The Single Market Programme in Europe (EC-92) was an ambitious attempt to create a common market. While EC-92 directives appeared to be outward-looking in their orientation, developing countries feared that the programme might lead to considerable trade diversion. Usually based on estimates by the Cecchini Report, ex ante analysis of the implications of EC-92 for the exports of developing countries generally suggested a positive effect, with strong EC-92-generated import growth swamping any diversion of developing-country exports due to the enhanced price-competitiveness effect. In this article, we revisit the question using two ex post approaches, that is an import-growth approach and a gravity model. According to our results, EC-92 directives appear to have had a negative effect on the exports of ASEAN and China, but not on those of the Asian NIEs. As expected, the effects were concentrated in light manufacturing and in some cases electrical machinery. The effect on ASEAN and on developing Asia in total was supported by the gravity-model approach.
Conventional wisdom1 holds that one of the factors responsible for the 1997/98 Asian financial crisis was the rise in the current account and trade deficits of the afflicted countries.2 In turn, it is plausible that regional integration schemes in Europe and North America diverted enough exports from ASEAN and South Korea to contribute to their trade deficits. The same may be said about Mercosur in South America. Diversion of Direct
Foreign Investment (DFI) originating in the industrial countries is also possible if not probable. Thus, regional integration schemes may have contributed, albeit marginally, to the economic crisis in: Thailand, Indonesia, Malaysia, the Philippines (namely, the big ASEAN four) and South Korea.
This article, a part of a larger study, explores part of the puzzle: what diversionary effect did the Unified Market Programme in Europe, known as EC-92, have on the exports of developing countries in Southeast Asia and elsewhere? EC-92 consists of 284 directives issued by the EC secretariat designed to remove all restrictions on the flows of trade, capital and people within the 12 members of the Community, and provide for market unification through other means. The directives began coming into force in the later part of the 1980s. While they were all supposed to be implemented by 1993, some were delayed until the mid-1990s. Theoretically, EC-92 could have had both positive and negative effects on nonmembers, including Asian developing countries; hence the need for empirical assessment.
II Possible Effects of EC-92
While the Cecchini Report of the European Community claimed that the internal growth effect of the EC-92 programme would be so strong as to swamp static diversionary effects on non-member countries, this cannot be taken at face value and needs to be estimated empirically. A number of studies have postulated what might be the effects of EC-92 on third-country exports. Bleaney, Greenaway, and Hine (1995) delineate a number of possible changes stemming from the EC-92 programme that might serve to reduce non-partner country penetration of EC markets: improved competitiveness of EC firms in EC and non-EC markets; trade diversion induced by internal liberalization; market exclusion due to discriminatory harmonization of technical barriers; market exclusion due to discriminatory government procurement; market exclusion due to "reciprocity" requirements; protection induced by eliminating national quotas; and protection induced by adjustment pressures (p. 86).3 Because these effects go well beyond the traditional pricerelated effects of a free-trade area or customs union, empirical trade models are not well equipped to estimate their ultimate effect on trade. Hence, generating such estimates by traditional approaches has been difficult (as discussed below).
As the development of the EC-92 programme unfolded, non-member developed countries and developing nations became concerned about the possible emergence of a "Fortress Europe". This Fortress Europe might not stem from the traditional diversionary effects of a preferential trading arrangement; rather, it was feared that the creation of a community-wide set of commercial policies and the myriad compromises that would have to be made in order to push through the EC-92 programme would lead to a more protectionist policy environment affecting both trade and investment. Thus, even if the EC-92 programme itself were to be mostly free of any explicit discrimination against outsiders, the associated commercial policy of the EC might become more inward looking.
The European Community went through great pains to assure its non-member trading partners that it had no intention of becoming more protectionist. The programme itself appeared to be outward-looking, and the fears of EC trading partners were generally allayed.4 Nevertheless, except at the sectoral level, little ex post assessment has been made of the impact of EC-92 on policy formation in the EC.
