Financing of Social Protection

Report

9 September 2026

3. Securing the Revenue Base for Social Protection

Copy link to 3. Securing the Revenue Base for Social Protection

Abstract

Labour income, the main financing source of social protection, is under pressure: population ageing lowers the number of hours worked, while new technologies, in particular AI, may increase capital productivity and lower the labour share of income. This chapter quantifies the effect of population ageing on the workforce potential in the medium term, charts the development of the labour share of since the 1970s and discusses the potential impact AI may have in the future. It goes on to review how social protection funding could be diversified and how countries could strengthen the link between entitlements and contributions. It discusses the potential of closing contribution gaps in undeclared work and reports reform trends in EU and OECD countries.

About half of spending on social protection is financed through social contributions, which are mainly1 levied on labour incomes (see Chapter 1). Labour income, however, is under pressure from several directions: population ageing decreases the size of the potential labour force and thus the aggregate wage bill, assuming no large increases in migration or employment. And while Artificial Intelligence (AI) deployment does not yet seem to have led to (net) job destruction (OECD, 2026[1]), it may raise the share of capital in total income through higher capital productivity. At the same time, demographic, technological and environmental challenges will add spending pressures on social protection budgets. Given the mismatch between growing demands and a potentially eroding funding base, countries are considering options to use alternative funding sources.

This chapter starts out with quantifying the effect of population ageing on the workforce potential in the medium term and looks at the potential of increasing employment, in particular by extending working lives and hours worked. It also reviews policies countries can use to maximise employment in the context of population ageing (Section 3.1). It then charts the development of the labour share of income across EU and OECD countries since the 1970s and discusses the potential impact AI may have in the future (Section 3.2). It goes on to review how social protection funding could be diversified (Section 3.3) and how countries could strengthen the link between entitlements and contributions (Section 3.4). Section 3.5 adds a more operational perspective by looking at closing contribution gaps in undeclared work. Section 3.6 reports reform trends in EU and OECD countries drawing on a dedicated policy questionnaire.

3.1. Maximising the labour potential

Copy link to 3.1. Maximising the labour potential

Shifts in the age composition of the population reflect the combined effect of persistently low fertility rates and people living longer. While the rate of decline in fertility has slowed, fertility rates continue to fall across the OECD on average and are below the replacement level in almost all OECD and EU countries. At the same time, remaining life expectancy at age 65 increased at a pace of about 1.5 years per decade from the 1990s to the 2010s. This pace has slowed to around one year by decade across the OECD, and to 0.5 years in the EU between 2015 and 2024, but longevity is expected to continue to increase (OECD, 2024[2]; Eurostat, 2026[3]). This will lead to a decline in the working-age population in most OECD countries, and an even stronger decline of its share in the population, leading to an increase in the old-age dependency ratio across the OECD on average from 31% in 2023 to 50% in 2060, and across the EU on average from 37% to 56% in 2060 (OECD, 2026[4]).

Without significant increases in labour productivity – e.g. through technological progress, including AI (see Section 3.2), and without increases in employment or hours worked – GDP per capita growth is projected to fall significantly (OECD, 2025[5]).

3.1.1. Population ageing will result in fewer hours worked, with a direct effect on total labour income

Figure 3.1 shows a simulation of the total decrease in hours worked in OECD and EU economies until 2060, drawing on the medium scenario of UN population projections, combined with projected employment rates for each gender-age group2 according to the methodology developed in OECD (2025[5]). It converts these employment numbers into total hours worked by assuming average hours worked by age and gender to remain constant at 2025 values. The hours margin is important in the context of financing of social protection as tax rates and social contributions can influence labour supply decision at the intensive margin (hours worked). Migration is not considered in this simulation exercise because of a lack of country-specific data on hours worked by sex, age and migration status. Migration is also more difficult to forecast as it is less structural than fertility and life expectancy. Assuming the net migration rate of the UN central population scenario (that is below current levels in many countries), the impact of migration on employment is projected to be moderate (OECD, 2025[5]).3

According to this exercise, almost all EU and OECD countries will experience a decline in total hours worked if employment rates evolve in line with projected age‑ and gender-specific trends and hours worked by worker remain at their 2023 age‑ and gender-specific levels. This is mainly driven by a decrease in the size of the working-age population, but also by a composition effect: older workers, whose employment rate is overall lower than that of prime‑age workers, will account for a larger share of the working-age population in 2060. Across the OECD on average, hours worked are projected to decrease by around 9% by 2060; across EU countries, which are ageing faster, on average, the decline is projected to be almost 20%. Declines of 33% and more are projected in the Slovak Republic, Poland, Lithuania, Bulgaria and Korea (Figure 3.1).

Figure 3.1. Without significant increases in employment, total hours worked will fall in most countries
Copy link to Figure 3.1. Without significant increases in employment, total hours worked will fall in most countries

Projected decline in hours worked from population ageing and potential gains in total hours worked from combining various scenarios, 2024‑2060, in per cent of total hours worked in 2023

3.1.2. This decline could in principle be compensated by maximising the labour force potential

By increasing employment among groups with currently low labour force attachment and increasing working hours to countries with higher work intensity for all groups, total hours worked could, in principle, be held constant across the OECD and the EU on average. In some countries, however, the decline in hours worked is too big to be entirely compensated from maximising the employment potential. Figure 3.1 considers four distinct levers:

1. By increasing employment rates of older workers to the levels of the best performing countries,4 OECD countries could, on average, increase hours worked by 6 percentage points (p.p.) and EU countries by 8 p.p. Luxembourg, Spain, Italy, Romania and Slovenia could gain more than 12 p.p. in hours worked by increasing employment rates among older workers.

The main lever countries have used to increase employment at older ages is to increase statutory retirement ages and limit avenues for early retirement. Current labour market entrants, who will be older workers in the 2060s, will face a statutory retirement age of 66 years across the OECD on average, compared to 64 for those retiring today, assuming current legislation is implemented (OECD, 2025[7]). Similarly, current labour market entrants across the EU will face an average statutory retirement age of 67 years, compared to 64 years for those retiring today.

Higher statutory retirement ages need to be underpinned by the promotion of occupational health, including adapted working conditions where necessary, life‑long learning and other employment and activation measures to enable workers to extend their careers further (OECD, 2024[2]).

2. Closing the gender gap among older workers would increase aggregate working hours by 4 p.p. across the EU and OECD average – with highest potential gains (10+ p.p.) for Colombia, Mexico, Türkiye and Costa Rica. Among EU countries, gains would be highest in Greece, Romania, and Italy (6+ p.p.).

The gap in statutory retirement ages is already expected to narrow from 0.8 years across the OECD and EU on average to 0.5 years across the OECD, and to 0.3 years across the EU on average for those entering the labour market today and retiring in the 2060s. Only few countries still have lower retirement ages for women than for men. Enabling women to work until the statutory retirement age will require significant investments in the care service infrastructure. Demands for elderly care will increase with population ageing.5 In the absence of an increase in the provision of affordable and high-quality care services, as well as a fairer split of unpaid work between women and men, this care will most likely continue to fall disproportionately on women, especially women over the age of 50, limiting their career prospects and earnings potential (OECD, 2024[2]; OECD, 2025[7]). Stronger female labour force attachment during the prime‑age working years will also likely increase employment at older ages and result in higher pension entitlements for women who are on average at higher risk of old-age poverty. For instance, in the EU in 2024, women over the age of 64 were 1.3 times as likely to be at-risk of poverty and social exclusion than men, and poverty risks are especially high for women living alone (European Commission, 2026[8]; European Commission and SPC, 2024[9]).

3. Closing the gender gap in employment among young and prime‑age workers has the potential to increase hours worked by 3 p.p. across the OECD average, and by 2 p.p. across the EU on average. Potential gains across the OECD are highest in Türkiye, Mexico and Colombia with over 10 p.p.; highest potential gains among EU countries could be achieved in Romania, Italy, Czechia and Greece with over 4 p.p.

As with older women, raising employment among prime‑aged women will require investments in the provision of care services, as a lack of adequate childcare provision is still a major barrier to work for many working-age people, especially women (OECD, 2024[2]).

4. Finally, raising hours worked among prime‑aged and older workers to the level of high-hours countries6 has the potential to increase working hours by 5 p.p. across the OECD and 6 p.p. across the EU on average, with the highest gains in Denmark, the Netherlands and Finland, where part-time rates of both men and women are high (see Chapter 2).

This may be the most ambitious, and least realistic adjustment. While investments in the child- and long-term care infrastructure can enable those with caring responsibilities to increase their working hours, the trend towards part-time work may at least partly be driven by an income effect (workers increasingly prefer leisure at higher incomes). Countries should assess their Tax-Benefit systems with a view to increasing incentives for full-time work (see Chapter 2), but it is unlikely that this will lead to the high effect sizes depicted in Figure 3.1.

3.1.3. But keeping hours worked constant is unrealistic even with decisive government action

Facilitating such an increase in labour supply would require decisive and costly government action, including significant investments in the provision of child- and long-term care, and it is unlikely that the full potential would be reached even if countries implemented the necessary reforms (OECD, 2025[5]). Lowering effective tax rates on labour (especially social contributions that are more important at low incomes) by broadening the funding base for social protection (see Section 3.3) may increase labour supply, but this effect is very likely to fall short of reaching the full employment potential.

Without the hours lever that is least likely to be fully realised, hours worked in 2060 are projected to be significantly lower than in 2023 in most OECD and EU countries. Increasing employment will therefore only be part of the solution to secure social protection financing.

3.1.4. The effect of ageing on labour productivity is ambiguous

It is unclear whether population ageing will drag down labour productivity – while productivity appears to decline at some point as workers age, this point may be quite late in their careers and is likely to depend on individual- and workplace characteristics. Older workers often take on roles, such as management, that require social skills and experience, that do not deteriorate with age, or to a lesser extent (OECD, 2020[10]). Older workers are also less likely to switch jobs and firms than younger workers, which reduces the quality of the average match and lowers productivity growth (OECD, 2025[5]). Ageing may also lead to a less dynamic business environment, as ageing consumers retain familiar consumption patterns, leading to fewer firm entries and less competition. Ageing may also decrease productivity, because older people demand services such as long-term care, that experience lower productivity growth (OECD, 2025[5]). On the other hand, population ageing may increase labour productivity, because ageing-driven labour shortages can incentivise technological innovation (see Section 3.2), and a larger capital stock may also increase productivity (see Section 3.3).

3.2. Trends in the labour income share

Copy link to 3.2. Trends in the labour income share

Since social contributions are mainly levied on labour income, the share of GDP accruing to labour is the most important structural factor driving the sustainability of social protection financing. This section charts the development of the labour income share across EU and OECD countries since the 1970s. It discusses reasons for its decline up until the early 2000s and examines the potential impact technological progress, in particular AI, may have in the future. It closes with policy implications.

3.2.1. The labour share has declined from the 1970s to the 1990s, but has been largely stable since then

There is a consensus in the literature that the labour share has declined since the 1970s in most developed countries (Karabarbounis, 2024[11]), but the extent of the fall depends on the time period that is observed and varies across countries. Major European economies such as France, Germany and Italy, as well as Austria, Belgium and Sweden, the United States, the United Kingdom, Australia and Mexico, show a clear decline in one measure of the labour share (employee compensation as a share of Gross Value Added, GVA7) from the late 1970s to the early 1990s (Figure 3.2).

