Few statutes change their meaning as completely as the Patents Act, 1970 changed between the last day of 2004 and the first day of 2005. On the earlier date, an Indian drug company could lawfully manufacture and sell a medicine invented abroad so long as it used its own process to make it. On the later date, the same act of manufacture, for a medicine covered by a granted product patent, was infringement. The Patents Amendment Act 2005 is the instrument that carried India across that line, and it is usually described as the moment Indian patent law surrendered to the TRIPS Agreement. That description is half right and wholly misleading. The amendment did the one thing the Agreement compelled, which was to make product patents available in the fields India had kept closed since 1970. Almost everything else in the Act, including the provisions that have made Indian patent law internationally distinctive and internationally contested, was a choice India made inside the space the Agreement left open, and several of those choices were written into the Bill during its passage through Parliament in March 2005 rather than into the Ordinance that preceded it.

That is the arc this article traces. India reached product patents in three amendments spread over a decade: a first amendment in 1999 that built the transitional machinery the Agreement required, a second in 2002 that completed general compliance outside the closed fields, and the third in 2005, which opened the closed fields and closed the transition. Read in sequence, the three amendments show a legislature conceding the single point the treaty actually specified and contesting nearly every other point. The reader who finishes this article should be able to say what the Agreement required, what it merely permitted, what India did with the permission, and why the loudest argument about Indian patent law is an argument about choices rather than about compliance.

A word on method, because the genre of this article is an audit of a trail rather than a commentary on a single Act. An amendment trail treats the statute book as the transcript of a negotiation. Every insertion, substitution and omission in the Patents Act between 1999 and 2005 answers to a pressure: a treaty article, a dispute ruling, a coalition partner, a campaign, a court. The trail method reads the text in that order, asking of each provision what required it, what permitted it, and what nearly prevented it. Provisions that look eccentric in a bare Act, a queue that held applications unexamined for ten years, a marketing right designed to expire, a royalty shield for producers the patent system was about to overtake, become legible when they are read as moves in a sequence whose other players sat in Geneva, Washington and the Indian Parliament at once. That is how the sections that follow are organised, and it is why the article keeps returning to dates: in a trail of this kind, the date of a provision is usually the most informative thing about it.

The Act India started with: process patents and Section 5

The starting point matters because the 2005 amendment is unintelligible without it. The Patents Act, 1970 was designed, on the foundation of the Ayyangar Committee’s work, to subordinate patent rights in food and medicine to public access. Its central device was Section 5. For inventions claiming substances intended for use as food, medicine or drugs, and for substances produced by chemical processes, Section 5 allowed patents only on the methods or processes of manufacture. The substance itself could not be patented. A foreign originator could patent its route to a molecule in India; it could not patent the molecule.

The practical consequences shaped an industry. Indian manufacturers became exceptionally skilled at devising alternative processes for known molecules, and because the product was unprotected, several firms could lawfully sell the same medicine in competition with each other. Process patents in these fields also ran for a shorter term than patents elsewhere in the Act: the shorter of five years from grant or seven years from filing, against a general term that was itself modest by later standards. Prices for medicines in India fell to levels that made the country a supplier of affordable generics far beyond its own borders, and the industry that grew up under Section 5 became the constituency with the most to lose from its repeal.

Section 5 did not stand alone in the original design. The 1970 Act surrounded it with devices that weakened even the process patents it allowed in the sensitive fields. Patents for food, medicine and surgical or curative devices were to be deemed endorsed with the words licences of right three years after grant, so that any person could take a licence on terms settled by the Controller if agreement failed. Government use provisions allowed the state to use patented inventions for its own purposes on payment, and the compulsory licensing chapter stood behind the whole. The philosophy was consistent: in the fields of food and health, a patent was treated from the outset as a qualified and defeasible right, closer to a regulated franchise than to a property the state would simply defend. That philosophy is the deeper thing the TRIPS amendments had to displace, and the trail is best read as the state retreating from it field by field while salvaging as much of its machinery as the treaty allowed. The licences of right device itself did not survive the rebuilding of the licensing chapter, but the instinct behind it, that a medicine patent carries social conditions, reappears throughout the later settlement in modernised forms.

The full architecture of the parent statute, its exclusions, its licensing chapter and its enforcement machinery, is set out in the guide to the Patents Act 1970, which owns that ground in this series. What matters for the amendment trail is narrower: by the early 1990s the Indian settlement rested on a single load-bearing provision, and the international negotiation that produced the TRIPS Agreement was always going to strike at exactly that provision, because the Agreement’s central patent rule is that patents must be available in all fields of technology.

The industry at stake: why the world watched an Indian filing rule

It is worth pausing on what Section 5 had built by the time the treaty struck at it, because the international attention the 2005 amendment received is otherwise inexplicable. Under the process-only regime, Indian firms had industrialised reverse engineering. A company that could devise a cheaper synthesis for a known molecule could enter the market lawfully and immediately, and competition among several such entrants drove prices down in a way no regulator had to engineer. The capabilities compounded: bulk drug manufacture, formulation skills, regulatory filings in export markets, and eventually research capacity of the industry’s own. By the time the transition ran out, Indian companies supplied affordable generics across the developing world, and treatment programmes run by governments and charities had been designed around Indian prices.

The figure that made the point in the passage debates came from the HIV treatment programme. Médecins Sans Frontières, whose access campaign followed the amendment through Parliament, recorded that half of the 700,000 people then receiving antiretroviral treatment in developing countries relied on Indian generic products, and that the price of a year of such treatment had fallen from about 10,000 dollars to about 250 dollars under Indian competition. Whether or not every critic accepted the campaign’s conclusions, the arithmetic explained the diplomatic weather: an amendment to the Indian Patents Act was, in effect, a decision about the medicine supply of countries that had no seat in the Indian Parliament. That is why a Bill about filing dates and claim forms drew protests in several capitals, why the Ordinance of December 2004 produced a global day of action by treatment activists the following February, and why the safeguards inserted in March 2005 were reported abroad as news. The amendment trail in this field is a domestic legislative record with an international shadow, and both halves of that description have to be kept in view.

The obligation: what TRIPS actually required of India

India was a founding member of the World Trade Organization from the beginning of January 1995, and membership carried the TRIPS Agreement with it. The Agreement’s core patent command sits in its Article 27: patents must be available for any inventions, whether products or processes, in all fields of technology, without discrimination as to the field of technology, provided the invention is new, involves an inventive step and is capable of industrial application. A regime that refused product patents for medicines and chemicals could not survive that command indefinitely. The dispute was never about whether Section 5 had to go. It was about when, and about everything else.

