I. Introduction
SAARC is well reputed for limited achievements on core issues. The fact that the South Asian Free Trade Area (SAFTA) agreement was signed at the 12th SAARC Summit in January 2004 is in itself an achievement. SAFTA was long overdue, the turbulent South Asian regional politics having often delayed its finalisation. In fact, there was a time when it appeared that SAFTA would remain only a vision (Kelegama, 1996, Mukherji, 2002).

SAFTA was first mooted at the 8th SAARC Summit in Delhi (1995), it was suggested then that it should come into operation by 2005. This date was revised at the 9th SAARC Summit in Male (1997), where it was declared that SAFTA should come into operation by 2001. However, the 9th SAARC Summit also took a decision to appoint a Group of Eminent Persons (GEP) to draw up a vision and a roadmap for SAARC. Obviously, the GEP had to look at SAFTA and its feasibility by 2001. At the 10th SAARC Summit in Colombo (1998), the GEP report was presented which stated that a more realistic timetable for SAFTA is 2008 and, for the least developed countries in South Asia this date was extended to 2010 (GEP, 1998).

At the Colombo Summit, the date for SAFTA was postponed without specifying any time bound target, but a decision was taken to have a 'Framework Treaty' by the year 2001. Due to regional politics, the preparation of the Treaty got delayed and it finally came into shape by January 2004.

In this paper, an attempt is made to examine the SAFTA agreement that was signed by the Foreign Ministers of the SAARC member countries at the 12th SAARC Summit. First, a brief survey is made in Section II on SAPTA. Section III then makes an assessment of the SAFTA agreement in the light of the GEP report recommendations. Section IV has some concluding remarks.

II. SAPTA to SAFTA
While the debate for the SAFTA timetable was going on, there was some progress in the South Asian Preferential Trading Arrangement (SAPTA) which came into operation in December 1995. By the time of the 10th SAARC Summit in July 1998, two rounds of SAPTA had been completed and close to 2,126 products were under tariff preferences but the progress had been slow (IPS, 1999: 22). It adhered to four modalities for tariff negotiations, viz., (a) product-by-product approach, (b) across-the-board tariff reduction, (c) sectoral tariff reduction and d) direct trade measures. The third round of SAPTA was completed by November 1998. By early 2000, the results were far from satisfactory, and a case study for Sri Lanka observed: 'SAPTA…has had no significant impact in changing the existing trade pattern of Sri Lanka vis-à-vis its South Asian partners' (Weerakoon and Wijayasiri, 2001: 21). Mukherji (2002: 98) concluded: 'Except for India, none of the other contracting states has conceded meaningful tariff cuts. The effects of trade liberalisation are thus modest.'

The progress of economic cooperation under the SAPTA umbrella and the design of a SAFTA agreement got delayed during the period 1999-2001 due to the deterioration of Indo-Pakistan relations¹. An attempt was made in December 2000 by a newly formed South Asian Citizen's Commission to pressurise SAARC member states to get the SAFTA 'Framework Treaty' by late 2001 but to no avail. A Summit could not be held during this three-year period and the 11th SAARC Summit took place in January 2002 in Kathmandu. In this summit, a decision was taken to have the SAFTA Treaty ready by the 12th SAARC Summit. The SAARC Secretariat coordinated the work of the commerce ministries of the respective SAARC member countries in preparing the SAFTA agreement. Meanwhile, the 4th round of SAPTA negotiations took place in October 2002.

The four rounds of SAPTA had resulted in coverage of over 5,000 tariff line items (SAARC, 2002). Studies have shown that the SAPTA process contributed very little in stimulating intra-regional trade (Mukherji, 2002, SACEPS, 2002a, and others). Due to the slow progress of the regional initiative of promoting trade, a number of SAARC member countries decided to embark on bilateral free trade agreements (BFTA). The Indo-Lanka BFTA was signed in late 1998 and came into operation in early 2000. Long existing Indo-Nepal treaties were formalised as a BFTA in 1996 (RIS, 2004: 53). A number of other sub-regional initiatives such as growth quadrangles (Bangladesh, Bhutan, Nepal and India) and triangles (Sri Lanka, Maldives, South India) were mooted and some of them were initiated. These sub-regional initiatives were not considered for preferential trading but for sectoral cooperation.

