South Asia as a Free Trade Area

South Asia contains roughly a quarter of humanity inside a compact, contiguous, historically single economic space, and trades with itself less than any other recognised region on earth. The agreement meant to fix that has been in force for two decades. The interesting question is therefore not whether the agreement works — it plainly does not — but why a region whose members share rivers, railways, languages and a customs inheritance behaves, commercially, as though its neighbours were on another continent, and what the state that constitutes four-fifths of the region’s output has actually done about it.

The economics that make regional trade agreements work, and why they bite differently here

Free trade areas are not automatically good for the countries that join them, and the theory that established this is the first thing a serious argument about South Asia has to get right.

Viner’s distinction: trade creation against trade diversion

• Jacob Viner established that a preferential arrangement is not a partial move toward free trade but a mixture of liberalisation and new discrimination, and that its welfare effect can go either way.
  • Trade creation occurs when a member’s high-cost domestic production is displaced by a lower-cost partner’s output — production shifts to the efficient producer and the importing country gains.
  • Trade diversion occurs when imports shift from an efficient non-member to a less efficient member purely because the member’s goods now enter duty-free — the importing country loses tariff revenue and buys dearer goods.
  • The net effect is an empirical question, not a theoretical one, and it depends on how high the pre-existing tariffs were, how similar the partners’ cost structures are, and how much of the member’s trade was already with the partner.
• The South Asian application is uncomfortable and rarely stated.
  • Members’ external tariffs are high by world standards, which raises the risk that preferential entry simply redirects purchases toward a regional producer who is not the cheapest in the world.
  • The classic instance is India–Sri Lanka trade in copper and vanaspati in the first decade of the bilateral agreement, where the surge came from tariff arbitrage rather than comparative advantage, with thin domestic value addition on the Sri Lankan side.
  • This is precisely the outcome Viner’s framework predicts when preferences are deep and rules of origin are loose, and it is why rules of origin are not a technicality in this region but the substantive question.
• Balassa’s ladder — free trade area, customs union, common market, economic union, total economic integration — is the standard sequence, and SAFTA was framed from the outset as the first step toward a customs union, common market and economic union; South Asia has not completed the first rung.

Why similar factor endowments cap the gains

• The Heckscher–Ohlin logic is that gains from trade grow with the difference in factor endowments between partners, and South Asian economies are conspicuously alike.
  • All are labour-abundant, capital-scarce and land-constrained, with comparative advantage in similar low-skill manufactures.
  • The clearest illustration is ready-made garments: over four-fifths of Bangladesh’s global exports, sold into a region where India, Pakistan and Sri Lanka all run substantial textile industries of their own.
  • Bangladesh’s exports to the whole SAARC bloc have been stuck around US$2 billion for half a decade, and roughly 88% of that goes to India alone — the structural ceiling is visible in the data, not merely in theory.
• Intra-industry trade, the mechanism that carries most trade among rich economies, requires product differentiation and fragmented production, and South Asia has little of either.
  • The region is almost absent from global value chains relative to Southeast Asia, so the cross-border component flows that drive ASEAN’s intra-regional share simply do not exist here.
  • Jayant Menon makes the comparison directly: Southeast Asia’s integration followed from export orientation and value-chain participation, not from the free trade agreement, which came later and mattered less.
• But the pessimistic reading is overdrawn, and three things cut against it.
  • Complementarity is much stronger in services than in goods — tourism, health care, education, professional services and information technology — and it is precisely services that SAFTA never covered.
  • Energy is complementary by geology: Himalayan hydropower against Indo-Gangetic demand is as clean a case of comparative advantage as the region offers.
  • Sub-regional complementarity is real even where regional complementarity is weak — the eastern quadrant of India with Bangladesh, Bhutan and Nepal is an economically coherent unit that the political boundary bisects.

The gravity model and the size of the shortfall

• The gravity model, first applied to trade by Jan Tinbergen, predicts bilateral trade from the two economies’ size and the distance between them, and it fits world trade data remarkably well.
  • South Asia is the standing anomaly: large economies, short distances, small trade.
  • The World Bank’s gravity estimate is that the region trades about US$23 billion with itself against a potential of US$67 billion — roughly one third of what geography and size predict.
  • The gap is not evenly spread. India–Pakistan formal trade could plausibly be fifteen times its recent level, worth about US$37 billion; it has been effectively nil since May 2025.
• The residual that gravity models cannot explain is the border effect, and it is the analytically important number.
  • John McCallum’s finding that Canadian provinces traded vastly more with each other than with equidistant American states established that a political boundary is itself a large trade cost; Anderson and van Wincoop later showed the effect is partly a statistical artefact of multilateral resistance, but not remotely all of it.
  • In South Asia the border effect is compounded rather than merely present: the boundary carries tariffs, para-tariffs, testing requirements, transit refusal, visa refusal and, on some segments, fencing.
• Amita Batra has run the gravity apparatus on South Asia explicitly and drawn the conclusion that gives her book its title.
  • Her finding is that a shared border, when disputed, functions as a barrier rather than a facilitator — the opposite of what gravity assumes.
  • She locates the region’s “anti-home bias” — the tendency to trade less with neighbours than with distant partners — in conflict rather than in endowments, which reverses the usual causal story.

