This chapter assesses the need for different categories of workers across OECD countries to complement their mandatory pension in order to reach their retirement income objective, and calculates how much more they would need to contribute. It also evaluates the extent to which individuals already contribute voluntarily to complement their mandatory pensions and assesses the effectiveness of different policies to increase contributions. The chapter ends with policy guidance on how to ensure sufficiently high contributions to achieve retirement income objectives.
OECD Pensions Outlook 2026
Are We Saving Enough for Retirement?
Abstract
Many people may lack enough resources to finance their retirement. Studies in a wide range of countries suggest that a large proportion of the population may be unprepared for retirement (European Commission, 2024[1]; Wolf and Lorena, 2021[2]; Yin, Chen and Munnell, 2024[3]; Department for Work and Pensions, 2025[4]). For example, in the United Kingdom, 73% of the working-age population is projected to have a pension income that falls below a moderate retirement living standard, while in the United States, 39% of households may be at risk of being unable to maintain their pre‑retirement standard of living in retirement. The reasons vary, with workers failing to contribute to mandatory pension systems fully throughout their career or failing to complement mandatory pensions with voluntary savings. This is a major concern as population ageing increases the number of years that retirement savings need to cover and raises income needs to finance long-term care, on top of the financial constraints that it imposes on pay-as-you-go (PAYG) pension systems. The lack of retirement readiness may therefore affect a growing portion of the population in the future.
This chapter focusses on raising pension contributions to increase retirement income. Several policies are available to increase retirement income and reduce the share of the population unprepared for retirement. Governments may increase public deficits, reallocate budgets or increase general taxes to raise pension benefits from public PAYG pension schemes. Another approach is to delay retirement ages to balance the periods during which people contribute and receive benefits. A third approach is to increase pension contributions so that people can earn additional pension entitlements. This chapter focusses on the latter.
Increasing pension contributions involves trade‑offs. It may improve future retirees’ well-being and support economic growth if these contributions go into asset-backed pension plans and get invested in the real economy (Oxford Economics, 2026[5]). However, increasing pension contributions from employers may raise labour costs, undermine business competitiveness and reduce wage growth, while increasing pension contributions from workers may reduce disposable income and consumption.
This chapter assesses the need to increase pension contributions in OECD countries and provides policy guidance on how to ensure sufficiently high contributions to achieve retirement income objectives.1 The analysis focusses on the role of asset-backed pension plans as a complement to PAYG schemes. It assesses the need for additional contributions across countries and different categories of workers, evaluates the extent to which individuals already contribute voluntarily to complement their mandatory pensions, and assesses the effectiveness of different policies to increase contributions.
In all the countries analysed but one, certain categories of workers would need to complement their mandatory pension to reach their retirement income objective. On average across countries, contribution gaps are larger for average and high earners, women and the self-employed. Everything else being equal, the contribution gap is larger for workers with lower retirement ages, higher life expectancies, shorter contribution periods and lower investment performance in their portfolios. However, the saving behaviour of current workers suggests that some of them may fail to reach their retirement income objective, as young workers, low earners, women and self-employed workers tend to participate and contribute less to asset-backed pension plans, and contribution levels frequently align with low default or minimum contribution rates.
Behavioural biases, low levels of financial literacy, affordability concerns, career interruptions and selected design features may pose barriers preventing individuals from contributing enough voluntarily given their retirement income objective. Various policies exist to address the barriers that individuals face when saving for retirement and steer individuals towards contribution levels in line with their retirement income objective. The different policies include increasing mandatory contributions, encouraging voluntary employer contributions, nudging individuals, providing financial incentives, informing individuals, and addressing the needs of individuals.
The chapter first determines the contribution gap in OECD countries by calculating the contribution rate that would need to be paid into a defined contribution pension plan to offset the difference between the replacement rate promised by mandatory pension systems and a target replacement rate. Section 2.2 looks at the actual saving behaviour of different individuals to see whether they already make additional voluntary contributions to complement their mandatory pensions and to understand the barriers that may prevent them from doing so. Section 2.3 assesses the effectiveness of different policies to steer people towards contribution levels in line with their retirement income objective. The final section concludes and provides 11 steps that policymakers could follow to ensure sufficiently high contributions to achieve retirement income objectives.
The first step to assess the need for higher contributions is to calculate contribution gaps. The analysis shows that contribution gaps vary widely across OECD countries and tend to be larger for middle and high earners, women and the self-employed.
The analysis defines the contribution gap as the contribution rate that a worker would need to pay into a defined contribution (DC) pension plan to offset the difference between a retirement income objective and the expected retirement income from the mandatory pension system. It shows how far workers are from reaching their retirement income objective, expressed as the voluntary contribution rate needed to complement their mandatory pension. The calculation of the contribution gap involves three steps:
1. Defining the expected retirement income from the mandatory pension system. The analysis uses the theoretical gross replacement rates from the OECD Pensions at a Glance Database (OECD, 2025[6]) as a proxy for the retirement income people in different countries can expect from mandatory pension systems. These rates measure how effectively, under current rules, mandatory pensions replace pre‑retirement earnings after a full career for workers entering the labour market today (OECD, 2025[7]). They are available for different levels of earnings, sexes and types of employment.2
2. Defining the retirement income objective. The analysis uses different target replacement rates at different levels of earnings to reflect the retirement income objective of maintaining living standards. These common cross-country targets are 80% for low earners (half of average earnings), 70% for average earners and 60% for high earners (twice average earnings) (Box 2.1).
3. Calculating the contribution gap. The analysis calculates the difference between the target replacement rate and the theoretical gross replacement rate, and, if the difference is positive, it calculates the contribution rate that a worker would need to pay into a DC pension plan to reach a gross replacement rate equal to that difference. This calculation assumes that workers contribute a constant share of earnings every year to the DC plan, with earnings growing in line with 2% inflation and 1.25% productivity growth, contributions earning a constant real investment return of 2.5%, and fees of 1% of assets. It assumes further that workers contribute to the DC plan for a full career, from age 22 until the normal age of retirement in each country, and, at the time of retirement, they buy a lifelong annuity calculated using a fixed discount rate of 1.5% in real terms and survival probabilities from the United Nations’ World Population Prospects 2024 database.3
There is no universal retirement income objective. The analysis herein uses 80%, 70% and 60% as common target replacement rates for low earners (half of average earnings), average earners and high earners (twice average earnings), respectively. There is no authoritative source for using these specific targets, but they are consistent with the reviewed literature.
There are several factors that justify different retirement income objectives across countries and individuals:
Spending needs: Spending needs in retirement likely vary across countries. For example, out-of-pocket healthcare expenses for retirees depend on the health insurance system. Within countries, spending needs may also vary across individuals, for example according to homeownership, debt levels, household composition or place of residence.
Desired lifestyle: Individuals may have different preferences leading to different desired lifestyles in retirement. For example, some individuals may want to engage in new learning, volunteering or social activities, while others may want to travel, translating into different levels of expenditures.
Tax system: The extent to which retirement income is taxed varies across countries, leading to different levels of net disposable income. Overall, 28 OECD countries out of 38 have some concession for older people or pensioners under their personal income tax system (e.g. higher personal tax allowances, extra tax credits, specific tax rules for pension income) (OECD, 2025[7]).
The retirement income objective could be expressed as a consumption benchmark. Unfortunately, only some countries have developed consumption benchmarks reflecting the specific expenditure levels that would be required to achieve different lifestyles in retirement, such as for example modest, moderate or comfortable, and their methodology may not be comparable (OECD, 2024[8]).
The retirement income objective is therefore expressed as a target replacement rate in this chapter. The replacement rate is defined as the ratio between the initial pension income and the final earnings and is commonly used to reflect the objectives of maintaining standard of living and smoothing consumption over working and retirement years. Only a few countries may have an explicit target replacement rate (OECD, 2014[9]).
The analysis herein uses a common target for all OECD countries. Although country differences in terms of income levels, spending needs and tax systems would suggest using different target replacement rates for different countries, a review of 52 studies and reports covering a range of countries shows that most target replacement rates fall between 70% and 85% (Government Accountability Office, 2016[10]). The OECD average replacement rate would be a convenient common target for all OECD countries, but the value of 52% for an average earner falls significantly below the range observed in other studies (OECD, 2025[7]).
The retirement income objective herein accounts for one important source of heterogeneity in target replacement rates, i.e. the income level. Studies in several countries find that the replacement rate needed to smooth consumption declines with income, since lower-income earners need to replace a higher proportion of their pre‑retirement earnings to avoid falling into poverty in retirement (Binswanger and Schunk, 2011[11]; Schmied, 2023[12]; Yin, Chen and Munnell, 2023[13]). However, the range of target replacement rates across the income distribution varies across countries, for example from 95% to 45% in the United States, from 75% to 60% in the Netherlands, and from 80% to 50% in the United Kingdom (Binswanger and Schunk, 2011[11]; Broome and Mulheirn, 2024[14]).
In conclusion, the analysis uses different retirement income objectives measured by different target replacement rates for low earners (80%), average earners (70%) and high earners (60%).
The theoretical gross replacement rates from the OECD Pensions at a Glance Database used as a proxy for the expected retirement income from the mandatory pension system have two main limitations. First, they represent the maximum expected retirement income since the pension models calculate them assuming that workers participate in the labour market during a full career from age 22 until the normal retirement age. However, most people lack full careers due to career breaks, immigration, periods of informality or spells of unemployment. Assuming a shorter career would reduce both the expected retirement income and the length of the contribution period into the voluntary DC pension plan, resulting in a larger contribution gap.
Second, the theoretical gross replacement rates assume that current rules for calculating mandatory pensions will remain in place throughout the entire career of workers entering the labour market today. However, the challenges facing pension systems in some countries may lead to lower future replacement rates. Indeed, public pension expenditures are expected to increase from 8.8% of GDP in 2023-2024 to 10.0% of GDP in 2050 on average across OECD countries with available data, with increases of more than three percentage points (p.p.) of GDP in Korea, Lithuania, Luxembourg, Slovenia and Spain (OECD, 2025[7]). Some countries may therefore implement pension reforms to limit the rise in expenditures, which could lead to lower future replacement rates from mandatory pension systems and increase the contribution gap. The calculated contribution gaps should therefore be interpreted as minimum estimates, as shorter careers and possible future pension reforms could make the true gaps larger.4
The contribution gap is 10% or more for male employees on average earnings in 17 countries out of 41. In other words, these workers would need to contribute at least 10% of their earnings to a voluntary DC pension plan to reach the 70% target replacement rate, on top of what they already contribute to their mandatory pension system. The contribution gap varies across countries, ranging from less than 5% in Bulgaria, Costa Rica, Finland, Mexico, New Zealand, the Slovak Republic, Sweden and the Republic of Türkiye (hereafter Türkiye), to more than 15% in Croatia, Ireland, Korea, Lithuania and Poland (Figure 2.1). In nine countries, the contribution gap is zero since the mandatory pension system already delivers an expected replacement rate of at least 70% for male employees on average earnings. In Switzerland, the 12% contribution gap may be partially offset by mandatory contribution rates in occupational pension plans above the legal minimum.
Note: Male employees on average earnings have a contribution gap of zero in Austria, Colombia, Denmark, Greece, Italy, Luxembourg, the Netherlands, Portugal and Spain, since the mandatory pension system already delivers an expected replacement rate of at least 70% for these workers.
The contribution gap is the largest for high earners in half of the countries, but in 12 other countries low earners need to contribute the most to a voluntary DC pension plan to reach their target replacement rate. Figure 2.2 shows contribution gaps for male employees split across three panels, with 12 countries in which there is a higher gap for low earners, 7 countries in which there is a higher gap for average earners and 19 countries in which there is a higher gap for high earners. The average contribution gap across countries with positive gaps is 8.6% for low earners and 9.7% for both average and high earners (among male employees). The progressivity in the mandatory pension systems could explain the larger contribution gaps for average and high earners. However, large contribution gaps for low earners may be of concern as they often face greater barriers to voluntary savings. Overall, there are 13 countries where low earners would need to contribute at least 10% of earnings (Canada, Croatia, France, Germany, Hungary, Ireland, Japan, Korea, Latvia, Lithuania, Poland, Switzerland and the United States).
Note: Male employees have a contribution gap of zero in Greece, Luxembourg and the Netherlands, since the mandatory pension system already delivers an expected replacement rate of at least 70% for these workers, irrespective of the earnings level.
The contribution gap is larger for women than for men in 17 countries. Among employees on average earnings, the average contribution gap, across countries with positive gaps, is 9.7% for men and 10.8% for women (Figure 2.3). Two main factors explain this difference. First, women live on average longer than men. While insurance companies must use gender-neutral mortality tables to price annuities in EU Member States, annuity prices in other OECD countries reflect gender differences in life expectancy. In these countries, higher life expectancy increases the contribution gap, since assets accumulated in the voluntary DC pension plan must finance a longer retirement period.
Second, women can retire earlier than men in some countries. Earlier retirement shortens the contribution period in DC plans, increasing the share of earnings that one should save voluntarily each year to reach the target replacement rate. In Costa Rica, Hungary, Israel, Poland and Türkiye, women can retire two to five years earlier than men, leading to lower expected replacement rates from the mandatory pension system, shorter contribution periods in the voluntary DC pension plan and, therefore, larger contribution gaps.
Note: The figure only shows countries where the contribution gap differs for male and female employees on average earnings.
The contribution gap is larger for the self-employed than for employees in 31 countries. This mainly reflects lower expected replacement rates from mandatory pension systems for the self-employed. In some countries, they are not required to participate in at least part of the mandatory pension system (e.g. Australia, Germany, Japan, Mexico, the Netherlands, Switzerland). In others, they contribute less than employees through lower contribution rates or flat-rate contributions and therefore accrue lower pension entitlements (e.g. Greece, Israel, Norway, Poland). The average contribution gap is 9.7% for employees and 13.1% for the self-employed, among male workers on average earnings. The contribution gap exceeds 15% for the self-employed in 13 countries (Figure 2.4).
Note: The figure only shows countries where the contribution gap differs for employees and the self-employed on average earnings.
All but one of the 41 countries analysed have contribution gaps. For each category of workers, there are between 32 and 38 countries with a contribution gap, and when combining the different categories of workers, there are 40 countries with a contribution gap. The only OECD country where the mandatory pension system promises a replacement rate after a full career of 60% to 80% depending on income for men, women, employees and the self-employed is Luxembourg.
The calculated contribution gaps are sensitive to several assumptions. First, it may be too optimistic to assume that workers save in a voluntary DC pension plan throughout their entire career. Some people may only start saving for retirement later in life because early-career resources are devoted to repaying student loans, buying a first home or covering child-related expenses, for example. Starting voluntary contributions at age 45 instead of 22 would roughly halve the contribution period in a DC plan in most countries and double the contribution gap. Second, the contribution gap depends on the investment performance of the portfolio in a DC plan. Returns depend on the investment strategy chosen by or allocated by default to plan members and may vary across generations given different capital market conditions. Lower returns increase the contribution gap since they reduce the assets accumulated at retirement and, in turn, the complementary income provided by the voluntary DC plan.
The second step to assess the need for higher contributions is to explore whether individuals are currently saving enough voluntarily to close the contribution gaps identified previously. This section looks first at voluntary contributions to asset-backed pension plans across different individuals in selected OECD countries and shows that certain categories of workers may not contribute enough voluntarily to close their contribution gap and reach their retirement income objective. The section then looks at the barriers that may prevent sufficient voluntary contributions. It ends with a discussion about other forms of savings that individuals may tap into to finance retirement.
