This policy note is based on “Saving for retirement is hard – but the EU can make it easier”, CEPS. The views expressed are those of the authors and not necessarily those of the institutions the authors are affiliated with.
Abstract
Europe’s retirement savings gap cannot be closed by tax incentives and information campaigns alone. Differences in household wealth across Member States are largely driven by structural factors (e.g. labour market conditions, pension system design, implicit state guarantees) which shape responses to policy tools. Tax incentives tend to raise savings mainly among older, higher-income and already engaged savers, while doing little for younger and lower-income groups, and fragmented national tax regimes undermine portability and scale. Automatic enrolment and defaults sharply increase participation in the short run, but effects weaken when liquidity needs or job changes allow balances to be cashed out. Evidence highlights strong demand for liquidity and the effectiveness of conditional access combined with matching incentives. Financial education has modest, short-lived effects, whereas simple digital tools and ‘future self’ interventions can meaningfully boost engagement. A shift toward behaviourally informed, equity-focused policies – centered on occupational pensions, smart defaults, conditional liquidity and better-targeted fiscal incentives – is therefore essential to broaden participation, improve adequacy and support the EU’s Savings and Investments Union agenda.
Differences in household wealth across EU Member States are large and cannot be explained by behavioural factors alone. Structural labour market conditions are central. Countries with higher wages, stable employment and limited informality naturally create more room for long-term saving, while fragmented or precarious labour markets with high income volatility tighten liquidity constraints and make households less willing or able to commit to illiquid retirement products. Pension system design also matters. Where pay-as-you-go schemes are generous and comprehensive, expectations of adequate state pensions reduce the need to accumulate private financial wealth. Where public pensions replace only a modest share of pre-retirement income, households are more likely to build occupational or personal pension savings. Expectations of state support can create implicit moral hazard when people believe government will ultimately guarantee a minimum retirement income even if private saving is low.
Cultural attitudes and preferences further shape how households save and invest. Trust in financial institutions, familiarity with capital markets and preferences for housing versus financial assets differ markedly across the EU. In many Southern European countries, housing remains the main store of wealth, while in Northern Europe financial saving and participation in occupational pensions play a larger role. These patterns show that behavioural frictions such as inertia or present bias interact with deeper institutional and cultural structures. EU policy must therefore operate on both fronts: improving incentives and defaults for retirement saving while also addressing structural barriers and trust deficits that limit private wealth accumulation.
Undersaving for retirement is a growing concern across advanced economies. In the US, an estimated 39% of working-age households are unlikely to maintain their standard of living in retirement (Yin, Chen, and Munnell 2024). In the EU, almost 20% of people aged 65 and older are at risk of poverty or social exclusion, with women disproportionately affected (Eurostat 2025). Researchers have long described this as a ‘retirement savings crisis’ (Benartzi and Thaler 2013). In response, governments have deployed tax incentives, automatic enrolment and retirement education, with mixed results.
Encouraging retirement saving is a strategic priority for the EU as ageing populations strain pay-as-you-go systems. Tax incentives are widely used to promote voluntary private saving, but their design and effectiveness differ sharply across Member States, raising questions about efficiency, distributional impact and behavioural effectiveness (Beshears et al. 2017; Horneff, Maurer, and Mitchell 2023; Szapiro 2024). Individuals often struggle to estimate future needs, value long-term benefits or act on intentions, leading to chronic under-saving. These frictions are reinforced by complex products, uneven financial advice and large differences in income and job security. From a Capital Markets Union (CMU) perspective, well-designed incentives can mobilise long-term capital, support inclusion and foster retail markets. Within the Savings and Investments Union (SIU), they also help reduce fragmentation and make individual products more attractive. Yet the wide diversity of tax treatments, combined with uncertain behavioural responses, still constrains the EU’s ability to scale up retirement saving in a coherent and inclusive way.
Countries differ widely in how they tax retirement savings (see Figure 1). A common benchmark is the EET system (Exempt–Exempt–Taxed), where contributions and investment income are exempt and withdrawals are taxed. Other systems reverse this sequence or use mixed combinations, altering both the overall tax advantage and how savers respond. In the US, for example, 401(k) plans and traditional Individual Retirement Accounts (IRAs) follow the EET model, while Roth IRAs apply a TEE model (Taxed–Exempt–Exempt), with contributions taxed upfront and withdrawals exempt. These design choices matter for take-up and for who ultimately benefits.