Since then, the labour share seems stable or even rebounding: at the OECD and EU average (only available since 1995) the labour share seems roughly constant over the past three decades, with a dip during the Global Financial Crisis (GFC) and the COVID‑19 crisis caused by a lower wage sum because of higher unemployment, followed by a rebound. See also (European Commission, 2023[12]). Some European countries, such as France, Germany, Austria, Belgium and Sweden, show a strong increase in the late 1970s, then a decline followed by a rebound after the GFC. In the United States and Australia, the decline seems more sustained and continuing, while in the United Kingdom, the labour share seems to have fully recovered from the mid‑1990s dip. Hungary and Poland show some movement but no clear trend in the labour share since 1997 (start of available data), while Japan follows a cyclical trend.

Estimates of the labour share are quite sensitive to statistical definitions and tax incentives to shift income between tax bases, which explains some of these cross-country differences, see Box 3.1. For instance, Figure 3.2 attributes all income of self-employed workers to capital because a split in capital and labour income is not available. Countries with a higher share of self-employment therefore have a lower labour share. The incidence of self-employment is much higher in Italy (21%) and Poland (19%) compared to countries such as in France (13%), Hungary (12%) or Germany (9%, see Chapter 2).

Figure 3.2. The labour share of income has been relatively stable in recent years
Copy link to Figure 3.2. The labour share of income has been relatively stable in recent years

Employee compensation in per cent of Gross Value Added (GVA) in selected OECD economies, three‑year moving average, earliest available to 2024

Box 3.1. Accounting changes can influence the labour share

Copy link to Box 3.1. Accounting changes can influence the labour share

The income of sole proprietors / unincorporated businesses is a mix between returns to capital and own labour input (e.g. self-employed taxi- or uber drivers operate their own car). Income of unincorporated enterprises cannot be isolated from GVA in Figure 3.2 due to data constraints. Figure 3.2 therefore assumes that all income of unincorporated businesses accrues to capital, which is clearly not the case and biases the estimate of the labour share downward. Estimates of the labour share therefore often look at the corporate sector only (Karabarbounis, 2024[11]).

Accounting changes – e.g. business owners shifting income from labour to profit for tax reasons – can also impact the labour share. Smith et al. (2022[13]) show that in the United States, accounting shifts from labour income to profits by owner-managers accounted for about one‑third of the decline in the labour share between 1978 and 2017.1

Another question is whether the real estate sector should be excluded from GVA when calculating the labour share. GVA includes rents and imputed rents of owner-occupiers which are subject to significant measurement errors. As the only labour cost in the real estate sector is that of real-estate workers, actual and imputed rents for homes dominate the real estate sector and can lead to fluctuations in the labour share driven by housing market developments, especially if the real estate sector is large. These fluctuations are decoupled from the labour market and labour productivity, which is why they are sometimes excluded. For instance, the decline in the labour share across the OECD on average between 1995 and 2022 declines from 2.3 to 1 p.p. when excluding primary and real estate activities (OECD, 2024[14]).

However, when using the labour share as an indicator of the overall distribution of economic output into returns to capital and labour, and for assessing how each contribute to public financing, it is important to include rents accruing to owner occupiers.

1. Two factors contributed to this change: owner-managers shifted their income from labour to profits following a 1986 tax reform. Additionally, some labour-intensive activities – mainly financial, legal or consulting services, as well as doctors’ or dentists’ practices – were increasingly organised in tax-preferred partnerships, and not in the corporate sector where income is divided in capital and labour. These capital-poor activities outstripped the growth of the corporate sector, especially since around the year 2000.

3.2.2. There is no consensus in the literature on the causes of the labour share decline

Several explanations have been suggested for the fall of the labour share, including de‑unionisation and an associated decline in the bargaining power of workers. Declining worker bargaining power is consistent with wage growth falling behind productivity growth. But the evidence of it being a major driver in the decline of the labour share in the 1990s is inconclusive.

First, labour share declines also happened in countries where union density remained high such as Austria, Belgium and Sweden (Figure 3.2). The labour share also recovered in recent years in several countries without an increase in unionisation. Second, across countries, the labour share declined more in the higher union-density manufacturing sector while it increased in the lower union-density services sector (Cho, Manaresi and Reinhard, 2025[15]; Autor et al., 2020[16]). Nevertheless, effective collective bargaining and social dialogue remain an important mechanism to enable effective profit sharing between workers and firms (OECD, 2026[1]). Similarly, international trade may drive down the labour share through labour market competition and the shift of production of higher labour-share products and services to countries with a lower labour cost. But if this were the case, the labour share would be expected to rise in the countries the production has moved to, which is not the case (Karabarbounis, 2024[11]).

Another explanation is that value added has shifted towards high-productivity, low labour-intensity firms (so-called “superstar firms”) causing the labour share to decline mechanically (Autor et al., 2020[16]). This “winner takes most” dynamic may be caused by increased product market competition driven by globalisation and technological change: the internet enables customers to compare prices more effectively and to obtain goods at a distance, leading them to choose higher quality options. Another potential factor in the rise of superstar firms may be that scale favours the acquisition and effective deployment of new technologies, making larger firms more productive. Cho et al. (2025[15]) show that, over the period of 1995 – 2017, reallocation of value added towards high-productivity, low-labour-share firms can explain just over half of the average reduction in the labour share of the manufacturing sector across 26 OECD countries.

3.2.3. Technological progress could drive down the labour share even if it does not lead to aggregate job losses

In general, technological progress improves labour efficiency by (partially) taking over/speeding up tasks performed by workers. This leads to an increase in output per effective labour input and a reduction in production costs. The employment effects of this process are ex-ante ambiguous: employment may fall as tasks are automated (substitution effect). On the other hand, lower production costs may increase output if there is sufficient demand for the good/service, leading to constant or even increasing employment, as has been the case for previous waves of technological progress (productivity effect, (Georgieff and Hyee, 2022[17])).

Higher labour productivity through technological progress tends to lower prices and can lead to higher wages, but this is not necessarily the case and depends on the effect on aggregate output and labour demand. But even if wages increase, automation will still depress the aggregate labour share as capital accounts for a higher share of aggregate production (Acemoglu, 2024[18]; Acemoglu and Restrepo, 2018[19]; Acemoglu, Autor and Johnson, 2026[20]). This is effect is similar to a decrease in the price of capital in that it increases capital’s relative productivity compared to labour (Karabarbounis, 2024[11]).

However, technological progress can also generate new tasks that require labour input (e.g. the occupation of software technicians emerged following the emergence of information technology), and productivity gains may create space for new goods and services, as has been the case in the personal service sector (e.g. nail technicians (Autor et al., 2024[21])). This effect can be large: (Acemoglu and Restrepo, 2018[19]) estimate that “new jobs” accounted for about 60% of the employment growth in the United States between 1980 and 2015. The introduction of new tasks and new jobs could therefore counterbalance decreases in the labour share following the introduction of new technologies.

It is too early to assess to what extent AI will generate new tasks, or which education or skill groups will benefit from it. Without the emergence of new tasks, the labour share of income may fall as AI is deployed throughout the economy. It is also important to note that even if new jobs are created through AI, this job creation may not fully compensate for jobs lost due to automation, and may take time, leading to a temporary fall in the labour share. It may also not benefit the same groups of workers who see their tasks automate, and adjustments may require social spending for income support and reskilling as the labour market adjusts (OECD, 2024[2]).

Empirical evidence suggests that technological progress has driven down the labour share in the past: using data on the adoption of robots and software across sectors, (Acemoglu and Restrepo, 2022[22]) estimate that about 45% of the observed changes in sector-level labour shares in the United States between 1987 and 2016 can be attributed to automation. (Minniti, Prettner and Venturini, 2025[23]) provide direct evidence of the effect of AI on the labour share: exploiting differences in the incidence of AI patents in European Regions, they estimate that AI may have reduced the labour share by up to one‑third of a p.p. between 2000 and 2017. While this effect is small, their period of observation is significantly before advanced Large Language Models (LLMs) were released.

3.2.4. Policy insights

The intuition that new technologies can reduce the labour share even with constant or rising employment is important for financing of social protection: relying disproportionately on labour earnings to finance social protection may mean forgoing potentially sizeable productivity gains from AI deployment. Countries may want to consider shifting their financing mix in response to a fall in the labour share (see Section 3.3).

Lower relative taxation of capital compared to labour can also directly incentivise the deployment of labour-saving technologies over and above what they would be in the absence of this imbalance (OECD, 2024[2]; OECD, 2023[24]). Capital is generally taxed at lower rates than labour in many OECD countries (Hourani et al., 2023[25]). Reducing this gap – by some combination of lowering non-wage labour costs and raising effective tax rates on capital – could weaken the incentives of firms to invest in automating technologies.

Public policies can also help workers to capture a larger share of productivity gains by setting the framework for sustainable wage increases and thereby stabilising the labour share and helping safeguard the funding source of social protection. Policy options include supporting skills development for all age groups including and especially mid-career and older workers, both within firms and through formal education and training institutions for workers and jobseekers. Depending on the national policy context, increasing the scope as well as the frequency of collective bargaining agreements can also help boost wages (OECD, 2026[1]).

3.3. Diversifying funding sources

Copy link to 3.3. Diversifying funding sources

3.3.1. Ageing will put downward pressure on labour income

Tax bases are age dependent. Employee earnings peak between ages 40 and 50 in most countries and then drop off as individuals retire (see Figure 3.3, Panel A). In a static environment, population ageing will mechanically reduce labour income as a tax base, as the working-age population shrinks, particularly in the high-earning middle‑age group.

This effect can be partly counterbalanced by higher labour force attachment in response to policy action (especially among older workers and women, see Section 3.1). A shrinking labour force may also tighten labour market, which could put upward pressure on wages, partly offsetting the decline in the aggregate wage bill resulting from a smaller workforce (Koutsogeorgopoulou and Morgavi, 2025[26]). Higher wages could also further increase labour supply, again offsetting mechanic decreases in hours worked from population ageing. The extent of these effects is difficult to predict.

Figure 3.3. Tax bases are age dependent
Copy link to Figure 3.3. Tax bases are age dependent

Distribution of tax bases per capita in USD by age across OECD countries, most recent available year

3.3.2. While capital income may be more stable

Capital income8 grows throughout adulthood and peaks at retirement age (65‑69), when retirees start dissaving, leading to a slight drop. However, on average, capital income at age 80+ is still higher than under 55 (Figure 3.3, Panel B). Wealth-to‑income ratios have increased across OECD countries, driven by gains in equity and housing markets, with strong gains among older cohorts who purchased homes before strong price increases started and benefited from appreciation (OECD, 2025[5]). The aggregate wealth stock is expected to further increase mechanically because of population ageing since older people hold more wealth, and for longer (see below).