The timing was governed by the transition architecture. The Agreement gave developing country members a general period before full application, and it gave a member that did not provide product patent protection in a field of technology a further period, running to the end of 2004, before it had to open that field. India used the whole of that additional period. Product patents for pharmaceuticals and agricultural chemicals were therefore a 2005 obligation, not a 1995 one, and accounts that date Indian product patents to 1995 mistake the filing machinery for the right itself.

Two interim duties attached to the delay, and they explain the strangest features of Indian patent law in the decade that followed. The first, in Article 70.8 of the Agreement, required a member using the additional period to provide a means by which applications for patents for such inventions could be filed from 1995, so that novelty and priority could be preserved against the day the field opened. This is the mailbox: applications received, date-stamped, and held unexamined. The second, in Article 70.9, required the grant of exclusive marketing rights for a product that was the subject of a mailbox application, where a patent and marketing approval had been obtained in another member and marketing approval obtained in India, for five years or until the patent application was granted or rejected, whichever period was shorter. The treaty framework in which these duties sit, and the way the same Agreement structured India’s obligations across the whole intellectual property field, is mapped in the guide to intellectual property law in India.

Which questions did TRIPS leave unanswered?

It left the operative ones. The Agreement never defines novelty, inventive step or industrial application, sets no grounds on which a compulsory licence may or may not issue, does not prescribe opposition procedure, and leaves exhaustion of rights to each member. Its silences, not its commands, are what Indian law was later built inside.

Two features of the treaty text show how deliberate the openness was. The first is the treatment of compulsory licensing in Article 31. The Article does not list the grounds on which a licence may be granted at all. It imposes conditions on the manner of granting: authorisation must be considered on its individual merits, the proposed user must ordinarily have tried first to obtain a voluntary licence on reasonable commercial terms, the use must be predominantly for the supply of the domestic market, and adequate remuneration must be paid. A member may therefore license for non-working, for excessive pricing, for public health, or for dependent patents, and the Article is satisfied so long as the procedure is honoured. The second is exhaustion, where Article 6 removes the question from dispute settlement altogether, leaving each member free to treat rights as exhausted by a first sale anywhere, which is the legal foundation of parallel importation. Even the exclusions the Agreement itself demonstrates, in Article 27, are instructive: it permits members to exclude inventions whose commercial exploitation must be prevented to protect ordre public or morality, and to exclude diagnostic, therapeutic and surgical methods and plants and animals other than micro-organisms. The negotiators who wrote those permissions understood patentability to be a field of national judgment bounded by a floor, not a uniform code.

The transition itself was built in two stages, and the trail’s periodisation follows the construction. The first stage, in Article 65.2, gave developing country members five years, to the end of 1999, to apply the Agreement generally. The second, in Article 65.4, addressed the specific case of a member that did not provide product patent protection in an area of technology: such a member could delay the product patent provisions for that area for a further five years, to the end of 2004, on condition that the mailbox and marketing right duties of Articles 70.8 and 70.9 were performed from the start. The least developed members stood on a different footing again, with longer periods that the membership has repeatedly extended, the pharmaceutical transition for those members running, under the extensions agreed, to 2033. India’s case was the middle one, and its difficulty came from the condition rather than the delay: the further five years were not a holiday but a supervised deferral, in which applications accumulated under foreign scrutiny and interim exclusivity had to be granted on demand. Every date in the Indian trail, 1995, 2000 and 2005, is one of these treaty dates arriving in the statute book.

Nor could India have declined the floor while keeping the rest of the trading system. The WTO agreements were concluded as a single undertaking: membership meant accepting the package, in which the disciplines on intellectual property were the price of secured access for agricultural and textile exports and of a dispute system that bound stronger economies as well as weaker ones. Indian negotiators had resisted the inclusion of intellectual property in the trade round at all, and the transition periods, the mailbox conditions and the silences described here were what resistance obtained. That provenance matters for reading the amendments fairly. They were not the work of a government converted to a foreign doctrine; they were the work of a state performing an unwelcome bargain at the last dates and the narrowest scope the bargain allowed.

That division between command and silence is the analytical key to everything that followed, and it deserves to be stated precisely. The Agreement fixes the floor of protection: patents available in all fields, a twenty-year term, defined exclusive rights, enforcement procedures, and a rule that a court must be able to shift the burden of proof in process patent cases in specified circumstances. It does not fix the ceiling, and on the questions that determine how a patent system actually behaves, including what counts as an invention at all, it is deliberately open-textured. Article 1 of the Agreement states that members are free to determine the appropriate method of implementing its provisions within their own legal system and practice. India’s three amendments were drafted in continuous awareness of that openness. The first two amendments dealt with what the Agreement specified. The third, while performing the one act of compliance that remained, legislated extensively in the spaces the Agreement had left unspecified, which is why the Act of 2005 can be simultaneously a compliance instrument and the foundation of a restrictive patent regime.

The first amendment: the mailbox and exclusive marketing rights

The transition machinery arrived late and under legal compulsion of a specific kind. The Government of India first attempted to create the mailbox by an Ordinance issued at the end of December 1994, days before the Agreement took effect. That Ordinance lapsed without being replaced by an Act, leaving the filing facility without a statutory basis. The United States then brought the first dispute ever filed under the TRIPS Agreement, contending that India had failed to provide the means of filing and the exclusive marketing rights that Articles 70.8 and 70.9 required. The panel ruled against India, the Appellate Body upheld the substance of the ruling, and a parallel complaint by the European Community on marketing rights succeeded as well. India accepted that the arrangements had to be placed on a sound statutory footing.

The reasoning of the two rulings deserves a sentence each, because they fixed the drafting standard the 1999 Act had to meet. India’s defence had been that instructions to patent offices to accept the applications, and to date-stamp them, satisfied the filing duty in substance. The panel and the Appellate Body rejected the equation of administrative practice with legal obligation: a means of filing had to exist under the municipal law of the member in a form that gave applicants legal certainty, and a discretionary administrative arrangement that a government could withdraw, and that the statute itself contradicted, did not qualify. The European Community’s complaint, decided by a panel the following year, applied the same logic to marketing rights, which the lapsed Ordinance had created and nothing had replaced. Together the two disputes taught a lesson the Indian drafting of the period absorbed thoroughly: under the TRIPS Agreement, transitional promises had to be enacted, not announced. The 1999 Act is written in that shadow, which is why a measure about queueing applications bothers to validate four years of filings retrospectively and to repeal and save the two Ordinances, the lapsed one of 1994 and its own predecessor of 1999, in terms.

