Employment and Social Developments in Europe (ESDE) 2024

Chapter 3 - The role of social investment
© GettyImage
The role of social investment

4. Impact of housing policies on poverty reduction and upward social convergence

Access to social housing or housing assistance of good quality for those in need is enshrined in Principle 19 of the European Pillar of Social Rights and is a relevant determinant of people’s labour market outcomes.A number of social policies can complement and further support social investment. These include, for example, social protection and housing policies. Access to affordable and adequate housing is crucial for reducing poverty and social exclusion. It also broadens people’s ability to access opportunities in education and the labour market, and positively influences people’s physical and mental health. In its 2018 report, the High-Level Task Force on Investing in Social Infrastructure in Europe underlined the importance of investment in housing – particularly affordable housing – for promoting upward social convergence in the EU. Similarly to housing policies, social protection can also act as enabler for effective social investment. A focus on the role of social protection is provided in Box 3.7.

House prices, together with the available stock of housing and mortgage rates, affect the affordability of housing for prospective homeowners and tenants. Easier access to mortgages, including via low interest rates, usually increases demand for houses which can subsequently drive-up house prices. These reflect the value of real estate transactions, including second homes, holiday homes and dwellings used for investment, and thus have a limited direct impact on monthly housing costs for tenants in the short term. (195)However, by reducing the affordability of homeownership, higher house prices make renting more appealing, increasing demand for rentals and hence contributing to higher rents. Higher house prices and interest rates can also be taken into account by landlords when setting rents for new lease contracts, contributing to higher rents over time thus increasing housing costs for tenants.

Housing prices and rents have increased considerably over the last decade.After the substantial increase recorded between 2014 and 2022 (51.8%), house prices started to moderate in 2023 (-0.3%), due to weakening borrowing capacity following the increases in interest rates. However, rents continued their upward trend (+11.3% since 2014, and +3% in 2023), resulting in higher housing costs for tenants (Chart 3.14)

Housing policies can help to reduce the share of housing costs in household disposable income.(196)The share of housing costs in household disposable income decreased from 22.4% in 2014 to 18.5% in 2020. While still below 2014 levels, the share of housing costs increased again to 19.7% on average in the EU by 2023 (Chart 3.14) (197) This was likely driven by the increase in housing costs, mostly due to higher electricity and gas prices, as well as increases in rents and interest rates of mortgages. Member States with the highest shares of housing costs in household disposable income exhibited stronger reductions over this period. (198)

Chart 3.14
Share of housing costs in household disposable income declined but house prices increased sharply between 2014 and 2023

Weighted EU average share of housing costs in household disposable income and variation across countries (in standard deviation, left chart), and house price index and rent index (right chart), 2014-2023

Share of housing costs in household disposable income declined but house prices increased sharply between 2014 and 2023 Share of housing costs in household disposable income declined but house prices increased sharply between 2014 and 2023

Note: In the left chart, 2021 data exclude France, which did not report housing costs in 2021. Standard deviation is a measure of cross-country variation, the higher the standard deviation, the higher the cross country variation.

Source: Eurostat [ilc_mded01], [prc_hpi_a] and [prc_hicp_aind].

Housing policies can help improve housing affordability. At national level, housing policies can include the provision of social housing, housing allowances, tax deduction for housing costs, and utility subsidies, as well as rent regulation or subsidies for improving energy efficiency. (199) These policies often target specific population groups (e.g. low-income households, tenants vs homeowners) and can have other objectives, such as improving the quality of housing or increasing homeownership. Some housing policies, in particular housing allowances, need to be targeted to help contain the fiscal cost and ensure they are not translated into higher housing or rental prices. The impact of housing policies on housing cost reduction depends on many factors. These include their generosity, coverage and take-up, regulatory environment, stock of housing supply and requirements for renovation, as well as employment, environmental, urban and spatial policies. The analysis below examines how two housing policies, - housing allowances and social housing - contribute to poverty reduction. These two policies have been selected on the grounds of data availability. Housing allowances are intended to compensate for housing costs based on a means-test and can be granted to both tenants and owner-occupiers. (200)As a proxy for social housing, this report uses social rent subsidies for tenants, thereby capturing the difference between social rent and estimated market rent for the dwelling.(201)

Housing policies are relevant for the labour market outcomes of individuals. The lack of affordable housing may affect the accessibility to good quality education, training and job opportunities, and discourage labour mobility, increasing the probability of higher unemployment rates and contributing to labour and skills shortages. (202) Living in poor quality accommodation can negatively affect health and well-being, likely reducing productivity. In turn, worse labour market outcomes (e.g. being unemployed, frequently moving in and out of employment, working on a fixed contract, receiving a lower wage) might prevent housing conditions from improving in the absence of buffers. For example, they might reduce the ability to pay rents and the chances of obtaining a mortgage, especially in more expensive areas with better job opportunities, limiting mobility, or not providing sufficient means to improve housing quality. (203)The design of housing policies and their complementarities to other policies also matter. For instance, in the presence of high transaction costs, housing policies supporting home ownership might reduce labour mobility, increase the acceptance of lower wages or increase demand.