In addition, several South Asian countries joined wider regional groupings in Asia such as the Indian Ocean Rim Association for Regional Cooperation (IOR-ARC initiated in 1997) and BIMSTEC (Bangladesh, India, Myanmar, Sri Lanka, Thailand Economic Cooperation initiated in 1997). Both these groupings were not preferential trading blocs- IOR-ARC was based on open regionalism where unilateral trade liberalisation was advocated, while BIMSTEC was initially based on sectoral cooperation. Membership in such pan-Asian regional groupings was obtained by some South Asian countries in the hope of gaining more economic benefits, which the SAPTA process was not delivering.

Clearly, the SAFTA agreement has come at a time when the trading environment in South Asia is complicated by the slow progress of SAPTA and a number of parallel regional and pan-regional initiatives are in place. It would, therefore, be pertinent to examine whether the SAFTA agreement has taken into account the factors governing the slow progress of SAPTA and the complications created by parallel initiatives.

The SAPTA framework was basically guided by the 'positive list' approach, though there were instances where sectoral tariff preferences were considered between member countries². The process adopted by SAPTA was extremely time consuming and slow. Moreover, at least in the first two rounds of SAPTA, non-tariff barriers (NTBs) were not considered for removal with the granting of tariff preferences. In a nutshell, besides these problems: (1) the tariff cuts were not deep enough, (2) a wide range of goods was not subject to preferential tariffs and (3) some actively traded goods were left out from preferential tariffs. These problems were visible in the first preferential trading arrangement in Asia, i.e., the Bangkok Agreement (BA) and were highlighted before SAPTA came into operation but the same mistakes were repeated³.

III. SAFTA Treaty and the GEP Report
Jean Monet spoke of European integration in the 1950s and some ridiculed him as a dreamer at that time. However, Monet's dream was realised in a 40-45 years' time period. Likewise, some commentators have expressed various reservations on the GEP vision of a South Asian Economic Union by 2020. However, it is not an impossibility as the realisation of the European Union clearly indicates.⁴ The GEP Report on 'SAARC Vision Beyond the Year 2000' still awaits adoption by the SAARC Council of Ministers. This is in contrast to other regional groupings such as APEC, where their GEP Report was adopted and put into practice soon after its submission.

The SAARC GEP Report has many suggestions, and with regard to movement towards a free trade area, the report, after taking cognizance of the problems encountered by SAPTA, recommended a 'negative list' approach for tariff reduction with an annual 12.5 per cent tariff reduction by member states, removal of all NTBs within a time frame, and a number of other trade facilitating policies. The SAFTA agreement is a far cry from the recommendations of the GEP Report. The tariff reduction process, timetable, additional measures (or direct trade measures) etc., differ significantly from the GEP recommendations. Moreover, there are inherent shortcomings in the agreement.

Many important items critical for the success of SAFTA are left for negotiations such as the rules of origin (Article 18), negative list, areas for technical assistance, etc. This could cause considerable delay and might make it difficult to have the SAFTA process fully operational by 01 January, 2006 (the declared date).

It is stated in Article 07 that for non-least developed countries (non-LDCs) the existing tariffs should be reduced to 20 per cent in two years and thereafter in a five year period, tariffs should be reduced to 0 - 5 per cent (Sri Lanka is given a period of six years). Least developed countries (LDCs) should reduce their tariffs to 30 per cent in two years and thereafter should bring down the tariffs to 0-5 per cent within an eight year period. If however, tariffs are below 20 per cent for non-LDCs, it is stated that an annual 10 per cent reduction should be made for two years. And for LDCs, if tariffs are below 30 per cent it is stated that an annual 5 per cent reduction should be made for two years. There is a 10-year period commencing 01 January 2004 for the FTA to become fully operational. Once the existing tariffs are reduced in accordance with the above format and completed by 01 January 2006, the subsequent tariff reduction process is aimed at achieving 0-5 per cent tariff rates by the end of eight years for LDCs and by the end of six years for the non-LDCs. The non-LDCs (and LDCs) are encouraged to adopt reductions in equal annual instalments of not less than 15 per cent (and 10 per cent for LDCs) annually.

