Speech

Making Regional Economic Integration Work, Speech by Agustín Carstens, Deputy Managing Director, IMF

January 12, 2005

Good afternoon. Thank you for that kind introduction. Let me first of all thank Dr. Kemal for inviting me to deliver this prestigious lecture. This is my first visit to Islamabad, although I have followed developments in Pakistan closely over the past few years. I am very pleased to be here today and I am grateful for the warm hospitality shown by the Pakistani government and the Pakistan Institute of Development Economics.

The topic of this conference, regional cooperation and economic growth, is a timely and important one for this region, and for Pakistan in particular. The government of Pakistan has in the past few years made good progress in turning the economy around from crisis to strong growth. But the challenges ahead are still large, and strong, noninflationary economic growth will be needed to help reduce poverty. Trade plays a key role in achieving more rapid growth. In this regard, the Fund strongly welcomes the Framework Treaty on creating a South Asian Free Trade Area (SAFTA) that was signed a year ago, as it provides a major opportunity to enhance trade and thus economic growth. So I want to use this opportunity to say something about how to make regional integration work, drawing on economic theory and from experiences with regional integration around the world.

I would like to start by observing that parallel with the globalization that has been in full swing over the last two decades, the world also witnessed a new wave of regionalism. Almost all countries have sought integration with other countries, and not necessarily always with neighbors. Preferential trade arrangements (PTAs) have been the central elements of all regional agreements. Furthermore, successful trade agreements have in some cases evolved to a customs union, a common market, and finally to an economic and monetary union.

In this context, it is not surprising that South Asian countries are also striving for greater regional integration. Despite geographic proximity and cultural affinity, South Asian countries, including Pakistan, currently barely trade with each other. Regional trade is substantially lower than in South East Asia and even the Middle East and North Africa. The Framework Treaty on creating SAFTA signed in January 2004 represents an important effort to enhance regional cooperation. Against this background, it may be useful for policy makers, academics, and society in general to look at the growing international experience with PTAs and the implications for the design of SAFTA.

Despite the trend toward regionalization, many economists argue that regional integration can harm member countries if the evolving PTA does not pursue a policy of openness towards other countries. By restricting market access of more efficient producers from outside the PTA, they can lead to welfare losses at home and abroad. But economic theory also offers some broad insights for conditions under which PTAs are likely to be welfare enhancing. The main conclusion of the analysis that I will present to you is that regional integration, if done right, can work.

Moreover, if done right, PTAs can be stepping stones toward multilateral liberalization. Once formed, a PTA can be thought of as an entity participating in international trade just like individual countries constitute such entities. From international trade theory, it follows that the best trade policy for this larger entity towards the rest of the world is free trade. Thus, it is in a PTA's own interest to pursue multilateral trade liberalization.

In my presentation, I will first provide some basic facts on the extent of regionalism, and discuss motives for entering into regional arrangements. Then I will review the impact of regional integration on member countries based on traditional and new trade theories, and I will look at relevant empirical findings. I will also discuss the relationship between PTAs and the exchange rate system and the challenges for the International Monetary Fund (IMF) arising from regional integration. Finally, I will propose some lessons for making regional economic integration work.

Please allow me to start by providing some information on the remarkable proliferation of regional and bilateral PTAs. Since 1990, the number of PTAs has more than quadrupled to about 230 by late 2004. The World Trade Organization (WTO) estimates that another 60 arrangements are in various stages of negotiation. If all these come into effect, there will be nearly twice as many PTAs as there are WTO members.

Today, nearly every country belongs to at least one PTA and numerous countries are party to several agreements. The average number of PTAs per country according to recent World Bank estimates is six. Industrial countries have concluded the most PTAs, with an average number of 13 per country. A substantial number of developing countries have signed bilateral preferential agreements with an industrial country. Interestingly, there are as yet no countries in South Asia that have signed a bilateral agreement with an industrial partner.

Indeed it is interesting that an increasing number of PTAs include a large industrial country. This suggests the growing importance of a "hub-and-spoke" structure in world trade. In a hub-and-spoke system, the largest countries sign bilateral agreements with many small countries. Such a system could marginalize the spokes, where market access conditions are much less advantageous than in the hub, which enjoys improved access to all of the spokes. At the same time, however, a growing number of overlapping PTAs has resulted in a "spaghetti bowl" of trade relations. As a result, market access among smaller countries has been improved, but a complex myriad of trade rules and procedures has been established that can be difficult to administer and that may actually hinder trade.