2.2.1. Voluntary contributions to asset-backed pension plans
There are significant portions of the working-age population in many countries that lack voluntary savings in asset-backed pension plans. The first step to close the contribution gap is to be a member of an asset-backed pension plan. However, few individuals participate in asset-backed pension plans on a voluntary basis to complement their mandatory pensions. Participation in voluntary pension plans varies across countries, from less than 10% of the working-age population in Costa Rica, Greece, Luxembourg or Mexico, to around 50% to 60% in Belgium, the Czech Republic (hereafter ‘Czechia’), Germany, Iceland, Ireland, Japan and Poland (Chapter 1).5 In 9 OECD countries out of 19 where only voluntary plans are available, fewer than half of the working-age population participates.
Participation rates in voluntary pension plans tend to be lower among younger workers, lower-income earners, the self-employed and women. Participation rates tend to increase with age and reach a peak towards the middle to the end of the career (OECD, 2012[15]). As a result, many workers may not contribute to a voluntary pension plan for a full career. Participation rates tend to increase with income, reaching a plateau from the 7th or 8th income decile (OECD, 2012[15]). Additionally, fewer women than men tend to participate in voluntary pension plans, in particular when occupational pension plans are not available in industries employing mostly women (OECD, 2021[16]). Finally, self-employed workers are less likely to participate in asset-backed pension plans than employees (OECD, 2019[17]).
Among individuals participating in an asset-backed pension plan, some may not actively and continuously contribute to it. For example, in Belgium, 78% of the labour force participates in a complementary pension plan, but this proportion falls to 59% when only considering individuals who currently contribute and accumulate pension entitlements (PensionStat.be, 2025[18]). In Ireland, around 20% of individuals with a personal pension plan are not contributing to it (Central Statistics Office, 2024[19]). In New Zealand, 30% of KiwiSaver members aged 18 to 65 were not contributing as of 31 March 2025 (Financial Markets Authority, 2025[20]), and the self-employed are more likely to be non-contributing members than employees (Te Ara Ahunga Ora Retirement Commission/Hnry, 2025[21]). In the United Kingdom, 58% of adults have a pension plan in the accumulation phase, but this proportion falls to 45% when focussing on pension plans into which contributions are currently being paid (Financial Conduct Authority, 2025[22]).
Contribution rates tend to be around a few and relatively low values. This is particularly the case when default or minimum contribution rates are present (Figure 2.5). In the United Kingdom, the most common employer and employee contribution rates to occupational DC plans correspond to the minima established for automatic enrolment, i.e. 3% of qualifying earnings for employers and 5% for employees. Similarly, in New Zealand, approximately 63% of KiwiSaver members contribute at the minimum and default rate, while 91% received employer contributions at the minimum rate. These figures include members who did not contribute for the full year and therefore have an annualised contribution rate below 3%.6 In Norway, 30% of employees receive the statutory minimum 2% contribution rate from their employer, while 50% receive an additional rate of 9.1% on average. Even in the absence of a minimum or default contribution rate, the distribution of contributions rates tends to be skewed towards low values. For example, in Belgium, 66% of employees who actively participate in an occupational pension plan have a total contribution rate below 3%. In Spain, 54% of occupational pension accounts did not receive any contribution in 2024, and among those with positive contributions, 68% received no more than EUR 900 during the year.
Notes: Data refer to complementary pension plans for employees for Belgium, occupational pension plans managed by Irish Life for Ireland, KiwiSaver plans for New Zealand, occupational DC schemes for Norway and the United Kingdom, occupational pension plans for Spain and occupational pension plans managed by Vanguard for the United States.
Source: Based on Sigedis (2025[23]); Irish Life (2018[24]); Kirkpatrick, L., L. Meehan and G. Pacheco (2024[25]); Finance Norway (2025[26]); DGSFP (2025[27]); ONS (2024[28]); ONS (2024[29]); Vanguard (2025[30]).
Average voluntary contributions tend to vary with income. In Belgium, New Zealand and the United States, contribution rates increase with income among individuals participating in a plan (Figure 2.6). By contrast, average contribution rates to Riester plans decline with income in Germany, as these personal pension plans benefit from government subsidies that are more attractive for lower-income earners. For Australia, data suggest that average contribution rates may be higher for lower-income earners than for middle‑income earners. This could be the result of the super co-contribution programme, which encourages lower-income earners to make voluntary contributions as the government pays matching contributions only to individuals earning below a certain threshold. In Canada and Denmark, average contribution levels are very similar up to the 6th or 7th decile and increase more sharply with income afterwards.
Notes: Data refer to personal contributions to superannuation funds for Australia, complementary pension plans for employees for Belgium, registered pension plans, registered retirement savings plans and tax-free savings accounts for Canada, personal pension plans for Denmark, occupational and Riester pension plans for Germany, KiwiSaver plans for New Zealand, occupation DC schemes for the United Kingdom and occupational pension plans managed by Vanguard for the United States.
Source: Based on Australian Taxation Office (2025[31]), Sigedis (2025[23]); Statistics Canada (2023[32]); F&P (2025[33]); Federal Ministry of Labour and Social Affairs (2025[34]); Kirkpatrick, L., L. Meehan and G. Pacheco (2024[25]); Department for Work & Pensions (2025[35]); Vanguard (2025[30]).
Men and women tend to have relatively comparable voluntary contribution rates or levels. In the United States and Belgium, contribution rates are higher for men than for women, but differences are below 60 basis points of earnings (Figure 2.7). In Australia and Denmark, women contribute more than men on average annually, respectively by around AUD 1 300 and DKK 1 200.
Notes: Data refer to personal contributions to superannuation funds for Australia, complementary pension plans for employees for Belgium, personal pension plans for Denmark, occupational (“occ” in the figure) and Riester pension plans for Germany, KiwiSaver plans for New Zealand and occupational pension plans managed by Vanguard for the United States.
Source: Based on Australian Taxation Office (2025[31]); Sigedis (2025[23]); F&P (2025[33]); Federal Ministry of Labour and Social Affairs (2025[34]); Kirkpatrick, L., L. Meehan and G. Pacheco (2024[25]); Vanguard (2025[30]).
Finally, voluntary contributions tend to increase with age, reaching a maximum when people are in their 50s or 60s, and sometimes even later. In most countries, average contribution rates or levels are higher for older people, indicating that individuals may increase their voluntary contributions as they get closer to retirement age (Figure 2.8). However, in Germany, individuals aged 25 to 35 contributed slightly more as a percentage of their earnings to their Riester plan (2.9%) than those aged 35 to 55 (2.6%) in 2021. In the United Kingdom, median contribution rates are relatively constant between 26 and 65 years old. Increasing contributions with age is in line with what the literature suggests is optimal for individuals to do, given higher liquidity constraints at the beginning of a career (e.g. repaying student loans, saving for a deposit to buy a home) and higher earnings at the end (Schlafmann, Setty and Vestman, 2024[36]; Fischer, Jensen and Koch, 2023[37]). However, needed late‑career contribution rates may end up being too large and individuals may not have the discipline to raise contribution rates to that level.
Note: Data refer to personal contributions to superannuation funds for Australia, registered pension plans, registered retirement savings plans and tax-free savings accounts for Canada, personal pension plans for Denmark, occupational and Riester pension plans for Germany, KiwiSaver plans for New Zealand, occupational DC schemes for the United Kingdom and occupational pension plans managed by Vanguard for the United States.
Source: Based on Australian Taxation Office (2025[31]); Statistics Canada (2023[32]); F&P (2025[33]); Federal Ministry of Labour and Social Affairs (2025[34]); Kirkpatrick, L., L. Meehan and G. Pacheco (2024[25]); Department for Work & Pensions (2025[35]); Vanguard (2025[30]).
This analysis suggests that low earners, women and self-employed workers, who have the largest contribution gaps in some countries, are also the ones least likely to participate in asset-backed pension plans and the ones contributing the least. Additionally, the pattern of rising participation and contributions by age suggests that voluntary contributions may need to increase significantly at the end of the career to close the overall contribution gap.
2.2.2. Barriers preventing individuals from contributing enough voluntarily
Several barriers may prevent individuals from contributing enough voluntarily to achieve their retirement income objective. These include behavioural biases, low levels of financial literacy, affordability concerns, career interruptions and selected design features.
Behavioural biases may affect individuals’ contribution decisions. Inertia and procrastination, due to the complexity of retirement planning and regret aversion, often lead individuals to delay or fail to make decisions that would be in their best interest, such as increasing contributions (Money & Pensions Service, 2020[38]; OECD, 2018[39]). Saving for retirement requires self-control, but due to present bias, individuals tend to put more emphasis on immediate gratification (e.g. through consumption) rather than future benefits (e.g. through savings), thereby failing to act on their good intentions. Distance to retirement makes future pension benefits more abstract and harder to evaluate for individuals, who may underestimate how much income they will need in retirement. This also relates to the fact that individuals think in nominal terms and fail to consider the impact of inflation on future spending. Loss aversion may also affect pension contributions as individuals may be reluctant to reduce take‑home pay to increase contributions.
Individuals may have low levels of financial literacy affecting their decision on an appropriate contribution rate. Determining the appropriate contribution rate requires knowledge, numeracy and financial skills (OECD, 2018[39]). Individuals should figure out how much money they will need in retirement and understand how contributions paid into a DC pension plan affect future retirement income. This requires numeracy and financial skills to assess whether the contributions paid will eventually translate into an income that will cover their needs in retirement. That assessment should be done given the level of retirement income that they may expect from other components of the pension system, which requires knowledge of how benefits are calculated in other pension schemes. Individuals also need to understand that income from DC pension plans is uncertain due to financial, demographic and labour market risks. A given level of contribution may be sufficient to achieve a retirement income objective only under certain circumstances, depending on how the different risks materialise. Moreover, many adults may have insufficient understanding of basic financial concepts and insufficient financial skills to take sound financial decisions (OECD, 2023[40]).
Low levels of financial literacy may lead to the use simple rules of thumb or heuristics to determine contribution rates. For example, individuals may select contribution rates at multiples of 5%, based on the maximum contribution rate allowed by the plan or by tax rules, or based on the minimum necessary to get the full employer matching contributions (Benartzi and Thaler, 2007[41]). However, such rules may not result in an appropriate level of contribution given their personal circumstances.
Another consequence of behavioural biases and low levels of financial literacy is a lack of engagement of people participating in a pension plan. As a result of this lack of engagement, they know little about how much they have accumulated, how much they should contribute and how their money is invested. In 2024, only 24% of UK adults actively contributing to a DC pension plan were considered as highly engaged with their pension based on a combination of nine indicators. While 62% knew how much was contributed to their plan, only 20% had thought a lot about how much they should be contributing each year to maintain a reasonable standard of living in retirement (Financial Conduct Authority, 2025[22]).
Willingness to trade current pay for future pension benefits may vary across different individuals. In the United Kingdom, workers are willing to forego 0.3% of their wage on average for a 1 p.p. increase in future retirement income (Deschacht, Guillemyn and Vujić, 2025[42]). The willingness to forego wages for pension benefits is larger for older workers, workers actively saving for their pension, higher-income earners, and workers with higher financial literacy and higher risk aversion. This indicates that willingness to contribute to a pension plan would differ across workers with different characteristics to the extent that such contributions would reduce their take‑home pay.
Competing financial goals and affordability concerns may prevent people from saving for retirement. People may prioritise saving towards a major expense (e.g. a house, a car, education) and paying down debt over saving for retirement (Pew, 2025[43]). Low-income earners may feel that pension contributions are not affordable and prioritise short-term budgeting (Department for Work and Pensions, 2024[44]). Moreover, the economic environment may affect individuals’ ability to save for retirement, with people more likely to reduce or stop contributions in times of economic uncertainty and rising costs of living (Allianz Center for the Future of Retirement, 2026[45]; Living Wage Foundation, 2024[46]). In the United Kingdom, low-income earners on precarious jobs feel their income is too unsecure to save for retirement (Department for Work and Pensions, 2024[44]).
Mandatory pension contributions may crowd out voluntary contributions. The average effective contribution rate to mandatory or quasi-mandatory pension schemes equalled 18.8% for private‑sector employees on average earnings in the OECD in 2024 (OECD, 2025[7]).7 It exceeded 25% in Czechia, France, Greece, Italy, Spain and the United Kingdom. This may leave limited room for voluntary contributions.
Employers may be unwilling to contribute more than the minimum on behalf of their employees. In New Zealand, only 9% of employees receive employer contributions to their KiwiSaver plan above the minimum rate of 3% (Te Ara Ahunga Ora Retirement Commission, 2025[47]). In the United Kingdom, 37% of employers only offer the minimum contribution rate established in the automatic-enrolment legislation, i.e. 3% of a qualifying band of earnings, to all their employees (Nest Insight, 2022[48]). This proportion increases to 59% among employers with less than 50 employees. Affordability is the main reason cited by 51% of employers offering the minimum, while 31% believe this is the amount recommended by the government, 26% prioritise putting money into base salary, and 24% say this is the norm in their sector. Employees may also lack interest in occupational pension plans, thereby discouraging employers to offer such plans or to contribute beyond the minimum (OECD, 2022[49]).
Moreover, trade unions may not push for higher contributions to occupational pension plans negotiated through collective bargaining. Trade unions may fear that occupational pensions may undermine the role of PAYG pensions and affect wages. For example, in Austria and Belgium, unions show little interest in raising contribution rates to occupational and sectoral pension plans respectively, as they tend to prioritise the public pension scheme as the main vehicle to provide retirement income, favour the maintenance of wage levels in the context of low wage growth, and consider that investment risk makes occupational pension benefits insecure (Keune, 2018[50]).
Career interruptions may also reduce individuals’ ability to pay steady contributions. Temporary employees may not always be able to contribute to an occupational pension plan when the plan imposes a minimum length of employment or contract duration for participation (OECD, 2019[17]). Additionally, in some countries, employer contributions to voluntary pension plans stop during maternity leave (e.g. Austria, New Zealand, Poland, the United States) (OECD, 2021[16]).8 Unemployment spells also prevent continuous contributions.
Even when employers contribute, employees may lose some of these contributions by leaving the employer before the end of the vesting period. In several countries, employer contributions to occupational pension plans may not vest immediately (OECD, 2019[17]). Employers mainly use vesting periods as a retention tool and to reduce costs. However, vesting periods affect the retirement savings of employees. In the United States, the proportion of participants leaving their employer before the end of their vesting period increased from 19% in 2010 to 29% in 2018. The average forfeiture represents 37% of the account balance of departing individuals, mostly affecting younger and lower-income employees (Carranza and Goodman, 2024[51]).
Contribution caps may constrain voluntary contributions. Many countries limit the level of contributions attracting tax relief (OECD, 2025[52]). Some countries express the cap as a percentage of income up to a nominal ceiling, while others express the cap as a fixed amount (OECD, 2024[53]). For example, fixed caps represent less than 10% of the average wage in seven OECD countries (OECD, 2024[53]). These caps are useful to limit the fiscal cost of tax incentives for the government and avoid rewarding higher-income earners disproportionately as they are more likely to save for retirement and have a higher capacity to save. However, they may be constraining for these individuals when they have significant contribution gaps. Some countries also calculate contributions on capped earnings. In the United Kingdom, the minimum total contribution rate under automatic enrolment must be at least 8% of a band of earnings, but the effective contribution rate as a percentage of total salary follows a hump shape with salary level and reaches a peak of 7% of total salary for employees earning GBP 50 000 (Pensions Policy Institute, 2024[54]). This means that higher-income earners may need to make the active decision to contribute beyond the default minimum to achieve their retirement income objective.