The above format of tariff reduction is a substantial departure from what is recommended in the GEP Report and is somewhat close to what the SAARC Chamber of Commerce and Industry (SCCI) suggested⁵. The problem with this format of tariff reduction is that, despite reduction of average tariffs, distortions will prevail in the form of high tariffs in particular products in some countries. Perhaps it would have been more efficient if convergence was achieved before embarking on lowering tariffs further. In order to ensure that the benefits incur to all member countries, achieving convergence as in the case of the ASEAN-FTA is always better.

Although Quantitative Restrictions (QRs) will be removed as soon as the tariff levels reach 0-5 per cent, it is not clear from Articles 06 & 07 (4 & 5) whether other NTBs will be removed with the QRs. Moreover, if the Treaty is going to strictly adhere to this method of removal of various NTBs, it will be difficult to exploit the full gains from various phases of tariff reductions, thus defeating the objective of preferential tariffs. There is no reference to a movement towards a Customs Union after the FTA in the SAFTA agreement- a key recommendation of the GEP Report⁶.

Article 7.3 (a & b) refers to the negative list. It is stated that the number of products in the negative list shall be subject to a maximum ceiling that is mutually agreed upon among the member states and will be reviewed every four years. There is no deadline for determining the negative list and there is no format for phasing out the negative list over the years. All these issues seem to have been left for discussion by the SAFTA Ministerial Council established under Article 10. If the negative list becomes too long, the agreement may not become compatible with Article XXIV of GATT. In fact, SAPTA comes under the Enabling Clause of the GATT, which is considered by some commentators as non-serious in commitment for it to cover 'substantially all the trade' as stipulated in the GATT (Kelegama and Adhikari, 2002).

When items critical for the success of an FTA are left for negotiation, the finalisation can get delayed. This was the case with the Indo-Sri Lanka BFTA where the negative list was open for negotiation. Consequently, it took nearly one year and two months for the agreement to become effective after it was signed. Various disagreements had to be sorted out before finalisation. The phasing out of the negative list could have been based on the ASEAN FTA model where there was a clear strategy with a tariff line classification based on an Inclusion List, Temporary Exclusion List, Sensitive List and a General Exception List for implementing a Common Effective Preferential Tariffs (Mukherji, 2002; SACEPS, 2002a). This dimension has been completely ignored in the SAFTA agreement.

In addition to the differential tariff reduction format, the agreement makes a number of provisions for according special and differential treatment to the LDCs in the region (Article 11). There are provisions for non-LDCs considering direct trade measures in favour of the LDCs, such as long and medium-term contracts containing import and supply commitments in respect of specific products, buy-back arrangements, state trading operations and government procurement. LDCs will get special consideration for technical assistance, in particular, to compensate for revenue shortfalls from tariff reductions. These measures do not go far enough to ensure that LDCs will be able to derive equitable benefits from SAFTA. It was this concern that made Bangladesh hesitate till the last minute before signing the agreement. Bangladesh wanted non-LDCs to refrain from imposing anti-dumping and countervailing measures against LDCs, and rightly so, since no such provision exists in any other existing FTA. This concern was partially accommodated by stating: 'The Contracting States shall give special regard to the situation of LDCs when considering application of anti-dumping and countervailing measures'. [Article 11 (a)]

The agreement does not consider the suggestion of the GEP for creation of a large fund for development of infrastructure, human resources and improvement of export supply capacity of LDCs. Without significant structural changes in the production structure, LDCs are unlikely to derive equitable benefits from SAFTA. SACEPS (2002a) has shown that in the EU for raising the level of development in the less developed member countries such as Spain, Portugal and Ireland, the European Commission had created a development fund for each of them amounting to 3- 5 per cent of their GDP. Such arrangements have not been considered in the agreement, perhaps due to reservations expressed by non-LDC member countries.