Because of this increase in the number of agreements, the share of global trade taking place between PTA members has also risen and has reached almost 40 percent of total world trade. Intra-regional trade is particularly large within Europe and Asia: in the European Union (EU), about 60 percent of the countries' external trade remains within the region and in East Asia the share is about 50 percent.

Despite these trends, the share of global trade induced by PTAs is believed to be declining. This reflects the overall trend toward lower tariffs, especially in the United States and the EU. In fact, almost two-thirds of the decline in average tariffs in developing countries during the last two decades is estimated by the World Bank to have come from unilateral reductions, compared with one quarter coming out of the Uruguay round and the remaining 10 percent from PTAs. Thus, while 40 percent of global trade takes place within PTAs, the World Bank estimates that only about half of this intra-PTA trade is on a PTA-related preferential basis.

I will now speak about the motives for countries entering regional arrangements. The most obvious one is to enhance trade and thus to achieve higher levels of economic growth and material welfare. Countries participating in PTAs often seek to secure access to large markets, such as the United States or the EU. However, developing countries often already enjoy considerable access to these markets because trade barriers on industrial products are typically low. Or they benefit from unilateral measures, such as the Generalized System of Preferences, or the U.S. African Growth and Opportunity Act. At the same time, agricultural products, an area where developing countries could reap major benefits, are usually excluded from regional agreements. Still, the motive for increased market access, however partial, remains powerful for many, especially developing, countries.

A second reason for entering into a regional agreement may be to give a small country an advantage over other similar countries in attracting foreign direct investment (FDI). Raising the level of FDI-or domestic investment, for that matter-requires making a country attractive vis-à-vis other countries. Increasing market size helps in this regard. Ensuring market access to a major market by entering a PTA may be one way of achieving this.

Thirdly, a regional agreement can also help in dealing with region-specific issues, such as border controls, transit, migration, or movement of labor. Countries recognize that other, more opaque barriers than tariffs, can hinder trade. These include border controls, fito-sanitary restrictions, weak transport systems, and regulatory differences. PTAs therefore increasingly cover some of these issues, which are more suitably addressed at the regional level. Some PTAs have also included dispute resolution mechanisms, which, in the implementation phase of the arrangement, have proven to be extremely useful.

Fourthly, and very importantly for North-South arrangements, PTAs can reinforce internal regulatory or structural reforms. This can be done through external treaty obligations and visible political commitments. Often, small countries participating in a PTA have just made, or are trying to push ahead, major reforms. Locking in such reforms clearly motivated agreements between the EU and countries in Central and Eastern Europe. Another example is the case of Mexico in the context of NAFTA.

Turning to large industrial countries, trade in goods per se no longer appears to be the dominant factor for participating in PTAs. A growing number of PTAs includes provisions on liberalizing services (including financial), investment, protecting intellectual property rights, labor and environmental standards, and dispute resolution. Industrial countries are keen to include such issues to counter what they regard as unfair competition due to, for example, piracy or poor labor standards. They also desire to open up markets for their services sectors, where they have a comparative advantage.

Leaving aside economic goals for a moment, for some countries political objectives are another important reason to enter into a PTA. Countries that may have far-reaching integration as a goal typically start out with trade agreements as a first step toward a deepening of political relationships. The EU is a clear example of this: initial agreements covered trade and investment, then member countries formed an economic and monetary union, and now a process is in train towards a fully fledged political union with a common constitution. Similarly, forging bilateral and regional trade ties is often linked to geopolitical and security considerations. Trade policy is a key instrument of foreign policy for the EU and the United States, which use PTAs to secure regional stability by promoting the development of participating countries. South-South agreements also tend to reflect a political desire to participate in a regional initiative, such as the Association of South-East Asian Nations (ASEAN) and MERCOSUR in Latin America.