Minimum and default contribution rates for voluntary contributions tend to be relatively low. For example, default contribution rates used in the context of automatic enrolment are below 5% in Quebec (Canada) and Türkiye. Low default contribution rates may be useful to encourage people to save and reduce opt-out rates. However, they may be an issue as people tend to stick with default contribution rates even when these are not appropriate for them (Madrian and Shea, 2001[55]; Choi et al., 2004[56]). People may also opt for the minimum due to affordability concerns or lack of understanding of the impact of contributions on future retirement income (OECD, 2018[39]).
Poor performance records may also put individuals off saving for retirement in asset-backed pension plans. If conservative investment strategies lead to low or negative returns, due for example to guarantees, high fees and high inflation, individuals may question the value of saving in asset-backed pension plans and look for alternative saving vehicles. For example, in Germany, the number of Riester contracts declined from 16.6 million in 2017 to less than 15.0 million in 2024, and 20% to 25% of contracts no longer receive contributions (Federal Ministry of Labour and Social Affairs, 2025[57]). This is due to several shortcomings of Riester plans, including low returns, high costs and insufficient transparency. In particular, the statutory capital guarantee limits investment in asset classes with higher risks and higher return potential, often leading to negative real returns net of costs (Deutsche Bank Research, 2026[58]).9
2.2.3. Other forms of savings
Other forms of savings can play a role in financing retirement, reducing the need for voluntary contributions. Indeed, failure to contribute enough to asset-backed pension plans may not be an issue if individuals can tap into other sources to finance retirement. However, other forms of savings are not perfect substitutes for pension savings.
Housing is the main asset in households’ portfolios. The main residence represents 50% of total assets on average in the OECD when looking at the asset decomposition of households’ portfolios, excluding public and occupational pension plans (Figure 2.9). This share ranges from 25% in Germany to 72% in the Slovak Republic. Even among lower-income earners, the share of homeowners is relatively high in many countries, reaching 54% on average in the OECD (OECD, 2024[59]). Other real assets (e.g. secondary residence) and deposits are also relatively important in household’s portfolios, representing 17% on average.
Note: Public and occupational pension plans are not included in financial assets.
Source: OECD (2021[60]), “Measuring the Role of Housing in the Distribution of Wealth” in Brick by Brick: Building Better Housing Policies, https://doi.org/10.1787/b453b043-en.
Given that homeownership is widespread in many countries, home equity release products could be a valuable additional source to finance individuals’ needs in retirement for many retirees. However, home equity release products are not suitable for every situation (OECD, 2024[61]). They are better suited for older individuals who plan to stay in their homes for the rest of their lives and for those who do not prioritise leaving a bequest to their heirs. Moreover, the use of these products remains very limited due to lack of awareness, lack of trust in financial providers, debt aversion and emotional attachment to homes. There is also limited supply of home equity release products because financing and capital requirements are a major barrier to entry, and providers may not always be able to effectively mitigate their own risk exposures.
Even after factoring in home equity, some individuals may still have a remaining contribution gap. In the United Kingdom, 34% of the working-age population may fail to meet their target replacement rate (as defined by the 2004 Pensions Commission) when summing up their expected income from public pensions, private pensions, financial assets and home equity (Pensions Policy Institute, 2025[62]). In the Netherlands, employees without an occupational pension do not compensate by accumulating more other assets (including home equity, business equity and financial assets). In 2020, the median net total assets equalled EUR 132 802 for employees with an occupational pension and EUR 90 445 for employees without an occupational pension. However, the self-employed without an occupational pension had accumulated more total assets, with a median value of EUR 196 622, with home equity being an important factor (Biesenbeek, Heerma van Voss and Mastrogiacomo, 2020[63]).
There is a range of policies that could ensure that individuals contribute to asset-backed pension plans in line with their retirement income objective.10 The different policies can be grouped under six categories (Figure 2.10).
2.3.1. Increasing mandatory contributions
Increasing mandatory contributions may raise the chances that individuals will reach their retirement income objective. However, this may have negative implications. Increasing mandatory contributions from employers may affect labour markets, while increasing mandatory contributions from employees may affect household finances and consumption. The impact on total savings may be uncertain as individuals could lower their voluntary savings to offset the increase in mandatory contributions.
Increasing mandatory pension contributions or introducing new mandatory contributions may be the most efficient way to raise pension benefits and reach individuals’ retirement income objective. This may be particularly the case when barriers to voluntary savings are difficult to overcome. For example, Norway introduced mandatory occupational pension plans for private‑sector employees with a minimum employer contribution rate of 2%, as trade unions and employer representatives were unable to reach an agreement on collectively bargained occupational pension schemes (Trampusch, 2013[64]; Pedersen et al., 2018[65]). In general, increases in mandatory contributions tend to be gradual to allow employers and employees to adjust to the change. This is currently the case in Chile and Mexico, for example.
Increasing mandatory pension contributions may have negative implications on labour markets, in particular employment, wages and formality. Mandatory pension contributions paid by employers increase labour costs, potentially reducing demand for labour. For example, in Finland, higher payroll taxes (pension and health contributions) reduced employment, with the most affected workers being those with lower levels of education and those performing routine and manual tasks (Benzarti and Harju, 2021[66]). In addition, employers may shift the cost of mandatory pension contributions to employees’ remuneration. Mandatory pension contributions may have a stronger impact on wages in more flexible and competitive labour markets (Kim, Kim and Koh, 2022[67]). The extent to which an increase in employer contributions is shifted to workers through lower wage growth indeed varies across countries, from 30% in the Netherlands (Bosch et al., 2022[68]), to 50% in the United Kingdom (Scarfe, Schaefer and Sulka, 2024[69]), 70% to 100% in Australia (Breunig and Sobeck, 2020[70]; Coates, Cowgill and Mackey, 2020[71]) and 100% in France (Bozio et al., 2025[72]). Moreover, mandatory pension contributions may decrease labour formality as they impose a liquidity constraint on low-income workers by reducing take‑home pay. Based on modelling and using data on Chile, Colombia and Peru, several studies show that mandatory pension contributions affect workers’ decision to take formal or informal jobs (Pardo, 2023[73]; Moreno, 2025[74]; Joubert, 2015[75]).
Increasing mandatory pension contributions may also have negative implications on household finances. Higher mandatory pension contributions may reduce household consumption, potentially putting serious constraints on lower-income earners. In Israel, the introduction in 2008 of mandatory pension contributions to pension funds led to a reduction in consumption, mostly due to a reduction in net wages (Frish, 2025[76]). Similarly, in Denmark, individuals subject to higher increases in pension contributions have lower consumption growth during working years (Larsen et al., 2025[77]). An increase in mandatory pension contributions may also lead to higher levels of household debt. Workers may take on higher debt, either because they feel more confident about their capacity to reimburse a loan due to the expected higher income in retirement (Andersen, Hansen and Kuchler, 2022[78]), or because they need to borrow to partially finance the increase in contributions (Beshears et al., 2025[79]; Choukhmane and Palmer, 2025[80]). However, higher mandatory pension contributions do not increase household debt in all countries (Sigurdardóttir, 2023[81]; Beshears et al., 2021[82]).
The impact of an increase in mandatory contributions on total savings depends on the extent to which individuals reduce their voluntary savings, and this varies across countries. In Iceland, the increase in the mandatory contribution rate to occupational pension funds between 2016 and 2018 led to an overall increase in household savings as private‑sector workers did not reduce significantly their voluntary savings, potentially due to a lack of awareness of the reform (Jensen et al., 2025[83]). In the United States, employees of a public university employed after an increase of 5 p.p. in mandatory pension contributions only had voluntary DC contribution rates 1.5 p.p. lower on average than employees hired before the change, showing that most people were not sensitive to the level of mandatory contributions (Friedberg, Leive and Cai, 2024[84]). In Australia, increases in employer contributions to mandatory superannuation funds reduce voluntary private savings, but the substitution effect is less than one, leading to an increase in household wealth (Ruthbah and Pham, 2020[85]). In Switzerland, workers subject to mandatory participation in occupational pension funds tend to reduce private non-pension wealth, except lower-income earners with liquidity constraints as they are less able to reduce their private savings to offset mandatory pension contributions (Burgherr, 2025[86]). In Denmark, increases in mandatory pension contributions are associated with lower savings outside the pension system (Larsen et al., 2025[77]).
2.3.2. Encouraging employers to contribute voluntarily on behalf of their employees
Employers hold a unique position in supporting their employees’ retirement savings. By linking occupational pension plan participation with employment contracts, enabling payroll deductions, and making contributions on behalf of their workers, employers can make it easier for staff to prepare for retirement (OECD, 2022[49]). However, cost may be a barrier for voluntary employer contributions.
In many OECD countries, employer contributions enjoy a beneficial treatment regarding social contributions. Social contributions are usually levied on gross income to finance social benefits such as healthcare, unemployment and pensions. Exempting employer pension contributions from social contributions or subjecting them to reduced rates makes it more cost-effective for employers to contribute to pension plans than to pay equivalent amounts as salary. Examples include Australia, Austria, Belgium, Ireland, Italy, Poland, Slovenia and the United Kingdom, where employer pension contributions are not counted towards calculating social contributions or receive preferential rates (OECD, 2024[53]).
Another way to reduce costs and encourage voluntary employer contributions is to provide tax incentives to employers. Employer contributions to asset-backed pension plans are always considered as tax-deductible business expenses for the purpose of corporate income tax. Some countries provide additional tax relief. In Germany, employers contributing at least EUR 240 per year to an occupational pension plan on behalf of an employee earning less than EUR 2 575 monthly get a tax credit of 30% of the contribution, up to a maximum contribution of EUR 960. In Spain, employers can deduct from corporate income tax 10% of their contributions to an occupational pension plan in favour of employees with an annual gross remuneration of less than EUR 27 000 (OECD, 2024[53]).11
Regulation can also ensure that employers contribute enough voluntarily on behalf of all categories of employees, in particular lower-income earners. In the United States, non-discrimination rules in traditional 401(k) plans ensure that contributions made by and on behalf of lower-earning employees are proportional to the contributions for higher-earning employees. If that is not the case, the employer needs to take corrective actions, such as making a taxable distribution of the excess contributions to higher-earning employees or contributing on behalf of all lower-earning employees. One way to satisfy the non-discrimination test is to implement automatic enrolment and automatic escalation of contributions to ensure high participation rates and sufficient contributions among lower-earning employees. Another way is to adopt a safe‑harbour plan offering all lower-earning employees an employer contribution equal to at least 3% of pay or offering an employer matching contribution using an approved formula (OECD, 2022[49]).
2.3.3. Nudging individuals
A nudge is an intervention adapting the decision environment of individuals to influence their behaviour and decision making. Thaler and Sunstein (2008[87]) define a nudge as “any aspect of the choice architecture that alters people's behaviour in a predictable way without forbidding any options or significantly changing their economic incentives.” In the context of pension contributions, nudges include default contribution rates, automatic increases in contributions, action-oriented messages, simplified processes and increases in retirement ages.
Default contribution rates
Choosing a default contribution rate is far from straightforward. A default contribution rate is a pre‑selected contribution rate that applies to all plan members who do not choose a rate. It can increase participation by simplifying decision making (OECD, 2018[39]). However, individuals tend to have different optimal contribution rates depending on their personal characteristics, such as income, age, gender and tenure (Choi et al., 2003[88]). When heterogeneity is high, any single default could be inappropriate for many workers. Moreover, it is important to consider the level of the default contribution rate carefully, as individuals may keep it for several years (Beshears et al., 2023[89]), either because they perceive it to be the implicit recommendation of the plan sponsor or the government (McKenzie, Liersch and Finkelstein, 2006[90]), or because of inertia. Lower-income earners may be less likely to depart from the default, in part because they face higher barriers to active decision making due to lower levels of financial literacy (Burke, Hung and Luoto, 2017[91]; Beshears et al., 2023[89]).
Low default contribution rates may reduce effective contribution rates. Default contribution rates tend to be low, typically below 5%, to avoid having too many people oversaving and to limit opt-out rates in the context of automatic enrolment. However, evidence suggests that some individuals would select a higher contribution rate in the absence of a default option (Nest Insight, 2025[92]; Carroll et al., 2009[93]). The average positive effect of automatic enrolment on contributions in 401(k) plans in the United States nearly disappears after three years of tenure as workers enrolled automatically stay at the default contribution rate, while those who opt in voluntarily increase their contribution rate over time (Choukhmane, 2025[94]). In the United Kingdom, the median total contribution rate to DC pension plans among participating private‑sector employees is lower now (8% of total pay) than before the rollout of automatic enrolment in 2012 (9% of total pay), because automatic enrolment brings in new savers contributing at the minimum, default levels (Department for Work and Pensions, 2025[35]).
Setting higher default contribution rates could increase contribution rates without significantly reducing participation rates, although individuals may be more likely to depart from the default. When individuals have heterogeneous preferences, it may be preferable to select a default contribution rate that is slightly higher than the optimal rather than lower, since it would tend to increase retirement savings, while it may be easier for individuals to reduce their contribution rate from a high default than to increase it from a low default due to loss aversion (Choi et al., 2003[88]). In the United States, setting default contribution rates between 7% and 11% leads to an increase in average contribution rates of 20 to 50 basis points of salary relative to a 6% default, with only the 11% default leading to an increase of 3.7 p.p. in the likelihood of not participating in the plan (Beshears et al., 2017[95]). In Italy, the number of self-employed workers contributing more than the minimum to their mandatory pension fund was multiplied by 2.7 between 2016 and 2017 as a result of increasing the default contribution rate from 10% (the minimum) to 20% and explaining the consequences of pension choices in terms of future pension income and tax savings (Rubaltelli and Lotto, 2021[96]). Moreover, higher default contribution rates may be less sticky than lower ones, as demonstrated by the United Kingdom, where only 25% of employees participating in a DC plan with a 12% default contribution rate were still contributing at this rate after 12 months of tenure (Beshears et al., 2023[89]).
Automatic increases in contributions
Increasing contributions automatically up to a pre‑set level, as an alternative to higher defaults, could help individuals to contribute more. In New Zealand, providing KiwiSaver members with the possibility to join a scheme with automatic increases of contributions could result in a difference between the initial and maximum contribution rate of 2.8 p.p. on average, without reducing the initial contribution rate (Hobbs and Feld, 2021[97]).12
Contribution increases could be linked to salary increases. This would reduce the feeling of loss of a lower take‑home pay and mitigate the affordability issue of increased contributions for low-income earners. In the United States, asking employees to commit to future increases in their contribution rate each time they get a pay rise up to a pre‑set maximum led participants in the Save More Tomorrow™ (“SMarT”) programme to almost quadruple their contribution rates over the course of 40 months, from 3.5% to 13.6% on average (Thaler and Benartzi, 2004[98]). In Spain, employees of a large financial institution who had just experienced an increase of minimum 1% in their salaries automatically participated in a programme whereby 50% of the salary increase was paid as a default contribution to the company pension plan and the default contribution was automatically increasing by 10% every year thereafter. Annual contributions increased by 19% and 31% after one year for employees enrolled in the programme in 2016 and 2017, respectively (García and Vila, 2020[99]).
