C E N T R E   D' E T U D E S   P R O S P E C T I V E S
E T   D' I N F O R M A T I O N S   I N T E R N A T I O N A L E S

FOCUS

Inequality and Openness

   Inequalities and their link to economic openness are among the major controversies provoked by the globalisation process. A first issue to be addressed, however, is related to the measurement of inequalities: according to some authors, inequalities are decreasing, while others state that they are increasing. How is this possible? Much depends on the concept of inequality which is used.

  In "L'Economie mondiale 2005" (2004), Isabelle Bensidoun sets out the different concepts of inequalities: between countries, between populations of countries and between individuals. This explains partly why results in the measurement of inequalities can appear so contradictory. She surveys the empirical and theoretical elements which may relate openness to inequalities.

   Loïc Cadiou and Pierre Villa (1998) compare results using the Gini coefficient and the Theil index. These indicators all more or less show that there has been a decrease in international inequality when weighted by population, but an increase in inequality when one country provides a single observation, whatever the size of its population.

   Isabelle Bensidoun, Agnès Chevallier and Guillaume Gaulier (2001) highlight the distribution of GDP per capita and its evolution between 1960 and 2000 throughout the world. If in 1960, many countries were concentrated around the average GDP per capita, by 2000 countries could be split into three different groups: some middle-income countries had converged towards the upper income group while others had fallen behind. Meanwhile, the gap between average incomes in the richest countries and average incomes in the poorest countries had grown (from 100: 6.9 in 1960, to 100: 5.3 in 2000).

   Is economic openness or isolation the main explanation for these trends?

   Proponents of liberalisation depict international trade as a strong factor favouring growth and thus economic catching-up by developing countries. Integration in international markets is seen as a way to get access to modern technology facilitating the modernisation of lagging economies, and so ultimately accelerating the convergence process. These countries could thus be the main beneficiaries of economic integration. But, as pointed out in the theoretical literature, countries lacking a minimum industrial base could actually suffer from economic integration and be caught in poverty traps as a result.

   In a seminal study, Dan Ben-David (1996) observed that trade-based groups of countries (groups comprising major trade partners) are more likely to exhibit convergence than groups of randomly selected countries. The robustness of this result is assessed in Guillaume Gaulier (2003). Ben-David’s finding of absolute (beta-) convergence within trade-based groups is confirmed but it is not sufficient to imply a significant reduction in standard of living differences (sigma convergence). Second, Guillaume Gaulier argues that the extent of convergence within trade-based groups depends on the concept of trade intensity which is used to build these groups. Little evidence of either beta- or sigma- convergence is found when the size of countries in trade groups is controlled for. Finally, the evidence for trade as a channel for convergence remains limited.

   Furthermore, as suggested by the modern trade theory, trade specialisation matters (Isabelle Bensidoun, Guillaume Gaulier and Deniz Ünal-Kesenci, 2001): exporting goods facing weak demand (generally primary products like minerals or food & agriculture), and having a low technology content will hardly lead to positive, long-term growth performance. According to these results, policies supporting specialisation in growth-engine sectors, in which a country lacks comparative advantage, might have a positive impact on long-term growth.

      The widespread increase in "within countries" inequalities in the developed countries has been well documented, while developing countries have experienced different outcomes, broadly characterised by increased inequalities in Latin America, and decreased inequalities in East Asia. Understanding the link of such evolutions with foreign trade is the subject of on-going research at the CEPII (Isabelle Bensidoun, Sébastien Jean and Aude Sztulman).

ON THE RESEARCH AGENDA

China’s Integration Into Asian Production Networks

   This research project analyses the rapid progress and the consequences of China’s participation in the international segmentation of production processes and of its integration into Asian production networks. The first section of the research points out how globalisation provides new opportunities for latecomers to enter international trade through production sharing, which has been an especially widespread phenomenon in Asia. The second section investigates how China has taken advantage of the globalisation process and assesses the consequences for bilateral trade imbalances of China’s integration in Asian production networks.

   An in-depth analysis of trade flows (using Chinese customs statistics) leads to the following conclusions: international processing activities, based on inputs imported from Asian countries, have been the engine of China’s trade expansion, allowing for a rapid diversification of its manufacturing export capacities; foreign affiliates have been responsible for a major and ever-growing part of China’s export growth, especially with Asian countries; China’s position in “triangular” trade is characterised by growing deficits with Asia in intermediate goods, especially in parts and components, and by rising surpluses with “Western” countries in final goods, mainly consumer goods but also increasingly in capital goods; imports of parts and components from Asia have been the major channel of technology transfers, which have helped China to improve rapidly the high-tech content of its exports. However the technological upgrading of China’s trade has remained quite circumscribed to foreign firm's export bases.

   The third part of the project focuses on the changes in Asian country trade patterns over the last twenty years. It highlights the intensification of intra-regional trade in intermediate goods and the rise of late-comers (China and the ASEAN countries) as suppliers of final goods both to the region and to the rest of the world. China’s participation in the Asian division of labour follows a pattern which appears to be quite similar to that of the ASEAN countries, but the size of China’s economy gives the country a trade potential which has considerably affected “triangular” trade, as Chinese exports to Western markets have displaced Japanese and NIE exports at a fast pace and are overtaking all other Asian exporters in sectors such as electrical and electronic goods.

   China’s emergence thus carries the mechanisms of triangular trade to extremes. Moreover, in the late 1990s, it was accompanied by a levelling off of exports from the second tier of new industrialised economies (the ASEAN countries excluding Singapore), suggesting that the latter may be crowded out of these markets by China. The fourth part of this work sets out the implications of China’s rise in triangular trade for Asian regional integration and for international trade imbalances.

