There are very few studies on the incidence of urban poverty and malnutrition in developing countries. The few studies that exist treat all urban areas – small towns, medium-sized towns and large metropolises – as a single urban entity. These studies conclude that urban poverty and urban malnutrition in developing countries is likely to grow in the near future (Haddad 1999, Ravallion 2002).

In this paper, we examine both the incidence of hunger-poverty – as measured by the inadequacy of calorie intake – among Indian metropolitan cities (urban agglomerations) in 1999-2000 as well as the change in hunger-poverty between 1993-94 and 1999-2000. The recent evidence from India (e.g., Meenakshi and Vishwanathan, 2003) suggests a divergent trend in the incidence of consumption-poverty and hunger-poverty; while the headcount index of consumption poverty has steadily declined since the 1970s, the incidence of hunger-poverty is reported to have increased.

There are two possible explanations for the divergent trend. First, the normative calorie norm that has been used to calculate hunger-poverty has remained the same since the 1970s (2,100 calories per person per day). Second, urban areas – comprising both small towns with of population of 5,000 persons population and large cities with over ten million population – are treated as a single entity by all the empirical studies. Dubey et al. (2001) have reported that the incidence of poverty in metropolitan cities is only about one-half of that in the smaller towns.

Both of these assumptions are not likely to be valid. The normative calorie requirement could change over time due to inter-temporal changes in the structure of employment. Further, the normative calorie requirement could differ significantly a cross cities of varying size, as the economic activities in these cities would be organized differently.

In this paper, we estimate hunger poverty by basing normative calorie requirements on household- level data on employment and occupational tasks, using the quinquennial rounds of a large national survey (National Sample Survey) during the 1990s. Using these estimates, we then analyze the socioeconomic and policy determinants of hunger-poverty in the metropolitan areas of India. The policy determinants include certain types of city characteristics and amenities.