While the stock of wealth is set to increase, returns to capital could fall due to ageing, because of diminishing returns to capital as the capital stock grows. This is particularly the case as people do not tend to dissave strongly in retirement, keeping the capital stock high as baby-boomers age (see below). Ageing has been identified as one of the reasons for the decline in real interest rates in past decades, and interest rates are expected to further decline because of population ageing (Auclert et al., 2026[28]). However, technology-driven boosts to capital productivity could counteract this effect (see Section 3.2). Furthermore, the effect of ageing on the stock of capital is unambiguously positive (Auclert et al., 2026[28]), which could result in rising aggregate capital income, even with falling returns.

3.3.3. Tax revenue is set to decline because of population ageing

OECD simulations estimate that, under a “no-policy-change scenario”, tax revenues as a share of GDP would be 1 p.p. lower on average across the OECD under the projected age structure in 2060. Larger declines would be expected in countries with faster population ageing and / or a stronger reliance on tax revenues from younger age groups (Sicsic and Hourani, 2026[27]).

Further increasing labour taxation would i) put an additional tax burden on younger cohorts, ii) increase the tax wedge on labour and therefore disincentivise employment at a time when mobilising groups with low labour force attachment is crucial (see Section 3.1) and iii) potentially be regressive given that labour income is less concentrated than capital income.

About half of all tax revenue is already levied on labour income. While it is not straightforward to disentangle taxes on incomes, profits and capital gains from individuals (second row of Table 3.1), a recent OECD personal income tax split exercise found that, on average across 29 countries with available data, taxes on capital gains amounted to only about 7% of taxes on income, profits and capital gains of individuals (OECD, 2025[29]). Together with social security contributions, this implies that around half of total tax revenue is levied on labour income.

The base of consumption taxes is expected to be resilient during the demographic transition: consumption grows with age, but increases are modest as individuals move towards late middle age. There is a slight dip after retirement (as retirement income is lower than labour earnings), but consumption at age 75‑79 is still higher than consumption for people under the age of 30 (Figure 3.3, Panel C).

However, the space to increase Value Added Tax (VAT) rates is limited in many countries. Value added Taxes (VAT) already account for over one‑fifth of government revenue across OECD and EU countries (Table 3.1) and were increased in the years following the GFC (European Commission, 2021[30]). In some countries, in particular in Europe, VAT rates may be close to the revenue‑maximising level, such that a further increase could lower revenue (Crowe et al., 2022[31]). However, scaling back reduced VAT rates and exemptions that benefit higher income households could raise revenues and improve the neutrality of consumption taxes in some countries. For instance, exemptions for health expenditure will increase in volume with an ageing population, but they are not the most effective way to support older low-income households – targeted transfers can typically achieve this goal at lower cost (Sicsic and Hourani, 2026[27]). Environmental taxes, in particular taxes on energy consumption and fuel, have declined in importance over the last decade across the EU, following special measures during the 2022 energy crisis, and more recently the war in Iran. As environmental taxes aim to disincentivise the consumption of the goods and services they are levied on, their revenues are expected to decline over time by design (European Commission, 2026[32]).

3.3.4. Broadening the taxation of capital income

Given that capital income is expected to be more stable than labour income in ageing societies, countries may consider increasing the share of capital income taxation in their tax mix (European Commission, 2021[30]; Sicsic and Hourani, 2026[27]; Arnemann et al., 2025[33]; Koutsogeorgopoulou and Morgavi, 2025[26]). Income from capital – including income from dividends, capital gains and interest – tends to be more favourably taxed than labour income in most OECD countries:

Dividend income is often taxed at a flat rate that tends to be lower than the average rate on labour income, or subject to a more favourable progressive tax schedule. It also often benefits from exemptions or special deductions. In most OECD countries, dividends are taxed more favourably than labour income for most income levels, even when accounting for corporate income taxes (CIT) paid by firms prior to distributing dividends. Some countries, including Denmark, Korea and Switzerland, however, have higher effective tax rates on dividends than labour income at all income levels when accounting for firm-level taxes (including CIT, (Hourani et al., 2023[25])).
Most OECD countries offer favourable tax treatment for capital gains: they are typically only taxed upon realisation, and many countries offer lower rates compared to other types of income, in particular labour income, or provide full or partial exemptions. Some countries also fully reset the capital gains cost base upon the death of the owner.9 Most countries exempt main residences from capital gains, and other housing assets are treated preferentially subject to minimum holding periods. Closely held businesses also benefit from preferential taxation in many countries (Hourani and Perret, 2025[34]).

The main reason for a preferential treatment of capital income vis-à-vis labour income is to encourage saving, investment and economic growth. The empirical evidence on the effect of capital gains taxes on savings and investment is mixed, however, and the empirical literature does not support a strong relationship between capital gains taxation and economic growth (Hourani and Perret, 2025[34]). Encouraging savings may also be less urgent since population ageing tends to increase the aggregate capital stock (see above).

Another reason for offering preferential rates on capital gains is preventing the “lock-in effect”: the fact that capital gains taxes are only paid when assets are sold can lead to owners holding the assets too long to avoid paying the taxes, which can lead to misallocation of capital. For instance, business owners may delay selling their business to a more effective successor. Empirical evidence does suggest that capital gains relief can reduce this effect, although effect sizes vary. Proposals exist to minimise this effect (Hourani and Perret, 2025[34]).

Higher taxes on capital would be progressive, as capital income is highly concentrated. Narrowing the gap between tax rates on labour and capital income may also reduce the incentive for shifting income from labour to capital (e.g. among self-employed workers) and therefore lower regulatory arbitrage (Crowe et al., 2022[31]), see also Chapter 2.

Revenue category	EU	OECD
Taxes on income, profits and capital gains of individuals and corporations	31.6	36.4
Taxes on income, profits and capital gains of individuals	22.0	23.7
Taxes on income, profits and capital gains of corporations	9.6	11.9
Unallocable between taxes on income, profits and capital gains of individuals and corporations	0.3	0.8
Social security contributions (SSC)	30.7	25.5
Taxes on payroll and workforce	1.1	1.3
Taxes on property	3.3	5.1
Recurrent taxes on immovable property*	1.5	2.9
Recurrent taxes on net wealth	0.4	0.4
Estate, inheritance and gift taxes	0.4	0.4
Taxes on financial and capital transactions	0.8	1.2
Other non-recurrent taxes on property	0.0	0.1
Taxes on goods and services	32.8	31.2
Taxes on production, sale, transfer, leasing and delivery of goods and rendering of service	29.4	29.1
General taxes	22.1	21.1
Value added taxes (VAT)	21.8	20.5
Sales taxes	0.0	0.4
Turnover and other general taxes on goods and services	0.3	0.2
Taxes on specific goods and services	8.2	8.1
Taxes on use of goods, or on permission to use goods or perform activities	1.9	2.0
Unallocable between taxes on production, sale, transfer, leasing and delivery of goods and rendering of services and taxes on use of goods, or on permission to use goods, or perform activities	0.1	0.1
Other taxes	0.5	0.5

Note: * Includes both taxes on the net wealth of individuals and corporations. Since a breakdown is not available for all countries who levy such taxes, it is not shown here. Data refer to 2023 for Romania. The EU average excludes Cyprus due to data unavailability. Information on some revenue sub-categories is missing for Estonia, Greece and Poland. Therefore, the EU average at the subcategory level sometimes does not contain all countries, which can lead to small deviations of the category and sub-category sums. The EU and OECD averages are unweighted.

Source: OECD Revenue Statistics.

3.3.5. Inheritances are set to increase

There is an increasing trend among retirees to hold on to their wealth until later in life and to die with higher levels of wealth than predicted by standard lifecycle models (the “retirement savings puzzle”). Reasons for slower-than-expected dissaving include precautionary savings due to longevity risks, the high share of housing in total wealth combined with housing market frictions that make it difficult to dissave housing wealth, and the desire to leave inheritances (OECD, 2025[5]). The inheritance motive seems to dominate for households at the upper end of the wealth distribution (De Nardi et al., 2025[35]).

In recent decades, asset appreciation has furthermore increased the stock of wealth, that is set to be transferred when the baby boomer generation dies. Additionally, the average value of individual bequests is further expected to increase as lower fertility rates imply fewer successors per bequest. Using data from the European Central Bank’s Household Finance and Consumption Survey, (Krenek et al., 2022[36]) estimate significant increases in revenue from inheritance taxes in five European countries under a no-policy-change scenario, driven by the increased value of bequests, and the smaller number of heirs, which makes inheritance tax exemption thresholds less relevant.

3.3.6. Inheritance taxes are unpopular

These trends point to the potential of inheritance taxes to play a bigger role in countries’ tax mix (OECD, 2021[37]). In 2024, inheritance and gift taxes only accounted for 0.4% of total government revenues across the EU and OECD on average (Table 3.1). 18 out of 41 countries with available data did not levy any gift or inheritance taxes, and even in the countries where inheritance taxes were most important (Korea, France and Belgium), they still accounted for under 2.5% of total government revenue (see Annex Figure 3.A.1). One reason for this is that inheritance taxes are unpopular: across 27 countries that participated in the OECD Risks that Matter Survey (RTM), fewer than one‑third of respondents considered inheritance taxes “fair or very fair” (Figure 3.4, Panel A).

Surprisingly, respondents at the bottom of the income distribution did not find them fairer than respondents at the top, even though they are less likely to pay inheritance taxes. This may be connected to misperceptions about inheritance taxes: studies consistently show that the majority of respondents think that more households are liable for inheritance taxes, and that tax rates are higher and less progressive, than they really are (OECD, 2021[37]). Differences in the perceived unfairness between income groups were statistically insignificant, except in the United States and Canada, where higher income respondents were significantly more likely to find inheritance taxes fair, and in Finland, where higher income groups were less likely to agree with them (Figure 3.4, Panel A).

Agreement with inheritance taxes was lowest in Belgium, France and Korea, which are among the countries where inheritance taxes account for the highest share of government revenue (over 1.5% of total tax revenue). Agreement in Latvia was also very low, even though inheritance taxes were below the OECD average and accounted for only 0.1% of total tax revenue (see Annex Figure 3.A.1). Inheritance taxes had the biggest support in Austria, the United States and the United Kingdom. The relationship to the tax incidence is less clear for these countries, as the United Kingdom and the United States have an above‑average incidence of inheritance taxes, while Austria does not have inheritance taxes.

This survey evidence suggests that earmarking may increase the public support for inheritance taxes. If they were earmarked for health expenditure, over 40% of respondents would support the introduction or increase of inheritance taxes, over 10 p.p. more than the share who find the tax fair (Figure 3.4, Panel B). Support is lower if inheritance taxes were earmarked for long-term care, support to low-income households and especially climate change mitigation policies.

While revenue earmarking can resonate with the public, it can also crowd out resources from other spending areas and reduce budgetary flexibility, as well as limit the role of the legislature in the budgetary process (Immervoll, 2024[38]). In addition to revenue earmarking, improving information on inheritance tax schedules may further improve support as individuals often underestimate minimum thresholds and overestimate rates (OECD, 2021[37]).