The chronology of the compulsion is worth fixing precisely. The United States requested consultations on 2 July 1996, in what became the first dispute ever adjudicated under the TRIPS Agreement; the panel and the Appellate Body ruled against India across 1997, and the ruling was adopted in January 1998 with a compliance period that ran to April 1999. The European Community’s complaint on marketing rights followed, with consultations in 1997 and a panel ruling in 1998 to the same effect. The 1999 Act, assented to on 26 March 1999, therefore landed inside the allowed period with weeks to spare, the single instance in the whole trail of India legislating ahead of a deadline rather than on it, and the exception proves the rule this article has pressed: the calendar that governed Indian drafting was set by obligation, and only an adverse ruling could advance it.

The result was the Patents (Amendment) Act, 1999, Act 17 of 1999, deemed to have come into force on 1 January 1995. Its retrospectivity was functional rather than decorative: it validated the applications that had been filed in the interval, including those filed while the lapsed Ordinance had purported to authorise them. The Act inserted a new sub-section into Section 5 providing that a claim for a patent for a substance intended for use as medicine or drug could be made and would be dealt with in the manner provided by a new chapter, and it inserted that chapter, Chapter IV-A, containing Sections 24A to 24F.

Section 24A directed the Controller not to refer mailbox applications for examination until the end of 2004. The marketing rights provisions then set their conditions. An applicant who had filed the mailbox claim, who had been granted a patent and marketing approval for the same invention in a convention country, and who had received approval to sell or distribute the article or substance in India from the authority the Central Government specified, could receive the exclusive right to sell or distribute it in India. That approval route ran through the medicines regulator, and the regulatory statute under which such approvals sit is treated in the article on the Drugs and Cosmetics Act 1940. The right lasted five years from the Controller’s approval or until the grant or rejection of the patent, whichever came first. Products based on the Indian systems of medicine and articles already in the public domain were carved out of the scheme.

Measured against the political temperature of the subject, exclusive marketing rights turned out to be a small instrument with one enormous illustration. On figures released by the office of the Controller General of Patents, Designs and Trade Marks, fourteen applications for exclusive marketing rights were filed and four were granted across the whole transitional decade. The grant that mattered was made in November 2003 to Novartis for imatinib mesylate, sold as Glivec, a treatment for chronic myeloid leukaemia whose mailbox application had been filed in 1998. Novartis enforced the right in the Madras High Court and obtained injunctions against generic manufacturers, who had been selling the medicine at roughly a tenth of the branded price: the Cancer Patients Aid Association, which later opposed the patent application itself, put the branded price at about 120,000 rupees for a month’s treatment against some 10,000 rupees for the generic versions. The episode demonstrated, in a single medicine, exactly what product protection would mean when it arrived in full, and it fixed the political terms on which the 2005 Bill would be debated: the argument was about the price of cancer treatment, conducted through the machinery of filing dates and marketing rights.

The repeal of the marketing rights chapter in 2005 raised no transitional puzzle, because each subsisting right was already bounded by the fate of its underlying application: it endured for five years or until the patent was granted or refused, whichever came first. The Glivec right accordingly ran on after the chapter that created it had been deleted, and was extinguished by the refusal of the application in January 2006 rather than by the repeal. There is a drafting lesson in the episode that the trail repeatedly teaches. Transitional rights designed with their own expiry built in need no savings clause and generate no vested-rights litigation when the transition closes; the marketing right, for all the controversy of its one great exercise, wound itself up on schedule and left the register clean for the regime that replaced it.

One further feature of the marketing right deserves emphasis, because it inverts the usual assumption that interim arrangements are weaker than final ones. The exclusive marketing right issued without any examination of the Indian application’s merits, it was subject to no opposition, and it was enforceable by injunction in the ordinary courts against producers who enjoyed no statutory shield, the proviso to Section 11A(7) being addressed to patents and not to marketing rights. For the five years or less of its life, in other words, the holder of a marketing right possessed a purer exclusivity than any patentee in the later system, whose grant must survive examination, opposition and the efficacy threshold before it can be enforced at all. The Glivec injunctions of 2004 were granted under this interim instrument. A trail audit should mark the paradox plainly: the strongest medicine monopoly of the Indian transition was created not by the compliance amendment of 2005 but by the transitional device of 1999, and the politics of 2005 are unintelligible unless that is understood, because Parliament in March of that year was legislating in the shadow of an exclusivity it had already seen operate at full strength.

The 1999 amendment did nothing else of substance, and that restraint was its design. It performed the two interim duties the Agreement imposed, retrospectively and completely, and it left Section 5(1) standing. India’s process-only regime for medicines survived the first amendment intact; what changed was that a queue of product claims had begun to form behind it, date-stamped and waiting for 2005.

The route to the 1999 Act itself ran through an ordinance. A Patents (Amendment) Bill introduced in 1998 to enact the same machinery had not completed its passage when the period allowed for compliance began to run out, and the government filled the interval with an Ordinance in January 1999, which the Act then repealed while saving everything done under it. The episode explains the Act’s doubled retrospective architecture. Twice in five years, in 1994 and again in 1999, an ordinance carried the mailbox across a gap that parliamentary time had opened, and twice a validating Act had to reach back to 1 January 1995 to make the filings good. Ordinance, delay and retrospective validation are not decorations on this trail; they are its recurring mechanism whenever a treaty deadline and a parliamentary calendar collide, and the pattern completes itself a third time in 2005.

The second amendment: general compliance in 2002

The Patents (Amendment) Act, 2002, Act 38 of 2002, which came into force on 20 May 2003, is the least discussed of the three and in some ways the most consequential for the ordinary working of the system. The general transition period for developing countries had expired at the beginning of 2000, and the 2002 Act brought the statute into line with the obligations that were already due, leaving only the product patent question for the final round.

Its term provision ended the patchwork under which process patents in the food and drugs fields ran for a shorter period than patents elsewhere. Section 53 was recast so that the term of every patent became twenty years from the date of filing, the period the Agreement prescribes. Its enforcement provisions inserted Section 104A, under which a court trying an infringement suit over a process for obtaining a product may direct the defendant to prove that its own process is different, where the patented process produces a new product or where the statutory likelihood conditions are met and the patentee has been unable through reasonable efforts to discover the process actually used. That reversal of the burden of proof tracks a specific requirement of the Agreement and it mattered greatly in a system built on process claims, where the defendant’s factory was the one place the evidence sat.