The Regionalisation of International Trade and Investment

   Regionalisation is a widespread feature of international trade. Among the eighty countries considered, all but ten carry out more than half of their foreign trade with one of the world’s three major trading regions (the Americas, Asia-Oceania and Eurafrica). This regional polarisation is especially strong in Eurafrica, where the region accounts for more than 75% of foreign trade for most countries therein. The regional polarisation of foreign trade is more limited in the Americas, but it is still significant, particularly for the US’s neighbours, Mexico and Canada. Asia-Oceania is the region exhibiting the weakest polarisation. However, intra-regional trade accounts for around 55 to 60% of total trade, except for the largest three economies in the region (Japan, China, South Korea).

   In order to qualify these observations further, we use relative trade intensities (RTIs). The RTI index characterises the intensity of trade relationships between a pair of partners by comparison with the total trade of each of these partners. As such, it refers to the geographical orientation of trade flows, controlling for total trade flows of both partners. The analysis points to the especially intense trade links of former communist countries with each other, and with Western Europe. It also emphasises the relatively intense trade links within Latin America, and between many Southeast Asian countries.

   Since distance is an obstacle to trade, countries trade more intensively with their neighbours. The regionalisation of foreign trade is to some extent a “natural” phenomenon, in the sense that countries tend to trade in large part with other countries belonging to the same "region". But there is more to regionalisation than natural trading relationships with neighbours. Regional trade arrangements (RTAs) might also have contributed to strengthening trade relationships within regions.

   The number of RTAs surged dramatically in the 1990s. By 2003, up to 189 agreements had been notified to the WTO. While the increase during the last decade has been spectacular, RTAs are not a new phenomenon. It has become usual to distinguish three waves of regionalism: the first beginning with economic integration of Western Europe, and the second with economic integration in North America. The surge observed since the mid-1990s, resulting mainly from agreements signed by former communist countries, corresponds to the third wave. Major agreements are briefly reviewed, emphasising inter alia the importance of agreements signed in Latin America.

   In order to gain further understanding of how these agreements may have shaped trade patterns, we provide an empirical analysis for a large number of countries covering the whole world, for the period 1967-2001, based on the CHELEM-CEPII database. This analysis updates and extends the work carried out in Freudenberg et al. (1998).

   The most salient particularity of the analysis is probably to control for country-pair, fixed-effects when assessing the impact of RTAs on trade flows. In other words, the estimates only focus on the intensity of the trade relationship between any country-pair, compared to its mean level across the period. Working exclusively for any country-pair, in comparison to the mean level of trade intensity, makes it possible to control for any pairwise-specific, time-independent country links, such as distance, language, former cultural ties (colonial for example), etc. This is a significant improvement in methodological terms, which has never been carried out before on a world-wide basis for such a long time span, to the best of our knowledge, except by Freudenberg et al. (1998), Fontagné et al. (1999) and Cheng and Wall (2003). Introducing these country-pair, fixed-effects is found to have a very strong influence on the results.

   The results allow differences across agreements to be characterised. The only important feature, common to all the agreements studied, is to strengthen trade relationships between members, but this is almost a tautological result. The EU, EFTA and MERCOSUR, strongly boosted trade between members as far as final goods are concerned. With regards to third countries, they induced trade creation (except in the case of the EU regarding primary products). NAFTA has also spurred significantly trade between its members, but it is difficult to reach general conclusions about its effects on third countries. Nevertheless, the mixed impact of NAFTA on trade with third countries stems mainly from the poor export performance (in particular of the US) during the 1990s, compared to the predictions of the model. In terms of imports, NAFTA is unambiguously found to have induced trade creation. The ASEAN induced strong trade creation with third countries, and this trade creation consisted of imports of primary goods and exports of transformed goods, pointing to the increased division of labour across member countries.

   In contrast to many previous studies, no clear evidence is found of the trade-diverting impact of RTAs. The Australia-New Zealand CER ends up associated with lower trade with third countries, but this is mainly linked to lesser exports, and it is not clear whether this should be interpreted as trade diversion. The Andean Community induced trade diversion for transformed goods, but this seems to have been more than compensated for by trade creation in the other production stages. The COMECON is found to have substantially diverted imports in final products, and this effect was apparently only partially balanced by import creation in primary and intermediate goods.

   Given the contrasting results across agreements, it is not possible to make any forecasts about the possible effects of future agreements. It is necessary to proceed by analogy with existing agreements in order to gain further understanding of any particular agreement’s likely impact. Broadly speaking, the results show that large RTAs generally create new opportunities for foreign producers, given the possibilities they offer to access a large market in a single country. This is mainly the case when the RTA is also a customs union: market access to member countries is harmonised, and re-exports within the customs union are not restricted, thus creating a single market from the point of view of the exporter. When an FTA is not a customs union, each member country maintains its own pattern of protection, and therefore imposes local content requirements on its imports from other member countries. This limits the advantages third-country exporters can draw in terms of easier market access. These conclusions should, however, be considered as tentative: strictly speaking, the estimates do not provide insights about the impact of future agreements, they only assess the impact of past agreements.

   The dynamics of regionalisation are also studied. In this respect, prima facie evidence points to the outstanding increase in the share of intra-regional flows in Southeast Asian foreign trade. This raises the question of whether it is due to any trend specific to this region. The estimates proposed allow this evidence to be put into perspective. In fact, the increase in intra-regional trade in Southeast Asia is inferior to what would have been predicted, based on the changes witnessed in economic size, income levels, and similarities in trade specialisation, as well as the enforcement of the ASEAN FTA. In North America, an upward sloping, unexplained regionalisation trend is found until the mid-1970s, while no such trend is found in Western Europe. These results emphasise the leading role played by the institutional process of regionalisation in North America since the late 1980s, and in Western Europe with its periphery, throughout the whole period of 1967-2001.