Figure 1. Tax treatment of retirement savings in OECD countries in 2023
Within the EU, Member States use a broad range of tax treatments for voluntary pension products, from pure EET to TEE, ETE and other hybrids. Even under the same nominal model, there are important differences in the generosity of relief, ceilings on deductible contributions and the taxation of investment income and payouts. In some countries withdrawals are taxed at full income tax rates, as in Germany and France where pension benefits are generally treated as regular income. In others only capital gains are taxed, as in Italy where some withdrawals are taxed solely on accrued returns. Some Member States apply reduced pensioner rates, such as Finland’s preferential treatment for retirement income. Investment income inside pension accounts is fully exempt in the Netherlands and Sweden, partially taxed in Austria, and in a few cases subject to social contributions. Belgium, for instance, may levy solidarity contributions on larger pension assets.
This heterogeneity has two main consequences. It generates large differences in the net-of-tax return on retirement saving, influencing participation and contribution levels in ways that may not align with policy goals. It also weakens portability and comparability across borders and complicates the development of scalable pan-European products such as the Pan-European Personal Pension Product (PEPP). Without some degree of tax coordination or mutual recognition, products like PEPP face a ‘lowest common denominator’ problem, since their attractiveness is constrained by the least generous national tax regime they encounter.
French experience offers a clear test of how tax incentives shape saving behaviour. The Loi Pacte, in force since 2019, introduced a tax deduction for voluntary contributions to employer-sponsored retirement plans. Using data on nearly 1.4 million workers across 2,679 firms, Brière, Poterba, and Szafarz (2025) show that the reform generated a modest but statistically significant rise in long-term contributions, of about EUR 150 per person per year on average, without reducing medium-term saving. This indicates limited substitution and a genuine net increase in retirement saving.
The effects, however, were highly uneven. Higher-income workers, those with larger pre-existing balances and those closer to retirement were much more likely to use the new deduction, while younger and lower-income workers reacted only weakly (see Figure 2). This mirrors international evidence: voluntary tax incentives are mainly taken up by those already inclined to save, whereas more passive savers remain disengaged. Behavioural mechanisms help explain this pattern. Time-inconsistent individuals, who procrastinate or heavily discount future costs, may prefer pre-tax contributions because the benefit is immediate and the future tax burden is psychologically distant (Bohr, Holt, and Schubert 2023). Misperceptions about pension taxation can further blunt the impact of tax ‘nudges’ (Blaufus and Milde 2021).
Figure 2. Pre-tax voluntary contributions by wealth- and age-based quartiles
The French reform is useful because it targeted both employees and small firms. Employers with fewer than 250 workers received tax advantages for contributing directly to their employees’ retirement plans, in addition to the new deduction for employees. This dual design highlights an important distinction also visible in international evidence. Studies from Denmark show that when pension tax rules changed around 1999-2000, most people (81%) behaved as passive savers and did not adjust contributions, while only a small minority (19%) responded, largely by shifting money between products rather than increasing overall saving. Later reforms in 2009, which capped deductibility for some short-term products but preserved it for annuities, again changed the composition of savings without raising total contributions. Taken together, these findings suggest that voluntary tax incentives mainly reshuffle portfolios among active savers rather than generate substantial new saving across the population.1
By contrast, automatic employer contributions can raise accumulated retirement wealth even when employees do nothing. Evidence from Denmark indicates that passive defaults set at the employer level are much more powerful than individual tax incentives alone (Chetty et al. 2014). When tax breaks primarily benefit ‘active’ savers who are already inclined to save, and passive savers remain disengaged, the overall effect is modest and often skewed toward higher earners. This raises broader questions about effectiveness and fairness. If a large share of the fiscal cost accrues to a relatively small, high-income and financially literate group that would mostly have saved anyway, the value for money of current approaches is doubtful. A more strategic use of public resources would focus on tools that create genuinely new saving and reach under-served groups, combining behavioural targeting with fiscal efficiency so that public spending on incentives translates into meaningful increases in long-term savings across a wider share of the population.
Many people do not save for retirement even when tax incentives and suitable products exist. Inertia often reflects behavioural frictions rather than lack of interest. Procrastination, difficulty handling complex choices and a strong focus on the present make it hard to engage with long-term planning. These biases are particularly acute for retirement, where horizons are long, outcomes uncertain and information complex. Automatic enrolment responds by changing the default. Instead of asking individuals to opt in, they are enrolled in a plan at a preset contribution rate and must actively opt out. Behavioural research shows that such defaults substantially increase participation, especially among passive savers, and do so at relatively low cost. Evidence from the US and the UK indicates that when schemes move from voluntary sign-up to automatic enrolment, participation in workplace plans almost doubles (Madrian and Shea 2001; Beshears et al. 2015; Berk et al. 2025), which explains why many reforms now include default mechanisms.
