Catalog

	Table of Contents
p. 75 of 81 BACK | NEXT I. Developing Asia and the World II. Economic Trends and Prospects in Developing Asia III. Preferential Trade Agreements in Asia and the Pacific Overview Trends in Trade and the Expansion of PTAs: Trends in World and Asian Trade Flows Diversity in of PTAs The Economic and Broader Effects of PTAS: Theoretical Arguments Effects of PTAs on Trade in Asia and the Pacific: Some Evidence Trends in Intrabloc Trade Trade Intensity Indices Gravity Model Estimation Beyond Gravity Models: Computable General Equilibrium Models Conclusions References

Asian Development Outlook 2002 : III. Preferential Trade Agreements in Asia and the Pacific : Effects of PTAs on Trade in Asia and the Pacific: Some Evidence

---

Gravity Model Estimation

Gravity model estimation provides a useful multivariate framework for assessing the impact of PTAs on the level and direction of trade.⁸ Gravity models estimate the level of trade between two countries and assess the impact of policy characteristics—such as membership of a given PTA—on trade flows. The model is based on the idea that trade between two countries, like the gravitational force between two objects, is a function of the countries’ “mass” (in this case population size and GDP) as well as the distance between them. While lacking microeconomic foundation, the basic gravity model has been widely applied as a predictive model for estimating the level of trade between two countries—particularly when other control variables such as measures of the country’s remoteness from major trading centers—are added.

The model used in the present research included an important innovation that distinguishes it from other attempts to use gravity models to explore the effects of PTAs on trade. Extending the model of Soloaga and Winters (2001), the new model is able to capture the specific effects of PTAs on trade in the Asia and Pacific region. Its estimates can determine whether PTA membership influences a given country’s trade with Asia in a different way from the way in which it affects its trade with the rest of the world.

It is beyond the scope of this chapter to discuss the estimates in detail, and instead this section will summarize the main results of the estimates and indicate the policy implications of the results.⁹

Data Issues. Eighty-three countries are included in the gravity model estimates. Bilateral exports for every pair of these countries are extracted from the IMF Direction of Trade Statistics database from 1980 to 2000. The number of observations varies per year. The export values are expressed in real terms based on a merchandise price index (base year 1990) obtained from WTO’s International Trade Statistics 2001. Estimates were made using single-year cross-sectional data, as well as panel data constructed at 5-year intervals (1980, 1985, 1990, 1995, and 2000).

Basic Determinants of Trade Flows. The estimation results for the basic variables identified in the gravity model as driving trade flows—namely GDP, distance between capitals of trading partners, physical area, and sharing a common border—have the expected signs and are highly statistically significant. This supports the validity of the gravity model, and model estimates generally explain cross-country trade flows well. Distances, the size of the economy, and whether countries share a common border all exert a statistically significant influence over trade flows in the direction expected. Between 68% and 73% of the variance in trade flows was explained by the basic gravity model across the various new estimates.¹⁰

PTAs fostering intrabloc trade, except APEC, had positive and statistically significant coefficients for intrabloc trade, had hardly any change in their overall imports, and increased their exports to the world over the period covered in the data. The Andean Pact, ECO, Mercosur, and SPARTECA had the largest effects on intrabloc trade and on trade between the PTA-member countries and the rest of the world. In the Asia and Pacific region, ECO, SAPTA, and SPARTECA had the largest effect in increasing intrabloc trade, while AFTA and CER showed no significant intrabloc trade effect.

ECO countries tended to trade more intensely among themselves at the expense of trade with the rest of the world. Estimates showed that intrabloc trade was higher at a statistically significant level in 1995 and 2000 than would be expected if the countries were not members of ECO. However, it must be recognized that after SAPTA, ECO had the lowest intrabloc trade share of the PTAs considered as its historical orientation was to export to the Russian Federation and, despite the adoption of market-oriented policies, the subregion has not yet become well integrated into the world economy.

The large intrabloc trade effect estimated for SPARTECA can be attributed to the heterogeneity between Australia and New Zealand on the one hand, and the rest of the members in this PTA on the other. Trade tends to flow more intensely among SPARTECA’s smaller member economies, but the potential for more trade among them is limited by the small size of their domestic markets. SAPTA and EFTA are also included in this category, even though these PTAs have lower gross intrabloc trade effects than the other PTAs.

This strength of the intrabloc trade effect is evidence of the size of the trade-diverting effects of PTAs and this evidence is generally in line with the results reported by Soloaga and Winters (2001). The effects of SAPTA membership on intra-SAPTA exports calculated in the estimation model are all statistically significant and positive. At the same time, it should be recognized that trade levels between some members of SAPTA are extremely low. For example, Frankel and Wei (1998) showed that India and Pakistan are trading at a 70% lower rate than would be expected given their size and proximity. In general, though, the estimates are merely indicators of how much more trade there is among SAPTA members than would be expected if SAPTA did not exist. It does not necessarily mean that overall levels of intraregional trade are high. The share of intra-SAPTA exports was only 4.8% in 2000, the lowest of any of the PTAs considered in this chapter.