A last reason to enter into regional agreements may be defensive. As more and more countries enter into regional agreements, the cost of nonparticipation rises. While some countries may prefer the multilateral route, they may also feel that not entering into regional agreements can lead to a competitive disadvantage relative to countries that have entered into PTAs.

I will conclude the discussion of motives by saying that while PTAs are signed for a variety of reasons, the impact on trade, growth, and employment seems crucial in determining the extent to which broader objectives are achieved. It is difficult to identify arrangements that have advanced wider political objectives, without having first achieved progress in enhancing trade, and having this seen reflected in higher rates of sustainable growth and employment creation. Thus, it appears that the willingness to accept trade liberalization and the accompanying economic adjustments is a first step that may be indicative of progress than can be made in other areas.

I will now discuss, the impact of regional integration on member and nonmember countries. PTAs can lead to increased trade among members, but also to reduced trade with nonmembers. In addition to pure trade effects, PTAs can strengthen investment and growth of member countries, and may create pressures and opportunities for further integration, particularly in the financial area.

As you may know, the theoretical literature on regional trade integration is huge and covers every side of the argument. One way of looking at this is that economists are clever at fine-tuning assumptions to reach a particular conclusion. Another way is to read these diverging results as giving some indications under which conditions a PTA can be welfare enhancing. The diverging nature of results stems largely from different approaches to modeling. Models based on traditional trade theory tend to find PTAs being potentially welfare-reducing. However, models based on newer trade theories find PTAs to provide avenues for welfare improvements that arise in a dynamic context.

I will begin by restating the basic findings of traditional trade theory. PTAs can be welfare-reducing if they shift trade from a low cost source outside the PTA to a higher cost source inside the PTA. Lowering trade barriers in a PTA leads to increased trade between members. This is trade creation. As these countries specialize according to their comparative advantage, their consumption possibilities increase. However, lowering trade barriers in a PTA can also lead to trade diversion from a cheaper source in a non-PTA country to a more expensive source in a PTA member who still enjoys a price advantage because of tariff protection. Consumers still get the good cheaper than before the PTA was introduced so that they realize a pecuniary gain. However, the importing country loses tariff revenue by shifting to an intra-PTA source. A priori, it is not clear whether the direct gain for consumers is larger or smaller than the tariff revenue loss, in particular if the government has to increase taxes to make up for the lost revenue. Thus, the welfare impact of a PTA can be positive or negative.

A possible risk with respect to PTAs is that they can lead to a deterioration of member countries' terms of trade. In case a PTA leads to trade diversion from a low cost third country to a higher cost intra-PTA source, the member country's import price increases. This happens even though domestic consumer prices are lower on account of the tariff reduction. However, there is also scope for terms of trade improvements over time through productivity gains on account of increased import competition and technology transfers.

While risks exists, I would like to emphasize that PTA members can take steps to minimize trade diversion effects. Trade diversion arises because an external tariff provides a price advantage to an intra-PTA supplier. Therefore, adopting low external tariffs will minimize trade diversion by creating a more level playing field for intra- and extra-PTA suppliers. One rule of thumb could be that PTAs should adopt the lowest external tariff in effect in a member country before the PTA was set up. Trade diversion effects will also be small for countries that enter a PTA with a large partner (in terms of absolute size and diversification) because the large partner's relative prices are closely aligned with relative world prices.

Similarly, it is worth noting that trade diversion may also be low if PTA members are natural trading partners. Natural trading partners are countries that engage in substantial trade independent of the tariff regime. For example, neighboring countries tend to trade more because of transportation costs, cultural affinity, or similar levels of development and thus consumption preferences. As such, a PTA among natural trading partners provides substantial scope for trade creation, but little scope for trade diversion because member countries trade largely with each other and not with nonmembers.

Moving away from traditional trade theory, newer theories based on imperfect competition and economies of scale suggest advantages of PTAs, in particular in a dynamic framework. Traditional trade theory is typically based on perfect competition, constant returns to scale, and abstracts from innovations. As such, dynamic gains from having a first mover advantage, acquiring new ideas, large scale production or agglomeration do not feature. New trade theory explores these factors and how they interact with PTAs.

