dentifying the long-run effects of in utero and early life conditions has become an important research topic. Since the seminal work by Almond (2006), a growing literature finds that in utero exposure to adverse environments may negatively affect health and educational attainment later in life (Almond, Edlund, and Palme 2009; Banerjee et al. 2010; Shah and Steinberg 2014). These studies document that conditions both in utero and in early childhood have a long-lasting impact on life expectancy, adult earnings, adult health, and cognitive development.

The British physician and epidemiologist David J. Barker argued that nutritional deprivation in pregnant women leads to impaired fetal development with consequences that persist after birth and even through adulthood (Barker 1990, 1995). The so-called fetal origins hypothesis can provide important insights into the understanding of low levels of human capital accumulation in resource-poor countries.

The challenge in identifying the causal impact of in utero conditions on later-life outcomes is to find a truly exogenous variation in such conditions. In recent years, economists have taken advantage of natural experiments that are quasi-random in nature to identify exogenous variation in early environmental conditions. Events such as pandemics (Almond 2006; Banerjee et al. 2010), famine (Kannisto, Christensen, and Vaupel 1997; Lindeboom, Portrait, and van den Berg 2010), armed conflict (Lee 2014; Akresh, Lucchetti, and Thirumurthy 2012), radioactive emissions (Almond, Edlund, and Palme 2009), air pollution (Currie, Neidell, and Schmieder 2009), Ramadan fasting (Almond and Mazumder 2011), government intervention—for instance, iodine supplementation, hookworm, and malaria eradication programs (Bleakley 2009; Field, Robles, and Torero 2009)—and extreme weather shocks (Maccini and Yang 2009; Shah and Steinberg 2014; Rocha and Soares 2015) have been used to estimate the causal relationship between the in utero condition and later-life outcomes.

We use the plausibly exogenous variation in rainfall to examine the effect of in utero exposure to a drought on the health outcomes of children aged under 5 years in India. Recent studies have examined the long-term effects of rainfall on the health and schooling of children. Maccini and Yang (2009) noted that higher rainfall in the year of birth leads to improved health, schooling, and socioeconomic status for women in Indonesia but not for men, although they discount the importance of in utero exposure to higher rainfall. Similarly, Shah and Steinberg (2014) assess the effect of rainfall variability in India on human capital accumulation and find that children who were exposed to higher rainfall in utero or between birth and age three score significantly better on literacy and numeracy skills, are less likely to repeat a grade, and are more likely to ever enroll in school. In a study in northeastern Brazil, Rocha and Soares (2015) report that exposure to drought in utero is correlated with higher infant mortality, lower birth weight, and a shorter gestation period. They identify lack of safe drinking water and a higher incidence of infectious diseases as potential mechanisms rather than the decline in agricultural production and lower nutrient intake. Even though neither Maccini and Yang nor Shah and Steinberg directly measure the channels through which weather affects children's health and cognitive development, they argue for the causal chain operating through the impact that weather variability has on agricultural production and, thus, income and nutrition for the rural population.