Essays on combining individual and aggregate data and marketing in emerging economies

This dissertation is comprised of three essays. In the first essay we discuss a new approach for combining individual and aggregate data Researchers often face the problem of limited data on individual characteristics and therefore, commonly augment available individual data with mean socioeconomic characteristics of the corresponding geographic group (census tract, zip-code, etc). We show that, in a linear model, coefficients associated with individual variables will be inconsistently estimated if those variables are correlated with the underlying individual variables summarized to obtain the aggregate group means. Consistent estimates can be obtained if the group means are conditioned on observed individual variables. We introduce a practically feasible approach to infer group level joint distributions to obtain such conditional estimates. The approach is illustrated in an application predicting the profitability of a bank's customers. In the second essay we study the phenomenon of private schooling in India.;Private schools serve a significant section of the Indian student population in the K-12 market, including the rural poor. This share is likely to increase with the country's growth as households, both rich and poor, clamor for quality education. In this essay, we present and estimate a discrete-continuous model of household school choice and educational spending decisions to understand the determinants of schooling (public and private) in rural India. Our estimates give insights on (1) households and individual determinants on schooling demand in rural India; (2) the tradeoff off between marketing mix variables such as fees and transportation facilities; and (3) how private-public partnerships through government subsidies for private schools can expand category consumption by increasing total school enrollment.;In the third essay we present a model of choice between modern and alternative medicine, for both exclusive and complementary use, by Indian consumers. We find that, in addition to demographics, consumers' underlying attitudes about modern science affect their decision to use alternative medicine. We also find, contrary to popular perception, that the use of alternative medicine is greater among higher income than lower income consumers. Our results have implications for targeting of alternative medicine based treatments.