arXiv2025
Children welfare is at the center of many welfare reforms such as cash transfers to families and training programs to parents. A key goal for policy-makers is to evaluate the costs and benefits of such reforms. The main challenge lies in that the outcome of interest, children welfare, is unobservable. To address this issue, I consider a collective labor supply model with children where adult members have preferences over their own leisure, expenditures, and children welfare. I show that the model nonparametrically partially identifies the impacts of parental inputs on children welfare in panel data. I then propose a novel estimation strategy that accommodates measurement error and can be used to efficiently construct valid confidence sets. Using Dutch data on couples with children, I investigate the structure of the expected production technology and how it varies with household characteristics. I find that the production of children welfare is characterized by decreasing returns to scale and large heterogeneity across household types. In particular, I find that children from disadvantaged households, whose parents have low education levels and are not homeowners, are significantly worse off. My results highlight the importance for welfare reforms to include policies targeted at improving children home environment.