Essays on Transaction Costs and Food Diversity in Developing Countries

This thesis addresses 3 different aspects of food and poverty related problems in developing countries. The first essay presents a new operational concept of transaction costs that firstly allows assessing the magnitude and secondly recognizing the non homogeneity of food products. This is realized by providing an estimate of the value of the good by means of a hedonic food price model. A model is proposed that decomposes unit values into spatial price factors and a value component that allows the comparison with a feasible value occurring in a situation without transaction costs. The model is estimated with a conditional mean stochastic frontier approach using data from Kenyan maize farmers. We find a magnitude of 12-18% for maize transactions in rural Kenya and identify drive time, market distance, education and counterparts in negotiations as main determinants. The second essay is concerned with the latent demand structure for food diversity in India using data from 68th round of the CES Consumer Survey. We assume that consumers facing subsistence concerns favor calories over food diversity and once passing the subsistence threshold substitute away from staples towards a more varied diet. Latent classes and consumption patterns are identified by means of nite mixture models. Therefore we examine the link between food diversity indices and socioeconomic indicators and explain component memberships in order to characterize latent classes and evaluate nutritional implications. Two clearly distinct demand patterns for diversity can be identified, consistent with our assumptions. The identified classes differ substantially in terms of income, household composition and nutritional adequacy ratios The third essay is concerned with the inference on nutrition from observed consumption. Measures of diversity have become popular tools to infer on nutritional adequacy from observed consumption. However the most common measures do not consider that equal distribution of food consumption does not re ect an optimal diet. The proposed index in this essay adjusts the existing concept of the healthy diversity index so that it is applicable for Indian dietary analysis and extends it for the analysis of household data. The results show that the modified HFD index is a superior predictor of nutritional adequacy compared to common measures like the Berry, Entropy or count index.