Figure 3.4. Inheritance taxes are unpopular, but earmarking may help
Copy link to Figure 3.4. Inheritance taxes are unpopular, but earmarking may help

Attitudes towards inheritance taxes in percentage of all respondents, 2024

Well-designed inheritance taxes can raise revenue while enhancing equity, at lower efficiency and administrative cost than other alternatives (Koutsogeorgopoulou and Morgavi, 2025[26]). Empirical evidence suggests that inheritance taxes distort savings decisions less than wealth taxes and are easier to administer than other taxes that need to be levied annually. Tax design matters: recipient-based inheritance taxes tend to be more equitable than taxes applied on total bequeathed wealth, and inheritance taxes should include exemption thresholds to protect small inheritances. Progressive tax rates are more equitable than flat taxes. Inheritance taxes should be designed to avoid opportunities for evasion by only applying reliefs where there is a strong rationale (e.g. main residences up to a threshold) and closely aligning inheritance and gift taxes (OECD, 2021[37]).

3.3.7. Taxes on immovable property are underutilised in many countries

House prices have increased strongly in recent decades in most OECD countries (OECD, 2025[5]), and are expected to increase further even if at a lower rate in countries that will experience a population decline due to ageing. Given the large potential base, taxes on immovable property seem underutilised at 2.9% of total tax revenue across OECD countries, and even more in the EU, where they only account for 1.5% of total tax revenue (Table 3.1). Because the tax base is immobile, property taxes are among the least distortionary taxes (Crowe et al., 2022[31]). As housing wealth is concentrated among older households, taxes on immovable property can contribute to intergenerational fairness in the context of population ageing. While it is less concentrated than other assets, it is still concentrated among high-income, high-wealth households, making property taxes progressive in principle (OECD, 2022[39]).

Property taxes can generate liquidity problems where household incomes are out-of-step with value of the house – this can happen when house prices are rising rapidly, for households who have owned a house for a long time, as well as upon retirement when income declines. This effect can be cushioned by deferring payments to the point of sale.

The main problem with property taxes is that they require frequent updating of house valuations, which is administratively costly. Indeed, outdated valuations of property are among the main drivers of the decrease in property tax revenue in the European Union over the past decade (European Commission, 2026[32]). Infrequent revaluations can lead to unequal treatment between recently revalued and other properties, as well as between properties that have not been updated in some time and experienced unequal changes in value. Sporadic updates cause revenue losses, as house price increases go untaxed. In practice, very few countries update property values regularly, resulting in very outdated values in some countries (OECD, 2022[39]). Even the use of digital tools does not seem to have made property valuations much cheaper for tax administrations. Property taxes are also unpopular because they are very visible: they have to be paid directly as they cannot be withheld at source and are not always matched by a revenue stream (Crowe et al., 2022[31]).

Since property taxes are typically levied at the regional and local level, higher property tax revenues could help local governments address spending pressures, including expenditure on long-term care, an expenditure item that is set to increase strongly due to population ageing, and that is often provided at the local level. This could also increase the incentive of local authorities to regularly update property values and improve the public acceptability of the tax.

3.3.8. Taxes on wealth can play a role in countries where the taxation of capital income and inheritances is low

Recurrent net wealth taxes are recurrent taxes on individual wealth stocks, including a wide range of movable and immovable assets net of debt. Recurrent taxes on net wealth only accounted for 0.4% of total tax revenue in OECD and EU countries in 2024 (Table 3.1).10 They used to be more widespread: in 1990, 12 OECD countries levied recurrent taxes on wealth, this number went down to four in 2020 (OECD, 2024[40]). Narrow bases, extensive exemptions and issues with the valuation of assets contributed to declining revenues and diminished political support (European Commission, 2026[41]). Recurrent taxes on net wealth still account for 6.8% of total government revenue in Luxembourg, and 5% in Switzerland.

Compared to capital income taxes, net wealth taxes tend to be more distortive: taxes are levied on a presumptive, instead of the actual return to investment, which can distort savings decisions. As the value of the asset is taxed and not its return, wealth taxes also favour holders of high-return assets.

Compared to inheritance taxes, wealth taxes may lead to double taxation if taxes on capital income exist, while inherited wealth is only taxed once upon receipt. Empirically, inheritance taxes have also been shown to have smaller effects on savings and investment. Inheritance taxes also only require wealth to be valued once, instead of on an ongoing basis, which is administratively easier. Overall, while taxes on capital income and inheritances may be a more efficient and less administratively costly way to tax wealth, wealth taxes may have a role to play where capital income is taxed lightly or where inheritance taxes do not exist or raise limited revenue (OECD, 2018[42]).

3.4. Strengthening the entitlement – contribution link

Copy link to 3.4. Strengthening the entitlement – contribution link

The financing mix of social protection spending reflects policy institutions (e.g. the role of occupational pension schemes in a number of countries, see Chapter 1), but also the relative importance countries place on different objectives of social protection, including insurance against income shocks and income smoothing, which are logically tied to contributions, versus redistribution and poverty prevention.

As such, benefits designed to replace lost earnings are often financed by contributions. This can be seen in Figure 3.5: in most EU countries with available data, the majority of old-age cash benefits was dependent on previous contributions in 2024 (Panel A). For unemployment benefits, the share was lower and varied by country: in some countries, unemployment benefits were almost exclusively insurance‑based, such as Italy, Belgium, Latvia, Bulgaria and Hungary. Other countries mainly relied on non-contributory benefits, such as Ireland. Others provided a mix, such as Germany and Austria (Figure 3.5, Panel C).

3.4.1. The link between benefits and contributions can be incomplete even for insurance benefits

However, also for insurance benefits that condition entitlements on prior contributions, the degree to which contributions determine entitlement levels varies across countries and programmes. In earnings replacement programmes, benefits are often closely tied to contributions, even if they contain an element of redistribution, for instance by providing higher replacement rates for those on low incomes. Examples include old-age and disability pensions, sickness cash benefits, and unemployment benefits. For other schemes, there may be no link at all, e.g. in the case of compulsory health insurance schemes where all insured have the same entitlement regardless of contributions paid (Immervoll, 2024[38]).

Even benefits that are designed to provide insurance against earnings loss may provide some benefits that are not directly tied to contributions:

Insured persons accumulate rights while they are not paying contributions. Examples include the crediting of pension contributions during times of unemployment or parental leave.
Insurance may extend to family members of insured persons, e.g. spouses or children in contribution-based health- and old-age pension schemes.
Claims may exceed contributions because of structural factors such as population ageing, or during extraordinary times (e.g. unemployment benefits during deep recessions).

Indeed, over 90% of old-age cash benefits (including survivor benefits) that households received across the European Union on average in 2024 were contribution based (see Panel A in Figure 3.5), a much higher proportion than the share of social contributions in the funding base of (cash and in-kind) old age benefits (63%, see Figure 1.8 in Chapter 1).11 This is because contributory pension systems receive general-revenue transfers, mainly due to population ageing (see Chapter 1).

3.4.2. For some benefits, the rationale for contribution-financing is weaker

Some benefits are not designed to insure against income loss, but to prevent poverty and redistribute income (e.g. social assistance and family benefits). Other schemes provide health and other services, e.g. long-term care, that are not economically tied to employment. For these benefits, the rationale for social-contribution financing is less evident, especially when labour income is under pressure.

For instance, while general revenue accounted for the majority of financing of (cash and in-kind) family benefits (see Chapter 1), in some countries, contributions accounted for a significant share of expenditure: in Estonia and Bulgaria, over 60% of family benefits were financed by social contributions, while under 50% of family benefits received by households were contributory. Similarly, in Austria, over 40% of family benefits were financed by employer social contributions (see Chapter 1), although all family benefits that households received were non-contributory (mainly universal child benefits, but also parental leave, see Figure 3.5, Panel C). The reason is that family benefits in Austria are mainly financed through the “fund for the equivalisation of family burdens” (Familienlastenausgleichsfonds), which mainly draws its revenues from employer contributions (Schratzenstaller, 2022[43]), even though benefits are not tied to work.12

When reviewing their financing mixes, countries could consider shifting financing from social contributions to general revenue for programmes where the mapping from contributions to entitlements is weak. This would decrease the burden on labour incomes, and may incentivise labour supply, crucial when working-age populations are shrinking (Arnemann et al., 2025[33]; Sachverständigenrat, 2026[44]). It could also decrease the incentive for employers and workers to misclassify work as independent contracting or self-employed work where social contribution obligations differ across employment forms (see Chapter 2), as well as for undeclared work (see Section 3.5).

Broadening the funding base away from labour income is also important to ensure intergenerational fairness, especially for long-term care, a risk that is not tied to work, and that will see significant spending increases due to population ageing (OECD, 2024[45]).

3.4.3. But financing is often tied to governance

In practice, it can be difficult to change the financing structure of a benefit, as governance and administration of social benefits often derive from financing sources. In many countries, contribution-financed social insurance historically developed alongside self-governing institutions, autonomous insurance funds and social-partner participation. These institutional arrangements can make shifts from contribution financing to general taxation administratively and politically difficult (Pierson, 1993[46]; Pierson, 2000[47]; Weaver, 2010[48])., In insurance‑based health systems, for instance, health-insurance funds as well as social partner representatives often have significant influence on the administration of health insurance schemes (OECD, 2024[49]). When the source of financing changes, administration and oversight may have to change as well.

For instance, in the early 1990s, French healthcare financing moved progressively from social security contributions towards an earmarked tax, the contribution sociale généralisée (CSG), levied on personal income including income from capital; it funds healthcare and means-tested benefits and is not levied on French residents who are not affiliated with the French social security system (Immervoll, 2024[38]; Bozio and Wasmer, 2024[50]). Parliament gained authority over health-spending targets in 1996, and a 2004 statute unified compulsory and complementary insurance funds under a government-appointed director. Formally, this preserved the historic co‑administration of health insurance by employer and worker representatives; in practice, it shifted the balance of power towards the government (Nay et al., 2016[51]).

Similarly, Spain’s health financing moved from social insurance contributions to general revenues in the late 1980s – a shift “from Bismarck to Beveridge” (Rodrı́guez, Scheffler and Agnew, 2000[52]; Bernal-Delgado E, 2024[53]). Until 1989, social contributions financed around 75% of public health financing, with general revenues covering the remainder. The 1986 Healthcare General Act and the 1989 General Budget Law reversed this balance: general revenues became the main source of financing (Bernal-Delgado E, 2024[53]; Rodrı́guez, Scheffler and Agnew, 2000[52]). General revenue financing accounted for 72% of total health expenditure in 2021, with only a residual contributory scheme for civil servants. Simultaneously, starting from the late 1980s, fiscal authority over public health expenditure was gradually transferred to the autonomous regions; this process was completed in 2002 (Bernal-Delgado E, 2024[53]).

An alternative to major reform could be increasing general revenue subsidies to insurance‑based programmes. It is important to note that using general-revenue funds to finance programmes that only benefit insured persons would raise equity concerns (Sachverständigenrat, 2026[44]). This is especially true where coverage gaps persist for some types of workers e.g. the self-employed. Subsidies that cover benefits that are not balanced by contributions – e.g. rights acquired during periods of non-contribution (such as parental leave or unemployment), are more defensible from an equity standpoint. The disadvantage of increasing general revenue subsidies to insurance‑based schemes is that it makes the financing structure of social protection less transparent (OECD, 2024[49]). It can also undermine the insurance logic.