Mercosur appears to be the PTA that has had the greatest effect in stimulating intrabloc trade and trade between its members and the rest of the world. This effect is brought out by the “trade multiplier,” which captures the extra trade associated with members of a given PTA. Across all PTAs, the trade multiplier averaged 7.8, compared with 20.8 for Mercosur. Preusse (2001) observed that intra-Mercosur exports expanded by 28.4% between 1995 and 1998. He attributes this outcome to the increasing significance of regional production. The effects of the Asian financial crisis and sociopolitical problems in the region cut short the impressive gains made following the formation of the customs union in 1995. Yeats (1998) noted that intrabloc trade in Mercosur increased at the expense of trade with nonmember countries and caused significant trade diversion. This stemmed from the group’s “discriminatory tariffs against nonmembers, which are four to six times higher” than those in other major PTAs. The trade bloc appeared to reinforce the inward orientation of Mercosur’s economies, enabling inefficient domestic producers to expand markets within the bloc while remaining uncompetitive on the wider international scene. The decline in Mercosur’s trade (both within and outside of the bloc) in the late 1990s reflects the impact of accumulated inefficiencies sustained through relatively high discriminatory protection. The Mercosur case highlights the importance of examining broader economic policies—rather than just looking at trade flows—when analyzing the impact of PTAs.

Andean Pact countries also exhibit strong tendencies to trade among themselves. Although there seems to be no evidence that this came at the expense of any particular trading partner of the Andean Pact members in the 1980s, there is evidence that trade diversion at the global level increased in the 1990s.

In terms of PTAs fostering greater trade with trading partners worldwide, membership of APEC and the EU was estimated to significantly expand trade both to other members of the PTA and to the rest of the world. The finding for APEC confirms results obtained in Frankel and Wei (1998) and Soloaga and Winters (2001). APEC is built on the principle of open regionalism, and estimation results for the PTA suggest the

The EU is sometimes considered to be trade diverting, but the estimates from the gravity model showed EU members as associated with increased intrabloc and total trade. Soloaga and Winters’ (2001) results indicate that EU membership has reduced the member countries’ imports from non-EU members although Sapir (1997) attributed this to the increase in the number of member countries in the EU during the 1990s.

As regards PTAs that reduce gross trade with the rest of the world while leaving intrabloc trade unchanged, Frankel (1997), Frankel and Wei (1998), and Soloaga and Winters (2001) found that AFTA expanded intrabloc and total trade. However, the estimation results showed that AFTA reduced trade outside of the trade bloc and did not change intrabloc trade relative to the level expected in the absence of the PTA. A possible reason why these results differ from those of earlier researchers is that these estimation results include data that cover ASEAN’s newest members—namely Cambodia, Lao PDR, Myanmar, and Viet Nam. Furthermore, the data include the period of the Asian crisis and subsequent recovery. Both could have had a possible impact on the trade-creating effects of AFTA. These newest members are economically less developed than the founding ASEAN countries and are, on average, less integrated into the world trading system. Their inclusion in the gravity model estimates may have diluted the effect of ASEAN on its trade within and outside of the PTA. Furthermore, during the Asian crisis the rate of growth in external trade of ASEAN fell significantly, creating a downward bias to the contribution of external trade in the estimates.

PTAs’ Contribution to Asian and World Trade. When PTAs are considered as a whole, empirical estimates show that PTAs both within and outside of the region have augmented trade in the Asia and Pacific region. Estimation results provided little support for the assertion that PTAs have diverted trade to member countries at the expense of non-PTA trade. The general effect of PTAs on Asia and Pacific trade is quite small compared with the effect of non-Asian PTAs in other regions. Overall, PTAs appear to have increased aggregate trade flows worldwide, and gravity model estimates provide evidence that PTAs create rather than divert trade globally.

The effect of the EU on Asian trade suggested by the estimates indicates a more complicated picture. The EU was found to have a strong positive effect on trade flows between member countries and the rest of the world—which includes greater trade to Asia. However, the specific effect of the EU on Asian trade was found to be negative and significant. Overall, the total world trade effect dominates the Asia-specific effect, making the net effect of the EU on Asian trade positive. The Everything But Arms initiative for EU member countries to grant nonreciprocal trade preferences toward least-developed countries could carry some negative consequences for the EU’s trade with a number of DMCs excluded from the initiative.

Back Trade Intensity Indices	Next Beyond Gravity Models: Computable General Equilibrium Models