loading stock data...
Finance Banking Featured World

Housing wealth across countries: The role of expectations, institutions, and preferences

Homeownership rates and housing wealth differ immensely across countries. Using a life-cycle model with housing wealth estimated on micro data from five large economies, this column provides a decomposition of the drivers of long-run, structural differences in housing wealth. It finds that expectations of house prices, the institutional setup of the housing market, and household preferences all matter, although preferences contribute less. Differences in homeownership rates are strongly affected by house price beliefs and the rental wedge, which reflects the quality of the rental market. Differences in the quantity of housing wealth are substantially driven by housing maintenance costs.

The main household residence is the largest asset for most households and an important determinant of wealth inequality (Paz-Pardo 2022, Daysal et al. 2023). Homeownership rates differ enormously across countries. For example, less than half of all households in Germany own their residence, while 80% of households in Spain are homeowners (Figure 1, Panel A). These striking cross-country differences in homeownership persist over the whole life cycle, with the homeownership gap between Germany and Spain staying around 30 percentage points, not narrowing with age (Figure 1, Panel B). In addition, homeowners in different countries accumulate substantially different amounts of housing wealth (Figure 1, Panels C and D).

Figure 1 The homeownership rate and mean housing wealth

Figure 1 The homeownership rate and mean housing wealth
Figure 1 The homeownership rate and mean housing wealth
Source: Household Finance and Consumption Survey, wave 2014; Survey of Consumer Finances 2016.
Note: The upper panels show the homeownership rate across countries and over the life cycle. The lower panels show the mean housing wealth of homeowners (in thousands of euros) across countries and over the life cycle. The homeownership rate is measured using the indicator of ownership of the household main residence.

Life cycle model of illiquid housing wealth

In a new paper (Le Blanc et al. 2025), we analyse how households accumulate housing wealth over the life cycle in a model (based on Li and Yao 2007) with risky labour income and house prices, illiquid housing wealth, and a discrete-continuous choice between owning and renting a house. We estimate the model on micro data on age profiles of homeownership, housing wealth, rents, and net wealth from a comprehensive set of five advanced economies: France, Germany, Italy, Spain, and the US. The paper complements existing work, which typically investigates cross-country differences in wealth using reduced-form regressions.

Our estimated model allows us to quantify three groups of explanatory factors for long-run, structural differences in housing wealth. First, in line with survey evidence, we allow for (persistent) differences in expectations of house prices across individuals within each country (Landvoigt 2017, Kuchler et al. 2023). Second, countries differ in the institutional set-up of the housing and rental market (maximum loan-value ratios, costs of renting, maintaining, and selling a house) (e.g. Greenwald and Guren 2021, Malmendier and Steiny Wellsjo 2024). Third, preference parameters such as impatience and the share of housing expenditures are allowed to vary across households (Epper et al. 2020, Calvet et al. 2024). Using micro and macro data, we also calibrate the remaining differences in house prices, incomes, and demographics. Our model thus includes several features that are important for modelling housing wealth: housing wealth is illiquid and subject to linear house selling costs, house size is continuous, house prices are risky, households face collateral constraints, and their beliefs about house prices differ.

Decomposing the cross-country differences in housing wealth

Our model fits empirical age profiles of homeownership rates and holdings of housing and total net wealth for each of the five countries reasonably well. Through the lens of the estimated model, we then interpret the substantial differences in homeownership rates and housing wealth across the five countries. We propose a decomposition in which, moving from one country (e.g. Germany) to another (e.g. Spain), we switch one by one from the German parameter values to the Spanish ones, in each step recording the effect of the given factor on the housing wealth variable (homeownership or mean housing wealth-income ratio). We find that all three groups of factors above contribute to explaining the cross-country differences in homeownership and housing wealth, although preferences contribute much less than house price beliefs and housing market institutions.

Explaining the differences in homeownership rates

As for the extensive margin of housing wealth (Figure 2), differences in homeownership rates are strongly affected by two variables: (i) house price beliefs and (ii) the rental wedge, the difference between rents and maintenance costs, which reflects the quality of the rental market and the segmentation between rental and owner-occupied housing markets. These two factors are key for the decision whether to buy versus rent a house: a higher rental wedge and higher expectations of house price growth make renting less appealing and increase the share of homeowners. Both elements contribute roughly equally to explaining the gaps in homeownership rates across countries and they both matter throughout the life cycle.