Over time, however, part of these gains erodes as defaults interact with liquidity needs and job mobility. In the US, workers who leave a job often transfer 401(k) assets into more flexible IRA accounts and then take partial or full withdrawals. In the UK, participation falls by around 13% when employees change employer and are not automatically re-enrolled, and after about three years cumulative contributions of auto-enrolled and voluntarily enrolled workers tend to converge. French employee savings schemes offer similar lessons. In many firms, 50% of the employer’s profit-sharing contribution is automatically invested in a long-term retirement account unless the employee opts out, which boosts long-term saving initially. Yet hardship provisions later allow substantial drawdowns. Data from Amundi show that under hardship conditions employees withdraw on average 92% of long-term (LT) balances compared with 68% of medium-term (MT) balances with a five-year lock-in (Brière, Poterba, and Szafarz 2022) (see Figure 3). In US 401(k) plans, early withdrawals and opt-outs from automatic escalation are also common, and many workers cash out before employer contributions are fully vested (Choi et al. 2024). Defaults thus remain powerful entry points, but their lasting effect depends on how they are combined with liquidity rules, re-enrolment practices and safeguards against ‘leakage’.
Figure 3. Share of account balance withdrawn from Medium-Term (MT) and Long-Term (LT)
retirement accounts when hardship occurs, by participant’s account composition
Similar patterns emerge in US 401(k) plans. Early withdrawals are common and many employees opt out of automatic contribution escalation. A significant share of workers cash out their balances before employer contributions fully vest, which undermines the build-up of retirement wealth and reduces the lasting impact of default-based interventions (Choi et al. 2024).
A core design choice in retirement saving is how much access people have to their money before retirement. Strict lock-in supports long-term accumulation, but if savings are too hard to access many, especially younger workers or those with volatile incomes and frequent shocks, will not participate at all. This trade-off sits at the heart of the CMU and SIU agendas, which aim to mobilise long-term capital while broadening participation across income groups and life stages. Growing experience suggests liquidity should not be treated as a binary choice. Instead of products being either fully illiquid or fully flexible, access can be structured along a spectrum, with carefully calibrated rules and incentives that reward long-term commitment while keeping some room for adjustment.
Countries implement this balance in different ways. Many systems restrict withdrawals to specific events such as disability, serious illness, unemployment or over-indebtedness, and some also allow early access for a first-home purchase. In the US, savers can roll 401(k) balances into more liquid IRAs when changing jobs, although early withdrawals before age 59½ typically incur tax penalties. In France, public pensions in pay-as-you-go schemes are inherently illiquid, while occupational products such as the PER (Plan d’Épargne Retraite) permit early withdrawals in narrowly defined cases, including home purchase, unemployment or over-indebtedness. This diversity shows there is no single ‘correct’ model and highlights the need to match liquidity design to savers’ economic realities and behavioural profiles. If access is too tight, participation and contributions may stay low; if it is too loose, accumulated savings risk being depleted well before retirement.
French employee savings schemes provide clear evidence of a strong preference for liquidity (Brière, Poterba, and Szafarz 2022). When employers direct profit-sharing into long-term (LT) retirement products by default, only 9% of employees remain in the LT option. When the default is instead a medium-term (MT) product with a five-year lock-in, 41% accept the default (see Figure 4). This gap indicates that perceived accessibility is a major determinant of participation.
Figure 4. Share of workers taking up the saving plan and the default option,
in firms with and without long-term (LT) retirement saving options
Incentives can partly counter this preference. Employees are far more willing to choose LT plans when contributions receive a higher employer match than MT plans. LT take-up is 38% when the LT match is lower than or equal to the MT match, rises to 63% when the LT match is higher but less than double, and reaches 72% when it is more than twice as generous. Revealed preferences tell a similar story. Among 150,000 employees who could freely split contributions between MT and LT vehicles, only 35% invested in LT products without matching, compared with 69% when matching was offered. Even then, the average share of contributions going to LT products was about 21%, well below the 37% implied by the default. These patterns point to limited intrinsic demand for illiquidity unless strong incentives are present. This is not simply short-termism. For many, especially younger or lower-income workers facing income risk, reluctance to lock in savings reflects rational caution. Policy efforts to increase long-term saving must therefore accommodate this need for flexibility rather than assume it away.