In particular, PTAs allow better exploitation of economies of scale. PTAs increase the market size for producers and allow them to move down the average cost curve by exploiting economies of scale. Assuming free market entry, this will benefit consumers in the form of lower prices. Moreover, investment can increase because the associated fixed costs are spread over a larger market size, which raises productivity and output. With heterogeneous goods, consumers will also enjoy more variety. And when allowing for innovations that involve significant up front fixed research and development (R&D) costs, a larger market also leads to increased R&D and thus innovation.

Continuing in the same vein, PTAs can also enhance investment and thus growth. A PTA can trigger increases in both domestic and foreign direct investment because of the increased market size. However, such investment is likely to be geographically concentrated within the PTA so that not all member countries may benefit alike. As such, a country with an initial advantage compared to other members could end up with most of the benefits resulting from a PTA. Creating a good investment climate then becomes important to attract investment and harness its benefits.

This type of analysis may be particularly relevant for South-South PTAs. These PTAs typically provide less scope for gains from trade on account of production complementarities or differences in factor endowments. However, South-South PTAs increase market size and can therefore lead to gains on account of economies of scale. Domestic investment as well as FDI may rise as a larger market makes higher up front costs profitable. Gains from specialization can be realized by breaking up the value added chain, including for commodities that will eventually be exported from the PTA.

The general lesson coming out of new trade theory is that it is advantageous for PTAs to offer open access to countries wanting to join. Once a PTA has been established, it has similar interests in regional and global trade liberalization as an individual country. As such, having a new member join the PTA provides scope for further trade creation. Therefore, building on the same rationale underlying PTAs in the first place, PTAs should welcome any country wanting to join as a general policy. Unfortunately, in practice, the process of joining can be quite complex and involve a variety of considerations.

I will now turn to of the impact of PTA. It is generally accepted that trade liberalization and openness are of course closely linked with economic growth. This proposition is widely supported by empirical studies. Therefore, one could also expect to find a strong link between PTAs and growth. However, the link between PTAs and growth is less researched. While some studies find that PTAs that substantially enlarge the market size have a positive effect on growth, others fail to find a positive impact. This may be an indication that PTAs do not always result in a significant increase in openness, or in some cases may even reduce the overall level of openness.

On balance, empirical studies seem to conclude that PTAs are not harmful, though not necessarily very beneficial either. Perhaps reflecting the theoretical uncertainties discussed above, empirical studies arrive at different results regarding the impact of PTAs. Very few studies find outright harmful effects from trade diversion. However, it also seems difficult to identify clear-cut positive effects. This may partly reflect the fact that regionalism, in particular in the 1990s, took place in parallel with multilateral opening and trends generally described as `globalization.' As such, PTA-specific effects may be difficult to separate from the multitude of other developments.

An interesting conclusion of the empirical studies is that many PTAs do not appear to lead to substantially increased trade flows between member countries. A large study covering 58 countries and nine PTAs found no indication that PTAs lead to an increase in trade volumes. Moreover, it did not find the new surge in regionalism of the 1990s to have resulted in increased trade between PTA members. Some more recent papers, however, did find some trade creation effects from NAFTA and MERCOSUR. And, looking at South-South type PTAs in central Europe and the Baltics, another recent study finds increased trade flows resulting from PTAs, suggesting that even in the absence of production structure and endowment complementarities, there is scope for trade creation.

Turning to trade diversion, there is mixed evidence of it. A number of studies analyzing the effects of NAFTA, MERCOSUR, and other PTAs find some trade diversion, although the effects are in most cases relatively small. Moreover, the conclusion seems to be that while recent PTAs are associated with trade diversion, this is typically dominated by trade creation.

But I have some even better news. There is some evidence that PTAs raise investment and therefore growth. Several studies find that the increased market size of PTAs has a positive impact on FDI received by member countries. This effect was more pronounced in the 1990s when PTAs became more widespread. Others, however, find only broad trade liberalization to lead to increased investment, while entering into a PTAs does not. A recent IMF study shows that FDI tends to be regionally concentrated within an PTA, emphasizing the need for countries to strengthen their investment climate when entering a PTA.

