80 SOCIAL SCIENTIST pite recording an annual growth rate in GNP per capita of only 0.2 per cent between 1980-92, achieved an annual rate of reduction in under-five mortality of 8.2 per cent.3 Within India itself, to belabour an old point, Punjab, with a per capita income that is more than twice Kerala's level, reports an infant mortality rate (IMR) of 55 deaths per 1,000 live births, more than four times higher than Kerala's IMR of 13 per 1,000 live births.4 Such a lack of correspondence between income levels and health outcomes should not come as a surprise. Income is useful only to the extent that it can be channeled for the provisioning of better health services, and also used by individuals to acquire better health care. Converting income into good health becomes so much more difficult, if not impossible, when public provisioning of health services is inadequate or of poor quality. Also, health outcomes are influenced by a number of other factors, such as the freedoms that women'enjoy, the levels of environmental cleanliness, the extent of political support, the quality of political leadership, the efficiency of government administration, dietary and child-caring practices, and so on, which are not necessarily related to income in any predictable manner. Similarly, it is true that enhanced basic education can result in improving health outcomes, and several pathways of influence have been identified. Education, especially of women, is postulated to improve their decision-making capabilities, enhance their position within the family, make their less fatalistic, and seek out appropriate health care. Once again, this is possible provided there is adequate public provisioning of good quality health care facilities. Even where such facilities are available, and people have adequate purchasing power and knowledge, very often socio-cultural factors, principally considerations of caste, class and gender, seriously prevent women and others from accessing health care services. All this is not to say that income and basic education are not important for influencing health outcomes. On the contrary, these illustrations draw attention to the importance of identifying and understanding factors that impede conversion of higher incomes and better education into good health. Formulating public policy interventions for health becomes difficult not only because of the complex inter-relationships that determine health outcomes, but also because it is a sector where equity and ethical considerations become as important as efficiency arguments. And these considerations significantly dominate, and influence, both private and public decision making. Take, for instance, a case where Delhi's infamous "redline" bus runs off the road, as it frequently does, and critically injures a pedestrian. The pedestrian is rushed to a nearby private clinic where it is found that he cannot afford to pay for his treatment. What should be done? If we think it is unethical for the private clinic to refuse treatment under such circumstances, who will meet the costs of treatment? Should the private clinic be left to device