A growing number of jurisdictions are testing liquidity options that sit between full lock-in and free access. In the United States, the SECURE 2.0 Act of 2022 introduced penalty-free emergency withdrawals from 401(k) plans. Early evidence suggests that most users of this feature continue to contribute afterwards, indicating that limited access need not trigger long-term disengagement. French experience points in the same direction: savers often draw on long-term accounts in hardship but resume contributions when incentives remain attractive.
The broader lesson is that liquidity can be part of the solution rather than only a threat to retirement adequacy. Carefully designed partial, conditional or time-limited withdrawals can provide flexibility without undermining the core purpose of long-term saving. Smart defaults, targeted matching and light ‘frictions’ such as short waiting periods or caps on withdrawals can preserve discipline while still accommodating
Behavioural design can raise participation and well-calibrated liquidity can support commitment, but neither is sufficient if products fail to deliver fair, transparent and sustainable outcomes. Saving for retirement means placing assets with intermediaries for decades, which brings core questions of prudential supervision, access to redress and long-term adequacy to the forefront.
Retirement products are currently governed by a patchwork of EU and national rules. The Institutions for Occupational Retirement Provision (IORP II) Directive sets prudential standards for occupational pension funds. The Markets in Financial Instruments Directive (MiFID II) regulates many investment products. The PEPP Regulation adds a portable, cross-border personal pension framework. Important gaps remain nonetheless, especially for cross-border providers and hybrid products that straddle regimes. Fragmented supervisory practices can result in uneven consumer protection and divergent levels of prudential control. Stronger EU-level coordination, potentially through an enhanced role for the European Insurance and Occupational Pensions Authority (EIOPA) and the European Securities and Markets Authority (ESMA), is needed so that all retirement products face robust oversight of governance, risk management and cost structures.
Although market risk ultimately lies with savers, investors must have effective channels for redress when losses stem from misconduct, excessive fees or unsuitable advice. Today most Member States rely on national ombudsmen, Alternative Dispute Resolution (ADR) schemes or the courts, and there is no mechanism specifically tailored to cross-border cases. This gap is especially problematic where the SIU and PEPP seek to expand cross-border provision. A European framework for collective redress or mandatory participation in EU-level ADR schemes could strengthen trust and help correct information and bargaining imbalances.
Prudential oversight is also critical for long-term performance. Very conservative investment rules, high charges or poorly designed liquidity features can erode returns as effectively as early withdrawals. A product that preserves nominal capital but consistently fails to keep pace with inflation can damage retirement security as much as one exposed to excessive volatility. Policymakers therefore need to calibrate safeguards so that they protect savers without preventing investment in productive, long-term assets. This balance between risk control and return is central to the broader EU capital markets agenda, where retirement savings should support both individual adequacy and real-economy investment.
Efforts to boost retirement saving through information campaigns and financial education have a long history, but results are mixed. Many national and EU initiatives rest on the assumption that better-informed individuals will make better financial choices, yet the evidence is more nuanced. Some measures deliver modest gains. In Germany, annual letters projecting future pension entitlements were linked to higher voluntary contributions and even increased labour supply (Dolls and Krolage 2019). In the US, providing income projections or small incentives to attend seminars raised participation and contributions slightly, particularly when peers were also exposed (Duflo and Saez 2003; Goda, Manchester, and Sojourner 2014).
These effects, however, are often context-specific and short-lived. Recent experiments show that peer comparisons or gamified incentives, such as lottery entries for checking pension information, can trigger short-term engagement but rarely produce lasting changes in saving behaviour (Bauer et al. 2022). People may become better informed yet still delay decisions or fail to act on intentions. Overall, financial education has only a small effect: one meta-analysis finds that literacy programmes explain about 0.1% of the variation in financial behaviours, with especially weak effects for lower-income groups (Fernandes, Lynch, and Netemeyer 2014). This reflects several behavioural mechanisms. Individuals often avoid information that may cause anxiety (the ‘ostrich effect’), forget content over time or revert to old habits without reinforcement (Karlsson, Loewenstein, and Seppi 2009; Sicherman et al. 2015). Time preferences, optimism bias and numeracy further shape how information is processed. Taken together, these findings suggest that traditional, information-heavy campaigns are unlikely to close retirement saving gaps, particularly for disengaged or vulnerable groups. A more promising approach must prioritise emotional salience, simplicity and ease of action rather than knowledge alone.
Given the limits of standard education programmes, behavioural research is shifting toward more immersive and emotionally engaging tools. Visual and interactive technologies are a key avenue. Experiments show that allowing individuals to interact with realistic digital avatars of their future selves increases their willingness to allocate money to long-term savings (Hershfield et al. 2011). The core insight is that people are more likely to save when they can emotionally connect with their future identity instead of viewing retirement as a distant abstraction.
