Work performed from the beginning to the end of the project is grouped under three work packages.
Work Package 1 (WP1).
The WP1's main aim is to construct bilateral trade costs incurred in international trade among countries and trade costs incurred in domestic trade of each country. This is done by developing a two-step methodological procedure. The first step is to compute domestic trade costs for a country using bilateral trade among regions of the country. The second step is to compute international and domestic trade costs using bilateral trade, amongst and within countries, relative to domestic trade costs of the chosen country from step one. The new method is a modern theory consistent one and require bilateral trade data to work. The procedure produces observable trade cost measures that fit actual trade data in use at the same time satisfies theoretical conditions set by the unique relationship among observable and unobservable elements suggested by a structural gravity framework.
As a result of the work under WP1, a research paper has been produced that describes a two-step method on measurement of international and domestic trade costs implicitly using a structural gravity framework. The paper provides with practical implementation of the method using actual trade data to show that it works. As a result, a new product level (GTAP classification) domestic and international trade cost data covering 134 countries has been generated. The analyses of the data show that domestic trade costs proportionally grow with economic size, and that international trade costs are highly asymmetric especially when trading pair sizes differ. It is shown in the paper that counterfactual general equilibrium simulations that ignore domestic trade costs and adopt symmetric international trade cost measures show over-predicted welfare and trade changes as a response to policy shock in larger sized countries rather than smaller countries. After the proper treatment of trade costs in the simulation, results improve.
Work Package 2 (WP2).
The WP2's main aim is to decompose constructed bilateral trade costs (from WP1) into discriminatory and non-discriminatory costs. The decomposition method is a top-down econometric approach performed at country-country-product level. Bilateral trade costs produced by the two-step procedure (from WP1) is used as the dependent variable. The Head-Ries Index (independent variable) represent an ideal measure of full discriminatory trade costs. Non-discriminatory trade costs borne by origin and destination captured with country fixed effect dummies. Tariff-equivalent level of discriminatory and non-discriminatory costs were reconstructed from the estimates by consistent aggregation.
The second paper solely focus on defining discriminatory and non-discriminatory costs and the decomposition procedure. In the paper, I provide theoretical foundation for the decomposition work and describe the data used for empirical section. The results of this work show that there is a link between (non-)discriminatory trade costs and economic sizes. As the size of an economy grows discriminatory trade costs it faces reduce while non-discriminatory trade costs instead increase. I also find that common trade cost ‘suspects’ (i.e. transportation, taxes, subsidies, tariffs etc.) explain only smallest part of (non-)discriminatory trade costs.
Work Package 3 (WP3).
The WP3's main aim is to introduce discriminatory and non-discriminatory costs in a CGE model and perform a CGE based quantitative assessment of costs and benefits of a RTA. For that I consistently merge the discriminatory and non-discriminatory trade cost measures (from WP2) in the GTAP 8 database. Further, I modify the standard GTAP model by introducing three new trade cost variables: bilateral unexplained discriminatory trade cost, non-discriminatory exporter trade cost and non-discriminatory importer trade cost variables.
In the final paper, I describe the changes undertaken in the GTAP data and the GTAP model. I also present the simulation results for the TTIP formation between the US and the EU. The simulation outcomes show that in Scenario 1 welfare gains are much higher than for Scenario 2. The similar results I obtain from using different model closures (Melitz, Armington or Krugman type). These imply that a cost and benefit analysis of a RTA formation in a CGE setting without non-discriminatory trade costs result in overprediction of the gains.