3.4.4. Policy insights

Countries could consider financing benefits that do not insure against work-related risks from general revenue where they do not already do so. In practice these are means-tested and universal cash benefits, as well as services not tied to employment, most importantly health and long-term care. This would take pressure off labour incomes and take advantage of funding bases that may be less affected by population ageing and automation. This may be politically challenging as the governance of these benefits is often linked to social contributions, with roles for the social partners (e.g. autonomous social insurance funds).

Increasing subsidies from general government revenue to insurance‑based schemes can be an alternative. Countries already subsidise contributory pension systems in many countries. For instance, the pension system balance (contributions minus expenditures) was negative in 16 out of 25 EU countries with available data in 2022, with an average deficit amounting to 1.6% of GDP, see Chapter 1 and (European Commission, 2024[54]).

However, countries should be cautious not to use general revenues to provide benefits only to insured persons as groups such as the self-employed or those with unstable careers benefit less from insurance‑based schemes in some countries. General-revenue subsidies should be limited to benefits that are not directly tied to contributions, e.g. crediting of pension contributions for times of parental leave (which is part of family policy) or for periods of unemployment (which belongs into general social policy for workers without entitlement to unemployment benefits).

Figure 3.5. A large share of cash unemployment- and family benefit receipt is non-contributory
Copy link to Figure 3.5. A large share of cash unemployment- and family benefit receipt is non-contributory

Average share of old-age, family and unemployment cash transfers in total gross household incomes in the EU, in per cent, by entitlement criterion, 2024

3.5. Closing contribution gaps in undeclared work

Copy link to 3.5. Closing contribution gaps in undeclared work

Undeclared work – defined as any paid activity that is lawful but not declared to public authorities (ELA, 2021[55])13 – and the under-declaration of income diminish the financing base of social protection systems. The EU definition of undeclared work excludes criminal activity and only refers to the non-payment of taxes and social security contributions, as well the disregard of labour law obligations. It is closely aligned with the International Labour Organization’s statistical concept of informal employment, more often used in the context of developing and emerging economies: work that is “not covered by formal arrangements such as … income taxation, labour legislation and social security” (ILO, 2023[56]).

Undeclared work takes three forms with distinct implications for financing and policy: i) fully undeclared work, ii) under-declared work, or “envelope wages”, where part of the pay of a registered worker is not declared and iii) misclassified self-employment, where a dependent employee is declared as self-employed to reduce labour costs mainly in the form of social contributions, see Chapter 2. Underreporting of income by self-employed workers is also common in many countries, see Chapter 2 and (Schoukens and Bruynseraede, 2026[57]).

Undeclared work is concentrated in labour-intensive, hard-to-monitor activities such as construction, hospitality, agriculture, and domestic, care and personal services. In the EU, the posting of workers and the cross-border provision of services can also obscure where work should be declared (ELA, 2021[55]).

3.5.1. Undeclared work remains widespread

By definition, undeclared work is difficult to measure. Macroeconomic models estimate the shadow economy at around 17% to 23% of GDP across the EU on average, depending on the model, with national values ranging from below 10% in Austria and Luxembourg to about 37% in Bulgaria. Other countries with high shares include Greece, with a share of almost 30%, and Croatia, Italy, Poland and Romania with shares around 23% (Kelmanson et al., 2019[58]; Schneider and Asllani, 2022[59]).14 However, the concept of the shadow economy is broader than undeclared work and these estimates may capture some illegal activity.

Franić et al. (2023[60]) compare labour-force‑survey estimates of hours worked with business statistics on declared work (the labour input method).15 They estimate that undeclared work amounted to around 15% of private‑sector gross value added in 2019 on average across the EU, down from 16% in 2013, with declines in 19 of 26 Member States. They find that there are large differences across countries: undeclared work is highest in Romania, Lithuania and Bulgaria, above 20% of private‑sector GVA, and lowest in Austria, Luxembourg, Sweden and Germany below 9% of GVA.

Direct survey evidence points in the same direction: in the most recent Special Eurobarometer survey on undeclared work from 2019, about one in ten Europeans reported buying goods or services that may have involved undeclared work, one in three knew someone working undeclared, and roughly half believed the risk of detection was low (European Commission, 2020[61]). Participation in undeclared work fell from 5% of survey respondents in 2007 to 4% in 2013 and has remained stable until 2019. Envelope wages similarly fell from 5% in 2007 to 3% in 2013 and 2019 (Williams and Horodnic, 2020[62]). The share of respondents buying goods or services that may involve undeclared work declined slightly from 11% in 2013 to 10% in 2019 (European Commission, 2020[61]).16 Note that these estimates are based on self-declarations and may therefore underestimate the real incidence.

Of workers performing undeclared work, 10% work for an employer, 42% are self-employed, and 48% perform “paid favours”: paid work for relatives, friends, neighbours or acquaintances that is undeclared (Williams and Oz-Yalaman, 2021[63]). These shares refer to the number of workers performing paid work. When measured by hours worked, undeclared dependent work accounts for the largest share of undeclared work in the EU (63%), followed by self-employment (36%) (Franić, Horodnic and Williams, 2023[60]).

3.5.2. Undeclared work narrows the financing base of social protection, weakens risk pooling and leaves workers without social protection coverage

Undeclared and under-declared work reduce contributory revenue directly. They also weaken risk pooling, particularly if workers facing lower risks (e.g. of unemployment or disability) evade participation in compulsory schemes. And they interact with the redistributive design of social insurance: because many social protection schemes provide higher replacement rates to lower earners, under declaring earnings (envelope wages) allow workers to benefit from this progressive design while withholding part of their contributions. Williams et al. (2015[64]) estimate that envelope wages amount, on average, to about a quarter of gross annual pay.

For workers, undeclared and under-declared work means that they do not have any, or only partial, entitlement to contribution-based social protection. They may therefore fall back on tax-financed, means-tested support in the event of job loss, injury or old age. In addition, undeclared work also lowers production costs for the firms that use it, giving them an unfair advantage over compliant competitors.

3.5.3. Overall, the incidence of undeclared work is higher among men, younger workers, the financially constrained and manual workers

Undeclared work is more common among men (5%) than women (3% (Williams and Horodnic, 2020[62])).17 It largely reflects gender segregation across sectors: men are more likely to work in construction, women in personal and household services. Undeclared work is difficult to address in personal and household services, as private households are very motivated to avoid non-wage labour costs and labour protections, and they are difficult to reach for labour inspectorates (OECD, 2021[65]). Undeclared work accounts for the majority of the non-care segment of household services, with up to around 90% of such work not being declared in some countries, while in the care segment its share is lower, at around a third. Migrants are over-represented in the household services sector: an estimated 17% of personal and household services workers are third-country nationals (ELA, 2021[55]; OECD, 2021[65]).

Incidence declines steadily with age: from 9% among those aged 15‑24, to 5% among 25‑39 year‑olds, 3% among 40‑54 year‑olds, and 2% among those 55 and over. Workers struggling with the cost of living are also far more likely to work undeclared: 9% of those struggling most of the time, against 3% of those who rarely struggle financially. Manual workers have the highest incidence by occupation (5‑6% of skilled and unskilled manual workers), consistent with the high incidence of undeclared work in construction, among the self-employed (6%), students and the unemployed. Envelope wages follow the same pattern (European Commission, 2020[61]; Williams and Horodnic, 2016[66]).

3.5.4. Undeclared work is more prevalent in small firms

In the EU, under-declaration is concentrated in small and micro firms: 6% of employees in firms with fewer than ten employees receive envelope wages, compared to under 1% in firms with 500 or more employees. Firms with fewer than 50 employees account for around 70% of all envelope‑wage recipients. This is consistent with weaker formal HR management practices in smaller firms making informal side‑arrangements easier to sustain (Williams and Horodnic, 2020[62]).

In some countries, including the Netherlands, Cyprus, Ireland and Finland, undeclared work is concentrated in even smaller units: 68‑90% of undeclared work takes the form of one‑person enterprises rather than a registered firms with off-book workers. In contrast, in Poland, Belgium, Bulgaria and Italy, over 90% of undeclared work is waged employment (Franić, Horodnic and Williams, 2023[60]):

3.5.5. Undeclared work may lead to lower wage growth over the life cycle

Direct evidence on the effect of undeclared work on earnings from EU countries is limited. Evidence from low and middle‑income countries suggests that informal workers tend to earn about 40% less than formal workers with similar characteristics (Ulyssea, 2020[67]; Ohnsorge and Yu, 2022[68]). Informality also influences wage growth over the life cycle. Estimates using household survey data from Mexico show that formal workers’ earnings rise to about 45% above their entry level after 25‑30 years of experience, whereas informal profiles plateau early, at around 13% (Siachoque, forthcoming[69]). Because contributory entitlements accrue over the entire career, a flatter profile reduces the value of lifetime contributions and depresses replacement rates at retirement. Lower wage growth is mainly caused by two factors. First, human capital: undeclared work offers limited access to on-the‑job training and involves short tenures that reduce the accumulation of firm-specific human capital (Ulyssea, 2020[67]). Second, it can be difficult to transition from informal to formal jobs, and informal jobs offer limited upward mobility (Bosch and Esteban-Pretel, 2012[70]; Donovan, Lu and Schoellman, 2020[71]).

3.5.6. High non-wage labour cost can incentivise undeclared work, but is not the main driver

A high tax wedge can push labour costs above the productivity of lower-skilled workers and create an incentive for workers and employers to leave earnings partially or entirely undeclared. Withdrawal of social benefits combined with social contributions for low-wage workers can generate an especially strong incentive to work undeclared. Using survey-experimental data from a sample of German benefit recipients, (Burgstaller and Feld, 2026[72]) find that one in five works undeclared, more than three times the incidence in the general population (6%). The interaction between means-tested transfer withdrawal and the tax and social-contribution system produces effective marginal tax rates of around 80% for many recipients of social benefits in the sample (Burgstaller and Feld, 2026[72]).18 Therefore, targeted reductions in contributions for low-income workers are a common instrument (Kolev, La and Manfredi, 2023[73]).

However, empirically, the relationship between non-wage labour cost and the incidence of undeclared work is weak. Franić et al. (2023[60]) find that among state intervention factors, social contributions and taxation have the lowest correlation with the incidence of undeclared work. Government effectiveness, corruption perceptions and indicators of development (e.g. human development index, social progression index) are most strongly associated with the incidence of undeclared work. Countries that score highest on government effectiveness such as the Netherlands, Denmark and Finland have the lowest levels of undeclared work. They also find that measures of trust in institutions, such as parliament and government, as well as labour inspectorates and tax and social security institutions, were moderately or weakly negatively associated with the incidence of undeclared work. This is consistent with the wider literature on informality in low- and middle‑income countries, which identifies institutional quality (including the effectiveness of labour inspections) and the level of development rather than tax rates as the principal drivers (La Porta and Shleifer, 2014[74]; Ulyssea, 2020[67]).