III Previous Studies
The most influential study regarding the possible (ex ante) effects of the EC-92 programme was the above-mentioned Cecchini Report (1988), which estimates a one-time potential increase in EC GNP of 2.5-6.5 per cent to result from the unified market. Most macroeconomic simulations and other ex ante models use the Cecchini estimates as growth scenarios. However, these studies tend to focus mostly on the effects of EC-92 on member countries; far less research has been devoted to estimating the effects of the programme on non-members, with some exceptions.
For example, Kreinin and Plummer (1992) presented ex ante estimates of the effect of the second enlargement of the EC and of NAFTA on ASEAN and South Korea. They found considerable trade diversion from Asia (they did not estimate trade creation), but it was impossible to assess and quantify the effect of EC-92 ex ante, using their technique. On the other hand, their highly-disaggregated data showed considerable commodity overlap between the exports of developing Asian countries and developed countries in Europe and North America. In a report for the Asian Development Bank (Staff Paper No. 48, 1991), Verbiest and Tang use a macroeconomic simulation model to estimate the effects of the completion of the internal market in the EC on Asian developing countries. While the paper suggests that "protectionist temptations" may increase as a result of EC-92, overall a positive impact on developing Asia should be expected over the medium- to long-run, assuming the Cecchini growth estimates are in the correct range. Verbiest and Tang look to the Uruguay Round and the multi-fibre negotiations as "more important than the Single Market operation in shaping the future trading relations between Asian developing countries and the EC".
Sung-Taik Han (1992) also uses the Cecchini results to simulate the effects of EC-92 on the Asian NIEs in a macroeconomic model. The price and income effects of EC-92 suggest that the Asian NIEs will initially experience a deterioration in exports due to the negative price (cost-reduction) effects, but this effect will eventually be swamped by the positive income effect. However, his microeconomic, industryspecific approach yields significant negative effects on important industries such as electrical goods and office machinery, a result which is consistent with the Kreinin and Plummer effects discussed above.
In a general way, given the stagnation of the European economy in the 1990s, it is difficult to assign high credibility to estimates based on the Cecchini growth projections. Indeed, Plummer (1994) in an ex post aggregative study, found considerable trade diversion to have resulted from EC-92.5 But his paper was based on 1991 data, before many EC-92 directives had been implemented. In contrast, the present study uses 1994 data (with results confirmed for 1996) - both in the aggregate, and with some industrial disaggregation. Another ex post study on the effects of EC-92 was undertaken by Sapir (1997), using data over the 1986-92 period.6 He focused mainly on the internal effects of EC-92 on member countries and the study is preliminary in nature as it does not include post-1992 data. Sapir finds that, at the sectoral level, EC-92 actually increased extra-EC imports in the case of sectors where public procurement was important, but penalized extra-EC imports in sectors subject to medium or high non-tariff barriers (see also Sapir 1997 for a survey of empirical models of EC-92).
Hence, apart from a few studies, the ex post effects of the EC-92 programme on non-partner countries have been sparse. In what follows, we apply two independent models and employ postEC-92 data to capture the effects of EC-92. These models do not allow us to distinguish between the direct effects of the EC-92 programme and policy changes occurring at the same time, emanating both from EC-92 and other sources. For example, any negative effects attributed to the EC-92 programme may have derived from an unrelated policy change. Nevertheless, EC-92 constitutes a fundamental change in the commercial policies of the EC, that is, a watershed in the process of creating the common market for goods, services, and productive factors. As such, it might be considered at least indirectly responsible for the resulting new configuration of the EC policy structure.
IV Methodology: Control Country Approach
As a first approximation we observe what happened to EC-12 imports from various developing countries between 1990 and 1994 and between 1990 and 1996. The year 1990 is a possible base year before many EC-92 directives were fully implemented, while the two later years should reflect the effect of the Unified Market Program in all its directives (1996 is the latest year for which the necessary data are available).