The exclusions from patentability were expanded and systematised in Section 3, including a clause excluding an invention which in effect is traditional knowledge or an aggregation of known properties of traditionally known components. The compulsory licensing chapter was revised, and the special provision for licences on government notification, Section 92, was recast to speak expressly of national emergency, circumstances of extreme urgency and public health crisis. The Act also made provision for international applications under the Patent Cooperation Treaty, which India had joined in December 1998, so that a single international filing could enter the Indian system. And it inserted Section 107A, which took two sets of acts outside infringement altogether: acts done solely for uses reasonably related to the development and submission of information required under a law regulating the manufacture or sale of a product, in India or elsewhere, which is the research and regulatory use exemption that generic manufacturers use to prepare a launch before a patent expires; and importation of a patented product from a person duly authorised to produce and distribute it, which is the parallel importation limb. Both limbs of Section 107A are flexibilities the Agreement preserves rather than requires, and both were in place before the 2005 Act widened one of them further.

The Patent Cooperation Treaty route regularised by the same Act changed the composition of the queue that 2005 would inherit. A single international application could thereafter be prosecuted into India as a national phase entry, claiming its foreign priority date, and an increasing share of the mailbox’s later filings arrived by that road, so that the applications waiting at the end of 2004 were substantially the Indian limbs of families already examined, and in many cases granted, in other jurisdictions. Nothing in the treaty system obliged India to follow those foreign outcomes, and the amended Act required independent examination under Indian standards, including the exclusions of Section 3 as they would stand after 2005. The divergence that resulted is the transition’s most quoted datum: the crystalline form at the centre of the Glivec affair was patented in dozens of other countries and refused in India, not because the Indian examiner disputed the foreign science but because the Indian statute asked a question, about enhanced efficacy over a known substance, that the foreign statutes did not. The PCT pipeline delivered the world’s filings to the Indian door; what happened next was determined by the Indian text, which is the trail’s whole argument in miniature.

Two further changes of the 2002 round belong in the record because they shaped how the mailbox applications would later be fought over. The Act tightened the disclosure obligations of applicants, requiring the source and geographical origin of biological material used in an invention to be declared in the specification, and making wrongful disclosure a ground of opposition and revocation; in a statute increasingly concerned with traditional knowledge and bioresources, this was the defensive counterpart of the expanded exclusions. And the Act began the migration of patent adjudication away from the High Courts toward the Intellectual Property Appellate Board, substituting the Board in provisions that had previously named the courts, a shift in forum that mattered when the first product patent grants were challenged a few years later and a patients’ organisation needed a tribunal willing to hear it.

By the middle of 2003, therefore, the Indian statute complied with the Agreement on every point except the one the additional transition period had postponed. The architecture of the eventual settlement was also already visible in outline: a twenty-year term, treaty-grade enforcement, widened exclusions, a revised licensing chapter and two statutory exceptions to infringement. What remained was the deletion of Section 5, and the question of what would be built around the deletion when it came.

The deadline arrives: the Ordinance of December 2004

The final amendment was attempted twice before it was enacted. A Patents (Amendment) Bill introduced in 2003 was considered by a Group of Ministers and was still pending when the Lok Sabha was dissolved in 2004, on which it lapsed. The incoming government, the United Progressive Alliance supported from outside by the Left parties, intended to bring a revised Bill in the winter session of 2004. The session ended without it, the transition period expired on 31 December 2004, and Parliament was not in session. On 26 December 2004 the President promulgated the Patents (Amendment) Ordinance, 2004, Ordinance No. 7 of 2004, expressed to take effect from 1 January 2005.

Reaching for the ordinance power at a treaty deadline raises the standing constitutional questions about legislation by executive decree, and the general doctrine on when a government may legislate without Parliament is treated in the article on the ordinance-making power. For this trail, the operative facts are two. The first is that the deadline was real and external: the additional transition period ran out on a date fixed in 1994, and a member state that simply defaulted invited dispute proceedings of the kind India had already lost once. The second is that an Ordinance is a temporary text. It would cease to operate six weeks after Parliament reassembled unless replaced by an Act, which meant the Ordinance of December 2004 opened a negotiation rather than closing one, and the negotiators on every side knew it.

The Ordinance performed the compliance core: it deleted Section 5, deleted the exclusive marketing rights chapter, and set the mailbox applications down for examination. But it also carried provisions that went further in the patentee’s favour than the settlement that finally emerged. It cut the grounds on which the grant of a patent could be opposed before grant from nine to two and removed the opponent’s entitlement to be heard in person. It proposed widening the patentability of computer programmes, so that a programme with technical application to industry, or combined with hardware, could be patented if the ordinary tests were met. Each of those features drew immediate objection from the access-to-medicines movement, from generic manufacturers and from the Left parties whose support the government needed in Parliament, and each became a bargaining counter in the passage that followed.

The Ordinance also mobilised the argument outside Parliament, and the state of opinion it created is part of why the passage took the form it did. Treatment activists treated the text as a verdict on the future of affordable medicines worldwide: a global day of action against the Ordinance was observed on 26 February 2005, and the access campaign’s literature followed the Bill clause by clause. Anand Grover of the Lawyers Collective, the unit that later carried many of the patent oppositions, argued that the Ordinance’s amendments were skewed in favour of Western multinational companies. The research-based industry answered that India was merely accepting the norm of every major economy and that its own innovative firms would be the long-term gainers; Harinder Sikka of Nicholas Piramal, quoted in the medical press at the time, asked how long Indian companies should be dubbed reverse engineers and copycats. The domestic industry itself was divided along the line of its own research ambitions, between research-based producers who saw product patents as an opportunity and the established generic houses whose exports the new regime threatened. The government’s public line, pressed by the Commerce Minister, was reassurance: that the Act carried enough safeguards to prevent a rise in drug prices, and that the flexibilities preserved the state’s room to act. Every element of that public argument reappears inside the Act that emerged, which is what makes the passage debate the best commentary on the text.

