What is the effect of large pension reforms on consumption? Despite the importance of this question for determining the macroeconomic and distributional consequences of pension reforms, existing evidence is remarkably scant. In this paper, I exploit large cohort-based pension reforms in a simulated instrument design to provide new evidence on their effect on retirement consumption, thanks to granular transaction and bank account data from one of the largest French retail banks. Taking advantage of large non-linearities in pension generosity across cohorts in a regression-kink-design, I estimate a precise zero effect of pension reforms on consumption during retirement. I further decompose this zero effect, showing that individuals did not respond by adjusting their stock of private savings. Instead, I document the presence of large labour supply responses close to retirement, that almost entirely offset the changes in the generosity of the pension system. Using a large panel of lottery winners within the bank to measure marginal propensities to consume out of income shocks, I finally show that pension reforms did not increase financial constraints nor the ability to smooth consumption during retirement. These findings reconcile two strands of the literature: (i) they confirm on the one hand that people are incredibly passive in their retirement saving behaviours even in the face of large nationwide pension reforms, (ii) they further substantiate that retirement timing strongly responds to salient parameters of pension profiles as in Seibold (2021). While individuals apparently fail to optimize on both labour supply and savings behaviours, this paper suggests that the welfare costs in terms of consumption smoothing of the heuristic behaviours they follow are likely limited on average.

What would the lives of women look like if they knew from an early age that they would not have children? Would they make different choices about human capital or early career investments? Would they behave differently in the marriage market? Would they fare better in the labor market? In this paper, we follow 152 women diagnosed with the Mayer-Rokitanski-Kuster-Hauser (MRKH) type I syndrome. This congenital condition, diagnosed at puberty, is characterised by the absence of the uterus in otherwise phenotypically normal 46, XX females. Using granular health registries matched with administrative data from Sweden, we confirm that MRKH is not associated with worse health, nor with differential pre-diagnosis characteristics, and that it has a large negative impact on the probability to ever live with a child. Relative to women from the general population, women with the condition have better educational outcomes, tend to marry and divorce at the same rate, but mate with older men, and hold significantly more progressive beliefs regarding gender roles. The condition has also very large positive effects on earnings and employment. Dynamics reveal that most of this positive effect emerges around the arrival of children in women in the general population, with little difference before. We also find that women with MRKH perform as well as men in the labor market in the long run. Results confirm that "child penalties" on the labor market trajectories of women are large and persistent and that they explain the bulk of the remaining gender gap.

How does information shape behavioral responses to pension reforms? To answer this question, I design a tailored survey, which is currently being fielded among French individuals aged 45-62. Preliminary results from a pilot wave highlight very imperfect knowledge about the pension formula. In particular, respondents systematically overestimate how much they would be penalized for retiring early, and underestimate how much they would benefit from delaying retirement. These findings are consistent with prior evidence on excess sensitivity of retirement timing to salient parameters of the pension system and behavioral attenuation due to cognitive uncertainty. I design an information provision intervention to understand whether personalized information about future pension benefits at different retirement ages, as opposed to general information about pension rules, can render individuals more responsive to the financial incentives embedded in pension systems. My experimental design takes advantage of a governmental pension simulator based on administrative records of individual career histories. This allows me to measure precisely the distribution of forecast error about expected pension benefits and how it is affected by information exposure. I then estimate the causal effect on intended retirement timing, intensive-margin labor supply decisions and savings choices. Lastly, I develop a simple model to interpret behavioral responses and assess the welfare effects of personalized information provision. 

This paper examines the relationship between parenthood, gender specialisation and wellbeing. Using a new matching methodology developed by Kleven (2022) and an event study framework, I leverage two large European wellbeing and values surveys to study the dynamics of life satisfaction and self-assessed affects in a pseudo-panel covering the five years leading up to first childbirth and the first ten years of parenthood. I find that both men and women experience a long-run decrease in wellbeing after becoming parents, but attribute most of this trend to financial stress. Across countries, larger child penalties are associated with worse wellbeing outcomes during parenthood, suggesting that gender specialisation does not stem from innate differences in preferences between men and women. I provide suggestive evidence instead that preferences acquired through the internalisation of gender norms are a powerful explanation for child penalties.

We use anonymised transaction and bank data from France to document the evolution of consumption and savings dynamics since the onset of the pandemic. We find that consumption has dropped very severely during the nation-wide lockdown but experienced a strong and steady rebound during the Summer, before faltering in late September. This drop in consumption was met with a significant increase in aggregate households’ net financial wealth. This excess savings is extremely heterogenous across the income distribution: 50% of excess wealth accrued to the top decile. Households in the bottom decile of the income distribution experienced a severe decrease in consumption, a decrease in savings and an increase in debt. We estimate marginal propensities to consume and show that their magnitude is large, especially at the bottom of the income and liquidity distributions.