This piece is extracted from Mordechai E. Kreinin Journal of Political Economy available at Jstor" Journal of Political Economy Vol. 72, No. 2, Apr., 1964 [link to this].
It has become customary to classify the implications of a customs union for world welfare 1 into static and dynamic effects. Following Meade, the static factors concern mainly changes in resources allocation and consist of trade creation, trade diversion and the terms of trade effects. The dynamic factors are the long-run consequences of increased market size for the growth rate of the integrating region 2. The relationship between market size and growth rate is thought to operate through the following avenues:
The dynamic consequences are considered to be favorable to world welfare, for they result in higher income in the customs union. Moreover, through their positive effect on imports into the region, they are supposed to counter — at least in part — the static diversionary effect on non-participating countries.
As far as most recent writers are concerned, the dynamic factors affect nonparticipating countries only indirectly, and in a favorable way, through the foreign trademultiplier. Thus, Lawrence 13. Krause states: "In discussions of the external impact of economic integration, the point is always made that the unfavorable impact of the formation of a customs union on nonmember country exports will be offset in whole or in part by the increase in member country growth rates stimulated by the establishment of internal free trade" 3. Likewise Erik Thorbecke writes: "From the standpoint of nonparticipating countries, the static effects usually will be negative and the dynamic effects usually will be positive 4." And Bela Balassa, after devoting four chapters to the dynamic factors operating inside the customs union, considers their indirect impact on non-member nations, and concludes: "Available information suggests that in the European Common Market, the growth effect may outweigh the preference effect in the course of a twenty year period if the national income of the Common Market countries increases at least by 120 per cent 5.
The following passage from Vanek's International Trade will serve to summarize current thinking on this subject:
Of course, to the extent that all, or some, of these dynamic factors do influence favorably either the absolute level of income at each period of time or the rate of growth of the union, the outside world must also necessarily benefit from it. Such gains must he used as an offsetting factor to the losses that almost unavoidably the customs union will inflict on the rest of the world, through its immediate terms-of-trade and income effects. It is possible to refer to such an effect as long-run external trade creation, as opposed to short-run external trade diversion and internal trade creation. These benefits may be very important if we assume a sufficiently long view of the customs-union problem 6.
I. DIRECT DYNAMIC EFFECTS ON NON-PARTICIPATING COUNTRIES
It is the contention of this paper that in its present approach, the theory of customs unions overlooks an important dynamic factor, the direct diversionary impact on nonmember countries. In a sense, it is the counterpart of the dynamic effect operating within the customs union. Each of the two effects has indirect repercussions through the foreign trade multiplier. This oversight has led economists to overstate grossly the beneficial consequences of dynamic factors for world welfare and for the welfare of nonparticipating nations.
It is the elimination of trade barriers in the integrating region and the resultant expansion of intra-union trade which is responsible for the expanded markets. The dynamic factors are thus superimposed upon the three static effects: both trade creation and trade diversion contribute to the increase in the size of the market, as do the improved terms of trade through their effect on national income. But, while trade creation has no unfavorable counterpart in nonmember countries, the latter two effects necessarily do. The markets of non-participating countries are contracted by trade diversion and by the deterioration of their terms of trade (which lower their income). And this direct contraction would affect the growth rates of these countries unfavorably. In short, non-member nations are subject not only to direct static, but also to direct dynamic diversionary effects. Although difficult to estimate in practice, the latter impact cannot be overlooked either by the theorist or by the policy-maker.
II. ON THE RELATIVE MAGNITUDE OF THE DYNAMIC EFFECTS
Can anything be said a priori about the relative magnitude of the favorable and unfavorable dynamic effects? Two considerations have direct bearing on this question. First, the enlarged market size of the integrating area is a consequence of all three static effects, while the contracted markets of non-member countries are a result of only two of these effects. Because of this asymmetry (other things being equal) the direct favorable dynamic impact would tend to exceed its unfavorable counterpart. But the indirect favorable effect on non-member countries (operating through the foreign trade multiplier) would not necessarily exceed the direct unfavorable impact upon them.
The second point is much less clear-cut. It relates to the size of the individual markets and to their proximity prior to integration—key variables in determining the magnitude of the internal dynamic effect.
While opinions concerning this magnitude vary widely 7 it is generally agreed that the smaller the integrating nations, and the lower the transport costs between them 8, the larger is the scope for internal and external economies resulting from integration. The scope for such economies would be quite limited in large nations which could realize them prior to the formation of a customs union. How large is "large enough" in this context is a moot question. Harry Johnson, for example, believes that a country the size of Britain is unlikely to benefit much from enlarged market size 9, and a similar observation would probably apply to the large members of the European Common Market (ECM). Even writers who do not share this view agree that the "Potential gains arc especially pronounced in the small and medium sized member countries: In Belgium-Luxemburg, the Netherlands, and Italy 10 (this point loses some of its cogency if Benelux is considered as one market even before integration.) It would be safe to assume that the four independent customs areas forming the ECM are not likely to benefit as much as would four equally contiguous but much smaller European nations forming a customs union.
On the other side of the ledger, the countries sustaining a direct diversionary dynamic effect are of all sizes and at all stages of economic growth. Many of them are small and medium-sized and will suffer badly; others are underdeveloped and their development may be thwarted by the diversionary impact. Thus, if the integrating countries are all fairly sizable, the dynamic diversionary impact on non-member nations may exceed its favorable counterpart, even allowing for the asymmetry between the two.
Because of its economic size, the United States may not sustain dynamic losses from the contraction of its domestic market caused by the formation of the ECM. But even the United States is unlikely to escape the direct consequences. In its case the main unfavorable impact takes the form of invest-meal diversion, a term which can be applied to the "non-economic" attraction of investment funds to the ECM in order to circumvent the discriminatory tariff wall. In the absence of the ECM, such funds would have been invested at home (or elsewhere) as dictated by considerations of comparative advantage. It is probable that the stronger the internal growth effect in the ECM, the larger would be the opportunities for American (and other) capital investments. The part of this capital which is attracted to Europe because of the discrimination constitutes the investment diversion. Its size would depend on the degree of tariff discrimination.
The total of the dynamic effects would depend on the final readjustment of world trade patterns following the establishment of a customs union. These effects do not readily lend themselves to empirical measurement. But it cannot be claimed a priori that their net effect on non-member countries is favorable in all or even most cases. That is, the "dynamic consolation prize" usually offered to non-participating nations which complain of the static discriminatory effect is not always valid. Furthermore, even with regard to the dynamic impact of a customs union on total world welfare, it cannot be concluded that the favorable effects exceed the unfavorable ones. Thus the commonly held view that the over-all dynamic impact of a customs union is favorable is also in error.
References
Mis en ligne le 01/08/2014 
pratclif.com