But many other supply and demand factors influence the change in EC imports between the two specified years. To allow for such effects we search for a "control country" - a country that is similar to the EC in all or most respects. We expect the change in imports into the EC, without the EC-92 programme, to equal the change in the control-country imports. Thus the difference between the change in EC imports and the change in the control-country imports is attributed to the EC-92 programme.7
While the United States appears as an appropriate control country, its own trade may have been distorted by the formation of the free trade area with Canada, even though the lion's share of bilateral trade between the two countries took place duty-free before the agreement went into effect (making the effects on trade small). Therefore we use the United States plus Canada, netting out their trade with each other, as control. In other words the control region is North America (labelled NA) in its trade with the outside world. Similar to Europe in population, real income, degree of industrialization, and other respects, NA has been used before8 as a control country to study European integration.
However, the control-country may require adjustments to the extent that it differs from the EC in the variables that determine imports: change in competitiveness and real income between the two years under study, that is 1990 to 1994 and 1990 to 1996. With respect to competitiveness, the differential inflation between the two regions is exactly offset by the differential exchange-rate changes. Hence no adjustment is needed. But the real growth rate in NA exceeded that of the EC by 4 per cent between 1990 and 1994 and by 6 per cent between 1990 and 1996. In making the adjustment for the differential growth we assumed an income demand elasticity of 1 for all imports, and of 2 for imports of several manufacturing categories. Indeed, estimates will be provided for aggregate imports, and (disaggregated) for six important classes of manufactures. In a later section the results will be checked against those obtained from a totally different approach: that of a gravity model.
In sum the estimating equations used in this section are:
where all notation is the same as in (1), with the addition of commodity k (SITC 5, 6, 71, 72, 73, 8). Recall that NA stands for North America.
While we considered Japan as another possible control country, it diverges so much from the EC over this period in terms of changes in degree of competitiveness, as proxied by changes in the real exchange rate, that its use was considered inappropriate.9
Table 1 presents the estimates for aggregate imports. The results are mixed. Japan sustained significant export diversion, which was considerably larger for the 1990-94 period than for the 1990-96 period. Among the developing countries there was no negative effect on the NIEs, including specifically South Korea, a country heavily afflicted by the 1997 financial crisis. If anything they benefited from the EC growth effect. Possibly the positive effect on Korea (especially in SITC 73) occurred because of EC quantitative restrictions against Japanese cars. On the other hand, China and the ASEAN group were negatively affected. And among ASEAN, the largest diversionary effect was on Malaysia and Thailand. In Latin America, Mexico and Chile were affected negatively, but the effect on Mercosur was negligible and in particular there was no effect on Brazil. Others affected negatively were North Africa, India, Pakistan and Israel.
In total, the negative diversionary effect far exceeds the positive. The EC-92 programme thus contributed at the margin to the trade deficits of Thailand and Malaysia, but not of Korea.
Table 2 presents the results for six manufacturing commodity categories. It shows that the diversionary impact on Malaysia is concentrated in electrical machinery (SITC 72) and miscellaneous manufactures (SITC 8), while in the case of Thailand it is in SITC 8. The NIEs lost ground in machinery (SITC 71, 72) but benefited elsewhere, while the loss of China appears everywhere but is concentrated in SITC 6 and 8 (probably labour-intensive manufactures). Brazil and Chile lost ground in SITC 6.
VI Gravity Model: an Alternative Approach
It is possible to check the direction of change, but not the magnitude, of our result by using a totally different approach, that of a gravity model. The model postulates that the intensity of trade between two countries depends on their respective gross domestic products (GDP), populations, and the geographical distance between them. It is a much criticized, yet widely used, approach to trade analysis. Our sample contained 5,112 bilateral trade flows each year and we estimated a gravity model for each such flow. To the above variables we added dummy variables for the ASEAN-EC trade, and for developing Asia-EC trade, where developing Asia includes ASEAN, the Asian NIEs, China, and South Asia (India, Pakistan, Bangladesh, and Sri Lanka), to find an EC-92 effect. In the overwhelming majority of cases, the estimates were robust, and the regression fit was good, with an adjusted R2 of between 0.45 and 0.75 - very respectable for cross-sectional data and on par with other gravity models.