Bhagwati’s spaghetti bowl, and South Asia’s version of it

• Jagdish Bhagwati coined the “spaghetti bowl” to describe the tangle of overlapping preferential agreements, each with its own tariff schedule, phase-out timetable and rules of origin, that a single consignment may have to navigate.
  • His objection was not to liberalisation but to discrimination that multiplies: each new agreement adds an origin regime, and administering them consumes the preference margin it was meant to deliver.
  • He argued the proliferation actively undercuts multilateral liberalisation by giving governments a cheaper-looking alternative and by creating exporters with a vested interest in preserving discrimination.
• South Asia has a small bowl with a lot of spaghetti in it.
  • A single consignment moving from Kathmandu to Dhaka may be governed by SAFTA, the India–Nepal Treaty of Trade, a bilateral transit treaty, SASEC instruments and national para-tariff schedules at once.
  • India alone operates SAFTA plus separate bilateral regimes with Nepal, Bhutan, Sri Lanka and Afghanistan, each with different origin criteria — so the regional agreement is often not the instrument traders actually use.
  • Utilisation of SAFTA preferences is correspondingly low, because a bilateral route or a sensitive-list exclusion frequently makes the regional preference worthless.
• The counter-argument from Krugman and Summers — that regional blocs among “natural trading partners” who already trade heavily are welfare-improving — does not rescue South Asia, because the premise fails: these states do not already trade heavily with one another.

“Both nature and man, but mostly man, have contrived to fragment the region, denying countries and people the benefits of proximity.”

Sanjay Kathuria

SAFTA: what was designed, and the compact audit of what went wrong

The instrument itself can be stated quickly, because its provisions are not where the interest lies. What matters is the gap between a tariff programme that largely did happen and a trade outcome that largely did not.

The instrument in outline

• SAPTA, signed at Dhaka in April 1993 and in force from December 1995, was the positive-list predecessor — four negotiating rounds covering over five thousand commodities, most of them traded in trivial volumes.
• SAFTA was signed at the twelfth summit, Islamabad, 6 January 2004, entered into force on 1 January 2006, and its Trade Liberalisation Programme began that July.
  • The programme required reduction to 0–5%, on staged timelines with longer periods for the least developed members — not the 20% floor the older accounts repeat.
  • It is a free trade area, not a customs union: no common external tariff, hence the necessity of rules of origin and the 40% value-addition requirement (30% for the least developed members).
  • Sensitive lists, revisable by the Ministerial Council, allow each member to exempt tariff lines entirely from the programme.
• The tariff programme was largely delivered. The trade was not. That single sentence is the whole diagnostic problem, and every explanation below is an answer to it.

The eight things that actually bind

Barrier	What it does	Where it bites hardest
Sensitive lists	Exempt whole tariff lines from liberalisation	About 35% of the value of intra-regional trade sits under sensitive-list tariffs; over 39% of India’s exports fall on partners’ lists
Para-tariffs	Duties applied only to imports, outside the tariff schedule, so untouched by SAFTA	Bangladesh, Sri Lanka and Pakistan — the region’s distinctive instrument
Non-tariff measures	Standards, testing, certification, port restrictions, licensing	Every land border; procedural opacity more than deliberate protection
No services regime in force	The agreement on services exists but schedules of commitments were never finalised	Where complementarity is actually strongest
No transit regime	Goods cannot cross a third member’s territory as of right	Afghan goods to India; Nepal and Bhutan to the sea; Bangladesh to Nepal and Bhutan
Connectivity and logistics	Few direct links, congested land ports, restrictive air service agreements	Petrapole–Benapole; the absence of direct flights between several regional capitals
Visas and movement	No usable business travel regime; the exemption scheme covers designated categories only	Services trade and buyer–supplier relationships
Political interruption	Trade suspended or restricted by executive decision	India–Pakistan since 2019 and wholly since May 2025; India–Bangladesh land-port restrictions from 2025

• The informal trade that all of this generates is the ninth item and belongs in its own section, because it is evidence about the barriers rather than a barrier itself.
• Two features distinguish this list from a generic account of trade barriers.
  • Tariffs are the part that worked, so an explanation resting on tariffs is refuted by the agreement’s own record.
  • The binding constraints are administrative and physical, which is why they survived a tariff agreement intact and why they cannot be removed by another one.

Sensitive lists: the exemption that swallowed the agreement

• A sensitive list is a schedule of tariff lines a member declares outside the liberalisation programme altogether — not slowed, but exempted.
• The lists are radically asymmetric, and the asymmetry is the opposite of what the region’s politics assumes.

Member	Sensitive list for least developed partners	Sensitive list for other members
India	25 lines — essentially alcohol and tobacco	Under 700
Bhutan	150	150
Maldives	681	681
Sri Lanka	1,042	1,042
Afghanistan	1,072	1,072
Pakistan	1,169	1,169
Bangladesh	1,233	1,241
Nepal	1,257	1,295

• India’s list for the least developed members is the shortest offered by any member, and it is the concrete form of the unilateral opening discussed below.
• The smaller members maintain the longest lists, which complicates the standard narrative in which India is the obstacle to regional trade.
  • Their defence is that revenue dependence on trade taxes is high and that infant industries need protection against an economy forty times larger.
  • The rebuttal is that the lists exclude precisely the products with regional trade potential — research on the exclusions identifies hundreds of lines with substantial unrealised trade, and consumer savings running into the high hundreds of millions of dollars annually.
• Sri Lanka’s position illustrates the self-harm: about 44% of its imports from South Asia and 23% of its exports to the region are covered by sensitive-list exclusions, and of its top twenty global exports, fifteen faced tariffs or negative listing inside the region.

Rules of origin: the question that decides whether preferences build anything

• Because a free trade area has no common external tariff, it must have rules that decide which goods count as originating in a member, or imports would simply enter through whichever member had the lowest external duty.
• SAFTA’s general test is a change of tariff heading plus 40% domestic value addition, relaxed to 30% for the least developed members, with a further concession for Sri Lanka.
  • The threshold is high for economies whose manufacturing depends on imported inputs, which is most of the region’s.
  • Bangladeshi garments are the illustration: an industry built on imported fabric struggles to demonstrate the value addition a preference requires, so the preference goes unclaimed.
• Cumulation — allowing inputs from other members to count toward the origin requirement — is the technical fix, and SAFTA’s regional cumulation provision is weak and little used.
  • Without effective cumulation, a preferential area cannot generate the cross-border production networks that make integration self-reinforcing elsewhere.
  • This is the precise mechanism by which Southeast Asia converted an agreement into a value chain and South Asia did not.
• The rerouting problem is the mirror image: where origin rules are loose or loosely policed, goods are transhipped through a member to capture a preference, as with palm oil moving through Nepal into India.
• The consequence for policy is uncomfortable but clear: loosening origin rules invites arbitrage and tightening them destroys utilisation, and the resolution is cumulation plus administrative capacity rather than a different threshold.

