The India–China border accords and what they bind

The Line of Actual Control is not a border. It is a line of separation that the two governments have never jointly mapped, and every agreement they have signed about it since 1993 is written to manage that fact rather than to remove it. Read together the instruments form a conduct regime rather than a settlement: they accept the LAC as the operative line without agreeing where it lies, cap what either army may do near it, and push disagreement into standing channels. The boundary question proper sits in a separate and much slower process. Knowing which document governs which behaviour is what makes an incident on the line legible instead of merely alarming.

The instruments, and what each one binds

1993 — Agreement on the Maintenance of Peace and Tranquility along the Line of Actual Control. Signed in Beijing on 7 September 1993. Its central move is procedural: pending a boundary settlement, each side will respect and observe the LAC, and neither will use force against the other. It also commits both to keep military forces in the border areas to minimum levels compatible with friendly and good-neighbourly relations, and refers questions about where the line runs to the joint machinery the two states had set up after the 1988 thaw. The significance is not the text but the concession inside it: for the first time both governments agreed to respect a line they disagree about.

1996 — Agreement on Confidence-Building Measures in the Military Field along the LAC. Signed in New Delhi on 29 November 1996. This is where the specific prohibitions live. Neither side is to use its military capability against the other. Major exercises near the line are limited and notified in advance. And within two kilometres of the LAC there is to be no firing of firearms, no blast operations, and no hunting with guns or explosives. That last clause is the one that later shaped how violence on the line actually happens.

2003 — the Special Representatives channel. A political-level negotiation on the boundary question itself, deliberately lifted out of the hands of surveyors and border commanders. It is the only track whose stated purpose is a settlement rather than management.

2005 — two documents on one day. On 11 April 2005 the two states signed a Protocol on Modalities for the Implementation of Confidence-Building Measures, which tells troops what to do when patrols meet — self-restraint, no escalation, banner drills, disengage and return — and an Agreement on Political Parameters and Guiding Principles for the Settlement of the Boundary Question, the framework for an eventual deal, including an undertaking to safeguard the interests of settled populations in the border areas.

2012 — the Working Mechanism for Consultation and Coordination on India-China Border Affairs. Established on 17 January 2012 as a standing channel staffed by diplomats and military officers, the WMCC is the routine venue where incidents are reviewed. The 2013 agreement names it as one of its own implementing mechanisms.

2013 — the Border Defence Cooperation Agreement. Signed in Beijing on 23 October 2013. Article VI obliges neither side to tail the other’s patrols in areas where there is no common understanding of the LAC. Article VIII requires maximum self-restraint and no use of force in face-to-face situations. Article IX states that the agreement operates without prejudice to either side’s position on the alignment of the LAC or on the boundary question — the clearest statement anywhere in the framework that managing the line and settling it are separate projects.

The design principle

Every one of these instruments is built on the same trade. India and China cannot agree where the line is, so they agree on what may not be done near it. That buys predictability at the price of permanence: conduct is constrained, claims are preserved, and nothing that either state regards as sovereign territory is conceded on paper. It also means the framework has no way of resolving the underlying dispute, and was never designed to. Treating it as a failed peace process misreads what it was for; treating it as a settlement misreads it in the other direction.

Why the violence in June 2020 was hand to hand

The clause in the 1996 agreement barring firearms within two kilometres of the line, and the self-restraint provisions of the 2005 protocol and the 2013 agreement, are the reason the deadliest border incident in decades was fought without rifles. Twenty Indian soldiers were killed in the Galwan Valley on 15 June 2020, with the Chinese toll left unconfirmed by Beijing. Two things are true about that at once, and both belong in any honest reading: the accords did what they were written to do, in that no exchange of fire occurred and the incident did not escalate into a shooting war; and they did not prevent men from being killed on the line. The framework constrains the form of violence more reliably than it constrains its occurrence. That is an IndiaStand assessment of the texts, labelled as such, not a claim either government makes.

The channels, and what each one can reach

Channel	Level	What it is for	What it cannot do
Special Representatives	Political	The boundary question and the 2005 political parameters	Move troops or settle a local face-off
WMCC	Diplomats and military officers	Reviewing border affairs and incidents, preparing higher-level talks	Bind either army by itself
Corps commander talks	Senior military, theatre level	Disengagement at named friction points	Address the claim line
Border personnel meetings and hotlines	Unit level	Routine contact, immediate de-confliction	Resolve a disputed alignment
Expert Group on Boundary Delimitation; Working Group on Border Management (under the WMCC)	Officials; tasked by the SRs in August 2026	An “Early and Substantial Harvest” on delimitation and on border management	Bind either government until terms of reference are agreed — their first task

The layering is deliberate. An incident that a hotline can defuse should never reach the Special Representatives, and a claim dispute that only the Special Representatives can address will not be solved by another round of corps commander talks — which is why rounds of the latter can continue for years without the boundary question moving at all.

The channels thickened in 2026. At their 25th round in Beijing on 25 August the Special Representatives tasked the two WMCC groups with an “Early and Substantial Harvest”, agreed two additional meeting points for general-level military talks and two additional military hotlines in the Eastern and Middle Sectors, and committed to discuss “ways to improve understanding of LAC in appropriate areas” (MEA). None of that is yet a new instrument with published text; it is an agreed agenda and additional plumbing.

What the framework does not cover

It does not delimit. There is no jointly agreed map of the LAC. Maps of the middle sector were exchanged in the early 2000s; the exercise was never extended to the western sector, which is where the 2020 crisis happened. Without a common map, “transgression” is a claim each side makes in its own cartography. The expert group on delimitation tasked in August 2026 is charged with exactly this question; as of October 2026 no terms of reference for it had been published.

It has no verification body and no penalty. Compliance is asserted, not inspected. The remedy for a breach is another meeting.

It does not reach third parties. The 2017 Doklam standoff sat on territory claimed by Bhutan, outside the bilateral instruments entirely — which is why the framework offered so little purchase on it.

Its most recent addition is unpublished. The patrolling arrangement announced in October 2024 for Depsang and Demchok has not been released as a text. What it obliges either army to do is therefore not publicly checkable, and any account of its contents is sourced to briefings rather than to a document. We note the gap rather than filling it.

Settled, and contested

Settled: the instruments exist, their texts are published by the Ministry of External Affairs, and they say what they say. The dates, the articles and the prohibitions above are all readable in the primaries linked below.

Contested: what the framework is worth. One position holds that it is a floor — a set of rules that has kept the most militarised disputed boundary in the world from producing a war for six decades. Another holds that it is a ceiling, freezing a status quo that has moved against India on the ground while the paperwork stayed constant, and that buffer zones created by disengagement have cost patrolling access that no agreement restores. A third holds that the 1993 framework was built for a border managed by small patrols and is poorly suited to one with permanent forward deployments and hard infrastructure on both sides. Those are the positions actually held. This brief attributes them and does not adjudicate between them.

How this connects on IndiaStand

The durable record sits in the India–China dossier, whose timeline carries each of these instruments with its source tier. The ministry that negotiates and publishes them is the Ministry of External Affairs; the force that lives with them is the Indian Army, under the Ministry of Defence. For the live thread on the current round of talks, see the dispatch on delimitation talks resuming.

Maintained topic brief. Analysis by IndiaStand — it describes the instruments and the range of positions actually held, attributes each claim, and makes no forecast and no recommendation.