Résumé : Household-level measures of income and consumption can conceal differences in family members' access to resources and influence over decisions. This thesis examines three dimensions of family life that matter for understanding these differences: preferences across household members, changes in household composition over time, and the legal rights governing access to and control over resources.The first chapter develops a revealed preference approach to comparing the intensity of preferences across individuals from observed consumption choices. In a two-person, two-good setting, it shows that single-crossing indifference curves are equivalent to a behavioural condition called choice dominance and derives a non-parametric test for finite choice data. It then extends the framework to household decisions, showing how differences in preference intensity restrict the response of household consumption to changes in bargaining power. Applications to experimental data on altruistic giving and household consumption data from Mexico illustrate the approach.The second chapter studies the duration of post-marital intergenerational coresidence in India. Using a decade of household panel data and a spouse-linking algorithm, it follows married couples from the bride’s entry into the husband’s parental household. Coresidence lasts longer when the husband works within the family and when parental household resources are greater, while larger households and higher levels of the wife's education are associated with earlier departure. Most departing couples settle within the same district, and approximately three-quarters of multi-son joint households divide before the father's death. These findings support a view of nuclear and extended households as stages in a household life cycle rather than fixed family types.The third chapter constructs a new dataset of women’s formal economic rights across 141 judicial districts in pre-Revolutionary France. It documents substantial regional variation and shows that stronger rights for married women tended to coexist with weaker protections for widows. Areas more suitable for plow agriculture before 1500 are associated with weaker rights during marriage and stronger protections in widowhood. A household model provides a possible explanation: a wider gender productivity gap reduces the productive return to married women’s control over resources while increasing the protection needed after a husband’s death.