Para-tariffs: the barrier that is distinctively South Asian

• A para-tariff is a border charge levied only on imports but sitting outside the customs tariff schedule — a regulatory duty, a supplementary duty, a development levy, a cess.
  • Because it is not a customs duty, it is not covered by SAFTA’s liberalisation programme, and a member can therefore comply fully with SAFTA while its effective protection is unchanged.
  • The instrument is almost absent in ASEAN, which is why comparisons that look only at bound and applied tariffs systematically understate South Asian protection.
• The scale is not marginal.
  • Bangladesh applies a regulatory duty to roughly 45% of tariff lines and a supplementary duty to about 23%.
  • Pakistan applies an additional customs duty across almost the entire tariff schedule and a regulatory duty on close to a fifth of lines.
  • Sri Lanka applies a ports and airports development levy to over 80% of lines and a cess to more than a quarter.
• Two consequences follow, and both matter for how the problem is framed.
  • Bangladesh’s real protection sits in para-tariffs, not tariffs — so an argument that Dhaka has liberalised because its customs duties fell is simply mistaken.
  • Any future negotiation that does not put para-tariffs inside the schedule will reproduce SAFTA’s outcome exactly, which is why their inclusion is the single most technically important item on any reform list.

South Asia’s tariffs came down and its trade did not follow, because the protection had already moved somewhere the agreement could not reach.

Borders thicker than distance

The deepest empirical finding about South Asian trade is not that the region trades little. It is that its members find it more expensive to trade with the country next door than with a country across an ocean — an inversion of the ordinary economics of proximity, and one that no tariff schedule can explain.

The measured inversion

• Average trade costs within South Asia run about 20% above comparable country pairs in ASEAN, and more than three times the costs prevailing among the North American partners.
• The pairwise comparisons are the ones that make the point unmistakable.
  • Trade costs are about 56% higher between Bangladesh and Nepal than between Bangladesh and Brazil.
  • They are about 84% higher between Sri Lanka and Nepal than between Sri Lanka and Brazil.
  • It is roughly a fifth cheaper for an Indian firm to trade with Brazil than with Pakistan — a figure that has been quoted so often it has stopped being surprising, which is unfortunate, because it is astonishing.
• The trade restrictiveness a South Asian exporter meets from a South Asian partner runs at several times what the same exporter meets from the rest of the world — on the World Bank’s index, India’s restrictiveness against regional imports has stood many times above its restrictiveness against imports from everywhere else.
• The asymmetry is deliberate in origin and habitual in persistence. It was built during decades when trade with a neighbour was treated as a concession to that neighbour rather than as a transaction between firms.

The region against its comparators

Indicator	South Asia	Southeast Asia	European Union
Intra-regional share of total trade	Under 6%	About 25%	About 60%
Intra-regional trade as share of regional output	Under 1%	Close to 10%	Very high
Common external tariff	None — free trade area only	None; a customs union was never the model	Customs union since 1968
Services liberalisation	Agreement in force, no schedules	Framework agreement with successive packages	Single market in services, incomplete but real
Transit rights	None regionally	Extensive	Free movement of goods
Para-tariffs outside the schedule	Widespread	Effectively absent	Absent

• The instructive column is the third: the European case rests on a customs union and a supranational court, neither of which South Asia has attempted or would accept.
• The instructive comparison is the second: Southeast Asia’s intra-regional share rose largely because of value-chain participation and unilateral opening to the world, and the agreement ratified a process rather than causing it.

The physical layer: land ports, air links and logistics

• Land ports are the choke point, and their condition is a policy choice rather than a constraint.
  • Petrapole–Benapole carries the great bulk of India–Bangladesh land trade through infrastructure that routinely queues trucks for days.
  • Warehousing, testing laboratories, parking and cold storage are absent or undersized at most crossings, so the cost is paid in time, spoilage and demurrage rather than in duty.
• Air connectivity is thinner than the region’s income levels would predict, and directly suppresses services trade.
  • There have been no direct flights between Colombo and Kathmandu, Kabul or Paro — capitals within short flying distance of one another.
  • Where air services were liberalised the response was immediate: India–Sri Lanka routes expanded to around a dozen Indian cities and roughly 147 flights a week, and India became Sri Lanka’s largest source of tourists.
• Transit is the structural absence, and it converts geography into a penalty.
  • Nepal and Bhutan reach the sea through a single corridor, and Nepali exporters have reported around six weeks to reach an export destination through Indian ports.
  • Afghan goods cannot move overland to India because Pakistan does not permit transit, so the trade is routed through Iran or does not happen at all.
  • Bangladesh’s trade with Nepal and Bhutan must cross Indian territory, which makes an Indian administrative decision a determinant of third-party trade.

South Asia’s problem is not that its neighbours are far apart. It is that its borders cost more to cross than its oceans do.