The acceptance of automatic increases in contributions varies and may decline over time. In the United States, among individuals who were newly entitled to participate in a Vanguard-administered plan with automatic enrolment and automatic annual increases in contributions in 2015, 54% were still in the contribution increase programme and 27% had reached their maximum pre‑set rate by 2020 (Proctor and Young, 2021[100]). In Spain, only 22% of employees of a large financial institution enrolled automatically in an automatic increase in voluntary pension contributions withdrew from the programme during the following year (García and Vila, 2020[99]). However, acceptance of the default automatic contribution increase may not remain that high in subsequent years. Among 21 US firms that implemented automatic increases, 43% of employees were still in the programme at the time of the first increase, going down to 29% by the third increase, with other plan participants not accepting the automatic increases (Choi et al., 2024[101]).
Moreover, automatic increases in contributions may not work for workers changing jobs frequently. Workers changing employers may not keep up with their initial contribution increase schedule. Indeed, the new employer may not offer a pension plan, there may not be a programme of automatic increases in contributions in the new plan, or the worker may start a new contribution increase schedule from a low initial contribution rate in the new plan (OECD, 2022[49]). Mobile workers may never reach the maximum contribution rate and therefore under-save for retirement during most of their career.
Action-oriented messages
Sending simple messages prompting pension plan members to increase their contributions can encourage higher contributions. In the United States, state government employees aged 50 to 69 participating in a DC pension plan who received a flyer by email with general information about DC pension plans and catch-up contributions increased their contributions by 1 p.p. more than those who did not receive the flyer one month after the intervention, and by 2.2 p.p. more after nine months (Clark et al., 2018[102]).13 Similarly, about 15% of employees participating in the pension plan of a US company but contributing less than 6% of earnings raised their contribution rate after receiving a form by mail allowing them to simply tick a box to increase their total contribution rate to 6% (Beshears et al., 2013[103]; Choi et al., 2017[104]). In comparison, only about 1% of low contributors raised their contribution rate to 6% when not receiving a similar form. Moreover, highlighting a higher potential increase in contribution rates rather than a smaller one in an email can nudge individuals towards higher contribution rates (Choi et al., 2017[104]).
When offering the option to increase contributions in the future, messages should associate the future date with a fresh start. Committing to increase contributions at a future date may help to overcome present bias. However, offering individuals an opportunity to delay savings for several months runs the risk of reducing contributions as people may interpret the intervention as evidence that saving for retirement is not an urgent priority (Beshears et al., 2016[105]; Reiff et al., 2023[106]). Expressing the future date as a fresh start, such as the individual’s next birthday or the first day of spring, can counteract this negative effect (Beshears et al., 2021[107]; Beshears et al., 2016[105]). This may be because fresh start moments are more salient than ordinary dates and can make people more optimistic about their ability to achieve their goals (Beshears et al., 2021[107]).
The effectiveness of messages including information about peer contributions may depend on how achievable individuals perceive the objective to be. Sharing peer information is expected to influence behaviour towards the peer-group average because it shows that a certain behaviour is relatively common and may encourage some individuals to do the same thing. In Norway, individuals receiving a message on the online portal of their pension provider describing how much other individuals in the same age group save per month increased their contributions to their personal pension plan in the following month (Døskeland et al., 2025[108]). Similarly, US Army service members who received an email encouraging them to participate in the Thrift Savings Plan and highlighting a contribution rate at which some coworkers start contributing, were more likely to contribute exactly at the highlighted rate (Goldin et al., 2020[109]). However, peer information may not significantly affect pension contributions when the goal is perceived to be challenging. Employees in a large US manufacturing firm who had low contribution rates and received information about the proportion of coworkers in the same age group contributing at least 6% of pay to the company pension plan were not more likely to raise their contribution rate to 6% than those who did not receive information about peers (Beshears et al., 2015[110]). This may be because low contributors may perceive the 6% contribution rate objective as hard to achieve.
Asking individuals to formulate a plan to save for retirement and sending reminders in line with this commitment can increase retirement savings. Reminders can increase retirement savings by helping individuals to keep their savings goals at the top of their mind. They may be particularly useful when individuals can flexibly decide how often and how much to save in their pension plans. In Colombia, sending monthly reminders personalised based on a savings plan formulated by participants in the Beneficios Económicos Periódicos (BEPS) programme increases the number of deposits without reducing the average amount deposited, compared to participants who did not formulate a savings plan and to participants who formulated a savings plan but did not receive personalised reminders (Gars et al., 2025[111]). Additionally, reminders with a savings goal perform better when framed as a monthly goal rather than as an annual goal because small amounts reduce the feeling of loss and increase the feeling of affordability (Azuara Herrera et al., 2021[112]).
Simplified contribution processes
Extending the mechanisms through which individuals can deposit contributions can remove some of the barriers to voluntary saving. In Mexico, there are eight different ways to make voluntary contributions, including through convenience stores (e.g. 7‑Eleven shops, pharmacies), mobile applications, providers’ websites and employers, to facilitate contributions across formal and informal workers in urban and rural areas. Digital platforms capture most of the flow of voluntary contributions (OECD, 2020[113]). Making it possible to deposit voluntary contributions at 7‑Eleven stores increased the number of workers making at least one contribution by 9% and the number of voluntary contributions by 6% after 43 months in municipalities with 7‑Eleven stores, compared to those without (Bosch and Rubli, 2021[114]). In Switzerland, the introduction of a digital pension application increased the probability of making additional voluntary contributions to mandatory occupational pension funds by 1.8 p.p., from an initial 2.8% prior to the introduction of the app. The simplification of the process of making voluntary contributions through the app was the main factor (Daminato, Filippini and Haufler, 2024[115]).
Automatic savings mechanisms can also simplify contribution processes by giving individuals peace of mind once the mechanism is set. Such mechanisms can harness the power of inertia for self-employed and informal workers the same way payroll deductions do for employees. In India, informal workers participating in the PM-SYM scheme contribute through auto-debit from a savings bank account. The contribution level is fixed during the entire accumulation phase, matched by the government and set at a higher amount as the individual joins the scheme at an older age. Pension savings may also be automatically associated with consumption. In Mexico, the application Miles for Retirement allows individuals to automatically save a percentage of their consumption when using their credit or debit card and to receive promotions from stores and brands in the form of contributions to their retirement account. This implies that even individuals who cannot afford to contribute on their own can receive contributions (Vižintin and Peksevim, 2021[116]).
Increases in retirement ages
Increasing the statutory age of retirement could push people to postpone retirement and increase the amount of pension contributions. Postponing retirement could have two opposite effects on voluntary savings. First, to the extent that workers work longer, they may continue to save for retirement, thereby accumulating more pension assets and entitlements than otherwise, i.e. the income effect. Second, some workers may anticipate the longer working life and reduce their voluntary contributions to keep the total amount of contributions constant, i.e. the substitution effect. In Denmark, people born after 1953 who experienced a six‑month increase in pension eligibility ages in 2011 had higher contributions to occupational and personal pension plans than those not affected by the reform (García-Miralles and Leganza, 2024[117]). This is mostly because people postponed retirement by extending employment and continued to contribute as before. In Japan, a reform delaying men’s eligibility for public pension benefits also raised savings, although this was not associated with an increased enrolment in private pension plans (Nakazawa, 2022[118]). By contrast, in Germany, women reduced their savings rate following the increase in the early retirement age in 1999 as they anticipated that their longer working life would raise their overall lifetime income (Etgeton, Fischer and Ye, 2023[119]).
2.3.4. Providing financial incentives to individuals
Financial incentives are monetary rewards offered to encourage a certain behaviour that may not take place otherwise. There are two types of financial incentives encouraging pension contributions. Tax incentives are indirect financial incentives provided by the government through the tax system. They come from a differential tax treatment applied to retirement savings as compared to other forms of savings (OECD, 2025[52]). Non-tax financial incentives are contributions paid directly into the pension account of eligible contributing individuals by the government and/or by employers. They can take the form of matching contributions or fixed nominal subsidies. Financial incentives provided by the government are more affordable when they create new savings, but they may still be a valuable mechanism to reallocate savings towards the long term and ensure that individuals have enough resources to finance retirement (OECD, 2018[120]).
Tax incentives
The possibility to deduct pension contributions from taxable income encourages individuals to contribute. Most OECD countries provide tax incentives for retirement savings. The most common tax treatment for retirement savings is the one where contributions are deductible from taxable income, investment income is not taxed and pension benefits are taxed (OECD, 2024[53]). Studies in a range of countries show that the tax-deductibility of contributions is effective at increasing contributions to asset-backed pension plans (OECD, 2018[120]). Higher-income earners are those who respond the most to such tax incentives, which benefit more those with higher marginal tax rates. For example, a recent study in France shows that voluntary employee contributions to occupational pension plans increased in firms introducing a tax-deductible option (Brière, Poterba and Szafarz, 2025[121]). This increase was not compensated by lower contributions to other occupational savings plans. Moreover, older workers and those with larger retirement saving balances and higher income were more likely to take advantage of the new tax-deductible option.
Tax incentives may also allow individuals to contribute more some years to compensate for periods with lower contributions. Contributions attracting tax relief are usually subject to annual caps. Some countries, such as Australia, Canada, Italy, Switzerland and the United Kingdom, allow some individuals to carry forward caps that were unused in the previous years (OECD, 2024[53]). The possibility to make “catch-up” or “retroactive” contributions aims to give flexibility to those who take time out of work, work part-time, or have irregular income and therefore have periods in which they make no or low contributions to their pension plan. An alternative way to allow for catch-up contributions is to set higher caps for older workers (e.g. Denmark, Ireland, Korea, the United States), acknowledging that they have a higher capacity to save (OECD, 2024[53]).
However, several factors may prevent individuals from taking full advantage of tax incentives:
Inertia: Most individuals are passive savers who do not react when there is a change in the tax treatment of pension plans. In Denmark, only 19% of individuals affected by a reform that reduced the tax-deductibility of contributions to a particular type of pension plan changed their saving behaviour by reducing or stopping contributions to these pension plans (Chetty et al., 2014[122]).
Lack of understanding of different tax treatments: Misunderstanding of tax rules may lead individuals to fail to contribute differently to pension plans with different tax treatments because they do not realise the impact on after-tax pension benefits (Beshears et al., 2017[123]). This is the case for example when offering choice between plans with upfront taxation and plans with deferred taxation,14 as individuals fail to understand that they need higher after-tax contributions with deferred taxation than with upfront taxation to achieve the same after-tax benefit in retirement, provided they are subject to constant tax rates over time (Blaufus and Milde, 2021[124]).
Lack of value for low-income earners: Non-refundable tax credits and tax deductions are not very valuable for individuals with low or no income‑tax liability (OECD, 2018[120]). In many countries, people earning below a certain threshold do not file income tax and thus cannot benefit from tax incentives. In the United States, less than 6% of taxpayers claim the Saver’s Credit, a non-refundable tax credit provided on voluntary contributions to occupational and personal pension plans.15 However, there are important differences across income levels, with claiming rates increasing from 0.03% among taxpayers with less than USD 10 000 in adjusted gross income to 63% among those with an adjusted gross income between USD 25 000 and USD 50 000 (Congressional Research Service, 2025[125]).
Lack of awareness of the incentives offered: One of the factors explaining the low take‑up of the Saver’s Credit in the United States is the lack of awareness of the programme, which remained below 50% in 2025 among workers in a surveyed company (Transamerica Institute, 2026[126]).
Non-tax financial incentives
Matching contributions are the most common type of non-tax financial incentives in the OECD and can take different forms (OECD, 2024[53]). An employer or government matching contribution may be expressed as a percentage of the individual’s own contribution or as a percentage of the individual’s salary. When expressed as a percentage of the individual’s own contribution, several tiers may be defined, such as 100% of the employee contribution up to 3% of salary and 50% of the employee contribution on the next 2% of salary. There may be a cap on the employee contribution rate subject to the match or on the nominal amount of the matching contribution.
The presence of a matching contribution can encourage individual contributions since individuals would otherwise leave free money on the table. In the United States, federal civilian workers eligible for an employer matching contribution contribute 3.5 p.p. more on average than those non-eligible (Falk and Karamcheva, 2022[127]). In Australia, eligibility for the government matching contribution, known as the co-contribution programme, increases the proportion of individuals making after-tax voluntary contributions to their superannuation fund (Chan et al., 2020[128]). Moreover, decreasing the income threshold above which individuals are no longer eligible for the government match reduces the percentage of individuals making a voluntary contribution by 0.9 p.p. and the amount contributed by 6.2% (Australian National University, 2021[129]).16 Among individuals who contributed voluntarily before the reform and lost eligibility for the government match following the decline in the income threshold, the decline was even larger, between 16% and 22% (Sobeck and Breunig, 2026[130]). However, individuals may reallocate their savings between different forms of voluntary schemes depending on eligibility for the government matching contribution (Chan et al., 2020[128]; Australian National University, 2021[129]).
A higher match rate may encourage higher contributions, although the impact may be hard to measure when match formulas vary on different parameters. In Australia, the increase in the proportion of individuals making after-tax voluntary contributions is larger when eligibility for the government matching contribution is associated with higher match rates, by 0.6 p.p., 0.9 p.p. and 2.7 p.p. for a match rate of 50%, 100% and 150%, respectively (Chan et al., 2020[128]).17 In Peru, offering a 100% match rate to informal workers raises the probability of contributing at least once by 2.3 p.p. in normal economic conditions, while a 50% match rate has no impact on contributions (Bernal, Galiani and Molina, 2025[131]). Evidence in the United States is mixed, with studies finding contradictory results on the link between employer match rates and employee contributions (Madrian, 2012[132]; Choi, 2015[133]). This could be because employer matching contributions use different combinations of match rates (i.e. the percentage applied on the employee contribution) and match thresholds (i.e. the maximum employee contribution matched by the employer). Moreover, employee contribution rates do not vary much according to the match formula on a sample of Vanguard-administered DC plans (Greig et al., 2024[134]).
The match threshold serves as an anchor that may lead some individuals to increase their contributions and others to reduce them. In the United States, the proportion of employees contributing at higher rates increased in a company that augmented the match threshold from 5% to 7% of salary for union workers, and from 6% to 8% of salary for management employees, without changing the match rate (50%), suggesting that the match threshold serves as an anchor for some employees to select their contribution rate (Choi et al., 2002[135]). However, this anchoring effect may also reduce the contribution rate of some employees who might otherwise contribute more (Huberman, Iyengar and Jiang, 2007[136]). Vanguard data show that the proportion of employees contributing more than the match threshold increases with income (Greig et al., 2024[134]).18 In Australia, the co-contribution programme increases the proportion of individuals contributing at or below AUD 1 000, which is the maximum individual contribution matched by the government, but decreases the proportion of individuals contributing at higher levels (Chan et al., 2020[128]). Indeed, many individuals contribute to maximise the government payment, while those who would have made voluntary contributions without the programme no longer need to contribute as much to reach the same overall contribution level.