The Ordinance was not the trail’s first attempt at the third amendment. A Patents (Amendment) Bill introduced in December 2003, under the preceding government, had already placed a complete product patent text before Parliament, with its own disposition of the flexibility questions, and had lapsed when the Lok Sabha was dissolved in 2004 with the Bill unpassed. The episode matters to the audit for two reasons. It establishes that the obligation’s arrival surprised no one: by the end of 2003 both major coalitions had committed themselves in drafting to the deletion of Section 5, and the argument of 2005 was never about whether product patents would come but about the safeguards that would accompany them. And it explains the Ordinance’s necessity in the new government’s hands. An administration formed in May 2004, facing a deadline seven months away and a lapsed Bill it had not written, had no parliamentary path that guaranteed enactment by 1 January 2005; the Ordinance power supplied the path, at the price of the passage fight that followed. The 2003 Bill thus stands in the trail as a dress rehearsal whose failure dictated the procedure, and much of the text, of the performance.

Two campaigns: the argument the amendment had to survive

No account of the trail is complete without the two organised campaigns that pressed on it, because the final text reads like their negotiated boundary. The access campaign, built from treatment activists, patient networks, public health organisations and sections of the generic industry, argued from the treatment programmes: a product patent regime in India would raise the price of new medicines everywhere Indian factories supplied them, and the campaign’s literature, down to its clause-level readings of the Ordinance, treated definitions and opposition grounds as the real battleground. Its specific demands were legislative: restore opposition in full, write the efficacy threshold into Section 3(d), keep compulsory licensing fast and export-capable, and protect the producers of the transition decade.

The counter-campaign, conducted through the research-based industry associations and amplified in the Special 301 process, argued from investment and reputation: that India’s regime singled it out among major economies, that mailbox claims would be defeated by procedure rather than merits, and that data protection and enforcement, not patentability alone, measured a serious innovation policy. Its specific demands were also legislative, in the opposite direction: confine Section 3(d), narrow opposition, and remove the uncertainties of the licensing chapter.

The United Progressive Alliance government stood between the campaigns with a parliamentary arithmetic that made the second set of pressures easier to resist than the first: it depended on Left support, and the Left parties carried the access campaign’s demands into the House as conditions. The Ordinance represented roughly how far the government would go toward the industry brief when it could act alone. The Act represents where it landed when it had to carry Parliament. The distance between the two texts, measured clause by clause in the next section, is the most precise record available of what Indian democracy did to a treaty obligation on its way into law.

Passage: how the Bill of March 2005 differed from the Ordinance

The Bill to replace the Ordinance was moved in the Lok Sabha by Kamal Nath, the Minister of Commerce and Industry, debated across sittings on 18, 21 and 22 March 2005, passed by the Lok Sabha on 22 March and by the Rajya Sabha on 23 March, and assented to on 4 April 2005 as the Patents (Amendment) Act, 2005, Act 15 of 2005. Most of its provisions were deemed to have come into force on 1 January 2005, closing the gap between the Act and the Ordinance it replaced, though a small set of provisions was reserved to be brought into force later by notification. Government amendments moved during the debate altered the text in respects that define the Act to this day.

Why did the government accept amendments to its own Bill in March 2005?

Because its majority depended on them. The ruling coalition governed with outside support from the Left parties, whose price for supporting a product patent law was a set of safeguards written into the statute itself rather than left to assurance. The amendments moved during passage were the negotiated cost of enactment, and they survive as the Act’s most argued provisions.

The changes made during passage cluster in five places. The definitions in Section 2 were tightened: a new definition of pharmaceutical substances confined the term to a new entity involving one or more inventive steps, a definition of new invention was added, and the definition of inventive step in Section 2(1)(ja) was put in the form it retains, a feature of an invention that involves technical advance as compared to the existing knowledge or having economic significance or both and that makes the invention not obvious to a person skilled in the art. Section 3(d) acquired its Explanation, listing the forms that count as the same substance unless they differ significantly in properties with regard to efficacy. The opposition provisions were rebuilt, as described in the next section. The export licence was widened, so that an importing country could qualify by having allowed the importation of the patented pharmaceutical product from India by notification or otherwise, instead of having to grant a compulsory licence of its own. The parallel importation limb was widened, from importation from a person duly authorised by the patentee to importation from a person duly authorised under the law. And the Ordinance’s excursion into software patentability was abandoned, leaving the exclusion of a computer programme per se where it stood.

Each delta deserves its practical translation, because together they are the difference between a permissive and a restrictive product patent regime. The added tail of the inventive step definition, the words that make the invention not obvious to a person skilled in the art, tied the new economic significance limb back to the classic obviousness inquiry; without it, a feature of no technical merit might have claimed an inventive step on commercial success alone. The definition of pharmaceutical substances, confined to a new entity involving one or more inventive steps, fixed the field in which the efficacy inquiry of Section 3(d) would operate, and denied applicants the escape of relabelling a derivative as something else. The Explanation to Section 3(d) was drafted by enumeration of the opposing playbook: salts, esters, ethers, polymorphs, metabolites, particle size, isomers and combinations are the standard instruments of pharmaceutical life-cycle management, and each is named and made presumptively the same substance. In the opposition clauses, the Ordinance would have confined the pre-grant representation to two grounds and treated the opponent as a source of information; the Act restored the full catalogue and the hearing, converting the examination of several thousand mailbox files into a litigable process with parties. In Section 92A, the substitution of notification or otherwise for the Ordinance’s narrower condition removed a requirement most eligible importers could never have met, since a country with no manufacturing capacity will often have granted no patent and possess no licensing machinery of its own. And in Section 107A(b), the move from authorisation by the patentee to authorisation under the law changed who controls the parallel import exception: under the Ordinance text the patentee’s own distribution decisions would have defined the exception’s scope, while under the enacted text the law of the place of sale does.

The abandoned software change deserves a paragraph of its own, because it shows the passage working on a question with no access-to-medicines valence at all. The Patents Act excludes a computer programme per se from patentability. The Ordinance proposed to qualify that exclusion so that a programme with a technical application, or combined with hardware, could cross the line, a change the software industry itself received with division, the larger product firms seeing advantage in it and the services firms and the free software movement seeing a threat to the interoperability on which their businesses rested. The Act dropped the change and left the exclusion in its unqualified form, referring the boundary question to examination practice, where it has been worked over ever since. For the trail audit the episode is a control case: even where no treaty article and no patient group was engaged, the Ordinance’s instinct was to widen patentability and the passage’s instinct was to decline the widening, which suggests the March amendments expressed a settled parliamentary disposition about the direction of Indian patent law and not merely a set of concessions extracted on one industry’s behalf.