Simple digital tools can also make a measurable difference. A recent study finds that offering a user-friendly pension app raises the share of people making voluntary contributions by 1.8 percentage points from a 2.8% baseline, mainly because the app makes contributing easier and less time-consuming rather than by providing more tax information (Daminato, Filippini, and Haufler 2024).2 When such tools are combined with pension simulators or robo-advisory interfaces, they make retirement planning more personally relevant and can support better investment decisions (Bianchi and Brière 2024). As digitalisation advances, these technologies could be embedded in EU initiatives such as PEPP, national pension tracking systems and financial literacy strategies, in line with the Retail Investment Strategy (RIS), which promotes interactive disclosures and digital tools to improve investor decision-making.
Digital tools and immersive interfaces are widening access to retirement planning, but traditional financial advisors remain a central – and underused – channel. More than 80% of financial decisions are made with some form of advice, yet many advisors are poorly prepared for retirement planning. In the US, for example, almost half of financial advisors (46%) reportedly lack a retirement plan of their own (Gladych 2013), which underscores the weaknesses of the current advisory ecosystem.
There is therefore ample room to improve both the coverage and quality of advice. Stronger training, the integration of pension-planning software into day-to-day practice and closer supervisory scrutiny could significantly raise the standard of guidance. Giving advisors access to up-to-date tools and personalised simulations would help them build structured decision environments that counter inertia and procrastination and nudge clients toward long-term saving. At EU level, ESMA and EIOPA could support this shift by issuing guidance on embedding behavioural insights into advice, including the use of default retirement products, standardised protocols for assessing retirement needs and simplified decision trees for long-term savings choices.
On November 2025, the European Commission tabled a proposal to amend the PEPP Regulation as part of a broader pensions package under the SIU (European Commission 2025). It seeks to address the very problems highlighted in the early PEPP experience and in recent assessments by EIOPA and the European Court of Auditors, namely low uptake, limited cross-border use and weak value for money. The aim is to make PEPP a more attractive and scalable third-pillar product by removing some of the rigidities of the original framework, strengthening value-for-money supervision and integrating PEPP more closely into workplace and digital infrastructures.
The review strengthens tax treatment by requiring Member States to give PEPP at least the most favourable tax relief available to comparable national products, it replaces the mechanical 1% fee cap with a value-for-money approach backed by better cost and performance data, and makes it easier to use PEPP in workplace settings, including in auto-enrolment schemes where national law allows this. It also reinforces data provision to pension tracking systems and EIOPA’s central register, which can support more transparent and digital engagement with savers.
Despite these steps, the review remains incremental and leaves key behavioural and structural levers barely untouched. It does not address the deeper fragmentation of national tax regimes, nor the regressive pattern whereby tax-advantaged saving is concentrated among older, higher-income and already engaged savers. The non-discrimination rule improves PEPP’s relative position but does not move the system towards flatter, more salient incentives or targeted support for liquidity-constrained households. Likewise, while the new framework makes workplace PEPPs possible, it does not create any EU-level expectations on auto-enrolment, re-enrolment or contribution escalation. The most powerful defaults therefore remain dependent on national second-pillar reforms and social-partner choices rather than on the PEPP architecture itself.
The proposal is also largely silent on conditional liquidity and on the behavioural design of engagement tools. Early access rules for PEPP continue to be driven mainly by national law, despite clear evidence that calibrated, needs-based liquidity is crucial for participation among younger and lower-income savers. The strengthened disclosure and tracking infrastructure could underpin simple simulators, apps and ‘future self’ tools, but their use is not mandated or incentivised. In that sense, the review improves a third-pillar product that will matter for mobile and financially engaged savers, yet it does not change the basic policy balance argued for in this article: a strategy that puts occupational pensions and well-designed defaults at the centre, and uses personal products like PEPP as a complementary layer within a behaviourally informed, inclusive retirement saving architecture.
The analysis points to a central conclusion: raising retirement savings in the EU requires more than tax breaks or information campaigns. It calls for a behaviourally informed, inclusive policy framework that supports the goals of the CMU and the SIU. Five priorities follow.
Tax incentives remain central in national strategies, but their impact is often modest and skewed toward older, higher-income and already engaged savers, as the Loi Pacte experience illustrates. To improve both effectiveness and equity, Member States should shift from marginal-rate deductions toward flat-rate credits and employer-based matching, which are more salient for lower and middle incomes and more likely to generate genuinely new saving.