The western closure and the price of a single veto

• Afghanistan’s trade with India is the purest case of a transit veto, and it shows what the absence of a regional transit regime costs a landlocked member.
  • Pakistan’s Afghan Transit Trade Agreement permits Afghan goods to reach India but has not permitted Indian goods to move the other way, so the corridor works in one direction at best.
  • India’s response was to build around the obstacle: the Chabahar port arrangement with Iran and the Zaranj–Delaram highway of 218 kilometres, which together create a sea-and-road route to Afghanistan that bypasses Pakistan entirely.
  • The route works, but it is a substitute for geography, and its costs are borne by Afghan and Indian traders rather than by the state that closed the shorter road.
• The same veto blocks the larger prize. Overland access from India to Central Asia runs through Pakistan and Afghanistan, and its closure is why India’s trade with a region it has cultivated politically remains negligible.
• In the east the mirror problem is trust rather than hostility. Bangladesh has granted India transit and transhipment access to the North-East in stages since 2010, and the arrangement has functioned — which shows that the eastern half of the transit problem is soluble and the western half is not.
• The asymmetry between the two halves is the strongest single argument for variable geometry: waiting for a regional transit regime means allowing one relationship to hold the other six hostage.

The trade the official figures do not record

Every statement that intra-regional trade is under 6% of the region’s total rests on recorded trade. A large share of what actually crosses these borders is not recorded, and the size of the unrecorded flow is itself the strongest evidence about what is wrong.

The estimated scale

• India–Bangladesh is the largest case. An official of Bangladesh’s revenue board put unofficial trade at about US$11 billion, against recorded two-way trade then in the mid-teens of billions — a ratio that makes the smuggled economy comparable in size to the legal one.
• India–Nepal informal trade has been estimated at roughly the same magnitude as formal trade, which the open border and the price wedge between the two markets make entirely plausible.
• India–Pakistan informal trade was estimated at around 91% of formal trade, much of it routed through third countries — Dubai principally — so that goods travel thousands of extra kilometres to avoid a border a truck could cross in an hour.
• India–Bhutan unofficial flows have been estimated at close to three times the recorded figure, and India–Sri Lanka at around 30%.
• Nisha Taneja’s body of work on this is the standard reference, and its central methodological point deserves emphasis: informal trade is estimated from trader surveys and mirror statistics, so every figure is a range, and the argument should rest on the order of magnitude rather than on any single number.

Why traders choose the informal channel

• The formal channel is slower and costlier than the duty saved. When clearance takes days, testing is arbitrary and documentation is opaque, a trader with a working capital constraint pays the informal premium and moves on.
• Sensitive lists and para-tariffs create the price wedge that makes the informal route profitable — a product excluded from liberalisation on both sides is a product that will move unofficially.
• Non-tariff measures are frequently the binding reason rather than duty: a consignment that cannot obtain a sanitary certificate does not become cheaper by paying duty.
• The open India–Nepal border and the porous Bangladesh boundary make enforcement expensive, so the informal channel is not a marginal leak but an established parallel market with its own financing and price discovery.

Why formalisation is worth more than the trade statistics suggest

• It is revenue. Informal trade is untaxed on both sides; on the Bangladeshi estimate the foregone duty is a material fiscal item, which is why the revenue authority raised it at all.
• It is measurement. A region cannot negotiate about trade it cannot see, and the low recorded share understates real integration while overstating how little the two economies need each other.
• It is regulation. Informal trade carries no standards, no phytosanitary control and no consumer protection, and it finances networks that also move other things.
• It is the strongest argument against the pessimists. The claim that South Asians do not want to trade with each other is refuted by the fact that they are already doing it in large volume, at high cost, illegally.

The political economy of protection: who blocks liberalisation, and why

Trade policy in this region is made by governments answering domestic constituencies, and the constituencies differ country by country. A list of barriers that does not name the interests behind them explains nothing.

Bangladesh

• The ready-made garment lobby is the most powerful private interest in the country and has no offensive interest in the Indian market, where domestic textile production is large and competitive.
• Revenue dependence on import taxation is high, which makes para-tariffs fiscally load-bearing rather than merely protective.
• The trade deficit with India is politically toxic — recorded bilateral trade runs at a ratio of roughly six or seven to one in India’s favour — and every liberalisation proposal is read domestically as widening it.
• Graduation from least developed country status removes the preferential margins Bangladesh currently enjoys and sharpens the case for a bilateral agreement with India, which is exactly why the partnership negotiations began.

Pakistan

• The military-security establishment has treated trade normalisation as a concession on Kashmir, and civilian governments that moved toward it have been checked.
• The 1996 grant of most-favoured-nation treatment by India was never reciprocated; Pakistan’s 2012 commitment to non-discriminatory market access was announced, prepared and then abandoned under pressure from agricultural, automotive and pharmaceutical interests and from the security establishment.
• Pakistan’s operative instrument has been a positive list of importable items — the inverse of a sensitive list, and a far more restrictive device.
• The sequence since 2019 is a complete closure: India withdrew most-favoured-nation treatment and imposed a 200% duty in February 2019; Pakistan suspended bilateral trade that August; and India banned all imports, direct and indirect, in May 2025.

Nepal

• Trade dependence is the grievance, not trade itself: India accounts for around two-thirds of Nepal’s total trade, and the deficit ratio has run at the order of fourteen to one.
• The Treaty of Trade renews automatically on a seven-year cycle, and the most recent renewal passed without renegotiation — Nepali commentary treats this as a lost opportunity rather than a favour.
• Nepali complaints are precise and mostly about implementation rather than the text: rules-of-origin criteria, quantitative restrictions on vegetable ghee, acrylic yarn, copper products and zinc oxide, inconsistent phytosanitary decisions, and the absence of an effective dispute mechanism.
• The domestic constituency for protection is real too — Nepali manufacturers want the option of duties against subsidised Indian agricultural imports, which the reciprocal duty-free arrangement forecloses.