The significance of the passage record is easy to understate. The flexibilities for which Indian patent law is known were not, in the main, extracted from a reluctant treaty in 1994. They were inserted into a domestic Bill, in public debate, in March 2005, by a government that needed the votes. That is why the Act’s restrictive features have a democratic provenance their critics abroad rarely credit, and why Indian governments defending them have been able to present them as a legislative settlement rather than an administrative evasion.

The demands that did not survive are as diagnostic as the ones that did. The access campaign’s maximal asks were refused: there was no statutory link between patent grant and price control, no blanket exclusion of mailbox applications claiming derivatives of molecules known before 1995, and no reversal of the decision to accept the product patent principle itself. The filter Parliament chose against weak derivative claims was Section 3(d) applied case by case in examination and opposition, not a categorical bar, and that choice led directly to the litigation of the following decade, because a case-by-case filter exports its hard questions to controllers and courts. Equally, the research industry obtained none of its structural asks: no data exclusivity, no patent linkage between the drug regulator and the patent register, no curtailment of Section 92A. A trail read through its failures shows a government that conceded procedure and thresholds to one side and conceded nothing structural to the other, a distribution that explains why the settlement, once enacted, proved stable enough to end the legislative argument for a generation.

The mechanics of the concession repay attention. An ordinance lapses unless both Houses approve the replacing Bill within weeks of the session beginning, so the government’s choice was never between the Ordinance text and the amended text in the abstract; it was between an amended Act and the collapse of the replacing legislation altogether, with the transition already expired and the mailbox already, in law, open. The United Progressive Alliance depended on the Left parties for its majority, and their support was priced in clauses. By accepting the amendments as its own, the government converted its critics into co-authors of the settlement, and the durability of the 2005 text owes something to that conversion: a statute carried by the votes of the parties that had campaigned against its Ordinance form is harder for any successor to disown. The Commerce Minister supplemented the textual concessions with assurances on the floor, that the Act carried safeguards against price rises and that the price monitoring machinery of the state would watch the patented medicines market. Such assurances are not law, and later argument has had to treat them carefully; but in the passage they functioned as the non-textual half of the bargain, the half addressed to members whose constituents bought medicines rather than read statutes.

The critics’ discounting of that assurance had a technical basis the trail should note. India’s price control machinery, the Drugs (Prices Control) Order made under the Essential Commodities Act and administered with the National Pharmaceutical Pricing Authority, operated on a scheduled basket: the Order of 1995 held 74 bulk drugs under control, and a newly patented medicine stood, in the ordinary case, outside the schedule and outside the ceiling. The assurance of March 2005 therefore promised vigilance without an instrument, unless the government chose to expand the schedule or to use the licensing chapter, and the access campaign’s insistence on statutory safeguards over ministerial assurances rested on exactly this gap. The point survived the passage: every later Indian argument about patented medicine prices has had to choose between the price control route, with its schedules and ceilings, and the patent route, with its thresholds and licences, and the choice was framed in the week the Act was passed.

The Ordinance-first sequence also inverted the ordinary burden of legislative inertia, and the inversion protected the settlement’s core. Opponents of a Bill not yet law need only delay it; opponents of an Ordinance already in force must assemble a majority for a different text within the session’s fixed weeks, against the knowledge that failure leaves not the old law but a vacuum, the transition having expired and the mailbox having opened on the Ordinance’s authority. Every amendment the Left and the access campaign won in March 2005 was therefore won against that structural headwind, which makes the scale of the passage changes more remarkable, and makes the survival of the core, the deletion of Section 5 on the treaty’s date with the treaty’s scope, correspondingly less surprising. The government conceded everything that could be conceded without reopening the obligation, because the obligation was the one element of the text that inaction would have preserved rather than defeated. Passage analysis that ignores the Ordinance’s head start will misread both the concessions and the refusals.

What the 2005 Act changed in the Patents Act

The centre of the Act is a deletion. Section 5 of the Patents Act, the provision that had confined food, drugs, medicines and chemical substances to process patents since 1970, was omitted altogether. With it went the foundation of the process-only regime: from 1 January 2005 an inventor could claim the substance itself, in every field of technology, and the exclusive rights in Section 48 attached to the product. Chapter IV-A, the exclusive marketing rights chapter inserted in 1999, was deleted with it, as a transitional device whose work was done; the transitional provisions dealt with rights already granted and suits already filed, as the next section explains.

What did deleting Section 5 change on the ground?

It changed what counted as infringement. Under the old law a competitor needed only a different process to be safe. After the deletion, a competitor needed a different product, a licence, an invalidation, or the shelter of a statutory exception. Every flexibility in the rest of the Act defines the room left inside that single change.

Against that deletion the Act set the new Section 3(d). The clause had existed before in a simpler form, excluding the mere discovery of a new property or a new use for a known substance. The 2005 substitution produced the provision in its operative form: the mere discovery of a new form of a known substance which does not result in the enhancement of the known efficacy of that substance, or the mere discovery of any new property or new use for a known substance, or the mere use of a known process, machine or apparatus unless the known process results in a new product or employs at least one new reactant, is not an invention. An Explanation added during passage provides that salts, esters, ethers, polymorphs, metabolites, pure form, particle size, isomers, mixtures of isomers, complexes, combinations and other derivatives of a known substance shall be considered to be the same substance, unless they differ significantly in properties with regard to efficacy. The provision’s target is the practice of extending protection through successive claims to marginal variants of a known compound, and its operative test is enhancement of known efficacy, a term the courts later had to construe. The construction, the litigation and the judgment that settled the clause’s meaning belong to the specialist article on the provision, and this article takes from it only what the trail requires: the clause was an Indian drafting choice of 2005, not a term of the treaty, and its efficacy language was put there during passage.

The opposition system was redesigned in both directions the Ordinance had failed to take it. Before grant, Section 25(1) permits any person to represent in writing against the grant once an application has been published, on the full set of grounds the statute lists, and the Controller must hear the opponent if a hearing is requested; the opponent is a party to the process rather than a supplier of information to it. After grant, Section 25(2) gives a person interested one year from publication of the grant to oppose before an Opposition Board. The redesign made the pre-grant stage a genuine filter on mailbox applications, several thousand of which were about to be examined at once, and it is the procedural reason the Glivec application received the scrutiny it did.