Four principles should guide reform: use collective channels (employers, sectoral schemes) rather than rely solely on individual action; ensure incentives do not disproportionately benefit higher earners; assess measures by their effect on net additional saving, not just product switching; and reduce fragmentation in national tax treatment, which undermines comparability and cross-border pensions. EU institutions can support this by issuing guidance on effective incentive design and encouraging gradual convergence.
Automatic enrolment and default contributions are powerful tools against inertia, but their long-term impact depends on how they interact with job mobility, liquidity needs and national pension architectures. Defaults should therefore be flexible and context-specific, with features such as vesting thresholds, automatic re-enrolment after job changes and gradual contribution escalation. Liquidity rules must be transparent and calibrated to prevent early erosion of savings.
At EU level, non-binding guidance and best-practice templates on enrolment, default fund design and contribution escalation could help Member States adapt these tools to local systems, while monitoring long-term outcomes and ensuring that portability gaps or access rules do not undermine defaults. ESMA and EIOPA are well placed to coordinate such guidance.
Loss of access to funds is a key deterrent to long-term saving, especially for younger and lower-income households, yet full flexibility conflicts with the purpose of retirement products. Conditional liquidity offers a middle path. Allowing partial or needs-based access in clearly defined circumstances (hardship, unemployment, major life events), combined with modest frictions such as waiting periods or withdrawal caps, can lower entry barriers while preserving the long-term nature of savings.
For products like PEPP, carefully calibrated access rules and liquidity-linked matching (for example, ‘liquidity ladder’ structures that reward longer holding periods) could help overcome fears of illiquidity and support both participation and commitment.
Information-heavy financial education on its own has limited and short-lived effects, particularly for disengaged or vulnerable groups. Behavioural tools such as pension simulators, robo-advice, ‘future self’ visualisations and intuitive digital platforms have shown greater promise in driving participation and contribution decisions. These should be scaled across the EU by embedding interactive planning tools into product interfaces and evolving national pension dashboards from static information sources into simulation engines. The Commission could back cross-border pilots and experiments to test and refine such approaches.
Financial advice remains underused as a lever for change. Many advisors lack the training, tools and protocols needed to support long-term retirement planning. Regulatory frameworks such as MiFID II and the Retail Investment Strategy should be adjusted to require retirement-specific advisory processes, promote the integration of digital planning tools and encourage simplified decision trees for long-term saving. ESMA and EIOPA could jointly develop guidance on embedding behavioural insights into advisory standards and licensing.
The greatest scope for expanding coverage and adequacy lies in occupational pensions, not in third-pillar products alone. Second-pillar schemes linked to employment are where defaults, automatic enrolment and collective arrangements can be deployed at scale, and where EU-level action can most effectively interact with labour mobility and the single market.
Building on IORP II, policy efforts should focus on: improving portability so that workers moving across Member States can transfer or consolidate entitlements without prohibitive costs; setting minimum funding and prudential standards that protect adequacy while allowing investment in productive assets; and supporting collective or sectoral funds that extend coverage to SMEs and non-standard workers. PEPP and other third-pillar products remain valuable, especially for engaged and mobile savers, but are likely to be niche compared with workplace schemes. A credible SIU strategy should therefore put occupational pensions at its core, aligning pension policy with labour market integration, capital market development and financial inclusion.
Fragmentation in tax rules, liquidity provisions, default designs and advice delivery continues to weaken participation and cross-border uptake. A coherent EU approach is needed that respects national diversity while establishing minimum structural and behavioural standards.
The PEPP reform now on the table makes progress on tax non-discrimination, value-for-money supervision and data infrastructure, but still treats personal pensions mainly as a niche third-pillar product for already engaged savers. Co-legislators should use the review to go further on behavioural design, fiscal attractiveness and transparency, including more salient support for lower- and middle-income households and clearer rules on conditional liquidity. Without movement in these areas, uptake is likely to remain limited.
More broadly, the SIU framework should be used to align national retirement saving strategies with EU-level goals: mobilising long-term capital, improving financial inclusion and enhancing individual resilience. The objective is not uniformity for its own sake, but an enabling ecosystem in which individuals can plan, save and invest for retirement with confidence, flexibility and trust in the system.
The EU’s retirement saving challenge cannot be solved by fiscal incentives or information campaigns alone. Tax deductions, default settings and education each matter, but they mostly reach the already engaged, while the passive majority remain constrained by inertia, uncertainty and liquidity needs. Evidence from France and other countries shows that even well-designed reforms have limited and uneven effects unless they are aligned with the behavioural realities of savers.