Depending on the formula, matching contributions may advantage higher-income earners. When matching contributions are expressed as a percentage of the individual’s income, all workers receive the same amount in proportion of their income provided they contribute at a minimum rate. However, matching contributions may advantage higher-income earners when they are expressed as a percentage of the employee’s contribution, in particular single‑tier match formulas, since higher-income earners are more likely to have the financial capacity to contribute up to the match threshold.19 A multi-tier match formula can partially address this regressive feature since a higher match rate is applied on the first tranche of employee contributions.20 Matching contributions can also be associated with a nominal cap to limit the absolute contribution made towards higher-income earners (OECD, 2024[53]). In the United States, match formulas with a nominal cap on the employer match are progressive while all the others are regressive (Greig et al., 2024[134]).
Irrespective of the formula used, higher-income earners may be more likely to respond to a matching contribution incentive than lower-income earners. In Australia, higher-income earners are more likely to increase their voluntary contributions when eligible for the government matching contribution (Chan et al., 2020[128]). This could be because lower-income earners have budgetary constraints, making it harder for them to save more for retirement. Similarly, in New Zealand, KiwiSaver members not receiving the government matching contribution or only receiving a partial payment tend to have lower incomes than those receiving the full contribution (Te Ara Ahunga Ora Retirement Commission, 2025[137]). Indeed, the proportion of contributing members falls from around 90% for members with an annual income above NZD 50 000 to 50% or less for those with an annual income below NZD 30 000.
Matching contributions can target lower-income earners specifically. In Australia, the government matching contribution is for individuals with a total income below AUD 58 445 making voluntary non-deducted contributions, and the maximum entitlement is reduced by 3.333 cents for every dollar that the individual earns above AUD 43 445. In Chile, workers aged between 18 and 35 with an income lower than 1.5 times the minimum wage are entitled to a government matching contribution for their first 24 monthly continuous or discontinuous contributions to the pension system. In Colombia, the BEPS programme focusses on low-income earners and matches 20% of their voluntary contributions by the time of retirement.
Fixed nominal subsidies provide a greater incentive for lower-income earners without the need to be targeted. Governments in several OECD countries provide fixed nominal subsidies to encourage individuals to make voluntary contributions (e.g. Germany, Lithuania, Poland) (OECD, 2024[53]). The fixed amount has a higher economic value for lower-income earners as it represents a higher share of their income. Fixed nominal subsidies may target women specifically to compensate for periods out of work to give birth and raise children (e.g. Chile, Estonia, Lithuania) (OECD, 2024[53]). However, lower-income earners may reallocate their savings when subsidies are available from a low contribution amount, such as in Germany (Corneo, Keese and Schröder, 2008[138]; Kolerus, Koske and Hüfner, 2012[139]; Pfarr and Schneider, 2012[140]).
2.3.5. Informing individuals
Providing information on pension plan design, expected benefits, how much income individuals may need in retirement and how to make voluntary contributions, may improve individuals’ understanding of the process to contribute and how pension contributions translate into retirement income. It could make them realise they need to contribute more and prompt them to increase their contributions or to ask trade unions to negotiate pension contributions as part of remuneration packages. Pension providers, plan sponsors, advisers and pension authorities may deliver information through brochures, pension benefit statements, pension dashboards, financial advice or financial education seminars.
Communication about pension contributions should be positive, plausible, plain spoken and personal to engage plan members. Messages framed in a positive way are more likely to engage individuals than those with a negative or loss frame (Nest Insight, 2020[141]). Credible expectations are also important as telling individuals they could reach a “comfortable” retirement is more likely to motivate them to revisit their contribution rate than telling them they could reach a “dream” retirement. Technical jargon should also be avoided with terms like “investment” and “pension pot” better understood than “asset classes” and “capital”, respectively. Moreover, communicating on future retirement income instead of future pension account size is more likely to prompt individuals to increase their contributions. Finally, messages should be personal, taking the perspective of the individual (“you” or “your”) rather than the one of employees or members.
Pension dashboards, as well as simplified brochures and pension benefit statements, can facilitate information processing but do not necessarily increase contributions.21 In the Netherlands, visiting a pension dashboard increases pension knowledge among pension plan participants but does not impact the willingness to save more even if there is a need to do so (Kramer, Lensink and Plantinga, 2024[142]). In the United States, receiving a shorter description of an occupational pension plan makes information more accessible and easier to process but does not affect the anticipated contribution rate (Cardella, Kalenkoski and Parent, 2021[143]). In Mexico, simplifying pension benefit statements by improving the readability and visual appeal of the information may have enhanced the efficiency of information processing for certain individuals (Shah et al., 2024[144]). In turn, this may have helped them to assess the relative performance of their pension provider, thereby encouraging or discouraging voluntary contributions (Shah et al., 2024[144]).
Including explanatory diagrams on pension benefit statements can favour higher contributions. In Ireland, individuals who saw, on a pension benefit statement, diagrams showing how projected fund sizes and income in retirement could be increased by contributing more now, were more likely to recommend raising the contribution rate compared to those who saw the same information presented in a table (McGowan and Lunn, 2019[145]). Diagrams improved individuals’ rationale for their decision. However, they did not improve recall and comprehension of information.
Providing personalised projections of account balances and retirement income can increase voluntary contributions. In Australia, individuals who received projected retirement income information increased their voluntary pension contributions during the two following years, compared to individuals who only received information about their current account balance (Smyrnis et al., 2019[146]). Moreover, showing projections of account balances in addition to retirement income projections had a larger impact on voluntary contributions. In Chile, individuals who received personalised projections of their future pension benefits, as well as projections of the difference in benefits resulting from three possible actions to increase benefits (increasing mandatory contributions, increasing voluntary contributions or delaying retirement), increased their voluntary savings by 10% on average in the first eight months after treatment, compared to those who only received general information (Fuentes et al., 2024[147]). The effect was concentrated among individuals who had previously overestimated their expected pension. In Germany, individuals who received a letter showing expected public pension payments increased their contributions to personal pension plans (Riester pensions) during the following three years without crowding out other forms of savings (Dolls et al., 2018[148]).
Providing context that enhances individuals’ understanding of the implications of different contribution rates can lead to higher contributions. Relative to simply listing different contribution rates and their associated expected retirement income projections, additionally associating the expected retirement income projections with retirement income standards (e.g. minimum, moderate, comfortable) and consumption baskets (i.e. a description of consumption possibilities for housing, transport, holidays and leisure) can lead to higher contribution rates (Barrett et al., n.d.[149]). Indeed, consumption baskets make future consumption possibilities more tangible, which in turn increases selected contribution rates as individuals are eager to achieve the related retirement consumption outcomes.
Financial advice may be more impactful for lower-income earners. Chapter 4 shows that financial advice can significantly contribute to improving financial outcomes in retirement by helping individuals to better plan for retirement, select more appropriate levels of investment risk and mitigate emotional reactions. However, individuals seeking financial advice tend to have higher income, be older, more educated, and more financially literate (Burke and Hung, 2015[150]). As these individuals are more likely to save for retirement in general, the positive impact of financial advice on retirement savings is more likely to be greater for lower-income earners. In addition, the impact varies across countries. In New Zealand, financial advice does not significantly influence contribution rates to KiwiSaver plans (Fang, Hao and Reyers, 2022[151]). By contrast, in Switzerland, individuals who discuss retirement-related matters with an adviser make higher contributions to voluntary personal pension plans in the days following the discussion (Hoechle et al., 2023[152]). In the United States, the impact of a financial adviser on the confidence of having enough money to live comfortably in retirement is larger among individuals with lower savings (Blanchett, 2025[153]).
Financial education seminars can have a positive impact on voluntary contributions. In Australia, working-age individuals who attended a financial education seminar organised by their pension fund were more likely to make voluntary contributions compared to the control group and made higher voluntary contributions of about AUD 15 000 on average over the span of two years, representing 6% of attendees’ pre‑seminar account balances (Ghafoori, Ip and Kabátek, 2021[154]). The seminars provided general advice related to retirement and basic financial skills and knowledge, and informed attendees about how much money one needs for retirement and how to make voluntary contributions, as well as their associated tax benefits.
2.3.6. Addressing the needs of individuals
Individuals may value certain design features that respond to their needs and could encourage them to contribute more to asset-backed pension plans. Adjusting the characteristics of pension funds to the preferences of plan members, providing an emergency savings option and allowing spousal contributions could help to increase voluntary contributions.
Certain characteristics of pension funds may influence the extent to which plan members may be willing to contribute. Individuals in six European countries would prefer to increase their contribution by 2 p.p. of their monthly salary in occupational pension plans with higher past performance, with boards considering the preferences and demands of plan members in their decision-making process, and with member representatives on the board (Fernández et al., 2024[155]). Higher employer contributions would also encourage higher contributions from individuals.
Providing an emergency savings option alongside a pension plan can improve individuals’ well-being without reducing retirement savings. Financial resilience is a strong predictor of saving for retirement (Suh, 2020[156]). Indeed, having emergency savings can reduce financial stress and provide confidence in saving for the long term. In the United States, employees participating in the emergency savings option within their 401(k) plan were twice as likely to increase their contributions to the retirement account of the 401(k) plan than other employees (Commonwealth, 2022[157]). The emergency savings option allows plan members to take early withdrawals to face contingencies without jeopardising future retirement income adequacy.
Permitting spousal contributions could allow couples to jointly plan for retirement and compensate for lower salaries or time taken out of work by one of the members of the couple to perform caretaking responsibilities. Several countries allow for contributions to the partner’s pension account or to a dedicated spousal retirement account (e.g. Australia, Canada, Spain, the United Kingdom, the United States). The same contribution limits typically apply for tax purposes, and spouse contributions count towards either the contributor’s or the recipient’s limit. Spousal contributions are particularly useful for women to accrue more pension rights and assets on their own to make up for unpaid activities performed for the household (OECD, 2021[16]).
The analysis shows that in 40 out of 41 countries, at least one category of workers would need to complement their mandatory pensions with voluntary pension contributions to reach their retirement income objective. The contribution gap shows how far workers are from reaching their retirement income objective. It is defined as the contribution rate that an individual would need to pay into a defined contribution (DC) pension plan to offset the difference between a retirement income objective and the expected retirement income from the mandatory pension system. The expected retirement income used is the theoretical gross replacement rates from the OECD Pensions at a Glance Database, and the retirement income objective used is a target replacement rate of 80% for low earners (half of average earnings), 70% for average earners and 60% for high earners (twice average earnings).
Contribution gaps are larger for average and high earners, women and the self-employed (Table 2.1). For a given distance between the expected and objective retirement incomes, the contribution gap is larger for workers with lower retirement ages, higher life expectancies, shorter contribution periods and lower investment performance.
Young workers, low earners, women and self-employed workers tend to participate and contribute less to asset-backed pension plans, making it more difficult for them to reach their retirement income objective. Behavioural biases, low levels of financial literacy, affordability concerns, career interruptions and selected design features are barriers that may prevent individuals from contributing voluntarily to achieve their retirement income objective. Additionally, contribution levels frequently align with low default and minimum contribution rates. Allowing individuals to use home equity release products reduces the need for additional pension contributions. However, these products are complicated, not easily available, and are not popular.
Various policies could address these barriers and steer individuals towards contribution levels that would allow them to achieve their retirement income objective. They are:
increasing mandatory contributions
encouraging voluntary employer contributions
nudging individuals through default options, automatic mechanisms, reminders or simplified processes
providing tax or non-tax financial incentives
informing individuals
addressing the needs of individuals.
2.4.1. Policy guidance
Policymakers could consider 11 steps to assess the need for additional pension contributions and ensure that individuals contribute enough to reach their retirement income objective (Figure 2.11).
Define appropriate retirement income objectives
Policymakers could define appropriate retirement income objectives for the population considering the spending needs of retirees. Retirement income objectives should account for the extent to which retirees have to self-finance certain costs, such as healthcare and long-term care. Retirement income objectives should also account for the population’s heterogeneity. This implies identifying the relevant groups of individuals who are likely to have different needs in retirement, for example distinguishing by income level, household composition (e.g. singles or couples, with or without children), homeownership and place of residence (e.g. in cities or in the countryside). The retirement income objective may be expressed as a target replacement rate or as a target amount in nominal terms depending on whether the policy objective is to smooth consumption or to reach a certain consumption standard (OECD, 2020[158]).
Assess the expected mandatory pensions
Policymakers could project the retirement income that the mandatory pension system is likely to deliver to future cohorts of retirees under realistic assumptions. Retirement income projections should consider the actual career patterns, earnings profiles and contribution densities of current workers, the effects of past pension reforms on different cohorts, and any behavioural changes and labour market transformations that may affect career paths and earnings. Assumptions should align closely to actual observations to provide a realistic view of future retirement incomes. It is important to model different categories of workers to account for different career patterns and earnings profiles, as well as the redistributive and means-tested features of the mandatory pension system.
Policymakers could then compare the expected retirement income to the retirement income objective across different groups of individuals to identify which ones would need additional pension contributions to reach their retirement income objective.
Policymakers could reflect on the financial sustainability of the mandatory pension system and on the level of diversification of the sources to finance retirement to evaluate the risks that the mandatory pension system may not eventually deliver as much as expected.
Derive the contribution gaps
Policymakers could estimate the level of additional pension savings that different categories of workers would need to contribute to reach their retirement income objective, when expected mandatory pensions fall short.
The calculation of this contribution gap could be tailored to the type of pension plan into which the additional contributions would be paid, i.e. mandatory or voluntary, with or without guarantees. Indeed, the mandatory or voluntary nature of the plan would affect the potential length of time during which the additional pension contributions would be paid as workers may not start saving from the beginning of their career in the case of a voluntary plan. The annual contribution effort to reach a given retirement income objective is indeed larger for shorter contribution periods. Moreover, contributions to a voluntary pension plan may be less constant during an individual’s career than those to a mandatory pension plan, as individuals may save less voluntarily when young as they have more demands on their income. The calculation of the contribution gap could consider the likely contribution pattern to derive what would be the maximum contribution effort during the career. Finally, it is also important to consider whether the plan receiving the additional contributions would provide guarantees or not (e.g. defined benefit versus defined contribution). In the absence of guarantees, the expected rate of return has a greater influence on the contribution gap, with a lower investment performance increasing the contribution effort.
The assumptions used to calculate the contribution gap could also consider variables such as inflation, productivity growth, expected return, interest rate, life expectancy, age of retirement and the level of fees and charges for managing pension assets.
Assess workers’ capacity to save more
Policymakers could assess the capacity of different categories of workers to contribute the amount needed to reach their retirement income objective. They could identify vulnerable groups that may be worse‑off if forced or pushed to save more for retirement. For example, it may be suboptimal to push individuals to pay additional pension contributions at the beginning of their career since they may face higher liquidity constraints linked to repaying a student loan, starting a family or buying their first home (Schlafmann, Setty and Vestman, 2024[36]; Fischer, Jensen and Koch, 2023[37]). Individuals with high-interest-rate loans may be better-off paying off their debt first, while workers on precarious jobs and low earnings may prioritise building precautionary savings. Moreover, migrant workers may have unaffordable contribution gaps as they may have lower entitlements from the mandatory pension system and shorter time to accumulate additional pension income. Policymakers could reflect on these situations and figure out whether increasing pension contributions is the best approach.