Increasing the supply of housing, including by investing in social housing, is very often essential to improve housing affordability, especially for households with lower incomes. Social dwellings are usually publicly owned (with countries investing in the stock of social housing, enabling them to benefit from an economic rent accruing to owners of property) or semi-privately owned (by housing corporations) but can also be privately owned. (204)Social housing usually takes the form of rental accommodation provided at below-market prices. In recent decades, social housing models in Europe have increasingly focused on lower-income households, primarily targeting older people and single-parents, while other vulnerable groups, such as single-person households, households without children, and migrants, are often low on the priority list or not eligible for social housing. (205) In addition to improving housing affordability, social housing can support macroeconomic objectives by easing supply bottlenecks, thereby mitigating house price pressures. This can in turn help to address labour shortages. (206) Social housing also plays a key role in housing first-type programmes for homeless people. In this respect, social and affordable housing policies underpin the EU social economy strategy. (207) More broadly, increasing the supply of housing can require a coherent mix of investments and reforms to address diverse issues such as the supply of basic infrastructure, spatial planning and other regulatory restrictions, or an insufficient supply of qualified labour.

The stock of social housing is quite low in the EU, with big investment gaps. The lack of affordable, good quality housing has several structural drivers, including low incentives, bottlenecks in construction and investment gaps regarding in particular energy efficiency and social housing production. The construction of social housing has decreased across the EU in recent decades, and some countries, particularly in Eastern Europe, have privatised large parts of the social housing stock. (208) Currently, social housing as a proportion of the overall housing stock exceeds 20% in only three Member States: the Netherlands (34.1% in 2021), Austria (23.6% in 2019) and Denmark (21.3% in 2022). It is between 10% and 20% in France (14% in 2018), Ireland (12.7% in 2016) and Finland (10.9% in 2021). (209) Long waiting lists are common in Member States with both larger and small social housing stocks, reducing the effectiveness of social housing in improving housing affordability for the most vulnerable households. (210) Long waiting lists are common in Member States with both larger and small social housing stocks, reducing the effectiveness of social housing in improving housing affordability for the most vulnerable households.

Chart 3.15
Public spending on housing increased but dispersion widened between 2014 and 2021

Weighted EU average government expenditure on housing and variation across countries (in standard deviation), 2014-2021

Public spending on housing benefits increased but dispersion widened between 2014 and 2021 Public spending on housing benefits increased but dispersion widened between 2014 and 2021

Note: Social housing is part of rent benefits and is included in the green line.

Source: Eurostat [spr_exp_fho].

Between 2014 and 2021, government expenditure on housing benefits per inhabitant, aimed at improving access to housing, increased in the EU.Taking into account differences in the cost of living between countries, it went from 109 to 120 PPS per inhabitant (211)on average, ranging from less than 1 PPS per inhabitant in Bulgaria, Croatia and Portugal to over 300 PPS per inhabitant in Ireland and Finland in 2021. It was strongly driven by expenditure on rent benefits, mostly comprising housing allowances (86.2% of rent benefits in 2021), with the exception of Belgium, Estonia, Ireland, Lithuania, the Netherlands, Romania and Slovenia, where social housing expenditure constituted a major share of rent benefits. Countries with the lowest initial levels of public spending on housing caught up more strongly between 2014 and 2021, (212)but the overall dispersion across countries widened (Chart 3.15). (213) In light of recent trends, many Europeans responding to a recent survey believe that addressing the high cost of living (48%) and the lack of social housing and homelessness (23%) should be a main priority in their country. Around 6 in 10 respondents think that their governments should increase their spending on housing, with variation across countries ranging from below 30% in Finland and Denmark to over 80% in Cyprus, Ireland and Greece. (214)

Chart 3.16
Housing allowances reduce the AROP rate

Differences in AROP and AROP rates before housing allowances, 2022

Housing allowances reduce the AROP rate

Note: AROP rate before housing allowances is calculated by excluding housing allowances from equivalised household disposable income and by calculating the share of people with an equivalised adjusted disposable income below 60% of the national median. AROP before housing allowances rate is missing for Romania, which does not collect data on housing allowances.