An influential literature also relates undeclared work to a gap between formal rules and social norms. This low “tax morale” is shaped by the trust of individuals in the government and the perceived value for money of social contributions, as well as a belief in the compliance of others (Williams and Horodnic, 2020[62]; Williams and Horodnic, 2016[66]). Over time the weight of deterrence in explaining behaviour has fallen while that of trust has risen. This is consistent with a broader body of evidence showing that voluntary compliance depends not only on enforcement but also on trust in public institutions and perceptions of the fairness of the tax and social protection system (OECD, 2019[75]; OECD, 2024[76]; European Commission, 2026[32]).

3.5.7. Policy insights

Effective responses combine enforcement with measures that make formal work more attractive. For workers and firms who are productive enough to operate in the formal economy, the priority is to ensure adequate legal, social-security and tax coverage and to enforce compliance. Digital technologies can help improve the effectiveness of enforcement. For instance, labour inspectorates increasingly use new digital technologies, including advanced data analytics and AI, to identify firms who are more likely non-compliant. This helps focus resources more effectively on non-compliant firms, while lowering burdens on complying businesses, thus encouraging overall compliance (OECD, 2026[77]).

For lower-paid, vulnerable workers, policy interventions often subsidise participation in contributory schemes and combine this with non-contributory social protection coverage, with the aim of poverty reduction. In general, additional financing should be mobilised through strengthened tax compliance and enforcement, while avoiding excessively high formalisation costs (in the form of taxes and social security contributions for low-paid workers, (OECD, 2024[78])).

The evidence indicates that enforcement alone is not sufficient. Deterrence raises the expected cost of non-compliance through risk-based inspections, data-matching across tax, social-security and labour records, and proportionate sanctions. A randomised field experiment on more than 15 000 Norwegian taxpayers found that both a letter raising the perceived probability of detection and one appealing to a sense of fairness and recalling wider societal beneﬁts of tax revenue increased self-reported income, but only the detection signal had a lasting effect (Bott et al., 2020[79]). At the same time, OECD analysis of tax morale finds that voluntary compliance depends less on rates and penalties than on trust in government, the perceived fairness of the system and how easy it is to comply, pointing to strategies that combine enforcement with simplification of administrative procedures and a stronger link between contributions and entitlements (OECD, 2019[75]; Kolev, La and Manfredi, 2023[73]).

Policies may lower the cost, and raise the benefits, of declaring work. Demand-side instruments are often used in personal and household services: service‑voucher schemes are the most common measure. Belgium and France reduce the price of household services through social vouchers, which households buy at subsidised prices. France, Sweden and Germany also grant tax credits, with France and Sweden refunding around 50% of household-service costs (wages and social security contributions) and Germany 20% (OECD, 2021[65]). Belgium’s scheme operates through a three‑way relationship between accredited providers, workers and households, while France’s CESU (Chèque Emploi Service Universel) system simplifies the hiring of domestic workers and makes its tax credit refundable for households with insufficient tax liability. Both instruments have been linked to a reduced incidence of informal work, with the French and Swedish systems being most effective at moving activity into the formal sector, though effectiveness depends on how far they lower the service price (OECD, 2021[65]). The design is decisive: using Italian administrative records and randomly timed inspections, Di Porto et al. (2026[80]) show that flexible‑work instruments intended to formalise casual work (labour vouchers) can instead be used to conceal it. Some firms activate a voucher for an undeclared worker only when an inspector arrives, raising voucher use by about 0.88 p.p. (from 18%) on the day of an inspection. The practice disappears once firms are required to notify voucher use in advance. Instruments must therefore be designed around enforcement, for example through advance notification and eligibility limits.

The appropriate policy also depends on whether undeclared work is dependent or self-employment (Williams and Oz-Yalaman, 2021[63]). Undeclared dependent employment is more often driven by exclusion from formal work, so active labour-market policies, and sanctions of non-compliant employers can be effective. In contrast, undeclared self-employment can be addressed by simplifying administrative procedures and making it easier to register a business (Williams and Oz-Yalaman, 2021[63]).

3.6. Looking ahead: what are countries doing to rebalance the mix of social protection financing?

Copy link to 3.6. Looking ahead: what are countries doing to rebalance the mix of social protection financing?

A short policy questionnaire “Financing of national social protection systems”, circulated to EU and OECD countries in September 2025 asked countries about current and future reform plans to change the funding mix of the social protection system, or to access new funding bases. This section lists some of the answers countries have provided, to highlight current policy trends.

Three broad themes emerge from the responses:

3.6.1. Several countries are adjusting the financing of their pension systems

Germany extended the existing guaranteed pre‑tax pension level (48% of average wages; Haltelinie) until 2031 in December 2025, and extended pension credits for child-rearing periods for children born before 1992 (so-called “mothers” pension’). These measures will be mainly general revenue funded (Deutscher Bundestag, 2025[81]). In June 2026, the German pension commission published its recommendations which the German coalition committee has committed to bring into legislation (Alterssicherungskommission, 2026[82]). These include introducing a statutory capital-funded pension (gesetzliche Kapitalrente) with a transitional supplement for near-retirement cohorts, as well as extending the mandatory contributory coverage to previously exempt groups, notably the self-employed. On financing, the commission recommends introducing a link between the pension age and life expectancy and maintaining a uniform pension contribution rate on wages without considering any other factors nor levying contributions on other forms of income. The tax-funded federal payment to the pension insurance should be clearly delineated to cover expenses that are not related to contributions and made transparent.
Latvia increased the ceiling for both mandatory and voluntary contributions to the pension system. Social security contributions will also be levied on income from intellectual property.
Romania increased the contribution rate to the second pillar pension system in 2024. It also reformed the indexation mechanism in the pension system.
Malta is increasing the minimum contribution period for the old-age pension from 41 to 42 years. Malta will also link increases of the contribution ceiling to the maximum pension and increases in the minimum social security contribution to the statutory minimum wage. This means that minimum contributions will increase automatically with the minimum wage.
Some countries are increasing contribution rates to their pension systems. Cyprus is planning to increase social security contributions stepwise from 24.1% in 2029 to 26.7% in 2039. Similarly, Bulgaria is increasing pension contribution rates by 3 p.p. between 2025 and 2028. Luxembourg increased the pension insurance rate in January 2026 by 1.5 p.p. (0.5 p.p. for each contributor: employees, employers and government). Spain is also gradually increasing its additional contribution to the public pension system.
Spain is also raising contribution rates to improve the sustainability of the pension system while the “baby boom” generation is retired. Contribution rates increased by 0.8% in 2025 and will be progressively raised to 1.2% in 2029. This raise is expected to remain in force until 2050.

3.6.2. Several countries are boosting occupational and private pension schemes

Ireland started to auto‑enrol workers in the retirement savings scheme in January 2026. Employer and employee contributions will be collected by the government directly from payroll. There will also be a public subsidy.
Spain is also considering auto‑enrolment in its occupational pension scheme. This is following iterative reform steps over 2022 – 2024 that simplified occupational pension plans for dependent employees, the self-employed as well as public sector workers. While enrolment did increase, especially in the construction sector, overall enrolment remained low.
Germany increased the subsidy for low earners in occupational pension schemes.
Luxembourg increased the tax deductibility for contributions to (private) third pillar pension from January 2026.
Romania increased the contribution rate to the second pillar (privately managed pension funds) of the pension system by one p.p. Within a wider reform effort, Romania is also working to strengthen the third pension pillar (voluntary pensions).
Malta is planning to expand the role of private old-age pensions by auto‑enrolling workers in the occupational pension scheme.

3.6.3. Another focus area is the financing of in-kind benefits and services

The Slovak Republic is considering reforming the financing of social services. In particular, a new bill aims to introduce a new system of state participation in the co-financing of personal assistance in the context of population ageing.
Luxembourg allocated additional EUR 59 million per year in general-revenue funds to the health system between 2026 and 2030 to close a funding gap in the national health fund.
Romania broadened the base of contributors to health insurance: since August 2025, pensioners earning above about EUR 600, persons on parental leave, unemployment benefit recipients as well as social assistance recipients as well as people protected by special laws (such as veterans) will be subject to a 10% compulsory health insurance contribution. Romania also abolished free health insurance for spouses and parents without their own income.
Romania plans to reform the public health law by creating a regulatory framework to expand the use of voluntary private health insurance in public hospitals. This should increase the available funds for public health insurance, both because it attracts private funds but also because healthcare costs for some patients are being paid by the private, instead of the public, health insurance.

References

[18] Acemoglu, D. (2024), “The simple macroeconomics of AI”, Economic Policy, Vol. 40/121, pp. 13-58, https://doi.org/10.1093/epolic/eiae042.

[20] Acemoglu, D., D. Autor and S. Johnson (2026), “Building pro-worker artificial intelligence”, NBER Working Paper, Vol. 34854, https://www.nber.org/papers/w34854.

[22] Acemoglu, D. and P. Restrepo (2022), “Tasks, Automation, and the Rise in U.S. Wage Inequality”, Econometrica, Vol. 90/5, pp. 1973-2016, https://doi.org/10.3982/ecta19815.

[19] Acemoglu, D. and P. Restrepo (2018), “The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment”, American Economic Review, Vol. 108/6, pp. 1488-1542, https://doi.org/10.1257/aer.20160696.

[82] Alterssicherungskommission (2026), “Empfehlungen der Alterssicherungskommission”, BMAS, https://www.bundesregierung.de/resource/blob/975228/2444780/0fcbf810775351c64007ee6cb2c7cad9/2026-06-24-bericht-alterssicherungskommission-data.pdf?download=1.

[33] Arnemann, L. et al. (2025), “Taxes, social security, contributions and redistribution, welfare state and public finances”, WeLaR Policy Brief 2025, https://projectwelar.eu/wp-content/uploads/2025/03/WeLaR_Policy_brief_6.5_Taxes-social-policy.pdf.

[28] Auclert, A. et al. (2026), “Demographics, Wealth and Global IMbalances in the Twenty-First Century”, Review of Economic Studies, p. Forthcoming, https://web.stanford.edu/~aauclert/demowealth21.pdf.

[21] Autor, D. et al. (2024), “New Frontiers: The Origins and Content of New Work, 1940–2018”, The Quarterly Journal of Economics, Vol. 139/3, pp. 1399-1465, https://doi.org/10.1093/qje/qjae008.

[16] Autor, D. et al. (2020), “The Fall of the Labor Share and the Rise of Superstar Firms*”, The Quarterly Journal of Economics, Vol. 135/2, pp. 645-709, https://doi.org/10.1093/qje/qjaa004.

[85] Bergolo, M. and G. Cruces (2021), “The anatomy of behavioral responses to social assistance when informal employment is high”, Journal of Public Economics, Vol. 193, p. 104313, https://doi.org/10.1016/j.jpubeco.2020.104313.

[53] Bernal-Delgado E, A. (ed.) (2024), Spain: Health system review., Copenhagen: WHO Regional office for Europe, https://www.consaludmental.org/publicaciones/Health-System-Review-2024.pdf.

[70] Bosch, M. and J. Esteban-Pretel (2012), “Job creation and job destruction in the presence of informal markets”, Journal of Development Economics, Vol. 98/2, pp. 270-286, https://doi.org/10.1016/j.jdeveco.2011.08.004.

