Peruvian society has achieved significant improvements in terms of lower fertility and mortality over the last forty years, which has brought down population growth rates to less than 1. 2% a year. These improvements have led, on average, to a demographic transition with lower dependency ratios. In general, this transition increases the ability of the society to take proper care of its non-working population groups, children and the elderly, which may be reflected in changes in household structure. We identify stylized facts about the implications of these changes at the micro level through the use of pseudo-panels from household-level data for Peru. We calculate age, cohort and year effects for variables related to household structure, educational attainment, labor force participation and savings. We find some evidence that suggests differences, by educational level, in the Peruvian demographic transition. Household size is smaller for the younger cohorts in all households but those with less educated heads. We argue that these different profiles are explained by the fact that reductions in fertility have not reached the less educated. On the one hand, these differences in household size patterns are similar to those in the number of children. On the other hand, cohort patterns in family living arrangements—i. e. , households with extended families—are similar across educational groups. However, family living arrangements change throughout the life cycle, in the sense that extended families are more common for households with very young (under 25) and elderly (over 60) heads. These changes in family arrangements over the life cycle add confusion to the meaning of headship, since in some cases the household reports as its head the older member and in other cases the main income earner. We also find that younger cohorts are more educated, are larger than older ones, and show lower returns to education. This is consistent with an increase in relative supply of educated workers that outpaces the increase in relative demand induced by economic growth, under the assumption of imperfect substitutability between equally educated workers of different cohorts. Finally, we show that intergenerational family arrangements over the life cycle limit the ability of the life cycle hypothesis (LCH) to explain household savings behavior. We find evidence that Peruvian households, especially the less educated, smooth consumption over the life cycle, not only through the typical saving-dissaving mechanism, but also by smoothing income. Net cash transfers, or living arrangements between parents and their offspring, play an important role in this income smoothing.