Nonhomothetic preferences as a cause of missing trade and the home market effect.

The dissertation examines two recently emerged topics in international trade: missing trade and the home market effect. These issues are approached from the perspective of the demand side, distinct from the traditional models that commonly neutralize the demand side by assuming homothetic preferences for its simplicity. By incorporating nonhomothetic preferences into traditional models, this dissertation provides an explanation for the missing trade puzzle and draws alternative implications for the home market effect. The first chapter shows that a simple modification of the Heckscher-Ohlin-Vanek (HOV) model, incorporating nonhomothetic preferences, can resolve the puzzle of missing trade. The puzzle stems from a recent empirical finding that the measured factor content of trade is too small compared to the theoretical prediction. This misprediction may be due to the strong assumption that each country consumes services of factors in proportion to its share of world income. This chapter shows, in a generalized model with arbitrary numbers of factors, goods, and countries, that nonhomothetic preferences will reduce the predicted factor service trade to the extent that income elastic factors are on average abundantly endowed in rich countries and income inelastic factors are on average abundant in poor countries. The second chapter applies this theoretical implication to data. A counterfactual exercise is adopted to calculate the factor content of trade pertaining to homothetic preferences and that pertaining to nonhomothetic preferences. These predicted values are respectively compared with the actual factor service trade obtained from more carefully generated data. The results show that the standard HOV model consistently over-predicts the factor content of trade. The third chapter examines the home market effect (HME), whereby countries export goods for which they have large home markets. The HME has been analyzed mostly on the basis of absolute market size. By incorporating nonhomothetic preferences into the model, however, this paper shows how cross-country per capita income differences can affect the implications of HME on the basis of both absolute and relative market sizes. Our empirical investigation shows that together with absolute size, relative market size plays a significant role in determining production location and the pattern of trade.PhDEconomicsSocial SciencesUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/131610/2/3057926.pd