Sri Lanka

• The bilateral agreement with India is politically contested in a way its economics do not warrant, and opposition to the services and investment successor has been organised by professional associations fearing Mode 4 movement of Indian professionals.
• Small and medium manufacturers fear scale asymmetry; the fishing dispute and the memory of the concentration of early gains in two tariff-arbitrage products supply the emotional case.
• The 1,042-line sensitive list and the ports and airports levy are the technical expression of the same politics.

India

• India’s protective interests are narrower but not absent — agriculture, dairy, and specific manufactures — and they are visible in the non-least-developed sensitive list and in the caution around comprehensive agreements.
• The more consequential Indian obstruction is not tariff policy but administrative and security practice: land-port restrictions, testing regimes, transit refusals and the securitised treatment of border trade.
• State governments hold effective vetoes over instruments the union negotiates — the pattern is familiar from water, and it recurs in border infrastructure and land-port operation.
• India’s willingness to use trade instruments as leverage in 2025 — the withdrawal of the transhipment facility in April, and the routing restrictions on Bangladeshi garments and consumer goods in May — confirmed for every neighbour that access to the Indian market is a political variable.

The structure common to all five

• The gains from liberalisation are diffuse and the losses concentrated, which is the standard political economy of protection everywhere, and nothing about South Asia changes it.
• What South Asia adds is a security overlay: trade concessions are read as concessions in an adversarial relationship, so the ordinary log-rolling that produces trade agreements elsewhere cannot operate.
• Asymmetry supplies the vocabulary: in every smaller capital, liberalisation is describable as absorption, and that description is electorally cheap and hard to answer.

India’s trade policy toward its own region, assessed

India’s regional trade record is usually argued about in the abstract. It is better argued about in the particular, because the particulars are more favourable to India than the rhetoric and more damaging than India admits.

Unilateral opening for the least developed members

• From 2011 India extended duty-free, quota-free access to the least developed members under SAFTA, reducing its sensitive list for them to 25 tariff lines — in substance alcohol and tobacco.
  • The offer is unilateral and non-reciprocal: Bangladesh, Nepal, Bhutan, the Maldives and Afghanistan obtained the concession without matching it.
  • It is the most open regional offer any member has made, and it is not remotely reflected in how India’s regional trade policy is described in its neighbours’ politics.
• The lesson of that gap is the important part. Tariff generosity did not change the trade outcome, because tariffs were not the binding constraint.
  • Bangladeshi exports to India did rise, but they remain concentrated in a narrow band and stuck around US$1.5–2 billion.
  • The barriers that continued to bind were testing and certification, land-port capacity, routing restrictions and rules of origin — none of which duty-free access touches.
• The lineage of the offer is the Gujral Doctrine’s principle of non-reciprocity, which held that India should give what it could to smaller neighbours without demanding equivalent return, and whose author was explicit about trade.

“We remain committed to a free trade area in South Asia by the year 2000 if possible, and latest by 2005.”

I.K. Gujral

The bilateral agreements that did the actual work

Arrangement	Status	What it actually does
India–Bhutan Agreement on Trade, Commerce and Transit	In force, current text 2016, periodically renewed	Free trade in all goods plus duty-free transit to third countries; India is about 80% of Bhutan’s trade
India–Nepal Treaty of Trade	In force, last amended 2016, auto-renews on a seven-year cycle	Duty-free access for Nepali manufactures with quantitative limits on four products; India is about two-thirds of Nepal’s trade
India–Sri Lanka Free Trade Agreement	Signed December 1998, in force March 2000	Sri Lanka’s first free trade agreement; asymmetric phase-out; trade grew roughly sixfold
India–Sri Lanka ETCA	Negotiations relaunched 2023, unconcluded	Would add services, investment and technology cooperation
India–Afghanistan Preferential Trade Agreement	2003, in force	Preferences on a limited schedule; transit is the binding constraint, not tariffs
India–Bangladesh CEPA	Study completed, negotiations not concluded and stalled since 2024	Would be the region’s most economically significant agreement
India–Maldives FTA	First round concluded July 2026	Eight negotiating areas; Malé targeting signature by end-2026

• The pattern is unambiguous: where South Asian trade has been liberalised in a way that traders use, the instrument has almost always been bilateral and Indian, not regional.
• The India–Sri Lanka agreement is the region’s one genuine success and its most instructive failure.
  • It was deliberately asymmetric in Sri Lanka’s favour — Sri Lanka received a longer phase-out and a much larger negative list than India took for itself.
  • Bilateral trade grew several-fold and India became a leading investor in Sri Lanka.
  • But the early export surge concentrated in copper and vanaspati, products where the gain came from tariff differentials rather than production, which discredited the agreement domestically at the moment it looked most successful.
  • The lesson is that rules of origin and value-addition thresholds determine whether a preference creates industry or merely arbitrage.
• The India–Nepal and India–Bhutan regimes predate SAFTA by decades and go further than it ever has — free trade plus transit — which is why neither country’s trade with India runs through the regional agreement at all.
• The India–Bangladesh partnership agreement is the largest unrealised item in the region. Studies on both sides project substantial gains; the negotiation has been suspended through the political transition, and Dhaka’s own commerce ministry has publicly noted the stall.

The 2025 reversal, and what it revealed

• India withdrew the transhipment facility on 8 April 2025, ending the arrangement under which Bangladeshi third-country exports moved through Indian land customs stations, ports and airports.
• In May 2025 India restricted Bangladeshi ready-made garments to two seaports and barred a range of consumer goods from entering through the north-eastern and West Bengal land ports.
• The stated rationale was reciprocity — Bangladesh had restricted Indian yarn and rice and intensified inspection of Indian goods, while treating the North-East as a market it could supply without granting matching access or transit.
• The revealed rationale is more important than the stated one. It demonstrated that market access in South Asia is revocable by executive decision, which raises the risk premium on every investment predicated on regional trade and is precisely the uncertainty a treaty is supposed to remove.
• The unwinding began quickly once the political context changed.
  • By March 2026 Dhaka’s commerce ministry was publicly seeking the reopening of land ports and border haats, only Benapole being fully operational.
  • By August 2026 the two sides were discussing land-port normalisation, the yarn restrictions and a joint business task force, with two-way trade around US$13 billion.

