We trace the effects of a devastating earthquake that occurred in Northern Pakistan in 2005. Using a new dataset from a survey conducted four years after the earthquake, we first show that the distance of the household from the fault line was not correlated with pre-existing household characteristics, while it was strongly predictive of earthquake-related damage and mortality. Through emergency relief aid, households living close to the fault line reported receiving substantial cash compensation that amounted to as much as 150% of their annual household consumption expenditure.

Four years after the earthquake, there were no differences in public infrastructure, household or adult outcomes between areas close to and far from the fault line. However, children in their critical first thousand days at the time of the earthquake accumulated large height deficits, with the youngest the most affected. Children aged 3 through 15 at the time of the earthquake did not suffer growth shortfalls, but scored significantly worse on academic tests if they lived close to the fault line. Finally, children whose mothers completed primary education were fully protected against the emergence of a test score gap. We estimate that if these deficits continue to adult life, the affected children could stand to lose 15% of their lifetime earnings. Even when disasters are heavily compensated, human capital accumulation can be critically interrupted, with greater losses for already disadvantaged populations.