In the US and UK in particular, new retirees make complex decumulation decisions regarding defined contribution pension wealth. We use detailed data on wealth portfolios of people aged 50-64 in the UK to understand how people access DC pensions. Accessing pensions by making (taxable) lump-sum withdrawal is most common, and is associated with impatience but not with financial literacy. Some individuals use liquidity of DC wealth to pay off formal loans. Finally, many seem unprepared for their decumulation decisions: over 40% of 50-64 year olds do not know how they plan to access their DC wealth, particularly women and those with lower levels of wealth.

This paper examines the welfare effects of pension reforms that increase financial incentives to delay claiming. We leverage the 2015 Brazilian pension reform, which allowed workers past an individual-specific threshold to claim early retirement benefits with full replacement rate - a 30% increase at the discontinuity. Using novel administrative records on early retirement claims and labor market data, we estimate the substitution elasticity of pension claiming and the income elasticity of retirement. We extend the model by Kolsrud et al. (2024) to show that these elasticities, along with the income elasticity of claiming, are sufficient statistics for welfare analyses whenever claiming and labor market exit are separate decisions. Lastly, we evaluate whether informational channels are behind the lack of responsiveness to incentives. We show that easier access to social security offices does not have a significant impact, but that local knowledge of social security rules can affect responsiveness to financial incentives.

I provide a comprehensive analysis of the impact of a pure change in pension wealth on retirement behavior. I exploit a 2010 Norwegian policy reform that retroactively extended childcare-related pension credits to mothers of children born between 1967 and 1991. I quantify the resulting increase in pension benefits and assess the policy's potential to narrow the gender pension gap, along with any distortionary effects on labor supply and pension claiming. Using population-wide administrative data from Norway, I estimate pension benefits for all beneficiaries had the policy not been adopted and implement a difference-in-differences design, leveraging the variation in the credits' impact on final pension benefits. My findings show that the policy increased pension benefits by an average of 2.5%, contributing to a 1.6 percentage point reduction of the gender gap in lifetime pension benefits with small employment effects.

The share of retirees and the size of benefits paid are critical to determine an economy's aggregate pension spending. Increasing the statutory retirement age is a popular policy proposal to decrease the burden of aging populations on pension systems by decreasing the share of households eligible for pension benefits. However, in a system in which pension benefits are a function of lifetime earnings, changes in both the extensive and intensive margins of labor supply in response to a reform impact pension costs. Using a heterogeneous-agent OLG model with endogenous retirement choice and pension benefits, I study a 2019 Brazilian pension reform which increased the retirement age. I find that while the reform decreases aggregate pension costs, it leads to welfare losses for high-income households. Moreover, I find that a policy limiting the link between intensive margin labor supply and benefits leads to higher aggregate pension costs than the 2019 reform.