The trade surplus that shapes everyone else’s politics

• India runs a surplus with virtually every neighbour, and in most cases a very large one — Bangladesh imports roughly six times what it exports to India; Nepal’s deficit ratio has run at the order of fourteen to one; Bhutan’s imports from India are triple its exports.
• This is the single most consequential political fact in South Asian trade, and it is largely a product of relative size rather than of policy.
  • A market of some 1.4 billion people adjacent to far smaller ones will absorb a modest share of its neighbours’ output while supplying a large share of theirs, whatever the tariff schedule says.
  • But the political reading in every neighbouring capital is that the imbalance is engineered, and liberalisation is therefore framed as deepening dependence rather than as widening opportunity.
• Two implications follow for Indian policy, and India has accepted the first and resisted the second.
  • It has accepted that concessions must be non-reciprocal, which is what the 25-line sensitive list represents.
  • It has resisted the harder conclusion, which is that a surplus state must actively import — India’s imports from the region remain a negligible fraction of its total imports, and no amount of tariff generosity substitutes for Indian firms sourcing regionally.
• Investment does what trade preferences cannot. Indian corporate investment in Bangladeshi, Nepali and Sri Lankan production creates an export interest inside the partner economy and converts a zero-sum trade argument into a shared one.

Sub-regionalism and variable geometry: the path that is actually open

If the regional agreement cannot be repaired because one member can veto its repair, the alternative is to proceed with the members who are willing. That is what has in fact been happening, and its record is mixed in an instructive way.

BBIN and the motor vehicles agreement

• The Bangladesh–Bhutan–India–Nepal Motor Vehicles Agreement was signed in June 2015, months after the equivalent SAARC instrument was blocked at the Kathmandu summit — the clearest demonstration in the region’s history of what variable geometry is for.
• Bhutan’s National Council declined ratification, on environmental and cultural grounds and on concern about vehicle volumes; the other three agreed in 2017 to proceed while leaving the door open.
• The protocol text was finalised in 2025, a decade after signature, providing for designated routes, facilitation committees, customs sub-groups and electronic vehicle tracking within two years.
• The delay is the finding. A four-country agreement among willing partners on a technically straightforward matter still took ten years to reach an operating protocol, which sets a realistic expectation for anything more ambitious.
• Bhutan participates selectively, which is the honest form of variable geometry — the arrangement proceeds without unanimity and without expelling the abstainer.

SASEC: the funded, quiet track

• South Asia Subregional Economic Cooperation is the least discussed and most productive regional trade instrument in South Asia, precisely because it is project-based rather than treaty-based.
• Founded by Bangladesh, Bhutan, India and Nepal, it added the Maldives and Sri Lanka in 2014, and Myanmar in 2017, and is financed and administered through the Asian Development Bank.
• As of December 2025 it had supported 99 projects worth roughly US$26.4 billion — around US$16.6 billion in transport, US$4.1 billion in energy, US$4.35 billion in economic corridor development and close to a billion in trade facilitation.
• Its design lesson is the important one: it has no charter, no summit, no unanimity rule, and no political agenda, so it cannot be paralysed by a bilateral dispute — it builds roads, border posts and transmission lines, and the trade follows.

BIMSTEC’s unconcluded free trade agreement

• The Framework Agreement was signed at Phuket in 2004 and its constituent agreements — goods, services, investment, customs cooperation, trade facilitation, dispute settlement — remain unconcluded after more than two decades.
• Bangkok Vision 2030, adopted at the sixth summit on 4 April 2025, contains a roadmap for early completion, and six working groups are active; Bhutan has publicly pressed for conclusion, noting that the grouping is among the least integrated trade regions in the world.
• The grouping’s attraction to India is that it delivers the regional agenda without the regional veto, and its limitation is the mirror image: it cannot reach Pakistan or Afghanistan and does not cover the western half of South Asia.
• Myanmar’s civil war has stalled the land connectivity on which the Bay of Bengal logic depends, so the trade case is currently weaker than the strategic one.

Energy: the one area where integration is measurably growing

• Cross-border electricity trade is the exception that proves what is possible, because it was built through bilateral instruments and technical regulation rather than through a regional treaty.
• India’s own regulatory framework did the enabling work. Guidelines and subsequent rules opened Indian transmission to cross-border trade, allowed neighbours to sell into the Indian power exchanges, and set the approval architecture — Bhutanese projects now trade energy on the Indian exchange.
• Bhutan is the mature model: four hydropower projects totalling 2,136 MW operational, with Punatsangchhu-II (1,020 MW) commissioned in 2025; hydropower exports are the largest single item in Bhutan’s trade with India and the largest single source of its revenue.
  • The model has also stalled: the 10,000 MW target has not been met, Punatsangchhu-I remains troubled by geology, and Bhutan has moved toward equity and joint-venture structures rather than the older intergovernmental grant-and-loan format.
• Nepal has moved from importer to seasonal exporter, and India has become the transit for a third country’s power.
  • A tripartite arrangement of 3 October 2024 among the Nepali, Indian and Bangladeshi utilities enabled 40 MW of Nepali power to reach Bangladesh across Indian territory, flowing first on 15 November 2024 over the Dhalkebar–Muzaffarpur line, seasonally from mid-June to mid-November.
  • The volume is trivial; the precedent is not. It is the first time India has acted as a power transit country, and it establishes that transit is administratively possible when the commodity is politically uncontroversial.
• The limits are worth stating plainly. India’s approval framework has excluded generation with investment from countries with which India lacks a bilateral power agreement, which in practice restricts Chinese-financed Nepali projects — so even the region’s most successful integration carries a geopolitical filter.