Closing the retirement savings gap requires a policy architecture that reflects how people actually decide. This means shifting from opt-in incentives to opt-out defaults, from rigid lock-in to conditional liquidity, and from abstract financial literacy to emotionally resonant, tech-enabled tools. It also means sharing responsibility: employers, intermediaries and public institutions must help create environments in which long-term saving becomes the path of least resistance. At the same time, behavioural design must be underpinned by robust prudential safeguards. Savers need products that are well-governed, fairly priced and subject to effective oversight and redress, and adequacy depends as much on long-term returns as on participation. Within this framework, occupational pensions are central. Second-pillar schemes embedded in the workplace remain the most powerful lever for expanding coverage, portability and adequacy, with PEPP and other third-pillar products playing a complementary role.
The PEPP review now underway, together with the broader SIU agenda, offers a timely opportunity to move in this direction. Their success will depend on whether co-legislators use the current reform window to tackle the structural and behavioural frictions identified in this study: fragmented tax rules, poorly calibrated defaults, inflexible liquidity provisions and weak engagement infrastructure. Addressing these barriers would not only improve individual retirement outcomes but also advance the EU’s wider objectives of capital market development, financial inclusion and economic resilience. The goal is not uniformity for its own sake, but a common behavioural and institutional foundation that enables all Europeans to save for retirement in a way that is fair, flexible and effective.
Bauer, J., S. Aarestrup, P. Hansen, and L. Reisch. 2022. “Nudging More Sustainable Grocery Purchases: Behavioural Innovations in a Supermarket Setting.” Technological Forecasting and Social Change 179 (June): 121605. https://www.sciencedirect.com/science/article/abs/pii/S0040162522001378.
Benartzi, S., and R. Thaler. 2013. “Behavioral Economics and the Retirement Savings Crisis.” Science 339 (6124): 1152-53. https://doi.org/10.1126/science.1231320.
Berk, S., J. Choi, J. Garg, J. Beshears, and D. Laibson. 2025. “Automatic Short-term Payroll Savings: Evidence from Four Large U.K. Experiments.” NBER Working Paper Series 32581. National Bureau of Economic Research. https://www.nber.org/system/files/working_papers/w32581/w32581.pdf.
Beshears, J., J. Choi, D. Laibson, B. Madrian, and K. Milkman. 2015. “The Effect of Providing Peer Information on Retirement Savings Decisions.” Journal of Finance 70 (3): 1161-1201. https://onlinelibrary.wiley.com/doi/abs/10.1111/jofi.12258.
Beshears, J., J. Choi, D. Laibson, and B. Madrian. 2017. “Does Front-Loading Taxation Increase Savings? Evidence from Roth 401(K) Introductions.” Journal of Public Economics 151 (July): 84-95. https://doi.org/10.1016/j.jpubeco.2015.09.007.
Bianchi, M., and M. Brière. 2024. “Human-Robot Interactions in Investment Decisions.” Management Science 0 (0). https://doi.org/10.1287/mnsc.2022.03886.
Blaufus, K., and M. Milde. 2020. “Tax Misperceptions and the Effect of Informational Tax Nudges on Retirement Savings.” Management Science 67 (8): 4643-5300. https://pubsonline.informs.org/doi/10.1287/mnsc.2020.3761.
Bohr, C., C. Holt, and A. Schubert. 2023. “A Behavioral Study of Roth versus Traditional Retirement Savings Accounts.” Department of Economics Working Paper Series 440. University of Zurich. https://www.econstor.eu/bitstream/10419/280776/1/1874532265.pdf.
Briere, M., J. Poterba, and A. Szafarz. 2022. “Precautionary Liquidity and Retirement Saving.” AEA Papers and Proceedings 112 (May): 147-150. https://www.aeaweb.org/articles?id=10.1257/pandp.20221022.
Briere, 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. Amundi. https://research-center.amundi.com/article/does-tax-deductibility-increase-retirement-saving-lessons-french-natural-experiment.
Chetty, R., J. Friedman, S. Leth-Petersen, T. H. Nielsen, and T. Olsen. 2014. “Active vs. Passive Decisions and Crowd-out in Retirement Savings Accounts.” The Quarterly Journal of Economics 129 (3): 1141-1220. https://www.jstor.org/stable/26372571.
Choi, J., D. Laibson, J. Cammarota, R. Lombardo, and J. Beshears. 2024. “Smaller than We Thought? The Effect of Automatic Savings Policies.” NBER Working Paper Series 32828. National Bureau of Economic Research. https://www.nber.org/papers/w32828.
Daminato, C., M. Filippini, and F. Haufler. 2024. “Digitalization and Retirement Contribution Behavior: Evidence from Administrative Data.” Review of Financial Studies 37 (8): 2510-2549. https://academic.oup.com/rfs/article/37/8/2510/7663471.
