Abstract
Since the 1970s, Armington has been the workhorse specification of trade in computable general equilibrium (CGE) models. Under Armington, agents substitute between products from different countries. Conceptually, Melitz provides a more attractive approach in which substitution is between products from different firms rather than countries. Other attractive features of Melitz are monopolistic competition and economies of scale from fixed establishment costs for firms and fixed set-up costs on trade links. In this paper, we show how, with little change to existing code, an Armington model can be converted to Melitz by adding a few equations and introducing closure swaps. We apply our Armington-to-Melitz method to the Armington-based Global Trade Analysis Project (GTAP) model to derive GTAP-Armington-to-Melitz (GTAP-A2M). We show how results from a CGE model with Melitz industries can be interpreted via back-of-the-envelope calculations. In simulations of the effects of a tariff imposed by North America on imports of apparel, we find greater welfare losses for exporting regions under Melitz than under Armington principally because contraction of apparel output in these regions under Melitz generates an increase in the cost to their households of domestic apparel. Finally, we review two other Melitz-based versions of GTAP: GTAP-HET and a recently published model by Bekkers and Francois.
Highlights
In this paper we describe briefly:(1) how an existing Armington-based model such as standard GTAP can be converted to a model with Melitz industries by the addition of some elementary equations and the use of closure swaps
In this subsection we describe the equations that we added to the end of the modified GTAP code to form the GTAP-A2M system
What are the advantages and disadvantages of GTAP-HET versus GTAP-A2M? The advantage of GTAP-HET is that it allows for the composition of inputs to current production in an industry to be different from that used in establishing firms and setting up on trade links
Summary
Thousands of economists are familiar with GTAP and many have completed courses on its Armington-based theory, data and application. There is a huge world-wide investment by economists in GTAP In these circumstances, achieving widespread adoption of a Melitz version of GTAP with new theory or data structures has the potential to impose large human-capital costs on the world’s community policy modellers. We wanted to give users the opportunity to adopt Melitz assumptions in a model consisting of familiar GTAP plus a small number of add-on equations. We designed these add-on equation so that through closure swaps they could be turned on or off to introduce Melitz assumption for user-chosen industries
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have