Chapter 1 studies the growth of trade. The U.S. trade share of gross output in manufacturing increased by a factor of 4.7 from 1970 to 2005 and grew slower before the early 1980's than after. As documented in Yi (2003), these patterns are difficult to explain by resorting only to observed tariff reductions. This paper argues the widespread adoption of Just-in-Time (JIT) logistics in the early 1980's provides the key to understanding the growth in the trade share. To do so, I first develop a dynamic trade model based on the logistics technology used in a firm's supply chain. Without JIT, a firm faces a newsvendor problem of stocking inventory before uncertain final demand is realized. Inventories constrain a firm's ability to respond to demand fluctuations. After adopting JIT, however, no such inventory constraint exists. Firms now respond to all demand fluctuations, as well as charge lower prices due to eliminating inventory costs. Thus, as firms adopt JIT with international suppliers, the volume of trade increases. In the model, firm characteristics, including the cost of airplane transportation, determine whether JIT with international suppliers is adopted. The model's predicted trade dynamics depend on how the set of firms using JIT with international suppliers changes over time. Numerical simulations show the model is capable of capturing the growth in the trade share beginning in the early 1980's. Moreover, I present evidence showing the theory is consistent with aggregate data as well as industry-level panel data.
Chapter 2 incorporates learning and reputation building into a simple dynamic stochastic model of international trade with asymmetric information. We use the model to study two bilateral trade flows influenced significantly by learning and reputation, namely U.S. imports of Japanese and French cars over the period 1961-2005. Numerical simulations of the model are capable of replicating these events in a robust fashion. In addition to matching these events, we explore further implications of our framework for understanding international trade patterns. Since learning and reputation building require time, predicted short run trade patterns can be quite different than those predicted in the long run. Sectorial differences in the speed of learning and reputation building affect predicted trade patterns. The extent of asymmetric information existing between importers and exporters also changes under different trade policies.
Chapter 3 is motivated by fear. Movie executives fear the collapse of Hollywood exports in the face of rising worldwide piracy rates. Yet box office sales growth remains stable. According to the Motion Picture Association of America, U.S. and international box office sales grew 7% and 17% between 2004 and 2008. The movie industry must be doing something right. This paper empirically analyzes several techniques the movie industry uses to reduce the replacement effect of piracy. We construct a unique data set on box office performance for our empirical analysis. The data covers a five year period, between 2004 and 2008, and 18 different countries with various piracy rates. Hollywood movies, those distributed by big studios, tend to be released in a foreign country earlier than other movies. The release gap, defined to be the lag between the U.S. release date and foreign release date, for Hollywood movies is shorter in countries with higher piracy rates. Higher piracy rates are also associated with higher box office share for action movies and lower box office share for comedies and dramas.