Figure 2 Decomposition of homeownership rates

Figure 2 Decomposition of homeownership rates
Figure 2 Decomposition of homeownership rates
Note: The figure illustrates how the gaps in the homeownership rate between Germany and other countries are explained by various factors. The dark blue bars show the fitted homeownership rate in the base country (Germany). The other bars reflect the impact of various factors on homeownership, averaged across the permutations of factors. The sum of all bars results in the homeownership rate in the second country (Spain, the US, France, or Italy). ‘Other factors’ include mortality, house selling costs, realised house price growth, and its variance and interest rate.

Quantitatively, the two channels are powerful: small differences in long-run house price beliefs and the rental wedge result in large differences in homeownership rates. The rental wedge ranges from around 2% in France and the US to 2.8% in Germany, 3.7% in Spain, and almost 5% in Italy, reflecting a less efficient rental market. Our model implies that the 2 percentage point difference in rental wedges roughly leads to a 25-30 percentage point difference in homeownership rates between Germany and Italy. Mean long-run house price beliefs range between 0% in Italy and 2.8% in France, reflecting the historical growth in aggregate house prices. Across countries, a one percentage point difference in house price beliefs results in roughly a 15 percentage point difference in the homeownership rate. These considerations imply that small differences in long-run house price beliefs − well within the range documented in survey data − are a powerful driver of homeownership in a model, substantially affecting important economic decisions of households.

As for other factors, tighter collateral constraints and steeply growing labour incomes in Germany and the US reduce the homeownership rate particularly among the youngest households, while the bequest motive affects the homeownership in particular among older households.

Explaining the differences in mean housing wealth

Regarding the intensive margin of housing wealth (Figure 3), differences in the value of housing wealth of homeowners (as measured by the mean ratios of housing wealth to income) are mostly driven by housing maintenance costs, which in effect reduce homeowners’ return on housing wealth. Quantitatively, the estimated maintenance costs for Germany (2.6% of housing wealth) are roughly half those in Spain, France, and the US (around 5% or more), implying higher housing wealth in Germany by a multiple of 2-4 times annual incomes. Other factors that matter for the accumulation of housing wealth (although less than maintenance costs) are the housing preference, house price beliefs, and the rental wedge. We estimate that Germany and Italy have a lower share of housing utility (around 0.20) than the other countries (roughly 0.30), which is reflected in a positive contribution of the parameter to housing wealth outside of Germany. Roughly twice as large as in Germany, the rental wedge in Italy reduces housing wealth as marginal buyers purchase smaller houses. Higher house price growth beliefs increase the amount of housing wealth in Spain and the US (compared to Germany) as existing homeowners upgrade to buy larger houses.

Figure 3 Decomposition of mean housing wealth-income ratios of homeowners

Figure 3 Decomposition of mean housing wealth-income ratios of homeowners
Figure 3 Decomposition of mean housing wealth-income ratios of homeowners
Note: The figure illustrates how the gaps in the mean housing wealth-income ratios of homeowners between Germany and other countries are explained by various factors. The dark blue bars show the fitted mean housing wealth-income ratios in the base country (Germany). The other bars reflect the impact of various factors on housing wealth, averaged across the permutations of factors. The sum of all bars results in the housing wealth-income ratios in the second country (Spain, the US, France, or Italy). ‘Other factors’ include mortality, transaction costs, realised house price growth, and its variance and interest rate.

Conclusions

Our paper focuses on the long-run, structural differences in housing wealth across countries and could be extended in several ways. Our setup could be used to analyse how various economies respond to shocks and economic policies (including fiscal, monetary, and macro-prudential) at higher, business-cycle frequencies. Our partial equilibrium model could also be embedded in a general equilibrium framework to investigate feedback between direct and indirect effects of shocks. It could be studied in more detail which supply-side or demand-side factors affect the rental wedge, for example, the history of institutions, cultural factors, and experiences of memorable inflation rates and housing returns. Future work could also zoom in on population groups, for example middle class or young households, and study how their homeownership status and accumulation of wealth are affected by shocks and housing market institutions.

Source : VOXeu

GLOBAL BUSINESS AND FINANCE MAGAZINE

GLOBAL BUSINESS AND FINANCE MAGAZINE

About Author

Leave a comment

Your email address will not be published. Required fields are marked *

You may also like

World

Openness to trade and regional growth: Evidence from Italy during the First Globalisation

The economic, social, and political consequences of globalisation have been a hot topic in the public debate over the last
Banking

Banks’ exposures to high-carbon assets may represent a medium-term vulnerability for the financial system

Climate change is rapidly being recognised as a potential source of financial risk by regulators and supervisors (Claessens et al.