When assessing workers’ capacity to save more for retirement, policymakers could also consider the role of employers. In most OECD countries, employers pay a significant share of total contributions to asset-backed pension plans (OECD, 2022[49]). Employers may be willing to contribute beyond the legal requirements on behalf of their employees to attract and retain skilled workers. However, not all employers may be able or willing to do so, in particular smaller employers who may lack sufficient resources.
Assess voluntary saving behaviours
Policymakers could assess voluntary saving behaviours before implementing measures to increase pension contributions when mandatory pensions fall short. They could check the extent to which workers are already saving enough voluntarily to complement their mandatory pension and reach their retirement income objective. Policymakers could gather evidence on the voluntary savings of the population across different categories of workers and savings vehicles, including asset-backed pensions, long-term savings vehicles and real estate. When considering savings that are not earmarked for retirement, studies could be conducted to assess the extent to which they may be used to finance retirement.
Consider the trade‑offs between increasing mandatory or voluntary contributions
Policymakers need to decide whether to increase mandatory or voluntary contributions when workers with contribution gaps are not saving enough voluntarily and have the capacity to contribute more. There are important trade‑offs to consider. Increasing mandatory pension contributions may be the most efficient way to raise pension benefits and reach retirement income objectives. However, mandatory pension contributions may reduce employment, wages and household consumption, while it may increase labour informality and household debt.
Identifying which categories of workers have larger contribution gaps could shed some light on which type of contribution to increase.
When contribution gaps are widespread across all categories of workers, including low-income earners, increasing mandatory pension contributions may be the best approach to lift everyone up. Effects on the labour market could be monitored and compliance measures reinforced to avoid an increase in informality.
When contribution gaps are concentrated among high-income earners, increasing voluntary contributions may be the best approach as these workers are the most likely to save voluntarily, and leaving mandatory contributions untouched would avoid forcing low-income earners to oversave.
When contribution gaps are concentrated among the self-employed and they contribute less than employees towards the mandatory pension system, increasing their mandatory pension contributions may be the best approach. It would level the playing field between employees and the self-employed.
Align contribution rules with identified requirements
Policymakers could ensure that the rules related to contributions to asset-backed pension plans are in line with the additional pension contributions required to reach retirement income objectives. This implies reviewing contribution limits, as well as default and minimum contribution rates. For example, default contribution rates may be raised if they are well below what most workers would need to save additionally. Still, when contributions attracting tax relief are capped, the cap may be set below the largest contribution gap identified across all categories of workers to limit fiscal costs. In particular, high-income earners may not need to receive full tax relief on their contributions to save at the required level as they are likely to save anyway.
Select the best policy to steer individuals towards the contribution level they need to achieve their retirement income objective
The best policy to steer individuals towards the contribution level they need to achieve their retirement income objective would depend on the proportion and categories of workers having a contribution gap, the level of the contribution gap and the level of financial literacy.
There are trade‑offs involved when selecting the best policy. On the one hand, relatively high default contribution rates, automatic increases in contributions and increases in statutory retirement ages are likely to affect the saving behaviour of many individuals as these policy options rely on inertia. However, inertia leads to low levels of engagement and individuals may passively accept contribution levels that are not optimal given their circumstances. On the other hand, policy options that rely on individuals or employers to actively respond to new information or financial incentives may not bring the expected result for many individuals as the expected behavioural response may not always happen. Nevertheless, better informed and engaged individuals may be more likely to select the contribution level that best matches their circumstances.
In any case, policymakers should avoid reducing mandatory pension contributions to PAYG schemes to make room for higher contributions to asset-backed pension plans. While this may diversify the sources of retirement income, it may not help individuals achieve their retirement income objective if PAYG pensions are reduced proportionately. It also creates a transition cost for the PAYG scheme. Instead, contributions to complementary asset-backed pension plans should come from new savings or reallocated savings, thereby increasing total pension benefits.
Ensure that other design features support or reduce the contribution effort
The other design features of asset-backed pension plans should not undermine the effect of increasing pension contributions on retirement income outcomes. Policymakers could minimise or eliminate vesting periods to reduce the number of workers forfeiting part of their employer contributions when changing jobs. They could also ensure that pension contributions are invested so as to reach the highest risk-adjusted returns possible as better performance reduces the contribution effort to reach a given retirement income objective. This involves some level of equity allocation to achieve better long-term performance and higher retirement income (OECD, 2024[159]). Additionally, policymakers could ensure that pension providers offer good services and value for the fees they charge to avoid unnecessary reductions to asset accumulations.
Implement gradual increases
Policymakers could implement gradual increases in contribution rates in the context of compulsory or automatic enrolment to allow workers and employers to adjust. Increases in mandatory or default contribution rates can be phased in over several years, such as in Australia (2.5 p.p. in four years), Chile (6 p.p. in nine years), Ireland (9 p.p. in 10 years), Mexico (8.5 p.p. in 10 years) and the United Kingdom (6 p.p. in seven years). Indicators could also be calculated to monitor the economic climate during the planned increase and trigger a temporary pause in the increase in case of economic turbulences affecting the capacity of workers and employers to contribute (WPI Economics, 2023[160]).
Monitor outcomes and adjust as necessary
Policymakers could monitor the effects of any new policy to steer individuals towards contribution levels that better match their retirement income objectives. They could calculate relevant indicators broken down for different groups of individuals to follow progress on contributions. If the impact on contributions is insufficient given broader adequacy objectives (OECD, 2020[158]), policymakers could adjust the policy based on the challenges faced or implement a new policy that would be more likely to succeed given past experience.
References
[45] Allianz Center for the Future of Retirement (2026), Americans Cutting Back on Retirement Savings, Allianz Life Study Finds, https://www.allianzlife.com/about/newsroom/2026-Press-Releases/Americans-Cutting-Back-on-Retirement-Savings.
[78] Andersen, H., N. Hansen and A. Kuchler (2022), “Mandatory pension savings and long-run debt accumulation: evidence from Danish low-wage earners”, Journal of Pension Economics and Finance, Vol. 22/3, pp. 371-399, https://doi.org/10.1017/s1474747221000482.
[129] Australian National University (2021), The impact of government funded retirement contributions (matching) on the retirement savings behaviour of low and middle income individuals, https://coilink.org/20.500.12592/zfv68n.
[31] Australian Taxation Office (2025), Taxation statistics 2021–22 - Individuals - Table 23, https://data.gov.au/data/dataset/4be150cc-8f84-46b8-8c61-55ff1d48a700/resource/545be24d-b8a2-4abd-9193-097e8114dcbc/download/ts22individual23contributionsbyagesextaxableincome.xlsx.
[112] Azuara Herrera, O. et al. (2021), Saving without Barriers: Lessons from the Retirement Savings Laboratory’s Pilot Projects, Inter-American Development Bank, https://doi.org/10.18235/0003177.
[149] Barrett, A. et al. (n.d.), “Can Evaluative Structures Assist Retirement Saving Decisions?”, https://doi.org/10.26481/dis.20240416ab.
[41] Benartzi, S. and R. Thaler (2007), “Heuristics and Biases in Retirement Savings Behavior”, Journal of Economic Perspectives, Vol. 21/3, pp. 81-104, https://doi.org/10.1257/jep.21.3.81.
[66] Benzarti, Y. and J. Harju (2021), “Using Payroll Tax Variation to Unpack the Black Box of Firm-Level Production”, Journal of the European Economic Association, Vol. 19/5, pp. 2737-2764, https://doi.org/10.1093/jeea/jvab010.
[131] Bernal, N., S. Galiani and O. Molina (2025), Does Matching Contribution Incentivize Informal Workers to Participate in Retirement Saving Plans? A Randomized Evaluation Interacted with a Natural Experiment, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w33925.
[95] Beshears, J. et al. (2017), “How Do Consumers Respond When Default Options Push the Envelope?”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3050562.
[79] Beshears, J. et al. (2025), “Does Pension Automatic Enrollment Increase Debt? Evidence from a Large-Scale Natural Experiment”, Working Paper, https://www.hbs.edu/faculty/Pages/item.aspx?num=65671.
[82] Beshears, J. et al. (2021), “Borrowing to Save? The Impact of Automatic Enrollment on Debt”, The Journal of Finance, Vol. 77/1, pp. 403-447, https://doi.org/10.1111/jofi.13069.
[123] Beshears, J. et al. (2017), “Does front-loading taxation increase savings? Evidence from Roth 401(k) introductions”, Journal of Public Economics, Vol. 151, pp. 84-95, https://doi.org/10.1016/j.jpubeco.2015.09.007.
[110] Beshears, J. et al. (2015), “The Effect of Providing Peer Information on Retirement Savings Decisions”, The Journal of Finance, Vol. 70/3, pp. 1161-1201, https://doi.org/10.1111/jofi.12258.
[103] Beshears, J. et al. (2013), “Simplification and saving”, Journal of Economic Behavior & Organization, Vol. 95, pp. 130-145, https://doi.org/10.1016/j.jebo.2012.03.007.
[107] Beshears, J. et al. (2021), “Using fresh starts to nudge increased retirement savings”, Organizational Behavior and Human Decision Processes, Vol. 167, pp. 72-87, https://doi.org/10.1016/j.obhdp.2021.06.005.
[89] Beshears, J. et al. (2023), “Automatic enrollment with a 12 percent default contribution rate”, Journal of Pension Economics and Finance, Vol. 24/1, pp. 152-182, https://doi.org/10.1017/s1474747223000148.
[105] Beshears, J. et al. (2016), “Framing the Future: The Risks of Pre-Commitment Nudges and Potential of Fresh Start Messaging”, Working Paper, https://static1.squarespace.com/static/5353b838e4b0e68461b517cf/t/583ca5acd2b8571174b28e40/1480369581625/48-Beshears_et_al_2016.pdf.
[63] Biesenbeek, C., B. Heerma van Voss and M. Mastrogiacomo (2020), “Workers without pension accrual”, DNB Occasional Studies, Vol. 20 - 3, https://www.dnb.nl/publicaties/publicaties-onderzoek/occasional-study/nr-3-2022-werkenden-zonder-pensioenopbouw/.
[11] Binswanger, J. and D. Schunk (2011), “What is an adequate standard of living during Retirement?”, Journal of Pension Economics and Finance, Vol. 11/2, pp. 203-222, https://doi.org/10.1017/s1474747211000618.
[153] Blanchett, D. (2025), Financial Advice and Retirement Confidence: What’s Wealth Got to Do With It?, https://www.kiplinger.com/retirement/retirement-planning/financial-advice-and-retirement-confidence-by-wealth-level.
[124] Blaufus, K. and M. Milde (2021), “Tax Misperceptions and the Effect of Informational Tax Nudges on Retirement Savings”, Management Science, Vol. 67/8, pp. 5011-5031, https://doi.org/10.1287/mnsc.2020.3761.
[114] Bosch, M. and A. Rubli (2021), “Increasing Retirement Savings through Access Points and Persuasive Messages”, Journal of Human Resources, Vol. 58/5, pp. 1752-1781, https://doi.org/10.3368/jhr.58.5.0220-10750r2.
[68] Bosch, N. et al. (2022), “The Incidence of Pension Contributions: A Panel Based Analysis of the Impact of Pension Contributions on Labor Cost, Wages and Labor Supply”, De Economist, Vol. 170/1, pp. 107-132, https://doi.org/10.1007/s10645-022-09401-z.
[72] Bozio, A. et al. (2025), “Does Tax-Benefit Linkage Matter for the Incidence of Payroll Taxes?”, Review of Economic Studies, https://doi.org/10.1093/restud/rdaf059.
[70] Breunig, R. and K. Sobeck (2020), The economic incidence of superannuation, https://treasury.gov.au/sites/default/files/2020-07/01-r2019-100554.pdf.
[121] Brière, M., J. Poterba and A. Szafarz (2025), “Does Tax Deductibility Increase Retirement Saving? Lessons from a French Natural Experiment”, Amundi Investment Institute Working Paper 166, https://research-center.amundi.com/article/does-tax-deductibility-increase-retirement-saving-lessons-french-natural-experiment.
[14] Broome, M. and I. Mulheirn (2024), Perfectly adequate?: Revisiting pensions adequacy 20 years after the Pensions Commission, Resolution Foundation, https://doi.org/10.63492/vmzf40.
[86] Burgherr, D. (2025), “Saving Responses to Mandatory Pension Plans”, Equality of Opportunity Research Series 78, https://www.urpp-equality.uzh.ch/dam/jcr:665c0dd7-bd77-454f-9155-5666feaa7d5f/78_Burgherr_SavingResponsesToMandatoryPensionPlans.pdf.
[150] Burke, J. and A. Hung (2015), Do Financial Advisers Influence Savings Behavior?, RAND Corporation, http://www.jstor.org/stable/10.7249/j.ctt19w72kb.
[91] Burke, J., A. Hung and J. Luoto (2017), “Opting out of Retirement Plan Default Settings”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.2910771.
[143] Cardella, E., C. Kalenkoski and M. Parent (2021), “Less is not more: 401(k) plan information and retirement planning choices”, Journal of Pension Economics and Finance, Vol. 22/3, pp. 331-351, https://doi.org/10.1017/s1474747221000445.
[51] Carranza, G. and A. Goodman (2024), Retention or regressivity? The empirical effects of 401(k) vesting schedules, Elsevier BV, https://doi.org/10.2139/ssrn.4876231.
[93] Carroll, G. et al. (2009), “Optimal Defaults and Active Decisions”, Quarterly Journal of Economics, Vol. 124/4, pp. 1639-1674, https://doi.org/10.1162/qjec.2009.124.4.1639.
[19] Central Statistics Office (2024), Pension Coverage 2024, https://www.cso.ie/en/releasesandpublications/ep/p-pens/pensioncoverage2024/personalpensions/ (accessed on 20 February 2025).
[128] Chan, M. et al. (2020), “How effective are Matching Schemes in enticing low-income earners to save more for retirement? Evidence from a national scheme”, IZA Discussion Papers 13939, https://hdl.handle.net/10419/232691.
[122] Chetty, R. et al. (2014), “Active vs. Passive Decisions and Crowd-Out in Retirement Savings Accounts: Evidence from Denmark *”, The Quarterly Journal of Economics, Vol. 129/3, pp. 1141-1219, https://doi.org/10.1093/qje/qju013.
[133] Choi, J. (2015), “Contributions to Defined Contribution Pension Plans”, Annual Review of Financial Economics, Vol. 7/1, pp. 161-178, https://doi.org/10.1146/annurev-financial-111914-041834.
[104] Choi, J. et al. (2017), “Small cues change savings choices”, Journal of Economic Behavior & Organization, Vol. 142, pp. 378-395, https://doi.org/10.1016/j.jebo.2017.08.010.
[101] Choi, J. et al. (2024), Smaller than We Thought? The Effect of Automatic Savings Policies, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w32828.
[56] Choi, J. et al. (2004), “For better or for worse: Default effects and 401 (k) savings behavior”, in Wise, D. (ed.), Perspectives on the Economics of Aging, University of Chicago Press, http://www.nber.org/chapters/c10341.
[88] Choi, J. et al. (2003), “Optimal Defaults”, American Economic Review, Vol. 93/2, pp. 180-185, https://doi.org/10.1257/000282803321947010.
[135] Choi, J. et al. (2002), “Defined Contribution Pensions: Plan Rules, Participant Choices, and the Path of Least Resistance”, Tax Policy and the Economy, Vol. 16, pp. 67-113, https://doi.org/10.1086/654750.