Source: EU-SILC scientific use files, 2022.

Social housing and housing allowances reduce the AROP rate in the short term, without accounting for the possible longer-term effects on housing prices and housing affordability. Comparing the adjusted AROP rate, which excludes social rent subsidies and/or housing allowances from equivalised household disposable income,(215) to the standard AROP rate provides a good measure of the impact of both policy instruments in reducing poverty. It does not however capture other relevant aspects of housing affordability such as supply and demand for affordable housing, underlying drivers or broader societal impacts. In 2022, housing allowances decreased the AROP rate by an estimated 1.4 pp (from 17.7% based on AROP rate before housing allowances to 16.3% of AROP rate) in the EU on average, with the biggest decreases in Finland (4.3 pp), Ireland (3.5 pp), France (3.1 pp) and Germany (3.0 pp) (Chart 3.16). (216) Similarly, in 2019, the provision of social housing proxied through social rent subsidies decreased the AROP rate in the EU by an estimated 0.4 pp on average (from 18.1% in AROP rate before housing allowances and social rent subsidies to 17.7% in AROP rate before housing allowances). (217) The highest decreases were observed in Ireland (3.7 pp) and Belgium (2.4 pp). The higher average impact of housing allowances could be due to relatively low public spending on social housing (Chart 3.15). While this analysis suggests that social housing and housing allowances reduce poverty, it does not account for the impact of other related policies or institutional setups and bottlenecks relevant for the functioning of rental and housing markets, thus likely reflecting the upper bound of the impact. For instance, the potential shift of (part of) housing allowances to landlords through higher rents might reduce the effectiveness of housing allowances in reducing poverty, i.e. in the absence of effective rent regulation policies. (218) Depending on the design, however, housing allowances might instead drive-up housing prices and reduce housing affordability over time, especially in markets with limited housing supply. (219)

The estimated direct impact of housing allowances on poverty reduction is particularly strong for vulnerable households. For example, households where no adult (+3.2 pp) or only one adult (+1.8 pp) is employed are more likely to be at risk of poverty after housing allowances are excluded from equivalised disposable income, compared to households where more than one adult is employed (Chart 3.17). Similarly, this probability is higher for tenants (+2.7 pp) and owners with a mortgage (+0.7 pp), compared to outright owners. Households with children (+2.3 pp), households with no adult with tertiary education (+0.9 pp), and households with at least one adult with a migration background (220) (+0.5 pp) are also more likely to be at risk of poverty after housing allowances are excluded from equivalised disposable income.

Chart 3.17
Housing allowances contribute substantially to reducing poverty among vulnerable households

Probability of a status change from not at-risk-of-poverty to at-risk-of-poverty after housing allowances are excluded from equivalised disposable income, 2014-2022

Housing allowances contribute substantially to reducing poverty among vulnerable households

Note: Excludes Romania, as it does not collect data on housing allowances. All estimates significant at 1% significance level. Reference categories in brackets: no adult working in the household, one adult working in the household (more than one adult working in the household); tenant, owner with mortgage (outright owner); household with children (household with no children); no adult with tertiary education in the household (at least one adult with tertiary education in the household); at least one adult with migration background in the household (no adult with migration background in the household). Model controls for country and year fixed effects.

Source: EU-SILC 2014-2022.

Box 3.7: Promoting upward social convergence and poverty reduction through social protection

Well-designed, inclusive social protection systems and social investment policies complement and reinforce one another as necessary components of modern welfare states. Social protection systems provide a safety net that protects individuals from economic and social risks by helping them maintain at least a basic level of income and thus contributing to macroeconomic stabilisation. They can mitigate the impacts on poverty and economic growth by increasing income and consumption, particularly through countercyclical spending during economic downturns. Evidence also points to complementarities between social protection and social investment, with spending on ALMPs and ECEC tending to have more positive effects where total social protection expenditure is more generous. (1)

Income stabilisers have the potential to contain poverty in times of economic downturn. The shock-absorption properties of effective tax-benefit systems can prevent diverging trends in inequality and social outcomes during recession and strengthen the resilience of a country. The analysis below (2) uses EUROMOD to quantify how the overall tax-benefit system mitigates the transmission of a market income shock to disposable household incomes. It simulates a 5% hypothetical reduction in gross market income (3) and estimates the share of income loss absorbed by a country’s tax-benefit system. As no changes in labour market status or prices are considered, (4) income stabilisation is driven by increases in means-tested benefits and reductions in taxes (5) and lower social insurance contributions (SICs). Stabilisation properties are measured through an income stabilisation coefficient (ISC), which varies between 0% and 100%, with higher values pointing to stronger stabilisation (Box A3.3 in the Annex)