It would be imperative to ask at this juncture why many SAARC members shy away from preferential tariff reduction. The first reason is rigid factor markets- in particular, labour and capital- prevalent in most SAARC member countries. These factors of production find it difficult to move out due to tight legislation governing them when the industries are subject to restructuring as result of tariff liberalisation.⁷ Consequently, instead of industry restructuring, what takes place is closing-down of industries with serious social and political consequences⁸. This is an issue for economies where the small and medium scale enterprises dominate the industrial and agricultural sectors in terms of employment. Thus, considerable structural adjustments also have to take place, particularly in LDCs to face tariff reforms with a greater degree of confidence.

Second, there is a fear among the smaller countries that the main beneficiary from tariff liberalisation would be the larger countries⁹. Irrespective of the theoretical viewpoint10, the perception of smaller countries needs to be recognised, and it was this realisation that led to the 'Gujral Doctrine' to be introduced by India in 1997/98. However, there is some dilution of the doctrine in recent years and giving vent to this, an editorial of the Economic and Political Weekly (EPW) stated: 'It is for India to ensure that smaller members of the region have a growing stake in regionalism …. This responsibility India has not taken seriously'. (EPW, January 10-16, 2004: 119)
To reduce the cost of structural adjustment, 'additional measures' are required. Additional measures in Article 8 do not measure up to the GEP report, where reference has been made for an establishment of a South Asian Development Bank, South Asian Development Fund and a South Asian Energy Grid. The GEP report has strongly recommended the creation of a SAARC Investment Area and vertical industrial integration. None of these receive mention in the agreement and it does not provide for creation of a mechanism for pursuing additional measures under Article 8¹¹.

There is a sizeable body of literature on South Asia Energy Grid (SACEPS, 2002b; RIS, 2002; and others). A SAARC Investment area has also been looked at in recent literature (SACEPS, 2002c; RIS, 2002; and others). In fact, the trade-investment nexus has come into effective operation in South Asian bilateral FTAs and RIS (2004) shows how the large trade deficits between two countries have been compensated by the capital account through significant investment flows. In the context of investment flows, horizontal and vertical integration of industries of South Asia becomes important to face the global competitive pressures¹². Even though a multitude of literature is available on these crucial issues, the agreement has completely overlooked these areas and solely focused on trade facilitating measures in Article 8.

Finally, the agreement is silent on how SAFTA is going to integrate the existing bilateral free trade agreements between some SAARC countries (such as the Indo-Lanka BFTA, Indo-Nepal BFTA, and the ones that are under consideration, for example, Pakistan-Sri Lanka BFTA and Indo-Bangladesh BFTA) into the SAFTA agreement. If integration is not an option, will SAFTA operate parallel to the existing treaties?. This seems to be most likely and will create a 'Spaghetti Bowl' type of phenomenon (a la Bhagwati, 2002) with parallel preferential tariffs, rules of origin and negative lists¹³. Thus the Customs and Commerce Departments in individual SAARC countries will have to be upgraded to meet this challenge.

Since trading in SAARC basically means to trading with India, this objective seems to have been met in the current trading environment by bilateral FTAs. Thus, SAFTA will basically boil down to trading between India and Pakistan. If the remaining bilateral FTA with India, viz., Indo-Bangladesh comes into operation soon, there are reasons to believe that there will be less enthusiasm among some SAARC countries about SAFTA.

The SAFTA agreement does not refer to liberalisation of trade in services. A regional trade arrangement should not only be deepened but widened as well. For example, the Indo-Sri Lanka BFTA has now been advanced to an Indo-Lanka Comprehensive Economic Partnership Agreement where liberalisation of services (in addition to investment) has been included. The BIMSTEC Free Trade Area that was signed in early February 2004 refers to liberalisation of services under a GATS-Plus framework in Article 4. In short, the SAFTA agreement is not futuristic¹⁴.

The above-mentioned technical shortcomings are obviously going to get aggravated by regional politics. Much will depend on whether Pakistan would offer MFN status to India in 2004¹⁵. Present indications are that there is a commitment to a peaceful settlement of the long-standing bilateral dispute between India and Pakistan. However, if the MFN issue between the two nations is not settled soon, it will be a major obstacle to the birth of SAFTA.

