country transfers, and, in the case of free trade areas, rules of origin. Countries may experience
welfare decline if there is failure to appropriately control one or more factors (tariffs, transfers,
rules of origin).
To place the discussion of rules of origin in context, recall that the modern theory of Pareto
superior preferential trading arrangements dates to the work of Kemp and Wan (1976) who
showed that any group of countries is capable of forming a welfare-enhancing customs union if the
common external tarif is chosen to freeze trade with the rest of the world and appropriate intra-
union transfers are used. Their result depended on showing that a Pareto-improving equilibrium
existed and has been extended in a number of important directions. The Kemp and Wan strategy
for the creation of customs unions did not provide a description of how to find the compensations
that support the welfare enhancement of the union households. Grinols (1981) identified the
self-financing necessary and sufficient transfers in the sense that they summed to union tarif
revenues and under certain circumstances were the only ones feasible to achieve a Pareto superior
allocation in a Kemp-Wan customs union. Hammond and Sempere (1995) proved the existence
of equilibrium in Kemp and Wan customs unions using commodity taxation to spread the gains
from freer trade. Kowalczyk and Sjostrom (2000), and Konishi, Kowalczyk and Sjostrom (2003)
showed that Kemp-Wan customs unions with the specified compensations bring global free trade
into the core and have other desirable properties.
Similar concerns apply to supporting welfare-enhancing free trade areas. With respect to rules
of origin—not an issue in free trade or customs union theory—the analysis shows that Assumption
2 rules are sufficient for welfare gains, and are also necessary in the sense that they are the most
restrictive that can be applied and still guarantee welfare-enhancement in any circumstance. This
extends a number of earlier results. For example, Panagariya and Krishna (2002) show that an
FTA equilibrium using Assumption 2 rules of origin will generate enough FTA income that every
household could be given enough that no one would be worse of. This extends Kemp and Wan in
the direction of free trade areas, but, as previously noted (Feenstra, 2004, p. 195), did not prove
that the described equilibrium exists.10 The fair sharing distribution rule, referenced in Section 4,
describes the necessary and sufficient compensations for welfare enhancement of union countries,
consistent with the existence of the equilibrium proved.
The conclusions regarding the results in this paper and those that precede it depend on one’s
10Ohyama (2002) also describes the implementation of an FTA where members would jointly enjoy greater
welfare levels, but does not address rules of origin. The lack of rules of origin in the equilibrium described could
prevent the existence of the assumed vector of prices and prevent the formation of a welfare-enhancing FTA in
the first place. Neither Ohyama nor Panagariya and Krishna described how to find the compensations that are
needed to support the gains of union households.
10