Bilateral versus Multilateral Free Trade Agreements: A Welfare Analysis
One sentence summary: International development policies
conducted for the removal of obstacles such as transportation costs and
differences in comparative advantages should be the main goal toward achieving
a multilateral FTA, which is shown to be the first-best solution to the
maximization problem of global welfare.
The corresponding paper by Demet Yilmazkuday and Hakan Yilmazkuday has been published at Review of International Economics.
Abstract
Why do we observe proliferation of bilateral free trade
agreements (FTAs) between certain types of countries instead of having progress
in attaining global free trade through a multilateral FTA? We answer this question
by exploring the enforceability of different types of FTAs through comparing
minimum discount factors that are necessary to sustain them in an infinitely
repeated game framework. We also search for the globally welfare maximizing
trade agreements that are sustainable under different conditions. The results
depict that transportation costs, differences in country sizes and comparative
advantages are all obstacles for having a multilateral FTA. Accordingly,
international development policies conducted for the removal of such obstacles
should be the main goal toward achieving a multilateral FTA, which we show to
be the first-best solution to the maximization problem of global welfare.
The lack of a forcing authority in trade relations of world
countries makes it difficult to achieve a multilateral free trade agreement
(FTA) that increases world welfare. This creates a structural problem of rules
in trade agreements that will self-enforce the trading countries to achieve
more liberal trade.
In this paper, we study these self-enforcing rules with
asymmetric countries from the perspective of bilateral and multilateral FTAs.
In particular, we attempt to find an answer to the question of "Why are
FTAs bilateral/regional/preferential rather than multilateral?" Before we
can answer this question, since many trade models find a multilateral FTA to be
the first-best solution, we need to know why bilateral FTAs exist in the first
place. To better understand the characteristics of bilateral FTAs, consider
Baier and Bergstrand (2004) who empirically find that the potential welfare
gains and likelihood of a bilateral FTA are economically and significantly
higher: (i) the closer in distance are two trading partners; (ii) the more
remote a natural pair is from the rest of the world (ROW); (iii) the larger and
more similar economically (i.e. real GDPs) are two trading partners; (iv) the
greater the difference in comparative advantages; and (v) the less is the
difference in comparative advantages of the member countries relative to that
of the ROW.
The empirical findings of Baier and Bergstrand (2004) above
are consistent with 85% of the 286 bilateral FTAs existing in 1996 among 1431
pairs of countries and 97% of the remaining 1145 pairs with no bilateral FTAs.
Hence, to have credible results, any trade model investigating bilateral FTAs
should be consistent with these empirical findings. We present such an
N-country-N-good partial equilibrium trade model by generalizing the
2-good-2-country model in Bond and Park (2002) through considering possible
interactions between country sizes, comparative advantage, and transportation
costs.
For simplicity, we assume that the downward-sloped demand
curve and upward-sloped supply curve of each good in each country are linear in
price of the good. We show that international trade between any two countries
is achieved through differences in supply and demand structures of the
countries. It follows that the equilibrium autarky price (for each good) in any
country depends on the individual country specific demand and supply structure,
while the equilibrium price (for each good) under international trade depends
on the demand and supply structures of all countries together with
country-specific tariff rates and transportation costs. We let each country to
receive all the tariff income for the good that the country imports.
Accordingly, the national welfare is defined as the weighted
sum of consumer surplus, producer surplus and tariff incomes. Using this
definition in a welfare analysis, we estimate the policy weights (assigned to
consumer surplus, producer surplus, and tariff income across countries) that
are consistent with the empirical findings of Baier and Bergstrand (2004)
regarding bilateral FTAs. The estimation results suggest that the policy weight
on producer surplus is much higher than the others. Specifically, the estimates
of policy weights on producer surplus range between 0.71 and 0.77, policy weights on consumer
surplus range between 0.01 and 0.07, and policy weights on tariff income are about 0.22. Using
the estimated policy weights, we investigate for the economic conditions of our
model under which bilateral versus multilateral FTAs may exist with
self-enforcing rules.
In order to incorporate self-enforcement agreements into the
model, we allow countries to get involved in a stationary dynamic tariff game,
i.e., countries play an infinitely repeated game for tariff rates. In this
game, each country is able to compare future payoffs out of a possible
collusion (cooperation) with future payoffs out of a possible deviation from
the agreement. In order to sustain collusion in a trade agreement, the
trade-off between the gains from deviating from an agreed-upon tariff policy
and the discounted expected future gains from collusion must be balanced in a
way that the latter should keep away countries from deviating.
Within this picture, we ask the more-detailed question of
this paper as "Why do we observe proliferation of bilateral FTAs between
certain types of countries instead of having progress in attaining global free
trade through a multilateral FTA?" We answer this question first by
exploring the enforceability of different types of FTAs through comparing the
minimum discount factor that is necessary to sustain them. After having such a
minimum-discount-factor analysis, we go one step further by searching for
globally-sustainable-Pareto-optimal trade agreements defined as the globally
welfare maximizing trade agreements that are sustainable under different
conditions.
Our results suggest that transportation costs, differences
in country sizes and comparative advantages are all obstacles for having a
multilateral FTA, which we show to be the globally first-best solution for
several different cases of our model. Therefore, removal of such obstacles
should be the main goal toward achieving a multilateral FTA. Accordingly,
international development policies conducted for countries to converge to each
other (e.g., increasing internal and international stability, reducing poverty,
investing in transportation technology, and international diffusion of
production technology) should be the main tools.