I would like to discuss now the impact of South-South PTAs. Some authors believe they may have less scope for gains from trade integration. First, they typically have much smaller markets than North-South or North-North agreements. Second, there is less scope for gains from comparative advantage because production structures and factor endowments tend to be similar. Third, South-South agreements may create less incentives to advance structural reforms than North-South agreements. Using a global model, the World Bank simulated the effects of South-South PTAs and found no overwhelming economic benefit from such PTAs for member countries. Since the theoretical gains in South-South arrangements may not be as large, such PTAs should put additional emphasis on lowering trade related costs such as customs procedures and diverging standards to enhance intra-PTA trade flows. In addition, this also underlines the importance for South-South PTAs to pursue an outward strategy of multilateral liberalization.

If these conditions are met, PTAs can be part of a virtuous circle of trade liberalization and trade expansion. The example of East Asia shows how increased intra-regional trade can go hand in hand with increased extra-regional trade, contributing to sustained high levels of growth. In addition, successful PTAs typically emerge in existing trading blocs rather than creating them.

I would now like to point out some more risks relating to PTAs. Restrictive rules of origin can erode advantages from PTAs. Rules of origin define the circumstances under which an import from outside the PTA can be traded within the PTA. Typically, these rules prescribe some local content requirement to allow an import to be traded under PTA conditions—rules of origin are considered necessary to prevent transshipping through the lowest external tariff PTA member. More complex, restrictive rules can be product specific and may involve multiple criteria. These rules can also be used as nontariff trade barriers to reduce intra-PTA trade. Customs administration faces particular challenges if treatment is not uniform across different products, or if a country is a member of more than one PTA, which can lead to lengthy clearance processes and imply significant costs for the private sector. By reducing intra-PTA trade in this manner, complex rules of origin erode the possible welfare gains from PTAs. It is therefore crucial that these rules are kept simple and that customs administration is well-equipped to implement them.

Another risk is that PTAs can lead to revenue losses for the member countries, at least in the short run. Tariff reductions associated with trade liberalization naturally entail the risk of falling government revenues. PTAs are no exception to this. Also, customs administration faces particular challenges to safeguard revenue by properly identifying imports from nonmember countries. Given heavy dependence on trade taxes in many developing countries, tariff reductions are likely to lead to significant revenue losses, in particular when intra-regional trade is important and common external tariffs are also reduced. Early measures to offset potential revenue losses are needed, such as strengthening tax administration, or reinforcing the consumption and income tax systems. Over time, though, there can be a positive impact from PTAs on revenue, if trade liberalization leads to a better resource allocation, stimulating growth, and, in turn, tax revenues.

I would now like to say a few words about the relationship between PTAs and the exchange rate system. At first glance, there does not seem to be a strong association in practice between PTAs and exchange rate regimes. Looking at existing customs unions, there are some currency unions (such as the Euro area and the West African Economic Monetary Union), but others are an eclectic combination of pegs or managed floats against each other's currencies and vis-à-vis outside currencies.

Despite this diversity, it does appear that in PTAs with high levels of intra-regional trade, the exchange rate regime is an important issue. When a PTA is successful in the area of trade liberalization, this can be expected to lead to further integration, particularly financial integration. Financial integration typically follows trade integration, as banks follow their customers. But with increasing cross-border financial flows, the threat to exchange rate stability grows as well.

As you know, uncertainty over exchange rates can affect trade directly, because it can wipe out profit margins, and indirectly because it can misdirect investment. Although the literature is ambiguous about the impact of exchange rate volatility on trade flows, the impact of exchange rate instability may actually be more damaging within a PTA. An exchange rate crisis in one member may adversely affect trade and FDI flows to other members, and possibly, through contagion, triggering currency crises in these other members as well. Members that lose competitiveness may resort to increased protectionism, either vis-à-vis the rest of the world by raising outside tariffs, resulting in more trade diversion, or vis-à-vis other members through less transparent non-tariff barriers. This defeats the purpose of the PTA. Countries may even choose to scale back or completely abandon their trade agreements. The 1999 devaluation of the Brazilian real, for example, strained the relationship between Argentina and Brazil, and gave rise to protectionist pressures, businesses threatening to relocate, and added to pressures on the Argentine peso.