As mentioned earlier this approach has been subject to much criticism, not the least of which is the absence of a relative price or real exchange rate variable. Additionally, the data may reflect supply-side influences within ASEAN (or developing Asia) rather than demand effects in the EC. To overcome such concerns and to legitimize the gravity approach at least to some degree, we ran separate gravity regressions with a binary variable for ASEAN-North American trade. It is the comparison between the coefficient of this variable and that of the ASEAN-EC variable that would indicate any EC-92 effect on ASEAN. The same procedure was used for developing Asia.
These two coefficients for the ASEAN dummy, all of which are statistically significant at the 95 per cent level, are plotted in Figure 1. It shows that in the 1980s both coefficients tracked the same path. But at the end of the decade, when EC-92 directives began to be implemented (gradually) the two series diverge, with the NAASEAN dummy advancing far ahead of that of the EC-ASEAN. This confirms the negative effect of EC-92 on ASEAN exports estimated by the normalizer approach.
Figure 2 plots the same binary variables for EC-developing Asia, and for NA-developing Asia, and a similar pattern appears. It confirms the negative impact of EC-92 on Asian exports. VII Conclusions
EC-92 directives appear to have had a negative diversionary effect on the exports of ASEAN (especially of Malaysia and Thailand) and China, but not on those of the Asian NIEs. As expected, the effects were concentrated in light manufacturing and in some cases electrical machinery. The effect on ASEAN and on developing Asia in total was confirmed by the gravity model approach. There remains the possibility of investment diversion which is a negative effect additive to that of trade diversion.
Finally, there is every reason to believe that the North American Free Trade Agreement (the United States, Canada and Mexico free trade area) would cause as much or more diversion of trade from Asia, because there is a great deal of commodity overlap between the exports of developing Asia and that of Mexico and Canada.10 In a more general way our results militate against the regional approach and in favour of the multilateral approach to trade liberalization. Finally, in as much as regional groupings are here to stay, the best strategy for Asia or for developing countries in general is to press in the WTO for open regionalism, and for a minimum of discrimination against outsiders.
The authors wrote this article as visitors at the Institute of Southeast Asian Studies (ISEAS). They are indebted to ISEAS for financial support, and to Prof. Chia Siow Yue, Manuel Montes, and Carolyn Gates for their generous help and substantive comments throughout the project. Ms Sangita Rao and the ISEAS staff provided excellent clerical assistance.
1. See Manuel F. Montes, The Currency Crisis in Southeast Asia (Singapore: Institute of Southeast Asian Studies, 1998) and Hal Hill, "Southeast Asia's Economic Crisis: Origins, Lessons, and the Way Forward", paper delivered at the ISEAS Thirtieth Anniversary Conference, Singapore, 30 July-1 August 1998. 2. Among the other proximate causes are the inflow (followed by outflow) of short term portfolio capital; high ratio of foreign liabilities to international reserves; a high proportion of non-performing loans of the banks; and fixed or targeted exchange rates set at unsustainable levels.
3. M.F. Bleaney, D. Greenaway and R.C. Hine, "The Impact of the 1992 Programme on Non-EC European Countries: An Overview", in Economic Integration in Europe and North America, edited by M. Panic and A. Vacic (New York: United Nations, 1995).
4. See, for example, Michael G. Plummer, "Economic Deepening and Widening in Europe: Implications for the Asia-Pacific Rim", in Effects of European Integration on Asia, edited by M. Toida and D. Hiratsuka (Tokyo: IDE, 1994).
5. Plummer (1994), op. cit.
6. Andre Sapir, "The Effects of Europe's Internal Market Program on Production and Trade Trade: A First
Assessment", Journal of Common Market Studies (1997).
7. M.E. Kreinin, "Effect of the EEC on Imports of Manufactures", Economic Journal (1972). 8. M.E. Kreinin, "Trade Creation and Diversion of EEC Enlargement", Kyklos (1984). 9. Also, it could be argued that Japan's import structure is so different from that of Europe as to make it an even less likely candidate.
10. See Kreinin and Plummer (1992), op. cit.
Mordechai E. Kreinin is University Distinguished Professor, Department of Economics, Michigan State University. Michael Plummer is Associate Professor of Economics and Director, MA Programs, Graduate School of International Economics and Finance, Brandeis University.