Editors’ note: This column is based on CEPR Discussion Paper 21645 “Pension income and post-retirement labor supply”. 

Across Europe, governments are simultaneously trying to achieve two objectives that are not always easy to reconcile. Ageing populations and rising fiscal pressures have increased interest in encouraging longer working lives. At the same time, concerns about old-age poverty have led many countries to expand minimum pension provisions for retirees with low pension income. Whether these objectives conflict depends on an empirically open question: how does additional pension income affect labour supply after retirement?

Our answer is that this conflict is real but modest. Higher pension income reduces work in retirement, but effects are more modest than labour supply responses to financial incentives at younger ages, and effects are concentrated among retirees who remain close to the labour market participation margin. 

A large literature has documented that pension generosity affects retirement timing (e.g. Hernæs et al. 2016, Manoli and Weber 2016, Giesecke and Jäger 2021), with higher benefits typically reducing labour force participation at older ages. A smaller literature examines how income shocks affect the labour supply of older individuals more directly. Gelber et al. (2016) and Fetter and Lockwood (2018) find sizable responses to such shocks among the elderly, whereas Johnsen and Willén (2022) find that negative pension shocks leave retirees’ labour market re-entry unchanged. Yet, the behaviour of individuals who have already retired remains poorly understood. To what extent do higher pensions affect whether retirees work at all, and do they affect how much they work when employed? In a recent paper (Kindermann et al. 2026), we study this question using Germany’s Earned Income Pension Credit (EIPC, Grundrente), a reform introduced in 2021 that increased pension income for retirees with long contribution histories but relatively low lifetime earnings.

A pension reform with complex eligibility rules

The EIPC tops up pensions for individuals with at least 33 years of contributions but below-average lifetime earnings. About 12% of pensioners are eligible, and roughly half of those receive the credit after a means test (Krolage et al. 2026). The average annual benefit is around €1,150, corresponding to roughly 9% of average pensions among recipients.

A key feature of the reform is that recipients were unlikely to anticipate the exact size of their benefit. Eligibility depends on detailed earnings histories spanning decades, while final benefits are also affected by household income three years prior to receipt and by annually updated thresholds. Overall, the reform generated a persistent and exogenous increase in pension income that left the marginal return to work unchanged once the claiming decision has been made. Due to the complexity of the formula, the EIPC generated substantial heterogeneity in permanent unearned income changes among retirees with similar employment histories.

We exploit this variation using administrative data on German retirees, comparing recipients with higher EIPCs to those with comparatively lower EIPCs in a difference-in-differences framework. This group assignment is based on average monthly earnings over the course of employment biographies and is therefore determined prior to the reform. 

Pension income reduces post-retirement labour supply

Every additional €100 of pension income reduces labour earnings by roughly €10. By 2024, individuals in our treatment group receive roughly €1,200 more annual pension income than our comparison group of retirees (see Figure 1a). This increase leads to a reduction in annual labour earnings of about €120 (Figure 1b). 

This marginal propensity to earn out of unearned income of -0.1 is relevant, but considerably smaller than many estimates of income effects from the literature, suggesting that retirees are less responsive to income shocks than working-age individuals (e.g. Gelber et al. 2016, Artmann et al. 2023, Giupponi 2024, Golosov et al. 2024). A natural interpretation is that labour supply at older ages is more discretionary: many retirees do not work at all, while most of those who work do so part-time, often in flexible or marginal jobs. As a result, the scope for adjustment is more limited.

Figure 1 EIPC effect on unconditional labour earnings

Notes: The figure displays difference-in-differences estimates contrasting individuals with higher and lower EIPC receipt. Panel (a) presents effects on annual pension income, confirming post-reform differences in pension income across groups, while Panel (b) shows the effect on unconditional labour earnings past retirement. The red dotted line marks the timing of the reform; capped vertical lines indicate 95% confidence intervals.

Most responses come from participation decisions, but earnings adjustments also play a role

A key advantage of our setting is that we can separate two margins of adjustment: whether retirees work at all (extensive margin), and how much they work conditional on employment (intensive margin). About two-thirds of the total earnings response comes from changes in participation, averaging €80. The remaining one-third reflects changes in earnings among those who remain employed, with full-year-equivalent earnings among the employed falling by approximately €204. 

A novel margin: Labour market re-entry after retirement

While the literature has mostly focused on labour market exits, we show that responses are driven by both those who were working and those who were not working prior to the reform. Those working at baseline reduce unconditional earnings by approximately €262. However, re-entry into employment also plays a meaningful role: initially non-employed retirees earn on average €85 less in 2024 in response to the EIPC. This means that additional pension income reduces both re-entry into employment, and the earnings at which those who enter join the labour market. While this channel has received little attention in the literature, it turns out to be quantitatively important in explaining aggregate labour supply responses among retirees. 

