Thesis: This thesis explores early childcare and activation policies, fundamental within the social investment paradigm, through two large field experiments in France, supported by the National Family Allowance Fund.In the first chapter, with Julien Combe, we consider access to daycare as a matching problem. We propose market design models to define assignment mechanisms and analyse the consequences of design choices in a field experiment. The problem is akin to school choice, but specific constraints affect the definition and scope of stable matchings. Our algorithms provide Student Optimal Fair Assignments (SOFA) in different versions of the problem. Our analysis focuses on the Matthew effect, demonstrating how design and policy choices influence it. Our tools promote fairness and transparency in assignment processes.Chapters 2 and 3 analyse data from an intensive experimental programme aimed at low-income single-parent families in France, implemented from 2018 to 2022.In Chapter 2, I analyse the effects on labour market participation and poverty, and how wrong we would have been not to use a randomised controlled trial. The analyses reveal initially negative effects that diminish over time. Participants have higher employment rates than other comparison groups, but this difference is entirely due to selection bias. This bias is so strong that estimates using the next best identification strategy – modern doubly robust differences-in-differences – fail to include experimental estimates within confidence intervals. Overall, the programme has no average effect on labour market participation and poverty after the end of the training. There are heterogeneous treatment effects by number of children at baseline.In Chapter 3, with Alexandra Galitzine, we challenge the narrative of "making work pay" for single-parent families in France. The 2019 reform of in-work benefits (Prime d'activité) was adopted contemporaneously with this programme. The intervention directly provided individualized and detailed information on the socio-fiscal system in a year-long support programme, likely to have further reduced various barriers to employment. We use this experiment to measure low-income single-parent families' reactions to incentives after the reform.Our primary contribution lies in estimating counterfactual distributions using experimental assignment variations. We find high labour income elasticities for participants, indicating significant disincentives to employment and increased in-work poverty. The programme's effects on family structure vary based on the number of children, highlighting the complex interplay between policy incentives and poverty dynamics. We coined the term "Assistaxation" to describe the phenomenon of heavily taxing the economic, physical, and mental resources of those accessing public assistance, leaving them with little means to escape.