[79] Bott, K. et al. (2020), “You’ve Got Mail: A Randomized Field Experiment on Tax Evasion”, Management Science, Vol. 66/7, pp. 2801-2819, https://doi.org/10.1287/mnsc.2019.3390.

[50] Bozio, A. and E. Wasmer (2024), Les politiques d’exonérations de cotisations sociales : une inflexion nécessaire, https://www.strategie-plan.gouv.fr/publications/mission-bozio-wasmer-politiques-dexonerations-de-cotisations-sociales-une-inflexion.

[72] Burgstaller, L. and L. Feld (2026), “How (Not) to Ask About Undeclared Work: Estimating Prevalence and Patterns Using a Survey Experimental Approach”, CESifo Working Papers 12728.

[15] Cho, W., F. Manaresi and M. Reinhard (2025), “Superstars, shooting stars, and the labour share: Cross-country evidence”, OECD Science, Technology and Industry Working Papers, No. 2025/01, OECD Publishing, Paris, https://doi.org/10.1787/ecb32e77-en.

[31] Crowe, D. et al. (2022), “Population ageing and government revenue: Expected trends and policy considerations to boost revenue”, OECD Economics Department Working Papers, No. 1737, OECD Publishing, Paris, https://doi.org/10.1787/9ce9e8e3-en.

[35] De Nardi, M. et al. (2025), “Why Do Couples and Singles Save during Retirement? Household Heterogeneity and Its Aggregate Implications”, Journal of Political Economy, Vol. 133/3, pp. 750-792, https://doi.org/10.1086/733421.

[81] Deutscher Bundestag (2025), “Entwurf eines Gesetzes zur Stabilisierung des Rentenniveaus und zur vollständigen Gleichstellung der Kindererziehungszeiten”, Drucksache 21/1929 Gesetzentwurf der Bundesregierung, https://dserver.bundestag.de/btd/21/019/2101929.pdf.

[80] Di Porto, E. et al. (2026), “The Perverse Effect of Flexible Work Arrangements on Informality”, Working Papers 786.

[71] Donovan, K., W. Lu and T. Schoellman (2020), Labor Market Dynamics and Development, Federal Reserve Bank of Minneapolis, https://doi.org/10.21034/sr.596.

[55] ELA (2021), Tackling undeclared work in the personal and household services sector.

[32] European Commission (2026), Annual report on taxation 2026 – Review of taxation policies in the EU Member States, Publications Office of the European Union, https://data.europa.eu/doi/10.2778/9336135.

[8] European Commission (2026), Commission staff working document: Poverty in the EU - key trends and policies. European Union Anti-Poverty Strategy., European Commission, https://employment-social-affairs.ec.europa.eu/policies-and-activities/social-protection-social-inclusion/addressing-poverty-and-supporting-social-inclusion/eu-anti-poverty-strategy_en.

[41] European Commission (2026), Wealth taxation, including net wealth, capital and exit taxes – Executive summary, Publications Office of the European Union, https://data.europa.eu/doi/10.2778/0913466.

[54] European Commission (2024), 2024 Ageing Report, https://doi.org/10.2765/022983.

[30] European Commission (2021), “Business Taxation for the 21st Century”, Communication from the Commission to the European Parliament and the Council, COM (2021) 251 final, https://taxation-customs.ec.europa.eu/system/files/2021-05/communication_on_business_taxation_for_the_21st_century.pdf.

[61] European Commission (2020), “Undeclared Work in the European Union”, Special Eurobarometer 498, https://doi.org/10.2767/018492.

[9] European Commission and SPC (2024), The 2024 pension adequacy report. Current and future income adequacy in old age in the EU. Volume I, Publications Office of the European Union, https://doi.org/10.2767/909323.

[88] European Commission Joint Research Centre (2026), Demographic transformation in the EU: understanding change and adapting to its consequences, Publications Office of the European Union, https://doi.org/10.2760/9943353.

[12] European Commission, D. (ed.) (2023), The future of social protection and of the welfare state in the EU, Publications Office of the European Union, https://doi.org/10.2767/35425.

[3] Eurostat (2026), Life expectancy by age and sex [demo_mlexpec__custom_22329321], https://ec.europa.eu/eurostat/databrowser/view/tps00205/default/table?lang=en (accessed on 6 August 2026).

[89] Felderer, B., M. Gstrein and L. Mateeva (2011), Familienlastenauslgeich in Österreich 2011. Rückblick, Status-quo und Zukunftsperspektiven, Institut für Höhere Studien, https://irihs.ihs.ac.at/id/eprint/2089/1/flaf_2011_071111.pdf.

[6] Fluchtmann, J., M. Keese and W. Adema (2024), “Gender equality and economic growth: Past progress and future potential”, OECD Social, Employment and Migration Working Papers, No. 304, OECD Publishing, Paris, https://doi.org/10.1787/fb0a0a93-en.

[60] Franić, J., I. Horodnic and C. Williams (2023), Extent of undeclared work in the European Union..

[86] Garganta, S. and L. Gasparini (2015), “The impact of a social program on labor informality: The case of AUH in Argentina”, Journal of Development Economics, Vol. 115, pp. 99-110, https://doi.org/10.1016/j.jdeveco.2015.02.004.

[17] Georgieff, A. and R. Hyee (2022), “Artificial Intelligence and Employment: New Cross-Country Evidence”, Frontiers in Artificial Intelligence, Vol. 5, https://doi.org/10.3389/frai.2022.832736.

[25] Hourani, D. et al. (2023), “The taxation of labour vs. capital income: A focus on high earners”, OECD Taxation Working Papers, No. 65, OECD Publishing, Paris, https://doi.org/10.1787/04f8d936-en.

[34] Hourani, D. and S. Perret (2025), “Taxing capital gains: Country experiences and challenges”, OECD Taxation Working Papers, No. 72, OECD Publishing, Paris, https://doi.org/10.1787/9e33bd2b-en.

[56] ILO (2023), “Defining informality for contributing family workers”, International Conference of Labour Statisticians 21st session/No. 4.

[38] Immervoll, H. (2024), “Financing social protection in OECD countries: Role and uses of revenue earmarking”, OECD Social, Employment and Migration Working Papers, No. 312, OECD Publishing, Paris, https://doi.org/10.1787/0d53155c-en.

[90] Joubert, C. (2015), “Pension design with a large informal labor market: Evidence from Chile”, International Economic Review, Vol. 56/2, pp. 673-694, https://doi.org/10.1111/iere.12118.

[87] JRC (2024), Choosing Europe’s future. Global trends to 2040, Publications Office of the European Union, https://doi.org/10.2760/180422.

[11] Karabarbounis, L. (2024), “Perspectives on the Labor Share”, Journal of Economic Perspectives, Vol. 38/2, pp. 107-136, https://doi.org/10.1257/jep.38.2.107.

[58] Kelmanson, B. et al. (2019), “Explaining the Shadow Economy in Europe”, IMF Working Papers, Vol. 2019/2019, https://doi.org/10.5089/9781513520698.001.

[73] Kolev, A., J. La and T. Manfredi (2023), “Extending social protection to informal economy workers: Lessons from the Key Indicators of Informality based on Individuals and their Household (KIIbIH)”, OECD Development Centre Working Papers, No. 350, OECD Publishing, Paris, https://doi.org/10.1787/ca19539d-en.

[26] Koutsogeorgopoulou, V. and H. Morgavi (2025), “Ageing populations, their fiscal implications and policy responses”, OECD Economics Department Working Papers, No. 1844, OECD Publishing, Paris, https://doi.org/10.1787/6aec03b3-en.

[36] Krenek, A. et al. (2022), “INTAXMOD - Inheritance and Gift Taxation in the Context of Ageing”, JRC Working Papers on Taxation and Structural Reforms No 4/2022, https://publications.jrc.ec.europa.eu/repository/handle/JRC128480.

[74] La Porta, R. and A. Shleifer (2014), “Informality and Development”, Journal of Economic Perspectives, Vol. 28/3, pp. 109-126, https://doi.org/10.1257/jep.28.3.109.

[23] Minniti, A., K. Prettner and F. Venturini (2025), “AI innovation and the labor share in European regions”, European Economic Review, Vol. 177, p. 105043, https://doi.org/10.1016/j.euroecorev.2025.105043.

[51] Nay, O. et al. (2016), “Achieving universal health coverage in France: policy reforms and the challenge of inequalities”, The Lancet, Vol. 387/10034, pp. 2236-2249, https://doi.org/10.1016/s0140-6736(16)00580-8.

[1] OECD (2026), OECD Employment Outlook 2026: Geographic Disparities in Jobs and Incomes, OECD Publishing, Paris, https://doi.org/10.1787/7e710f54-en.

[77] OECD (2026), “Piloting digital solutions for risk-based inspections in Italy: Regional case studies”, OECD Regulatory Policy Working Papers, No. 25, OECD Publishing, Paris, https://doi.org/10.1787/6c37962f-en.

[4] OECD (2026), Population projections (dataset), https://data-explorer.oecd.org/s/56q.

[5] OECD (2025), OECD Employment Outlook 2025: Can We Get Through the Demographic Crunch?, OECD Publishing, Paris, https://doi.org/10.1787/194a947b-en.

[7] OECD (2025), Pensions at a Glance 2025: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e40274c1-en.

[29] OECD (2025), Revenue Statistics 2025: Disentangling Personal Income Tax Revenue in OECD Countries, OECD Publishing, Paris, https://doi.org/10.1787/3a264267-en.

[78] OECD (2024), Breaking the Vicious Circles of Informal Employment and Low-Paying Work, OECD Publishing, Paris, https://doi.org/10.1787/f95c5a74-en.

[49] OECD (2024), Fiscal Sustainability of Health Systems: How to Finance More Resilient Health Systems When Money Is Tight?, OECD Publishing, Paris, https://doi.org/10.1787/880f3195-en.

[45] OECD (2024), Is Care Affordable for Older People?, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/450ea778-en.

[2] OECD (2024), Megatrends and the Future of Social Protection, OECD Publishing, Paris, https://doi.org/10.1787/6c9202e8-en.

[14] OECD (2024), OECD Compendium of Productivity Indicators 2024, OECD Publishing, Paris, https://doi.org/10.1787/b96cd88a-en.

[76] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.

[40] OECD (2024), Taxation and Inequality: OECD Report to G20 Finance Ministers and Central Bank Governors, OECD Publishing, Paris, https://doi.org/10.1787/8dbf9a62-en.

[24] OECD (2023), OECD Employment Outlook 2023: Artificial Intelligence and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/08785bba-en.

[39] OECD (2022), Housing Taxation in OECD Countries, OECD Tax Policy Studies, No. 29, OECD Publishing, Paris, https://doi.org/10.1787/03dfe007-en.

[65] OECD (2021), Bringing Household Services Out of the Shadows: Formalising Non-Care Work in and Around the House, Gender Equality at Work, OECD Publishing, Paris, https://doi.org/10.1787/fbea8f6e-en.

[37] OECD (2021), Inheritance Taxation in OECD Countries, OECD Tax Policy Studies, No. 28, OECD Publishing, Paris, https://doi.org/10.1787/e2879a7d-en.