[94] Choukhmane, T. (2025), “Default Options and Retirement Saving Dynamics”, American Economic Review, Vol. 115/11, pp. 3749-3787, https://doi.org/10.1257/aer.20210881.
[80] Choukhmane, T. and C. Palmer (2025), “The Effect of Increasing Retirement Saving on Consumption, Balance Sheets, and Welfare”, Working Paper, https://mitsloan.mit.edu/centers-initiatives/cfi/how-do-consumers-finance-increased-retirement-savings.
[102] Clark, R. et al. (2018), “Informing Retirement Savings Decisions: A Field Experiment on Supplemental Plans”, Economic Inquiry, Vol. 57/1, pp. 188-205, https://doi.org/10.1111/ecin.12731.
[71] Coates, B., M. Cowgill and W. Mackey (2020), No free lunch: Higher super means lower wages, https://grattan.edu.au/wp-content/uploads/2020/02/No-free-lunch-Higher-superannuation-means-lower-wages.pdf.
[157] Commonwealth (2022), A Case Study in Workplace Emergency Savings: How UPS and Voya Designed a Solution Resulting in $10 Million in Savings, https://buildcommonwealth.org/research/a-case-study-in-workplaceemergency-savings/.
[125] Congressional Research Service (2025), “The Retirement Savings Contribution Credit and the Saver’s Match”, In Focus IF11159, https://www.congress.gov/crs-product/IF11159.
[138] Corneo, G., M. Keese and C. Schröder (2008), “Can governments boost voluntary retirement savings via tax incentives and subsidies? A German case study for low-income households”, Economics Working Papers 2008-18, https://www.econstor.eu/bitstream/10419/27673/1/EWP-2008-18.pdf.
[35] Department for Work and Pensions (2025), Analysis of Automatic Enrolment saving levels, https://www.gov.uk/government/statistics/analysis-of-automatic-enrolment-saving-levels/analysis-of-automatic-enrolment-saving-levels.
[4] Department for Work and Pensions (2025), Analysis of future pension incomes, https://www.gov.uk/government/statistics/analysis-of-future-pension-incomes-2025/analysis-of-future-pension-incomes-2025.
[44] Department for Work and Pensions (2024), “Low Earners and workplace pension saving – a qualitative study”, DWP research report 1046, https://www.gov.uk/government/publications/low-earners-and-workplace-pension-saving-a-qualitative-study/research-report-low-earners-and-workplace-pension-saving-a-qualitative-study.
[42] Deschacht, N., I. Guillemyn and S. Vujić (2025), “Trading Pay for Pensions: Evidence from a Discrete Choice Experiment in the United Kingdom”, IZA Discussion Paper Series 18164, https://docs.iza.org/dp18164.pdf.
[58] Deutsche Bank Research (2026), Private pension reform in Germany – a bold move towards capital markets, https://www.dbresearch.com/PROD/IE-PROD/PDFVIEWER.calias?pdfViewerPdfUrl=PROD0000000000620489&rwnode=REPORT.
[27] DGSFP (2025), Seguros y Fondos de Pensiones: Informe 2024, https://dgsfp.mineco.gob.es/es/Publicaciones/DocumentosPublicaciones/Memoria%20informe%20del%20sector%202024.pdf.
[148] Dolls, M. et al. (2018), “Do retirement savings increase in response to information about retirement and expected pensions?”, Journal of Public Economics, Vol. 158, pp. 168-179, https://doi.org/10.1016/j.jpubeco.2017.12.014.
[108] Døskeland, T. et al. (2025), “The role of social norms in retirement saving: Evidence from two natural field experiments”, Journal of Business Research, Vol. 190, p. 115201, https://doi.org/10.1016/j.jbusres.2025.115201.
[119] Etgeton, S., B. Fischer and H. Ye (2023), “The effect of increasing retirement age on households’ savings and consumption expenditure”, Journal of Public Economics, Vol. 221, p. 104845, https://doi.org/10.1016/j.jpubeco.2023.104845.
[1] European Commission (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://data.europa.eu/doi/10.2767/909323.
[33] F&P (2025), Pensionsindbetalinger, https://fogp.dk/media/y3tjoqdj/pensionsindbetalinger.xls.
[127] Falk, J. and N. Karamcheva (2022), “The impact of an employer match and automatic enrollment on the savings behavior of public-sector workers”, Journal of Pension Economics and Finance, Vol. 22/1, pp. 38-68, https://doi.org/10.1017/s1474747221000366.
[151] Fang, J., W. Hao and M. Reyers (2022), “Financial advice, financial literacy and social interaction: what matters to retirement saving decisions?”, Applied Economics, Vol. 54/50, pp. 5827-5850, https://doi.org/10.1080/00036846.2022.2053654.
[57] Federal Ministry of Labour and Social Affairs (2025), Statistik zur privaten Altersvorsorge (Riester-Rente), https://www.bmas.de/DE/Service/Statistiken-Open-Data/Statistik-zu-Riester-Vertraegen/statistik-zusaetzliche-altersvorsorge.html (accessed on 1 April 2026).
[34] Federal Ministry of Labour and Social Affairs (2025), Verbreitung der Altersvorsorge 2023, https://www.bmas.de/DE/Service/Publikationen/Forschungsberichte/fb-653-verbreitung-der-altersvorsorge-2023.html.
[26] Finance Norway (2025), Statistikk: Slik er pensjonsordningene i ulike bransjer, https://www.finansnorge.no/tema/statistikk-og-analyse/pensjon-og-sparing/sparesatser-i-innskuddspensjon/.
[22] Financial Conduct Authority (2025), Financial Lives 2024 survey, https://www.fca.org.uk/financial-lives/financial-lives-2024.
[20] Financial Markets Authority (2025), KiwiSaver Annual Report 2025, https://www.fma.govt.nz/assets/Reports/KiwiSaver-Annual-Report-2025.pdf.
[37] Fischer, M., B. Jensen and M. Koch (2023), “Optimal retirement savings over the life cycle: A deterministic analysis in closed form”, Insurance: Mathematics and Economics, Vol. 112, pp. 48-58, https://doi.org/10.1016/j.insmatheco.2023.05.010.
[84] Friedberg, L., A. Leive and W. Cai (2024), “Does mandatory retirement saving crowd out voluntary retirement saving?”, Journal of Economic Behavior & Organization, Vol. 225, pp. 20-36, https://doi.org/10.1016/j.jebo.2024.06.024.
[76] Frish, R. (2025), “The effect of mandated savings on private consumption: Evidence from Israel’s pension reform”, Research in Economics, Vol. 79/2, p. 101038, https://doi.org/10.1016/j.rie.2025.101038.
[147] Fuentes, O. et al. (2024), “Personalized Information as a Tool to Improve Pension Savings: Results from a Randomized Control Trial in Chile”, Economic Development and Cultural Change, Vol. 72/2, pp. 725-770, https://doi.org/10.1086/720718.
[99] García, J. and J. Vila (2020), “Financial literacy is not enough: The role of nudging toward adequate long-term saving behavior”, Journal of Business Research, Vol. 112, pp. 472-477, https://doi.org/10.1016/j.jbusres.2020.01.061.
[117] García-Miralles, E. and J. Leganza (2024), “Public Pensions and Private Savings”, American Economic Journal: Economic Policy, Vol. 16/2, pp. 366-405, https://doi.org/10.1257/pol.20220019.
[111] Gars, J. et al. (2025), “Personalized reminders: Evidence from a field experiment with voluntary retirement savings in Colombia”, Journal of Behavioral and Experimental Finance, Vol. 48, p. 101112, https://doi.org/10.1016/j.jbef.2025.101112.
[154] Ghafoori, E., E. Ip and J. Kabátek (2021), “The impacts of a large-scale financial education intervention on retirement saving behaviors and portfolio allocation: Evidence from pension fund data”, Journal of Banking & Finance, Vol. 130, p. 106195, https://doi.org/10.1016/j.jbankfin.2021.106195.
[109] Goldin, J. et al. (2020), “How much to save? Decision costs and retirement plan participation”, Journal of Public Economics, Vol. 191, p. 104247, https://doi.org/10.1016/j.jpubeco.2020.104247.
[10] Government Accountability Office (2016), Retirement Security: Better Information on Income Replacement Rates Needed to Help Workers Plan for Retirement, https://www.gao.gov/products/gao-16-242.
[134] Greig, F. et al. (2024), Are employers optimizing their 401(k) match?, https://corporate.vanguard.com/content/dam/corp/research/pdf/are_employers_optimizing_their_401k_match.pdf.
[155] Henkens, K. (ed.) (2024), “Can pension fund design increase pension savings? Evidence from a conjoint experiment in six European countries”, Work, Aging and Retirement, https://doi.org/10.1093/workar/waae021.
[97] Hobbs, B. and J. Feld (2021), “What would happen if people could choose automatically increasing KiwiSaver contributions?”, New Zealand Economic Papers, Vol. 57/1, pp. 61-75, https://doi.org/10.1080/00779954.2021.1998198.
[152] Hoechle, D. et al. (2023), “Financial Advice and Retirement Savings”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4339769.
[136] Huberman, G., S. Iyengar and W. Jiang (2007), “Defined Contribution Pension Plans: Determinants of Participation and Contributions Rates”, Journal of Financial Services Research, Vol. 31/1, pp. 1-32, https://doi.org/10.1007/s10693-007-0003-6.
[24] Irish Life (2018), A guide to defined contribution plan contribution structures, https://web.archive.org/web/20240523001906/https://www.irishlifeemployersolutions.ie/report-launch-guide-dc-plan-contribution-structures.
[83] Jensen, S. et al. (2025), “Does Mandatory Saving Reduce Voluntary Saving? Evidence from a Pension Reform”, Review of Economics and Statistics, pp. 1-48, https://doi.org/10.1162/rest.a.264.
[75] 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.
[50] Keune, M. (2018), “Opportunity or threat? How trade union power and preferences shape occupational pensions”, Social Policy & Administration, Vol. 52/2, pp. 463-476, https://doi.org/10.1111/spol.12377.
[67] Kim, J., S. Kim and K. Koh (2022), “Labor market institutions and the incidence of payroll taxation”, Journal of Public Economics, Vol. 209, p. 104646, https://doi.org/10.1016/j.jpubeco.2022.104646.
[25] Kirkpatrick, L., L. Meehan and G. Pacheco (2024), Distributional analysis of KiwiSaver contributions, New Zealand Policy Research Institute, Auckland, https://assets.retirement.govt.nz/public/Uploads/Policy/Distributional-analysis-of-KiwiSaver-contributions.pdf.
[139] Kolerus, C., I. Koske and F. Hüfner (2012), “Selected Aspects of Household Savings in Germany: Evidence from Micro-Data”, OECD Economics Department Working Papers, No. 999, OECD Publishing, Paris, https://doi.org/10.1787/5k8zpt6w2k7h-en.
[142] Kramer, M., R. Lensink and A. Plantinga (2024), “Do online pension dashboards affect pension knowledge and expectations? Evidence from a randomized survey experiment”, Journal of Pension Economics and Finance, Vol. 23/4, pp. 504-527, https://doi.org/10.1017/s1474747224000040.
[77] Larsen, L. et al. (2025), “Mandatory pension contributions, household consumption, and savings”, Journal of Pension Economics and Finance, pp. 1-24, https://doi.org/10.1017/s147474722510005x.
[46] Living Wage Foundation (2024), Retirement realities: Understanding the impact of the cost-of-living on pension savings and retirement expectations, https://www.livingwage.org.uk/retirement-realities-understanding-impact-cost-living-pension-savings-and-retirement-expectations.
[132] Madrian, B. (2012), “Matching Contributions and Savings Outcomes: A Behavioral Economics Perspective”, in Hinz, R. et al. (eds.), Matching Contributions for Pensions: A Review of International Evidence, The World Bank, https://doi.org/10.1596/978-0-8213-9492-2.
[55] Madrian, B. and D. Shea (2001), “The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior”, The Quarterly Journal of Economics, Vol. 116/4, pp. 1149-1187, https://doi.org/10.1162/003355301753265543.
[145] McGowan, F. and P. Lunn (2019), “Supporting decision-making in retirement planning: Do diagrams on Pension Benefit Statements help?”, Journal of Pension Economics and Finance, Vol. 19/3, pp. 323-343, https://doi.org/10.1017/s1474747219000015.
[90] McKenzie, C., M. Liersch and S. Finkelstein (2006), “Recommendations Implicit in Policy Defaults”, Psychological Science, Vol. 17/5, pp. 414-420, https://doi.org/10.1111/j.1467-9280.2006.01721.x.
[38] Money & Pensions Service (2020), Barriers to planning for retirement and later life, https://legacymasassets.blob.core.windows.net/production-fincap-cms/files/000/000/524/original/Barriers_to_ret_MAPS_final.pdf.
[74] Moreno, C. (2025), “The impact of pension systems in labor markets with informality”, Working Paper, https://carlamoreno.org/files/jmp_moreno.pdf.
[118] Nakazawa, N. (2022), “The Effects of Increasing the Eligibility Age for Public Pension on Individual Labor Supply”, Journal of Human Resources, Vol. 60/1, pp. 102-128, https://doi.org/10.3368/jhr.0421-11627r1.
[92] Nest Insight (2025), Simplifying retirement savings for self-employed people, https://www.nestinsight.org.uk/wp-content/uploads/2025/06/Simplifying-retirement-savings-for-self-employed-people.pdf.
[48] Nest Insight (2022), Employer pension contributions in the UK, https://www.nestinsight.org.uk/wp-content/uploads/2022/10/Employer-pension-contributions-in-the-UK.pdf.
[141] Nest Insight (2020), Beyond the defaults, https://www.nestinsight.org.uk/wp-content/uploads/2020/11/Beyond-the-defaults.pdf.
[52] OECD (2025), Annual survey of financial incentives for retirement savings: Country profiles 2025, OECD Publishing, Paris, https://doi.org/10.1787/631aa80d-en.
[6] OECD (2025), OECD Pensions at a Glance Database, https://data-explorer.oecd.org/s/568 (accessed on 24 March 2026).
[7] OECD (2025), Pensions at a Glance 2025: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/e40274c1-en.
[53] OECD (2024), “Assessing the design of financial incentives for retirement savings in OECD countries”, in OECD Pensions Outlook 2024: Improving Asset-backed Pensions for Better Retirement Outcomes and More Resilient Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/51510909-en.
[8] OECD (2024), “Designing the payout phase for defined contribution pensions to better meet financial needs in retirement”, in OECD Pensions Outlook 2024: Improving Asset-backed Pensions for Better Retirement Outcomes and More Resilient Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/51510909-en.
[159] OECD (2024), “Does investing in equity markets bring better retirement income outcomes to members of defined contribution pension plans?”, in OECD Pensions Outlook 2024: Improving Asset-backed Pensions for Better Retirement Outcomes and More Resilient Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/51510909-en.
[61] OECD (2024), “Home equity release products for the elderly: opportunities and challenges”, in OECD Pensions Outlook 2024: Improving Asset-backed Pensions for Better Retirement Outcomes and More Resilient Pension Systems, OECD Publishing, Paris, https://doi.org/10.1787/51510909-en.
[59] OECD (2024), OECD Affordable Housing Database - indicator HM1.3 Housing tenures, https://oe.cd/ahd.