Vehicle	Membership	Instrument type	What it has delivered
SAFTA	All eight	Treaty, unanimity	Tariff reduction that did not move trade
BBIN	Four, operating as three	Treaty, coalition of willing	A motor vehicles protocol after ten years
SASEC	Seven	Project-based, bank-financed	99 projects, about US$26.4 billion
BIMSTEC	Seven, Bay of Bengal	Treaty, inter-regional	Charter in force 2024; free trade agreement still unconcluded
Bilateral FTAs	India plus one	Treaty, bilateral	The only liberalisation traders actually use
Cross-border power trade	Bilateral and trilateral	Regulatory and commercial	Growing volumes; first third-country transit, 2024

India’s regional trade policy has been bilateral in practice and multilateral only in rhetoric — and the bilateral half is the half that worked.

What greater economic cooperation would actually require

A list of desirable reforms is not an argument. The useful exercise is to sort the measures by who has to agree to them, because that determines what is available now and what has to wait — and the ordering turns out to be almost the reverse of the order in which they are usually proposed.

Tier one: what India can do alone, and should

• Import more. India’s imports from South Asia are a negligible share of its total imports; a deliberate sourcing policy, backed by public procurement and by Indian firms’ supply chains, changes the political arithmetic in every neighbouring capital faster than any tariff cut.
• Fix the land ports. Warehousing, testing laboratories, parking, cold chain and round-the-clock operation at Petrapole–Benapole, Raxaul–Birgunj, Phuentsholing and Agartala–Akhaura are unilateral investments with immediate bilateral returns.
• Make market access predictable. The 2025 restrictions demonstrated that access is revocable; a published, rule-based procedure for restrictions would cost India little and buy credibility that no concession can.
• Mutual recognition of test results, beginning with the products that dominate regional trade, removes the largest single non-tariff cost without requiring any partner to change its standards.
• Liberalise business visas unilaterally for the neighbours, on a multi-entry long-validity basis — the cost is administrative, the benefit accrues to services trade and to the buyer–supplier relationships that goods trade depends on.
• Invest in the neighbours’ production, since Indian equity in a Bangladeshi or Nepali exporter creates a domestic lobby for access to India where none exists today.

Tier two: what two or three willing partners can do

• Conclude the bilateral agreements already on the table — the partnership agreement with Bangladesh, the services and investment successor with Sri Lanka, the agreement with the Maldives — since each delivers more than a revived SAFTA would.
• Operationalise BBIN fully, including cargo movement under the finalised protocol, with Bhutan participating selectively.
• Extend the power trading architecture: more transmission interconnection, a functioning regional market design, and additional third-country transit arrangements of the kind the 2024 Nepal–Bangladesh flow established.
• Build the sub-regional corridor as a production network, not a road — the eastern quadrant has the density and the complementarity for genuine value-chain fragmentation if border frictions fall.
• Payment and settlement is the quietly transformative item: the Asian Clearing Union already exists and settles on a bimonthly cycle; rupee vostro arrangements, currency swaps, and the UPI and RuPay linkages with Bhutan, Nepal, Sri Lanka and the Maldives reduce the dollar constraint that binds small deficit economies hardest.
  • The reciprocity gap is the visible weakness — neighbours’ citizens have had far less access in India than Indians have had abroad — and closing it is cheap.

Tier three: what needs the region, and therefore waits

• Pruning sensitive lists on a time-bound schedule, with a residual exception category capped at a small share of tariff lines, is the core SAFTA reform and requires the ministerial machinery to function.
• Bringing para-tariffs inside the liberalisation schedule is technically the single most valuable change and politically the hardest, because it is where the protection actually lives.
• A services agreement with real schedules of commitments — the existing instrument has been in force since 2012 without them — is where the complementarity is, and where the region’s deficit countries would gain rather than lose.
• A regional transit regime is the one measure that would transform the region’s economic geography, and it is hostage to the India–Pakistan relationship in the west and to trust in the east.
• Harmonised standards through the regional standards body is the slowest and most durable of the reforms; its governing board met at Dhaka on 3–4 June 2026 with seven members present, which is a reminder that the technical machinery still convenes when the political machinery does not.

Measure	Whose agreement is needed	Political feasibility now
Import more from the region	India alone	High — a decision, not a negotiation
Land-port and testing infrastructure	India alone, bilaterally coordinated	High
Predictable, rule-based market access	India alone	High, and currently unmet
Business visa liberalisation	India alone	Medium — security objections are the obstacle
Bilateral comprehensive agreements	India plus one partner	Medium, and improving with Bangladesh and the Maldives
BBIN cargo operation	Three or four	Medium
Power trade and third-country transit	Two or three	High — already happening
Sensitive-list pruning	All eight	Low while the ministerial machinery is dormant
Para-tariffs inside the schedule	All eight	Low — this is where the protection sits
Services schedules	All eight	Low
Regional transit regime	All eight	Very low in the west, plausible in the east

“No serious connectivity initiative can ever be a one-way street.”

S. Jaishankar

How the argument is conducted: the main positions

The literature on South Asian integration is not a single body of opinion. It divides into recognisable positions with different diagnoses, and the policy prescription follows from which diagnosis one accepts.