The definitions package attracted a scholarly critique that belongs in the record because it cuts against the passage narrative. Srividhya Ragavan, one of the closest academic readers of the amendments, observed of the transition that it was shaped by a persistent concern to avoid the abrupt displacement of domestic public interest priorities; yet commentators also noted that the new definition of inventive step, with its alternative limb of economic significance, could in principle lower the inventive threshold for some claims even as Section 3(d) raised it for pharmaceutical derivatives. The two provisions pull in opposite directions, and examination practice, rather than the parliamentary record, decides which pull prevails in a given field. The trail therefore contains a deliberate tightening and an arguable loosening inside the same amending Act, and any verdict on the 2005 settlement that notices only one of them is incomplete.

The same Act made less discussed adjustments at the edges of the right. It aligned the statement of the patentee’s exclusive rights more closely with the language of Article 28 of the Agreement, and it strengthened the provisions bearing on national security: the secrecy directions that can be imposed on sensitive applications, and the controls on residents seeking patent protection abroad for inventions that might have defence relevance, a part of the Act the Patent Office itself summarises as strengthening the security provisions. A trail audit should record these movements because they show the 2005 Parliament using the compliance exercise for purposes of its own, tightening state control at the security margin while it loosened the patentee’s exclusivity at the access margin.

Compulsory licensing emerged from the 2005 round with its architecture completed. The general licence in Section 84 remained available three years after grant where the reasonable requirements of the public were not satisfied, where the patented invention was not available at a reasonably affordable price, or where it was not worked in India, with Section 83’s principles, including the direction that patents are granted to encourage inventions and to secure that they are worked in India on a commercial scale, framing the inquiry. Section 92 retained the emergency route on government notification. Beside them the Act inserted Section 92A, a compulsory licence solely for the manufacture and export of patented pharmaceutical products to a country with insufficient or no manufacturing capacity in the pharmaceutical sector, where that country has granted a compulsory licence or has by notification or otherwise allowed the importation from India. Section 92A is the domestic implementation of the solution the WTO membership built for the problem identified in paragraph 6 of the Doha Declaration, the predicament of countries that could use a compulsory licence in theory but had no factory in which to work one. The 2005 Act also strengthened the provisions guarding against the patenting abroad of inventions with defence implications and rationalised timelines and fee procedures across the Act, changes that attract less attention because they altered administration rather than rights.

The three grounds of Section 84 repay a plain statement, because they define what working means in Indian law. The reasonable requirements of the public are not satisfied where the patentee’s conduct prejudices an Indian trade or industry, where demand for the patented article is not met to an adequate extent or on reasonable terms, or where the article is not available at a reasonably affordable price; a patent is not worked where importation by the patentee or those deriving title from the patentee hinders or prevents working in India; and a licence may issue where the patent blocks the exploitation of another, later patent of importance. Section 85 adds that two years after a first compulsory licence, the patent itself may be revoked for continued non-working. The importation limb carries the deepest legal argument in the chapter: Article 27.1 of the Agreement forbids discrimination according to whether products are imported or locally produced, and critics have argued from it that penalising importation as non-working breaches the treaty, while India answers that Article 31 prescribes no grounds and that the Doha affirmation protects the choice. No panel has resolved the point, and the trail leaves it exactly where the 2005 Act placed it, in the statute, unlitigated at the WTO and available to every future applicant.

The licensing suite is best read whole, as an architecture of three doors. The first is Section 84, opened by an applicant who makes out non-working, unmet demand or unaffordable price three years after grant. The second is Section 92, opened by the state: on a Central Government notification of national emergency, circumstances of extreme urgency or public non-commercial use, the Controller may grant licences to applicants without the three-year wait, and the ordinary expectation of prior negotiation with the patentee falls away in the notified case. The third is not a licence at all but the state’s retained position as user, in Section 100, which permits government use of patented inventions on payment, and in the conditions written into every grant by Section 47, under which a patented medicine may be imported by the government for distribution in the hospitals and dispensaries it maintains. The later amendments left the three doors standing and adjusted their fittings; the 2002 recast of Section 92, speaking expressly of public health crises, oiled the second. The distinction matters to the compliance argument, because the doors answer to different parts of the treaty: applicant licences test the conditions of Article 31, state-initiated licences test its emergency logic, and government use rests on the oldest sovereign reservation in the statute, one the 2005 Parliament was careful not to surrender while it surrendered so much else.

Behind the doors stands a charter. Section 83, recast by the 2002 amendment, states the general principles the Controller is to apply across the working and licensing chapters: that patents are granted to encourage inventions and to secure that the inventions are worked in India on a commercial scale to the fullest extent reasonably practicable without undue delay; that they are not granted merely to enable patentees to enjoy a monopoly for the importation of the patented article; that the protection and enforcement of patent rights contribute to the promotion of technological innovation and to the transfer and dissemination of technology, to the mutual advantage of producers and users, in a manner conducive to social and economic welfare and to a balance of rights and obligations; that patents granted do not impede the protection of public health and nutrition; and that the Government’s power to take measures of public health protection stands unobstructed. The language of the section tracks the objectives and principles articles of the TRIPS Agreement itself, Articles 7 and 8, written into domestic law where a Controller must apply them case by case. It is the clearest single evidence in the statute that the flexibility project was not an oversight of the compliance exercise but its organising idea: the treaty was answered, in part, by quoting it back.

The flexibilities India kept, and the Doha foundation

Stand back from the detail and the flexibility suite has a shape. The exclusions in Section 3 define what cannot be monopolised at all: discoveries, scientific principles, methods of agriculture and horticulture, treatment processes, plants and animals, computer programmes per se, traditional knowledge, and, under clause (d), new forms of known substances without enhanced efficacy. The opposition provisions place the grant itself under adversarial scrutiny at two points. The licensing chapter subordinates the granted right to working, affordability and public requirements, and adds an export channel for countries without factories. Section 107A removes research, regulatory preparation and parallel importation from the scope of infringement. The transitional proviso, examined next, protected investment made in reliance on the old regime. Each element has a distinct function, and the suite is sometimes misdescribed as a single act of resistance when it is better understood as a set of valves at different points in the system: before grant, at grant, during the term, and at the boundary of the right.

Which flexibility in the 2005 settlement matters most in practice?

The one that operates before any right exists. Section 3(d) and pre-grant opposition decide which mailbox claims become patents at all, so they govern the stock of rights every later flexibility must then qualify. Licences and exceptions matter intensely in individual cases, but the filtration provisions shape the whole field continuously and silently.