The Indian Prime Minister stated in 2003 that South Asia should move towards a common currency. Obviously this would be possible if there is monetary cooperation among South Asian countries. Currently, under SAARCFINANCE a discussion on monetary cooperation is ongoing. Moreover, literature on the subject is also on the increase (Maskey, 2002; RIS, 2004; and others). RIS (2004), for instance, has suggested the introduction of a parallel currency as the first step towards moving to a common currency. The subject did not receive any attention from the 12th SAARC Summit. It appears that the futuristic proposals have been set aside in order to cover the backlog that has accumulated due to the frequent postponement of SAARC Summits.

IV. Concluding Remarks
Whether regional trade is the best available option for South Asia has been a subject of debate since the mid-1990s. Critics of promoting regional trade in South Asia via preferences have argued that South Asia would be better off focusing on trade with the rest of the world, in particular, EU and USA (Srinivasan, 1994; Pigato et al., 1997; Panagariya, 1999; and others). A recent report released by the World Bank (2003) argues: 'Because many tariffs in the region are very high, especially in India and Bangladesh, there are large potential trade diversion costs for the region as a whole if the various preferential trade agreements were ever to be seriously implemented. The consequent reductions in economic welfare would show up principally in reduced customs revenue and terms-of-trade losses. It is unlikely that benefits through increased competition, economies of scale, or improved operating efficiency of import competing firms would outweigh these overall economic costs. There are much larger gains for increased trade with the rest of the world (ROW), especially trade with the developed countries and with more advanced developing countries in South East Asia, including China. This is because the South Asian countries have a comparative advantage in relation to ROW in similar, mostly labour intensive products, and the volume of trade and the economic benefits from trading these products among themselves are limited by comparison.' (p. 22)

On the other hand, those who argue the case for regional trade state that substantial trade is already taking place in South Asia with informal trade amounting to a large proportion of formal trade. The exact intra-regional trade is estimated anywhere between 8-10 per cent. Although studies have shown that there are limited complementarities in the SAARC region, it is argued that this was also the case in ASEAN during the mid-1970s, and that dormant complementarities in the region could be invigorated by intra-regional investment and FDI¹⁶. They also argue the cost of non-cooperation to be quite high (RIS, 2004 and 1999; GEP, 1998; CUTS, 1996; and others). The debate is far from settled. Irrespective of the debate, there is a general belief that regional cooperation in South Asia should not be viewed only from the trade perspective, and that there are many gains from regionalism in other areas.

The past decade has seen the emergence of a number of regional trading blocs in different parts of the world and data shows that nearly 60 per cent of world trade is now conducted on preferential basis. The countries that are not part of a trade bloc face the risk of discrimination of their exports and loss of competitiveness. Thus, in the light of global trends, irrespective of the pros and cons of the academic debate, South Asia has been pushed to adopt regional economic integration. In the SAARC, promoting intra-regional trade is part of a large package of economic cooperation and SAFTA is a part and parcel of South Asian Economic Cooperation.

However, the movement to SAFTA is taking place in an environment where: (1) the precursor to SAFTA, i.e., the four rounds of SAPTA have failed to show concrete results, (2) several bilateral FTAs are well entrenched in the South Asian trading system, and (3) South Asian tariffs are already coming down under World Bank/IMF structural adjustment programmes. The third factor in effect is automatically reducing the preferential margin. Moreover, there are a number of shortcomings, clauses open for interpretation and items for further negotiations in the SAFTA agreement This shows that most of the research work that was done by the SAARC second-track 'think tanks' has not been fed in effectively to the SAARC first-track or the official process¹⁷.

Given this situation, not much can be expected from SAFTA. The initial euphoria that comes with the signing of the SAFTA agreement will soon taper away. The realities and the geo-politics of the region will once again determine the pace of negotiations in SAFTA. By that time, the bilateral FTAs would have delivered most of the results for the smaller South Asian countries and SAFTA will be an agreement mainly to promote India-Pakistan trade. Is it due to this realisation that the SAFTA agreement did not bother about a vision and ignored a number of worthy suggestions of the GEP Report?

(Dr Saman Kelegama is the Executive Director, Institute of Policy Studies of Sri Lanka).

End Notes

Bibliography