Thus, within a PTA, in the same way as trade integration can create pressures for financial liberalization, financial integration and liberalization may call for increasingly reliable exchange rate stability. One way to reduce the risks associated with exchange rate volatility within PTAs is of course to already take into account the potential divergence in exchange rate regimes when choosing partners. Countries with lower volatility (e.g. developed countries) and countries with similar exchange rate regimes and similar macroeconomic patterns would make better PTA partners.

But another important message could be that promoting exchange rate stability requires greater macroeconomic policy coordination. The need for policy coordination depends not only on trade integration, but also on the degree of factor mobility across borders, both labor and capital. The deeper the economic integration, the more coordinated should be macroeconomic policy. One extreme is the possibility of a monetary union, which would completely eliminate exchange rate instability. Recent empirical work suggest that monetary unions may have a significant impact on trade: other things the same, countries that share a common currency trade three times as much as countries whose currencies are independent of each other. One caveat, though: the effect of a monetary union on a country's trade is stronger if the partner is one with whom one already trades. This implies that the option of adopting a common currency may not be as beneficial for some countries as adopting the U.S. dollar or the Euro. Short of a monetary union, coordination between national currencies may take the form of pegging to the currency of the lead country in the PTA, or ensuring consistency with respect to an outside benchmark currency, or a basket of currencies, so that bilateral exchange movements within a PTA are limited.

I would be amiss not to say a few words on the IMF's view on regional integration. The IMF's Articles of Agreement call upon the Fund to "... facilitate the expansion and balanced growth of international trade and to contribute thereby to the promotion and maintenance of high levels of employment and real income and to the development of the productive resources of all members as primary objectives of economic policy." As such, the IMF supports PTAs that are consistent with this mandate. The main pillars of the IMF's involvement in regional initiatives are regional surveillance and technical assistance to regional institutions and to individual countries in a regional context, including through regional TA and other regional centers.

As regards surveillance, the IMF is increasingly conducting it at a regional level. Most notably, regional surveillance takes place for currency unions. In 1998, the IMF adopted formal procedures for surveillance of the euro area, making discussions with its regional institutions an integral part of the IMF's Article IV surveillance over its members. There are annual missions to the euro area to conduct these so-called Article IV consultations, similar to those of other member countries, though with a focus on monetary, exchange rate and trade policy. Fiscal and structural policies continue to be discussed largely as part of the individual Article IV consultations with the euro area member countries.

But the Euro area is not unique. Surveillance of other regional arrangements is also evolving. The IMF sends annual missions to the West African Economic and Monetary Union (WAEMU), and has prepared annual reports since 1998. Likewise, annual reports for the Central African Economic and Monetary Community (CEMAC) have been prepared since 1999. Most recently, regional consultations with the East Caribbean Currency Union (ECCU) have commenced in 2002, though policy discussions had been going on for some time longer. In all of these cases, regional surveillance focuses inter alia on monetary and exchange rate policy, competitiveness, convergence, trade policy and financial sector supervision. In the case of the Gulf Corporation Council (GCC), regional issues are taken up in the context of discussions with individual member countries but Fund management also meets collectively with senior officials of the GCC. In addition, research at the IMF has, for example, looked at labor market problems across GCC members.

But regional surveillance is not limited to monitoring currency unions and PTAs. The IMF produces semi-annual and annual regional outlooks that analyze macroeconomic developments in a regional context and cover specific cross-country issues. For example, the most recent regional outlook from the IMF's Middle East and Central Asia Department, which includes Pakistan, focused on financial sector developments. Another recent IMF study analyzed financial integration of Central American Free Trade Agreement member countries, drawing in part on country-specific Financial System Stability Assessments. Typically, IMF surveillance evolves when there are regional institutions as centralized counterparts.

Turning to technical assistance, the IMF also provides assistance to regional institutions. Technical assistance has addressed issues such as customs and domestic tax regimes. For example, the IMF has been involved in trade facilitation in Middle East and North Africa (MENA) countries such as in the Arab MAGHREB Union and the Gulf Cooperation Council. The IMF has also advised the East African Customs Union on harmonizing tax incentives to avoid harmful tax competition among member countries.