Who responds? Age matters, but gender differences are negligible

We next examine whether the effects are concentrated among particular groups of retirees.

A clear pattern emerges with respect to age. Individuals who are slightly younger at the time of the reform respond much more strongly than older cohorts: those born in 1953 reduce labour earnings by about €230, while for those born just three years earlier the effect is only approximately €70. This gradient is consistent with a simple intuition: labour supply responses after retirement are highly concentrated among individuals who are still close to the participation margin. Among older cohorts, the individuals who would otherwise be most responsive have already withdrawn from the labour market, leaving little remaining labour supply to adjust.

We find little systematic heterogeneity by gender, despite women being more likely to be eligible for the reform. This contrasts with a broader literature documenting gender differences in retirement behaviour (Blau 1998, Lalive and Parrotta 2017, Giupponi 2024, Giesecke and Jäger 2021) but is closer to findings showing limited gender differences once selection and income differences are accounted for (Danzer 2013). We attribute this to the EIPC’s eligibility rules, which select men and women with similar contribution histories and household incomes and thereby remove much of the income and selection variation that usually drives gender differences.

Policy implications

Our findings have several implications for the design of redistributive pension systems and pension systems in ageing societies. The central message is that higher pension income does reduce post-retirement labour supply: this response is meaningful, but it is modest in magnitude. A marginal propensity to earn out of unearned income of around -0.1 implies that roughly a tenth of every euro of additional pension income is offset by reduced earnings. This constitutes a notable cost in foregone labour supply, which should be factored into evaluations of the net effect of redistributive policy reforms. 

Second, this response operates through a broad range of margins and is concentrated among younger retirees still close to the participation margin. Retirees adjust not only by reducing earnings while employed, but also, and primarily, by changing whether they participate at all, including by exiting employment and by foregoing re-entry into work after an earlier exit. This last channel is easily overlooked: standard analyses focused on the initial retirement decision miss the retirees who would otherwise have returned to work, yet our results show this margin contributes substantially to the aggregate response. 

Taken together, the findings point to a real but limited trade-off between pension adequacy and work incentives. The response is genuine, economically relevant, and unevenly distributed across retirees, yet considerably smaller than what the literature finds for prime-age workers. Policies aimed at improving pension adequacy and those aimed at encouraging longer working lives therefore entail modest trade-offs. As many countries continue to reform their pension systems in response to demographic pressure, recognising these trade-offs is essential for designing pension systems that protect against old-age poverty without unnecessarily discouraging post-retirement work.

References

Artmann, E, N Fuchs-Schündeln and G Giupponi (2023), “Forward-looking labor supply responses to changes in pension wealth: Evidence from Germany”, IZA Discussion Paper 16132.

Blau, D M (1998), “Labor force dynamics of older married couples”, Journal of Labor Economics 16: 595–629.

Danzer, A M (2013), “Benefit generosity and the income effect on labour supply: Quasi-experimental evidence”, The Economic Journal 123: 1059–1084.

Fetter, D K and L M Lockwood (2018), “Government old-age support and labor supply: Evidence from the Old Age Assistance Program”, American Economic Review 108: 2174–2211.

Gelber, A M, A Isen and J Song (2016), “The effect of pension income on elderly earnings: Evidence from Social Security and full population data”, NBER Working Paper.

Giesecke, M and P Jäger (2021), “Pension incentives and labor supply: Evidence from the introduction of universal old-age assistance in the UK”, Journal of Public Economics 203: 104516.

Giupponi, G (2024), “When income effects are large: Labor supply responses and the value of welfare transfers”, Working Paper.

Golosov, M, M Graber, M Mogstad and D Novgorodsky (2024), “How Americans respond to idiosyncratic and exogenous changes in household wealth and unearned income”, Quarterly Journal of Economics 139: 1321–1395.

Hernæs, E, S Markussen, J Piggott and K Røed (2016), “Pension reform and labor supply”, Journal of Public Economics 142: 39–55.

Johnsen, J V and A Willén (2022), “The effect of negative income shocks on pensioners”, Labour Economics 76: 102175.

Kindermann, F, C Krolage, S Kunz, M Pannier and K Ströhlein (2026), “Pension income and post-retirement labor supply”, CEPR Discussion Paper 21645.

Krolage, C, F Kindermann, K Ströhlein and S Kunz (2026), Evaluierung der Grundrente, Bundesministerium für Arbeit und Soziales, Berlin.

Lalive, R and P Parrotta (2017), “How does pension eligibility affect labor supply in couples?”, Labour Economics 46: 177–188.

Manoli, D and A Weber (2016), “Nonparametric evidence on the effects of financial incentives on retirement decisions”, American Economic Journal: Economic Policy 8: 160–182.