[10] OECD (2020), Promoting an Age-Inclusive Workforce: Living, Learning and Earning Longer, OECD Publishing, Paris, https://doi.org/10.1787/59752153-en.

[75] OECD (2019), Tax Morale: What Drives People and Businesses to Pay Tax?, OECD Publishing, Paris, https://doi.org/10.1787/f3d8ea10-en.

[42] OECD (2018), The Role and Design of Net Wealth Taxes in the OECD, OECD Tax Policy Studies, No. 26, OECD Publishing, Paris, https://doi.org/10.1787/9789264290303-en.

[68] Ohnsorge, F. and S. Yu (2022), The Long Shadow of Informality: Challenges and Policies, The World Bank, https://doi.org/10.1596/978-1-4648-1753-3.

[47] Pierson, P. (2000), “Increasing Returns, Path Dependence, and the Study of Politics”, American Political Science Review, Vol. 94/2, pp. 251-267, https://doi.org/10.2307/2586011.

[46] Pierson, P. (1993), “When Effect Becomes Cause: Policy Feedback and Political Change”, World Politics, Vol. 45/4, pp. 595-628, https://doi.org/10.2307/2950710.

[52] Rodrı́guez, M., R. Scheffler and J. Agnew (2000), “An update on Spain’s health care system: is it time for managed competition?”, Health Policy, Vol. 51/2, pp. 109-131, https://doi.org/10.1016/s0168-8510(99)00080-9.

[44] Sachverständigenrat (2026), Frühjahrsgutachten 2026 [Spring report 2026 of the German Expert Council Social Insurance], Sachverständigenrat Wirtschaft, https://www.sachverstaendigenrat-wirtschaft.de/fruehjahrsgutachten-2026.html.

[59] Schneider, F. and A. Asllani (2022), Taxation of the Informal Economy in the EU.

[83] Schneider, F., A. Buehn and C. Montenegro (2011), “Shadow Economies All Over the World: New Estimates for 162 Countries from 1999 to 2007”, in Handbook on the Shadow Economy, Edward Elgar Publishing, https://doi.org/10.4337/9780857930880.00007.

[84] Schneider, F. and D. Enste (2013), “Methods to estimate the size of the shadow economy”, in Schneider, F. and D. Enste (eds.), The Shadow Economy, Cambridge University Press, https://doi.org/10.1017/cbo9781139542289.003.

[57] Schoukens, P. and C. Bruynseraede (2026), Social security coverage and its financing for the self-employed in EU Member States, https://employment-social-affairs.ec.europa.eu/social-protection-social-inclusion/mutual-learning-workshop-access-social-protection-coverage-and-contributions-self-employed-2026-04-28_en.

[43] Schratzenstaller, M. (2022), “2022”, Wifo-Monatsberichte, Vol. 95/2, https://www.wifo.ac.at/publication/pid/17807005.

[69] Siachoque, J. (forthcoming), “The wage growth penalty of informal employment.”, University of Nottingham.

[27] Sicsic, M. and D. Hourani (2026), “The impact of population ageing on tax revenues in OECD countries”, OECD Taxation Working Papers, No. 75, OECD Publishing, Paris, https://doi.org/10.1787/94cbc7e5-en.

[13] Smith, M. et al. (2022), “The Rise of Pass-Throughs and the Decline of the Labor Share”, American Economic Review: Insights, Vol. 4/3, pp. 323-340, https://doi.org/10.1257/aeri.20210268.

[67] Ulyssea, G. (2020), “Informality: Causes and Consequences for Development”, Annual Review of Economics, Vol. 12/1, pp. 525-546, https://doi.org/10.1146/annurev-economics-082119-121914.

[48] Weaver, K. (2010), “Paths and Forks or Chutes and Ladders?: Negative Feedbacks and Policy Regime Change”, Journal of Public Policy, Vol. 30/2, pp. 137-162, https://doi.org/10.1017/s0143814x10000061.

[62] Williams, C. and I. Horodnic (2020), Trends in the undeclared economy and policy approaches. Evidence from the 2007, 2013 and 2019 Eurobarometer surveys, https://doi.org/10.2767/69833.

[66] Williams, C. and I. Horodnic (2016), “Tackling the undeclared economy in the European Union: an evaluation of the tax morale approach”, Industrial Relations Journal, Vol. 47/4, pp. 322-340, https://doi.org/10.1111/irj.12142.

[64] Williams, C., I. Horodnic and J. Windebank (2015), “Evaluating the Prevalence and Distribution of Envelope Wages in the European Union: Lessons from a 2013 Eurobarometer Survey”, Journal of Contemporary European Research, Vol. 11/2, https://doi.org/10.30950/jcer.v11i2.664.

[63] Williams, C. and G. Oz-Yalaman (2021), “Explaining different types of undeclared work: lessons from a 2019 Eurobarometer survey”, Employee Relations: The International Journal, Vol. 43/6, pp. 1382-1396, https://doi.org/10.1108/er-12-2020-0544.

[91] Zhang, Y. and T. Van Overbeke (2026), “Measuring Fiscal Asymmetry: Effective Tax Rates on Capital and Labour in 29 European Countries 1998 - 2021”, EIQ Paper No 196 / 2026, https://www.lse.ac.uk/european-institute.

Annex 3.A. Additional Statistics

Copy link to Annex 3.A. Additional Statistics

Annex Figure 3.A.1. Inheritance, estate, and gift tax revenues as percentage of total tax government revenues in per cent 2024
Copy link to Annex Figure 3.A.1. Inheritance, estate, and gift tax revenues as percentage of total tax government revenues in per cent 2024

Notes

Copy link to Notes

← 1. Some replacement incomes are also subject to social contributions, but the majority is levied on labour income, see Chapter 1.

← 2. Fluchtmann, Keese and Adema (2024[6]) take cohort-specific observed labour force participation and unemployment rates as starting points and derive future rates by assuming constant future unemployment rates and that, for each age and gender category, future labour market entry and exit rates are equal to the most recent rates of the same age category (averaged over a 5‑year period and excluding the years of the COVID‑19 crisis). The labour market entry rate is defined as the percentage change in nonemployment rates from age t to age t+1 for the same cohort, when negative and zero otherwise (that is the percentage decrease when it is effectively a decrease). The labour market exit rate is defined as the percentage change in employment rates from age t to age t+1 for the same cohort, when this is negative and zero otherwise (that is, again, the percentage decrease when it is effectively a decrease).

← 3. The extent of migration will be influenced by demographic developments (e.g. the speed of the decrease of fertility in Sub-Saharan Africa), external factors such as the effects of climate change, as well as policy choices, but demographic imbalances make it likely that immigration will remain relevant in particular for Europe (JRC, 2024[87]).

← 4. Raising the employment rate of workers aged 55 and older by 2060, so that the rate of decline in employment rates in the 50‑54 age group will be as small as the 10th percentile of the distribution among OECD countries, see notes to Figure 3.1.

← 5. For instance, the number of people over the age of 50 potentially requiring long-term care in the EU is projected to rise from 36 million in 2025 to 48 million by 2070 – 11% of the EU population (European Commission Joint Research Centre, 2026[88]).

← 6. This scenario raises the working hours of the gender with the higher working hours in each age group (in most cases men) to the age 80th percentile across countries for the same age group. Working hours for the other gender are then raised to maintain the gender gap in hours worked. See also notes to Figure 3.1 for an example. Working hours of young workers are not raised as education participation may remain the same, or increase, until 2060.

← 7. GVA is GDP minus taxes on production, which cannot be attributed to capital or labour in the production process. See notes to Figure 3.2 for details.

← 8. Including dividend, interest, other property income, imputed rents and retained earnings of corporations.

← 9. This “step-up in basis” approach is very common: 13 of 27 OECD countries with available information apply it to inherited assets, at least for some asset classes (OECD, 2021[37]), as do 16 of 27 EU countries (Zhang and Van Overbeke, 2026[91]). The second most common approach is passing capital gains to heirs (“carry-over basis”), and only three countries – Denmark, Hungary and Canada – tax unrealised capital gains at death. This can create avoidance opportunities, especially if there are no inheritance taxes, or inheritance tax exemptions are very high (OECD, 2021[37]).

← 10. The item “recurrent taxes on net wealth” in Table 3.1 includes both taxes levied on the wealth of individuals and corporations. Since not all countries report the breakdown, it is not shown. Including recurrent taxes on net wealth of corporations, nine OECD and EU countries levy such taxes.

← 11. The two figures are not fully comparable, since Figure 3.4 only includes cash benefits received by households, whereas Figure 1.8 in Chapter 1 includes cash and in-kind benefits. The value of in-kind benefits is not included in the EU-SILC.

← 12. The reason for this financing structure is historic: when the family benefit (and the fund for its financing) were established in 1949, it was part of a collective bargaining agreement, that provided family benefits only for children of dependent employees. The family benefit was extended to all children in 1955 (Felderer, Gstrein and Mateeva, 2011[89]).

← 13. ELA: the European Labour Authority, an EU agency established by Regulation (EU) 2019/1149 and operational since 2019. Since 26 May 2021 it has also hosted the European Platform tackling undeclared work, originally established by Decision (EU) 2016/344 in 2016, as one of its permanent working groups.

← 14. The calculation of the size and development of the shadow economy is done with the MIMIC (Multiple Indicators and Multiple Causes) estimation procedure. The model estimates the unobserved shadow economy as a latent variable in a structural model, inferred jointly from its economic causes, the tax and social-contribution burden, regulatory intensity, the quality of institutions and tax morale and its observable indicators, such as cash demand, official GDP and labour-force participation. The resulting relative index is then scaled to a share of GDP using an external benchmark for a base year (Schneider and Enste, 2013[84]; Schneider, Buehn and Montenegro, 2011[83]).

← 15. The labour input method compares hours worked reported in the Labour Force Survey – which captures undeclared as well as declared work – and hours worked recorded in Structural Business Statistics, which captures only declared work. This labour gap is valued at factor cost and expressed as a share of GVA rather than GDP and is therefore not directly comparable with the macro estimates of the shadow-economy above.

← 16. Eurobarometer (2019) question: “Have you yourself carried out any undeclared paid activities in the last twelve months, either on your own account or for an employer?”. Envelope wages are measured by a separate question: “Has your employer paid you any of your income in the last twelve months in cash and without declaring it to tax or social security authorities?”. Across waves, the questions have been reworded slightly (Williams and Horodnic, 2020[62]).

← 17. This refers to the same question as in footnote 16 above, tabulated by gender.

← 18. This result is closely related to the broader literature on informality and means-tested transfer programmes in low- and middle‑income countries. Because eligibility for such programmes is typically withdrawn as declared income rises, they act as an implicit tax on formalisation and can lead workers to prefer informal or undeclared work to retain benefits. Evidence from Latin America finds that means-tested cash transfers reduce formal employment (Bergolo and Cruces, 2021[85]; Garganta and Gasparini, 2015[86]), and that if social contributions are perceived as a pure tax rather than as a link to future benefits informality rises. For instance, a 5 p.p. increase in Chile’s mandatory pension contribution rate was estimated to have raised the size of the informal sector by around 14% among men and 9% among women (Joubert, 2015[90]).

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment