Alternative Hypotheses and the Volume of Trade: Evidence on the Extent of Specialization
Public Policy Institute of California
David L. Hummels
Purdue University - Department of Economics; National Bureau of Economic Research (NBER)
In this paper, we provide alternative hypotheses against which to evaluate the trade volume predictions of complete specialization models, which include monopolistic competition. This allows us to identify which model predictions are substantive and which empirical regularities are surprising. There are four main insights. First, the model implies that consumers highly value variety, importing all goods that are produced. Trade flow data shows that importers purchase a very small fraction of available varieties. Second, the model predicts multilateral trade volumes that are much larger than we observe or can explain with barriers of reasonable size. Third, the use of a welfare-based measure of model fit indicates that the model implies enormous unrealized gains from trade. In each instance, the data suggest that the simple model considerably overstates either the extent of specialization (the degree to which goods are differentiated) or the degree to which consumers value that differentiation. Fourth, we demonstrate that the gravity equation (the most celebrated prediction of complete specialization models) has no power to distinguish the null hypothesis that production is completely specialization from an alternative that production is determined by factor endowments and incompletely specialized. A consequence is that stylized facts about trade costs associated with national borders and distance can be explained by a model with no trade barriers.
Number of Pages in PDF File: 45
JEL Classification: F12, F15working papers series
Date posted: April 26, 1999
© 2013 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo7 in 0.390 seconds