The theoretical literature on customs unions has tended to gloss over determination of the optimal level and structure for the common external tariff (CET) and the distinction between forming a new union and the addition of new members to an existing one. In Discussion Paper No. 368, Research Fellow Christopher Bliss examines the implications for the common external tariff of enlarging a customs union.
Bliss develops the idea of a `neutral' union, which sets its external tariffs so as to preserve the same volume and structure of net trade with the rest of the world as before. One can then consider separately an external tariff adjusted to an optimum which alters net trade flows subject to the constraint that the welfare of the rest of the world is not actually reduced, or a tariff adjusted to the union's selfish optimum, where the outside world may be harmed. Except in very stylized cases, however, it is not easy to say how even a neutral enlargement will affect the CET. For the EC, the joining countries in the 1980s and 1990s tend to be high-cost manufacturers and low-cost agricultural producers, more like other countries outside the Community than those already inside, so trade diversion is important. This suggests that, in order to get the neutral outcome, the CET on goods predominantly supplied by non-members should be reduced although treaties of association may already be serving to mitigate trade diversion anyway.
In many simple trade models, Bliss notes, the optimal tariff is zero, making the issue of its adjustment to enlargement trivial. But various cases can be identified where a positive CET is optimal. First, it is highly unlikely that a customs union could form or enlarge without some region or interest group within the union suffering losses. Although the EC has mechanisms to compensate losers, their effectiveness is doubtful and there will be a tendency to maintain losers' incomes by raising the CET on some factors or goods. Second, when the customs union has chauvinistic or other non-economic objectives, it may prefer intra- to extra-union trade. The consequences depend on whether chauvinism is national or `communautaire'. Bringing in low-cost farmers, for example, assists the Community-chauvinistic desire for agricultural self-sufficiency. With nation-chauvinism, however, high-cost farmers in existing members may press for an increased tariff to compensate for reduced agricultural prices.
The question of how the CET is affected by customs union enlargement is very complicated. Numerous effects operate and may well offset each other, even when each effect alone has an unambiguous direction. Enlargement can imply a higher or a lower CET, depending on the nature and reasons for existing tariffs and on the economic structures of the unenlarged union, the joining countries and the rest of the world. One important effect of enlargement, according to Bliss, is that coordination of tariff-setting policy will be possible: union members considering their joint tariff will take into account the effects of their own trade on the terms of trade of their partners.

The Optimal External Tariff in an Enlarging Customs Union Christopher Bliss

Discussion Paper No. 368, February 1990 (IT)