Key Takeaways

• While opt-out pension systems can be effective in balancing individual freedom with retirement security, Estonia’s 2021 reform revealed the risks of taking this design too far. Within five years of allowing full withdrawal from its mandatory second pension pillar, 37% of participants had exited, creating a two-track retirement system with significant long-term consequences.

• Withdrawal decisions were driven more by low institutional trust and political framing than by genuine financial necessity. Half of withdrawn funds sat idle in bank deposit accounts a year after the reform rather than being reinvested, underscoring that exit decisions for many were largely psychological rather than economic.

• The Estonian case study serves as a cautionary tale for policymakers globally: even with design features intended to limit exits, opt-out reforms that allow unrestricted early withdrawal can rapidly undo the benefits of automatic enrollment, with disproportionate harm to the most financially vulnerable.

Watch webinar recording here.

Summary

Motivation: Auto-enrollment has become an increasingly popular feature in pension systems, helping to encourage long-term saving while preserving the freedom to opt out. In 2021, Estonia tested the limits of this design when a politically-driven and contentious reform converted its mandatory second pension pillar into an opt-out system, allowing participants to withdraw their entire accumulated savings at any age, with no partial withdrawals permitted and a 10-year re-entry ban. The reform’s aftermath offers real-world insights into how pension design, political messaging, and institutional trust interact to shape retirement savings behaviour at scale.

Methodology: The research draws on data from the Estonian Central Bank and Ministry of Finance to examine the characteristics of withdrawers, such as income, gender, education, family size, and trust levels, as well as the subsequent uses of withdrawn funds. The author employs correlational analysis to identify key predictors of withdrawal behaviour, and compares Estonia’s case with parallel reforms in Lithuania, the UK pension freedom reforms, and 401(k) leakage patterns in the US.

Findings:

• Approximately 20% of second pillar participants withdrew in the first window in 2021, with cumulative withdrawals reaching 37% over five years. The first-wave total of €1.3 billion, equivalent to 4.5% of GDP, contributed an estimated 1-2% of additional inflation.

• Withdrawn funds were largely not reinvested productively: 50% sat in regular bank accounts a year after the reform, 30% went to loan repayments, and 15% was spent on consumption. Among younger withdrawers specifically, a further 9% was directed towards gambling.

• Low institutional trust was the strongest predictor of early withdrawal, surpassing financial necessity. Trust in the financial industry, media, and the public sector was correlated with the likelihood of exit, particularly in the large first wave.

• Withdrawal rates varied sharply by demographics, with 52% of women with basic education withdrawing compared to 24% of men with higher education. Given Estonia’s gender pay gap and women’s nine-year longer life expectancy, this asymmetry will translate into worse retirement outcomes for women.

• While the reform was framed as returning money to individuals, the state also benefited fiscally, generating €450 million in income tax on withdrawals. The 4% social tax contribution of every participant who exited, estimated at €200-250 million per year, was redirected to fund current pensions.

• Increased competition following the reform has driven down fees across the system, and nearly 100,000 participants have since increased their voluntary contribution to 4% or 6%, underscoring that flexibility can also encourage more engagement with retirement saving.

• Lithuania introduced a similar reform in early 2025 and again saw approximately 37% of participants withdraw within three months. US 401(k) plans face a structurally similar leakage problem when participants change employers.

Q&A Highlights:

• All-or-nothing design: The requirement to withdraw the full accumulated balance with a 10-year re-entry ban ended up punishing those most financially pressured to leave, while doing little to address the risk of higher earners extracting social tax contributions.

• Role of COVID: The reform was not framed as a COVID response, but withdrawals were higher among those who lost income or credit access during the pandemic, suggesting liquidity-driven exits for a subset of participants.

• What a better reform might look like: Retaining a mandatory second pillar pension locked until retirement under strong state oversight, while reframing the third pillar as a flexible, tax-incentivized long-term saving product would separate retirement security from accessible liquidity.

• Future regret: Qualitative evidence suggests regret is already present among some who withdrew, but the full consequences will likely not materialize until retirement, when the foregone compound growth becomes visible by comparison with peers who remained in the system.