non-camp vaccination sources and regional-level variation in vaccination levels. In both cases I again find evidence for a non-monotonic relationship between the level of vaccinations and the gender differences.
In the final section of the paper I consider whether these non-monotonicities in vaccination map into non-monotonicities in mortality. I use retrospective information on child mortality to construct a panel of death rates over the period from 1982-1993. I find a non-monotonicity in the relationship between increases in vaccination access and the gender imbalance in mortality. In areas that start with low levels of vaccination, excess female mortality increases over time; in areas that start with high levels of vaccination, excess female mortality decreases over time. This is true only in areas with increases in vaccinations over this period – i.e. those that appear to have had a decrease in vaccination cost.
Together, the results provide strong support for the non-monotonicity predicted by the theory. This result may have policy implications for India and other developing countries that face gender discrimination. policy-makers have argued that increasing the level of development is one of the key factors in ameliorating gender inequality. In 2001, a World Bank report on gender and development begins with the statement that poverty and gender inequality are closely linked: “Large gender disparities in basic human rights, in resources and economic opportunity ... are pervasive around the world ... And these disparities are inextricably linked to poverty,” (World Bank, 2001). One of the aspects of development cited as crucial to affecting gender inequality is access to health services (World Bank, 1991; Hill and Upchurch, 1995). It has been argued that increasing the level of health care will benefit women and reduce gender inequality (Grown, Gupta and Pande, 2005), although the link between development and inequality is not limited to health care (see, for example, Duflo, 2005). This argument is particularly salient in India, where poverty is often linked to gender ratios and excess female mortality by region (World Bank, 1991; Chatterjee, 1990). The results here suggest how interventions are introduced to developing areas may be meaningful. A program like the health camps analyzed here, which is not targeted to decrease gender inequality, may actually have the opposite effect: simply increasing access is not sufficient. Programs that are targeted to affect gender inequality may be necessary, although there are clear tradeoffs; for example, targeting gender inequality may be inconsistent with the aim of decreasing other forms of inequality.
The rest of the paper is organized as follows. Section 2 discusses the theoretical framework, both in general and with a specific application to investments by parents in the health of children. Section 3 describes the National Family and Health Survey data. Section 4 presents the results on