Chart 1
Substantial variation in the extent and composition of income stabilisation after a 5% market income shock across Member States

Country-level ISC, by policy instrument (%), 2022-2023 average

Substantial variation in the extent and composition of income stabilisation after a 5% market income shock across Member States

Source: JRC calculations based on EUROMOD, version I6.0+.

Tax-benefit systems can stabilise incomes, particularly through direct taxes yet with substantial heterogeneity across Member States. On average, tax-benefit systems in the EU in 2022-2023 would have absorbed almost half of the simulated 5% market income shock, with the ISC averaging 41.4% across the Member States. (6) Direct taxes would have absorbed 29.4% of a market income shock in 2022-2023, followed by SICs (9.8%) and means-tested benefits (<0.5%). However, the variation between countries is substantial, ranging from 17% in Bulgaria to 58.9% in Belgium (Chart 1). Within the different tax-benefit components, direct taxes would have absorbed 29.4% of a market income shock in 2022-2023, followed by SICs (9.8%) and means-tested benefits (<0.5%). Also for the various components, considerable differences emerge between Member States. In five Member States, SICs (rather than direct taxes) represent the largest component of income stabilisation. (7) The degree of income stabilisation provided by each component also varies by Member State, ranging from 6.3% in Romania to 45.6% in Denmark for direct taxes; from 0% in Denmark (8) to 32.4% in Romania for SICs; and from less than 0.5% in Hungary, Poland, Estonia and Latvia to 5.3% in France for means-tested benefits (Chart 1).

The degree and composition of income stabilisation after an income shock varies between income quintiles. On average, incomes at the bottom (43.5%) and top (42.6%) of the income distribution are more cushioned against a market income shock than those in the middle (39.7%) (Chart 2) While increased benefits lead to a larger shock-absorption among households in the bottom income quintile (providing a stabilisation of 18.6%), the reduction in taxes due to the income loss among households in the top income quintile leads to an almost equally high degree of income stabilisation as for the bottom quintile, despite almost no increase in benefits (0.2% stabilisation). (9) Similarly, personal income taxes absorb an increasingly higher share of the market income shock for the higher the income quintile, increasing from 12.3% for households in the first quintile to as much as 35.3% for households in the fifth quintile.

Chart 2
Relative significance of taxes and benefits in absorbing a 5% market income shock varies substantially across the income distribution

EU-level ISC, by income quintile and policy instrument (%) (2022/2023 average)

Relative significance of taxes and benefits in absorbing a 5% market income shock varies substantially across the income distribution

Source: JRC calculations based on EUROMOD, version I6.0+.


  • 1. (European Commission, 2016)
  • 2. While social protection addresses both the risk of economic downturn and risks that occur in the context of sickness or retirement, this section will focus on the role of social protection in stabilising incomes.
  • 3. Market income includes employment and self-employment incomes, investment and property incomes, pensions from individual private plans, and regular net inter-household transfers, all reported in gross terms.
  • 4. While social protection addresses both the risk of economic downturn and risks that occur in the context of sickness or retirement, this section will focus on the role of social protection in stabilising incomes.
  • 5. Due to the progressive nature of personal income taxation.
  • 6. EU-level ISC computed by aggregating changes in market income and disposable incomes across countries. The result is an EU-level weighted average, with the shares of country-specific market income shock out of total market income shock as the weighting factor. EU-level results in this analysis are thus more influenced by larger countries.
  • 7. Bulgaria, Croatia, Hungary, Romania, Slovakia.
  • 8. Result for Denmark follows from the classification of labour market contributions in EUROMOD as taxes rather than SIC and from the fact that unemployment benefit contributions and supplementary labour market contributions do not depend directly on earnings.
  • 9. Chart A3.3 in the Annex presents the relative significance of taxes and benefits in absorbing a 5% market income shock, by income quintile and Member State.