Position	Diagnosis	Prescription
The conflict school	Integration is blocked by unresolved political disputes; economics is downstream	Settle or contain the disputes first; expect little until then
The trade-cost school	The binding constraints are logistics, standards and border administration, not politics	Attack trade costs directly; they are technical and negotiable
The political economy school	Protection persists because concentrated domestic interests defend it	Build offsetting export and investment constituencies
The open-regionalism school	Preferential agreements are neither necessary nor sufficient; unilateral opening is	Liberalise on a most-favoured-nation basis and let the region benefit
The sub-regionalist school	The eight-member frame is the problem, not integration itself	Proceed with the willing; build corridors, not charters

• Amita Batra is the clearest exponent of the conflict reading, arguing that South Asia’s shortfall is measurable against gravity predictions and is explained by disputes rather than by endowments.
• Sanjay Kathuria and the trade-cost literature locate the problem in connectivity, para-tariffs, sensitive lists and the trust deficit, and argue that each is addressable without a political settlement.
• Nisha Taneja’s work on informal trade supplies the political economy school’s best evidence: the demand for regional trade already exists and is being met illegally.
• Selim Raihan and the Bangladeshi literature press the services case, arguing that the region’s complementarity is stronger in services than in goods and that the deficit economies would gain there.
• Jayant Menon makes the open-regionalism argument sharply, urging that any concession India grants a partner be multilateralised so the agreement does not become a discriminatory instrument.
• Prabir De and the connectivity literature argue the sub-regionalist case in its economic form: corridors, border infrastructure and trade facilitation deliver measurable returns while treaty negotiation does not.
• The positions are less opposed than they look. Each identifies a real constraint; what separates them is which constraint they think binds first, and the honest answer is that different constraints bind in different corridors.
  • In the west the conflict school is plainly right — no amount of trade facilitation moves India–Pakistan trade while the border is closed.
  • In the east the trade-cost and political economy schools are right — the political relationship permits far more trade than actually occurs, and the shortfall is administrative.

Is a South Asian free trade area worth the effort at all?

The strongest objection to everything above is not that the barriers are hard to remove. It is that removing them would not be worth very much — and that objection deserves a serious answer rather than a slogan.

The case against

• The complementarity is genuinely limited. These are similar economies at similar income levels producing similar goods; the textbook gains from trade are correspondingly modest.
• The arithmetic is unflattering. Even the optimistic estimate — from about US$23 billion to about US$67 billion — is a rounding error against India’s total trade, and a US$44 billion increment spread over eight economies does not transform any of them except possibly Nepal and Bhutan.
• India’s growth partners are elsewhere, and the revealed preference of Indian trade policy reflects it: comprehensive agreements with the United Arab Emirates, Australia, the European Free Trade Association bloc and the United Kingdom, and a deepening trade relationship with the United States and East Asia.
• The opportunity cost of negotiating capacity is real. A trade bureaucracy that spends years on a regional agreement worth tens of billions is not spending them on agreements worth hundreds.
• Trade diversion is a live risk, not a theoretical one, given the region’s high external tariffs — a preferential arrangement among high-tariff economies may simply move purchases toward inefficient regional producers.
• The historical record is discouraging. Two goods agreements in three decades, a services agreement that never acquired schedules, a motor vehicles protocol that took ten years, and a Bay of Bengal free trade agreement unconcluded since 2004.

The reply

• The volume argument attacks the weakest version of the case. Nobody serious claims a South Asian free trade area would transform Indian growth; the case rests on three things volume does not measure.
• Connectivity has returns far beyond the trade it carries. The eastern corridor’s binding constraint is that a coherent economic geography — the Ganga–Brahmaputra basin — is administered as four separate trade regimes; removing the friction raises returns on infrastructure India is building anyway, and the North-East’s development case depends on it entirely.
• Formalising informal trade is a gain the trade statistics do not show. If unrecorded India–Bangladesh trade approaches the recorded figure, the reform that captures it delivers revenue, regulation and measurement without needing a single new buyer.
• Interdependence has political value not reducible to its economic value. A neighbour whose firms depend on Indian inputs and Indian buyers has a domestic constituency for stability; a neighbour with no commercial stake in India has none, and hedging toward Beijing is correspondingly cheap.
• The counterfactual is not neutrality but substitution. Where India does not supply connectivity, finance and market access, another power does, and the terms on which it does so are visible across the region’s ports, transmission lines and debt schedules.
• The gains are concentrated where India’s strategic interest is. Bhutan, Nepal, Bangladesh and Sri Lanka would gain far more proportionally than India would — which is the point, because the purpose of an asymmetric concession is to buy an asymmetric outcome.
• The strongest institutional reply is that the alternatives are not substitutes. The Bay of Bengal grouping cannot reach Pakistan or Afghanistan; bilateral agreements cannot create transit across a third country; sub-regional projects cannot harmonise standards. Something regional is required for the residue, however dormant it is now.

“We are also willing to look at bilateral frameworks for trade liberalization in case SAPTA and SAFTA are prevented from taking-off.”

I.K. Gujral

Conclusion

The South Asian free trade area failed in a specific and instructive way: the tariffs came down and the trade did not move, because the protection had already migrated into para-tariffs, testing regimes, transit refusals and border infrastructure that no tariff agreement could reach. India’s own record inside that failure is better than its neighbours say and worse than it believes — the most generous tariff offer in the region, delivered alongside the administrative practices and the 2025 restrictions that made the offer worth less than it looked.

• The realistic programme is not a revived SAFTA. It is unilateral Indian opening on the measures that need nobody’s consent, bilateral agreements with the partners who want them, sub-regional operation among the willing, and the regional machinery kept alive for the transit and standards questions that only it can eventually answer.
• The measure of success is not the intra-regional share. It is whether a truck crosses Petrapole in hours rather than days, whether a Nepali exporter can obtain a test certificate that India accepts, and whether market access survives a diplomatic quarrel.
• The deepest obstacle is not economic and never was. A region whose members trade more cheaply with Brazil than with each other has made a political choice, sustained over seventy years, and it can be unmade the same way.

The case for a South Asian free trade area does not rest on the volume of trade it would create. It rests on what a region that trades cannot easily do to itself.