[40] OECD (2023), “OECD/INFE 2023 International Survey of Adult Financial Literacy”, OECD Business and Finance Policy Papers, No. 39, OECD Publishing, Paris, https://doi.org/10.1787/56003a32-en.
[49] OECD (2022), “How best to involve employers in the provision of asset-backed pension arrangements”, in OECD Pensions Outlook 2022, OECD Publishing, Paris, https://doi.org/10.1787/41f6fe2a-en.
[60] OECD (2021), Brick by Brick: Building Better Housing Policies, OECD Publishing, Paris, https://doi.org/10.1787/b453b043-en.
[16] OECD (2021), Towards Improved Retirement Savings Outcomes for Women, OECD Publishing, Paris, https://doi.org/10.1787/f7b48808-en.
[158] OECD (2020), “A framework for assessing the adequacy of retirement income”, in OECD Pensions Outlook 2020, OECD Publishing, Paris, https://doi.org/10.1787/67ede41b-en.
[113] OECD (2020), “Increasing the role of retirement savings plans for workers in non-standard forms of work”, in OECD Pensions Outlook 2020, OECD Publishing, Paris, https://doi.org/10.1787/67ede41b-en.
[17] OECD (2019), “Are funded pensions well designed to adapt to non-standard forms of work?”, in Pensions at a Glance 2019: OECD and G20 Indicators, OECD Publishing, Paris, https://doi.org/10.1787/b6d3dcfc-en.
[120] OECD (2018), Financial Incentives and Retirement Savings, OECD Publishing, Paris, https://doi.org/10.1787/9789264306929-en.
[39] OECD (2018), “Improving retirement incomes considering behavioural biases and limited financial knowledge”, in OECD Pensions Outlook 2018, OECD Publishing, Paris, https://doi.org/10.1787/pens_outlook-2018-8-en.
[9] OECD (2014), “Saving for retirement and the role of private pension provision in retirement readiness”, in OECD Pensions Outlook 2014, OECD Publishing, Paris, https://doi.org/10.1787/9789264222687-en.
[15] OECD (2012), “Coverage of Private Pension Systems: Evidence and Policy Options”, in OECD Pensions Outlook 2012, OECD Publishing, Paris, https://doi.org/10.1787/9789264169401-en.
[28] ONS (2024), Employee contribution bands by age group and pension type: Table P5, https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/workplacepensions/datasets/annualsurveyofhoursandearningspensiontablesemployeecontributionbandsbyagegroupandbypensiontypep5.
[29] ONS (2024), Employer contribution bands by age group and pension type: Table P9, https://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/workplacepensions/datasets/annualsurveyofhoursandearningspensiontablesemployercontributionbandsbyagegroupandbypensiontypep9.
[5] Oxford Economics (2026), Higher auto-enrolment contributions, pension adequacy and economic outcomes, https://www.oxfordeconomics.com/resource/higher-auto-enrolment-contributions-pension-adequacy-and-economic-outcomes/.
[73] Pardo, O. (2023), “Mandatory retirement savings in the presence of an informal labor market”, Journal of Population Economics, Vol. 36/4, pp. 2857-2888, https://doi.org/10.1007/s00148-023-00967-9.
[65] Pedersen, A. et al. (2018), “Trade unions and the politics of occupational pensions in Denmark and Norway”, Transfer: European Review of Labour and Research, Vol. 24/1, pp. 109-122, https://doi.org/10.1177/1024258917748873.
[62] Pensions Policy Institute (2025), Modelling of pension policy options, analysis based upon the Wealth and Assets Survey dataset and PPI individual modelling – updated for 2025 PLSA RLS Update, https://www.pensionspolicyinstitute.org.uk/research-library/research-reports/2025/modelling-of-pension-policy-options-updated-for-2025-plsa-rls-update/.
[54] Pensions Policy Institute (2024), Automatic enrolment contributions briefing paper, https://www.pensionspolicyinstitute.org.uk/media/mwal1sfp/20240620-ae-contributions-briefing-paper-final.pdf.
[18] PensionStat.be (2025), Taux de participation, https://www.pensionstat.be/fr/chiffres-cles/pension-complementaire/taux-de-participation.
[43] Pew (2025), “Workers Without Access to Retirement Benefits Struggle to Build Wealth”, Issue Brief, https://www.pew.org/en/research-and-analysis/issue-briefs/2025/06/workers-without-access-to-retirement-benefits-struggle-to-build-wealth.
[140] Pfarr, C. and U. Schneider (2012), “Choosing between subsidized or unsubsidized private pension schemes: evidence from German panel data”, Journal of Pension Economics and Finance, Vol. 12/1, pp. 62-91, https://doi.org/10.1017/s1474747212000170.
[100] Proctor, D. and J. Young (2021), Automatic escalation and DC saving rates, https://institutional.vanguard.com/content/dam/inst/iig-transformation/insights/pdf/2022/automatic-escalation-and-dc-saving-rates.pdf.
[36] Ramadorai, T. (ed.) (2024), “Designing Pension Plans According to Consumption-Savings Theory”, The Review of Financial Studies, https://doi.org/10.1093/rfs/hhae061.
[115] Ramadorai, T. (ed.) (2024), “Digitalization and Retirement Contribution Behavior: Evidence from Administrative Data”, The Review of Financial Studies, Vol. 37/8, pp. 2510-2549, https://doi.org/10.1093/rfs/hhae015.
[106] Reiff, J. et al. (2023), “Save More Today or Tomorrow: The Role of Urgency in Precommitment Design”, Journal of Marketing Research, Vol. 60/6, pp. 1095-1113, https://doi.org/10.1177/00222437231153396.
[96] Rubaltelli, E. and L. Lotto (2021), “Nudging freelance professionals to increase their retirement pension fund contributions”, Judgment and Decision Making, Vol. 16/2, pp. 551-565, https://doi.org/10.1017/s1930297500008664.
[85] Ruthbah, U. and N. Pham (2020), Household Savings and the Superannuation Guarantee, https://www.monash.edu/__data/assets/pdf_file/0005/2406893/Household-Savings-and-the-Superannuation-Guarantee.pdf.
[69] Scarfe, R., D. Schaefer and T. Sulka (2024), “The incidence of workplace pensions: Evidence from the UK’s automatic enrollment mandate”, Working Paper No. 2402, https://hdl.handle.net/10419/289841.
[12] Schmied, J. (2023), “The replacement rate that maintains income satisfaction through retirement: The question of income-dependence”, The Journal of the Economics of Ageing, Vol. 26, p. 100471, https://doi.org/10.1016/j.jeoa.2023.100471.
[23] Sigedis (2025), La constitution d’une pension complémentaire chez les salariés en Belgique: Participation et montant des cotisations en 2023, https://www.pensionstat.be/file/cc73d96153bbd5448a56f19d925d05b1379c7f21/20a8ad78d927df7ad0d4a0d22451c2ec04356edf/rapport-thematique---la-constitution-dune-pension-complementaire-chez-les-salaries-en-belgique.pdf.
[81] Sigurdardóttir, E. (2023), Do Mandatory Pension Savings Cause Debt Accumulation? An Event Study Approach, https://research.cbs.dk/en/studentProjects/do-mandatory-pension-savings-cause-debt-accumulation-an-event-stu/.
[146] Smyrnis, G. et al. (2019), “Motivated Saving: The Impact of Projections on Retirement Saving Intentions”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.3464813.
[130] Sobeck, K. and R. Breunig (2026), “Retirement Savings Incentives for Low‐ and Middle‐Income Individuals: Does Government Funded Matching Change Behaviour?*”, Economic Record, Vol. 102/337, pp. 208-226, https://doi.org/10.1111/1475-4932.70032.
[32] Statistics Canada (2023), “Recent trends in families’ contributions to three registered savings accounts”, Income Research Paper Series, https://www150.statcan.gc.ca/n1/pub/75f0002m/75f0002m2023008-eng.htm#n3-refa.
[156] Suh, E. (2020), Younger Adults’ Retirement Saving and Wealth Accumulation in Britain: A Quantitative Investigation, https://etheses.lse.ac.uk/4135/1/Suh__Younger-adults-retirement.pdf.
[47] Te Ara Ahunga Ora Retirement Commission (2025), KiwiSaver Employer Contribution Insights 2025, https://assets.retirement.govt.nz/public/Uploads/Retirement-Income-Policy-Review/2025-RRIP/KiwiSaver-Employer-Contribution-Insights.pdf.
[137] Te Ara Ahunga Ora Retirement Commission (2025), KiwiSaver Government Contribution Distributional Analysis 2025, https://assets.retirement.govt.nz/public/Uploads/Retirement-Income-Policy-Review/2025-RRIP/KiwiSaver-Government-Contribution-Distributional-Analysis-2025.pdf.
[21] Te Ara Ahunga Ora Retirement Commission/Hnry (2025), Improving the retirement savings of the self-employed, https://assets.retirement.govt.nz/public/Uploads/Improving-the-retirement-savings-of-the-self-employed_Final.pdf.
[98] Thaler, R. and S. Benartzi (2004), “Save More Tomorrow™: Using Behavioral Economics to Increase Employee Saving”, Journal of Political Economy, Vol. 112/S1, pp. S164-S187, https://doi.org/10.1086/380085.
[87] Thaler, R. and C. Sunstein (2008), Nudge: Improving decisions about health, wealth, and happiness, Yale University Press, https://yalebooks.yale.edu/book/9780300262285/nudge/.
[64] Trampusch, C. (2013), “Employers and collectively negotiated occupational pensions in Sweden, Denmark and Norway: Promoters, vacillators and adversaries”, European Journal of Industrial Relations, Vol. 19/1, pp. 37-53, https://doi.org/10.1177/0959680112474882.
[126] Transamerica Institute (2026), Less Than Half of U.S. Workers Know About a Tax Credit for Retirement Savers, https://www.transamericainstitute.org/in-the-news/press-releases/details/less-than-half-of-workers-know-about-tax-credit-for-retirement-savers.
[30] Vanguard (2025), How America Saves 2025, https://corporate.vanguard.com/content/dam/corp/research/pdf/how_america_saves_report_2025.pdf.
[116] Vižintin, Ž. and S. Peksevim (2021), “How to Increase Your Savings While Spending”, Behavioral Scientist, https://behavioralscientist.org/how-to-increase-your-savings-while-spending/.
[2] Wolf, I. and Lorena (2021), “Benefit Adequacy in Funded Pension Systems: Micro-Simulation of the Israeli Pension Scheme”, International Journal of Economics and Business Administration, Vol. IX/Issue 2, pp. 143-164, https://doi.org/10.35808/ijeba/694.
[144] Wood, S. and C. Morewedge (eds.) (2024), “The Beneficent and Maleficent Effects of Simplification on Retirement Savings”, Journal of Consumer Research, Vol. 52/2, pp. 244-265, https://doi.org/10.1093/jcr/ucae067.
[160] WPI Economics (2023), Raising the bar: A framework for increasing auto-enrolment contributions, https://wpieconomics.com/publications/raising-the-bar/.
[3] Yin, Y., A. Chen and A. Munnell (2024), “The National Retirement Risk Index: An Update from the 2022 SCF”, Issue in Brief 24-5, https://crr.bc.edu/the-national-retirement-risk-index-an-update-from-the-2022-scf/.
[13] Yin, Y., A. Chen and A. Munnell (2023), “The National Retirement Risk Index: Version 2.0”, Issue in Brief 23-10, https://crr.bc.edu/the-national-retirement-risk-index-version-2-0/.
← 1. The analysis covers all OECD countries as well as Bulgaria, Croatia and Romania.
← 2. For New Zealand and the United Kingdom, the voluntary schemes with an automatic enrolment mechanism are considered as quasi-mandatory for employees and therefore included in the expected retirement income from the mandatory pension system because they cover more than 85% of the target population.
← 3. These assumptions are in line with those used in the OECD Pensions at a Glance Database to ensure consistency with the theoretical gross replacement rates used.
← 4. By contrast, the contribution gaps estimated for Swiss employees are likely to be overstated, as contribution rates in mandatory occupational pension plans are often higher than the legal minimum. As a result, employees typically accrue higher pension entitlements than those reflected in the theoretical replacement rates used in this analysis. These additional contributions vary across pension funds and are therefore difficult to model.
← 5. In Belgium, the complementary pension system is generally classified as voluntary, even though participation is mandatory for some employees, such as those covered by industry-wide pension plans.
← 6. Among members aged 55 and older, 19% to 20% contributed at 4% and this could be because 4% was the default rate up to April 2009.
← 7. Pension contributions in some countries also finance disability or invalidity benefits, as well as other social security benefits.
← 8. In New Zealand, the government makes a matching contribution equivalent to the employer contribution for recipients of paid parental leave (PPL) who elect to continue KiwiSaver deductions from their PPL payments, since 1 July 2024.
← 9. The newly introduced retirement savings account (Altersvorsorgedepot) without mandatory guarantee aims to overcome the shortcomings of Riester plans and make the tax-subsidised pension system more attractive by offering greater flexibility, lower costs and improved long-term return potential.
← 10. To the extent possible, the analysis relies on empirical studies in the literature to assess the effectiveness of the different policies on contribution levels.
← 11. In the case of employees with an annual gross remuneration higher than EUR 27 000, the tax deduction applies only to the proportional part. The same deduction applies to self-employed workers with employees (employers who are liable for personal income tax).
← 12. The experiment allowed individuals to select an initial contribution rate, a rate of increase in their contribution rate (0.5, 1, 1.5 or 2 p.p.), an increase period (yearly or every second year) and a maximum contribution rate (from 3.5% to 10%).
← 13. In the United States, the catch-up provision raises the contribution limit for individuals aged 50 and over.
← 14. In a plan with upfront taxation, contributions are not tax deductible, while investment income and pension benefits are tax free. By contrast, in a plan with deferred taxation, contributions are tax deductible, investment income is tax free and pension benefits are taxed.
← 15. From 2027, the Saver’s Credit will be replaced by the Saver’s Match, a government matching contribution.
← 16. The maximum entitlement is phased out at a constant taper rate to zero between the lower- and upper-income thresholds. Therefore, individuals with an income above the upper-income threshold are not entitled to the government co-contribution.
← 17. The parameters of the co-contribution programme in Australia have changed several times since its inception. In particular, the match rate has varied, taking values of 50%, 100% and 150% depending on the years. As the maximum individual contribution matched by the government has remained at AUD 1 000 the whole time, the maximum entitlement has been equal to AUD 500, AUD 1 000 or AUD 1 500 depending on the years.
← 18. Higher-income earners still have an incentive to contribute beyond the match threshold because of the possibility to deduct contributions from taxable income in the United States.
← 19. Single‑tier match formulas may still be preferable to tax incentives when the objective is to encourage retirement savings from different income groups, as low-income earners with no tax liability can still benefit from matching contributions, while they do not benefit from tax incentives.
← 20. For example, consider a high-income employee who can contribute 8% of salary and a low-income employee who can only afford a contribution rate of 4%. With a 50% match rate up to 8% of salary, the high-income employee would receive the maximum employer matching contribution of 4% of salary, while the low-income employee would only receive an employer matching contribution of 2% of salary. With a 100% match rate on the first 3% of salary and 50% on the next 2% of salary, the high-income employee would still receive the maximum employer match of 4%, while the low-income employee would now receive 3.5% of salary from the employer.
← 21. The European Commission uses the term pension tracking services to refer to pension dashboards.