Notes

  • 195.(European Commission, 2019a).
  • 196.Housing costs include mortgage interest payments for main dwelling net of any tax relief (for owners), rental payments (for tenants), structural insurance, mandatory services and charges (sewage removal, refuse removal, etc.), regular maintenance and repairs, taxes on dwelling, and the cost of utilities (water, electricity, gas and heating). Mortgage interest payments and rental payments are considered gross of housing benefits (i.e. housing benefits are not deducted from the total housing cost). For tenants, only those costs paid by tenants are taken into account (as opposed to those paid by landlords). For more analysis of housing affordability, see (European Commission, 2019a).
  • 197.Housing costs analysed constitute average costs as a share of disposable income, thus mark considerable heterogeneity by different factors such as location, as well as household composition and potential adjustments in the choice of dwelling and behavioural responses (in terms of the type of housing chosen) to changes in housing costs or in income.
  • 198.Beta coefficient is negative (-0.04) and statistically significant at 5% significance level.
  • 199.See (Eurofound, 2023b) for a more detailed mapping of housing policies in the Member States.
  • 200.Housing allowances compensate for housing costs, such as rent, gas, electricity, heating, water or utility bills. They can be granted to both tenants and owner-occupiers, and do not include social housing policy organised through the tax-benefit system and capital transfers (notably investment grants).
  • 201.Social rent subsidies (as a proxy for social housing) are estimated using the imputed rents variable (HY030G) for social tenants (those who indicated paying below-the-market rent) from EU-SILC. The imputed value is the equivalent market rent that would be paid for a similar dwelling to that occupied, reduced by any rent actually paid.
  • 202.(Borg and Brandén, 2018).
  • 203.(Arundel and John, 2017).
  • 204.(Eurofound, 2023b) ; (European Commission, 2023f).
  • 205.(Eurofound, 2023b); (Scanlon, Fernández Arrigoitia and Whitehead, 2015); (Heylen, 2024).
  • 206.(Whitehead, 2017); (European Commission, 2023f).
  • 207.(European Commission, 2021c).
  • 208.(Scanlon, Fernández Arrigoitia and Whitehead, 2015) ; (Whitehead, 2017) ; (Eurofound, 2023b).
  • 209.See OECD Affordable Housing Database here
  • 210.(Eurofound, 2023b); (Pestel-Institut, 2024).
  • 211.PPS is the artificial currency unit that eliminates price level differences between countries, i.e. one PPS can buy the same volume of goods and services in all countries. Government expenditure on housing in PPS is derived by dividing the government expenditure on housing of a country in national currency by the respective PPPs, which are obtained by comparing price levels for a basket of comparable goods and services representative of consumption patterns in the various countries.
  • 212.Beta coefficient is negative (-0.03) and statistically significant at 5% significance level, primarily driven by the convergence in rent benefits (beta coefficient for benefits to owner occupiers is not statistically significant).
  • 213.While countries with the lowest initial levels catch up relatively to those with the highest initial level of public spending on housing, changes in the public spending of other countries contributes to increased variance thus increasing overall dispersion.
  • 214.2024 Eurobarometer on Social Europe; 2022 Eurobarometer on Fairness, Inequality and Intergenerational Mobility.
  • 215.Share of people with an equivalised adjusted disposable income below 60% of the national median, and further referred to as AROP before housing allowances and social rent subsidies, and AROP before housing allowances, respectively.
  • 216.Romania is excluded from both averages, as it does not collect data on housing allowances (variable HY070G in EU-SILC). The reported results assume that no housing allowances are shifted towards landlords through higher rents. The rather small impact of housing allowances in reducing the AROP rate in some countries is largely due to low spending on these policies. However, part of the differences in impact on poverty reduction between countries might also be driven by the design of housing allowances, with some countries targeting households at risk of poverty more effectively.
  • 217.Social rent subsidies (as a proxy for social housing) are estimated using the imputed rents variable (HY030G) for social tenants (those who indicated paying below-the-market rent) from EU-SILC. Since 2021, this variable is no longer part of the annual EU-SILC but will be provided on a three-yearly basis as part of the module “labour and housing conditions” (starting in 2023). Some countries (Denmark, Germany, Malta, the Netherlands, Slovakia, Sweden) did not collect data on imputed rents for social tenants for most years considered and are excluded from the analysis on social housing, in addition to Romania (where housing allowances do not exist). While data for social rent subsidies (imputed rents) are available until 2020, data for 2019 were used to avoid estimating the impact of temporary measures adopted during the COVID-19 pandemic.
  • 218.(Laferrère and Le Blanc, 2004); (Viren, 2013).
  • 219.(International Monetary Fund, 2024); (Hyslop, 2019). Defined as being born outside the country of residence and/or having a foreign citizenship.
  • 220.Defined as being born outside the country of residence and/or having a foreign citizenship.