Over the last couple of years, the IMF has established several regional centers for providing technical assistance, in Africa, the Middle East, the Caribbean, and the Pacific islands. These regional centers aim to enhance the effectiveness of the IMF's technical assistance to the participating countries. They allow for a more rapid response to emerging TA needs; for closer coordination with other TA providers and with regional institutions; and for more efficient and sustained assistance for regional integration initiatives. Technical assistance is provided by a team of IMF experts assigned to each center, supplemented by short-term specialists contracted to provide targeted advice and training.

I would like to end by drawing some lessons on how to make regional integration work. The bottom-line is that it can work, but positive outcomes depend on design and implementation. What are the key ingredients of success? Four major factors are:

(i) A large and diverse membership;

(ii) Continued reduction in external tariffs;

(iii) Comprehensive product coverage, with simple and transparent rules of origin; and

(iv) Effective implementation.

First, a large and diverse PTA membership can increase the likelihood that the production and trade structures of PTA members are complementary, and that members can take advantage of economies of scale. Comparative advantages may be easier to identify and exploit in North-South PTAs. But South-South agreements can have a large and broad membership as well, including in East and South Asia. In any event, the key lesson appears to be that PTAs should maintain open access and allow countries outside the PTA to join as easily as possible; membership negotiations thus need to be facilitated.

Second, continuing to lower external tariffs after establishing the PTA is essential for reducing trade diversion. This is particularly important when trade barriers against nonmembers are high (as in South Asia) and could result in large trade diversion. Experience suggests that members will experience gains in welfare, if after the establishment of a PTA, growth of imports is evident from all countries, not just PTA partners, as was the case with the EU. In addition, low external tariffs also improve the competitiveness of the domestic export industry, which is important for supporting growth.

Third, comprehensive product coverage within a PTA ensures that high as well as low tariffs are reduced. Otherwise, benefits to participants would be small, and welfare losses would result from firms moving towards less efficient industries where protection is enjoyed. In practice, textiles, agriculture and services are often exempt or subject to long adjustment periods. This results in large welfare losses, with developing countries often suffering the most. If exemptions are unavoidable, a short negative list, listing only those goods and services for which tariffs are not or less reduced, is preferable to a positive list, because with a negative list new goods and services would automatically benefit from the lower tariffs. In addition, liberal, transparent, clear, and consistent rules of origin are needed. Rules of origin are essential for implementing a PTA that is not yet a customs union. The more liberal rules tend to apply a general rule on the percentage of local value added that is required for preferential access. Increased cost of compliance on account of a proliferation of PTAs with complex rules of origin can create substantial problems, especially in countries with less capacity.

And fourth, implementation matters—signing a PTA is not enough. Agreements are often not implemented because interest groups stop them. In practice, PTAs are often signed with a political objective but, after lengthy negotiation, turn out much more restrictive than initially intended. In implementing PTAs, politicians need to bear in mind that regional integration can only be successful if it unleashes new competition that lowers domestic prices and introduces new technology. As with multilateral liberalization, exposing members to competition results in benefits, but this can also cause some pain. Governments need to recognize this and be prepared for it, including through providing an appropriate social safety net.

Turning finally to issues of direct interest to Pakistan, the IMF strongly supports efforts to enhance regional cooperation in South Asia, and we encourage Pakistan to pursue integration under the SAFTA framework. The potential for trade creation and increasing welfare is large in light of the current, very low level of regional integration. SAFTA can also prove to be an important vehicle for increased cooperation between Pakistan and India. As I mentioned before, succesful progress in trade liberalization is often a critical first step on the way advance wider political objectives. But trade liberalization under SAFTA needs to be complemented in Pakistan and other member countries with continued unilateral trade liberalization vis-à-vis the rest of the world. Unless the high external protection levels for some sectors are dismantled in parallel, regional tariff cuts could lead to substantial trade diversion. On the issue of how to make SAFTA itself as effective as possible, exemptions to the agreement need to be limited. The sensitive goods lists and the possibility of temporary suspension of concessions should therefore be minimized. Liberal rules of origin need to be negotiated under the agreement. Trade facilitation by means of harmonization of standards and customs procedures, and cooperation in transport infrastructure is also important. Back-loaded tariff cuts are possible under the agreement, but this would limit the economic impact of SAFTA.

Finally, the more Pakistan accompanies trade openness with furthering its structural reforms, the more positive the results will be in terms of employment and sustainable growth.

Thank you very much.