Dolls, M., and C. Krolage. 2019. “The Effects of Early Retirement Incentives on Retirement Decisions.” Ifo Working Paper 291. https://www.ifo.de/DocDL/wp-2019-291-dolls-krolage-retirement-incentives-decisions.pdf.
Duflo, E., and E. Saez. 2003. “The Role of Information and Social Interactions in Retirement Plan Decisions: Evidence from a Randomized Experiment.” The Quarterly Journal of Economics 118 (3): 815-842. https://academic.oup.com/qje/article-abstract/118/3/815/1943005.
European Commission. 2025. “Proposal for amending Regulation (EU) 2019/1238 on a pan-European Personal Pension Product (PEPP).” European Commission COM(2025) 840 final. https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52025PC0840.
Eurostat. 2025. “Living Conditions in Europe – Poverty and Social Exclusion.” Eurostat. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Living_conditions_in_Europe_-_poverty_and_social_exclusion.
Fernandes, D., J. Lynch, and R. Netemeyer. 2014. “Financial Literacy, Financial Education, and Downstream Financial Behaviors.” Management Science 60 (8): 1861-83. https://www.jstor.org/stable/42919641.
Gladych, P. 2013. “Almost Half of Financial Advisors Don’t Have Retirement Plans.” Benefitspro.com, December 13, 2013. https://www.benefitspro.com/2013/12/13/almost-half-of-financial-advisors-dont-have-retire/.
Goda, G., C. Manchester, and A. Sojourner. 2014. “What Will My Account Really Be Worth? Experimental Evidence on How Retirement Income Projections Affect Saving.” Journal of Public Economics 119 (November): 80-92. https://www.sciencedirect.com/science/article/abs/pii/S0047272714001819.
Hershfield, H., D. Goldstein, W. Sharpe, J. Fox, L. Yeykelis, L. Carstensen, and J. Bailenson. 2011. “Increasing Saving Behavior through Age-Progressed Renderings of the Future Self.” Journal of Marketing Research 48: S23–37. https://pmc.ncbi.nlm.nih.gov/articles/PMC3949005/pdf/nihms550109.pdf.
Horneff, V., R. Maurer, and O. Mitchell. 2023. “How Would 401(K) ‘Rothification’ Alter Saving, Retirement Security, and Inequality?” Journal of Pension Economics & Finance 22 (3): 265-283. https://doi.org/10.1017/s1474747222000105.
Karlsson, N., G. Loewenstein, and D. Seppi. 2009. “The Ostrich Effect: Selective Attention to Information.” Journal of Risk and Uncertainty 38 (2), 95-115. https://www.cmu.edu/dietrich/sds/docs/loewenstein/OstrichEffect.pdf.
Madrian, B., and D. Shea. 2001. “The Power of Suggestion: Inertia in 401(k) Participation and Savings Behavior.” The Quarterly Journal of Economics 116 (4): 1149-1187. https://www.jstor.org/stable/2696456.
OECD. 2023. “Annual Survey on Financial Incentives for Retirement Savings.” OECD Publishing EBooks. https://www.oecd.org/en/publications/annual-survey-on-financial-incentives-for-retirement-savings_2154cc08-en.html.
Sicherman, N., G. Loewenstein, D. Seppi, and S. Utkus. 2015. “Financial Attention.” Review of Financial Studies 29 (4): 863-97. https://www.cmu.edu/dietrich/sds/docs/loewenstein/FinancialAttention.pdf.
Szapiro, A. 2024. “Is ‘Rothification’ Coming for Your Retirement Account?” Morningstar, December 3, 2024. https://www.morningstar.com/retirement/is-rothification-coming-your-retirement-account.
Yin, Y., A. Chen, and A. Munnell. 2024. “The National Retirement Risk Index: An Update from the 2022 SCF.” Center for Retirement Research Number 24-5. Boston College. https://crr.bc.edu/wp-content/uploads/2024/02/IB_24-5.pdf.
However, the key structural driver of Danish pension savings was the tripartite agreement in 1988/89 between unions, employers and the state, which established occupational pension funds for blue-collar workers. By embedding contributions directly into collective wage agreements, participation became automatic and contribution rates quickly rose above 10% of wages. This institutional shift explains much of Denmark’s high pension coverage today, highlighting the transformative effect of collective defaults compared to voluntary incentives.
The app’s impact was stronger among men and higher-income individuals, indicating that digital tools may also inadvertently widen existing inequalities.