The Declaration had a pre-history in a single commercial offer, and the trail should record it because it explains the politics better than any communiqué. In February 2001 the Indian manufacturer Cipla, under Yusuf Hamied, offered a triple combination of antiretroviral medicines to the medical charity Médecins Sans Frontières at 350 dollars per patient per year, and to governments at a higher but still fractional price, at a time when the branded combination cost between ten and fifteen thousand dollars a year. The offer detonated the assumption on which the trading system’s health politics had rested, that patented prices were a fact of nature. Within the year, the African Group and its allies had carried the access question onto the WTO’s ministerial agenda, the pharmaceutical industry had withdrawn its litigation against South Africa’s medicines law, and the membership adopted the Declaration at Doha. The sequence matters for reading the 2005 Act: the flexibilities India legislated in 2005 were not abstractions retrieved from the treaty text by lawyers. They were the codification, in one member’s statute, of a political settlement that an Indian company’s price list had forced into being four years earlier.

The interpretive warrant for building these valves was supplied at Doha. The Declaration on the TRIPS Agreement and Public Health, adopted by the WTO Ministerial Conference on 14 November 2001, states in its fourth paragraph that the TRIPS Agreement does not and should not prevent members from taking measures to protect public health, and affirms that the Agreement can and should be interpreted and implemented in a manner supportive of WTO members’ right to protect public health and, in particular, to promote access to medicines for all. Its fifth paragraph records the included flexibilities: each member’s right to grant compulsory licences and the freedom to determine the grounds for them, the right to determine what constitutes a national emergency, and freedom on exhaustion of rights, which is the treaty basis of parallel importation. The Declaration is an interpretive instrument rather than an amendment, and lawyers argue about its precise legal weight; its practical weight is not in doubt. It converted India’s reading of the Agreement from a defensive position into a text endorsed by the whole membership, and the 2005 Act was drafted in its light, three years and a few months after its adoption.

Paragraph 6 of the Declaration identified a structural defect in the flexibility scheme, and its repair completes the Doha story inside the 2005 Act. Article 31 contemplated licensing for the licensee’s own market, its condition that use be predominantly for the supply of the domestic market assumed the licensing country could manufacture. Members with insufficient manufacturing capacity could not use the system at all, since no producer country could license mainly for export. The Declaration instructed the TRIPS Council to solve the problem, and the General Council decision of 30 August 2003 waived the domestic-market condition for exports to eligible importing members under safeguards, a waiver later written into the Agreement itself as Article 31bis. India accepted the protocol that made the amendment early, in 2007, and the amendment entered into force a decade later when the required number of acceptances had accumulated, a small datum that belongs in the trail because it shows the flexibility’s author-state behaving as its sponsor in the treaty system as well as its beneficiary at home. Section 92A is India’s enactment of that solution, and its drafting through the Ordinance and the passage, first narrow and then widened, tracks the international negotiation in miniature. A complete trail audit should mark the sequence: the political declaration of 2001, the legal waiver of 2003, and the Indian section of 2005 are three instruments of descending generality performing one repair, and the repair is meaningless in any country but a manufacturing one, which is to say that the flexibility’s value was always concentrated in the state whose mailbox this article is about.

The export section’s own record carries an irony the trail should register. The paragraph 6 system was first used anywhere not by India but through Canada, whose access to medicines regime supplied a generic HIV combination to Rwanda in 2008 after Rwanda notified the WTO of its need, an episode that demonstrated both that the machinery worked and how cumbersome its notifications and conditions made it. No licence under Section 92A has been granted, and Indian exports of generic medicines have continued to flow overwhelmingly under ordinary commercial arrangements and voluntary licences rather than under the section. The flexibility’s importance in the trail is therefore not measured by its use. It stands in the Act as a held option, evidence that the 2005 Parliament intended the licensing chapter to be complete in every direction the treaty permitted, including the direction its own exporters were least likely to need. In a negotiation conducted partly through signals, a complete chapter was itself a signal, addressed to the next round of argument about what Indian law was for.

The fields beyond pharmaceuticals: chemicals, food and plant varieties

Section 5 had a second limb that the medicines debate habitually forgets. Beside the food, medicine and drug clause stood a clause covering substances prepared or produced by chemical processes, the section naming alloys, optical glass, semi-conductors and inter-metallic compounds among its examples. The process-only rule therefore ran across the chemical industry and not merely across the pharmacopoeia, which is why the mailbox obligation in Article 70.8 extended to agricultural chemical products, why the American complaint of 1996 named agrochemicals beside pharmaceuticals, and why the queue of 8,926 applications held the filings of a chemical industry and not of drug companies alone. The 2005 Act deleted the section entire, both limbs at once, so product patents returned for chemical substances generally in a single stroke whose economic weight fell mostly outside the debate that had consumed Parliament. Agrochemical producers, speciality chemical firms and the makers of food substances acquired product rights on the same day as the pharmaceutical industry, on the strength of an argument conducted almost entirely about medicines.

Plants took the treaty’s other road, and the contrast is instructive. Article 27.3(b) permits a member to exclude plants and animals from patentability, other than micro-organisms, while requiring plant varieties to be protected by patents, by an effective sui generis system, or by a combination of the two. India kept the exclusion, in Section 3(j)’s bar on plants and animals and on essentially biological processes, and built the sui generis system outside the Patents Act altogether, in the Protection of Plant Varieties and Farmers’ Rights Act, 2001, a statute whose protection of farmers’ rights to save and replant seed has no counterpart in patent law anywhere. The 2005 amendment left both halves of that settlement undisturbed. The episode disciplines any account of the trail as a simple surrender: where the treaty offered a genuine choice of instrument, India had legislated once, early, on its own design, and the compliance amendments passed it by because there was nothing in it to correct.

The mailbox opened: examination, the proviso and the position of pre-2005 producers

When the transitional provisions closed at the end of 2004, the mailbox held, on the Controller General’s figures, 8,926 applications, of which 7,520 had been filed by foreign entities and 1,406 by Indian applicants. The composition of the box showed whose transition it had been: on a contemporaneous count of the pharmaceutical filings, Pfizer led the applicants with 373 applications, followed by Johnson and Johnson with 262, while the largest Indian filer, Dr Reddy’s, had 205, and applicants from the United States accounted for 2,324 filings across the regulated fields, the great majority pharmaceutical. The Indian presence in the box carries its own lesson. The largest domestic filer was a company already investing in research of its own, and Indian entities between them lodged more than one filing in seven: the domestic industry was not standing outside the new system protesting it, but was hedging into it, acquiring positions under the regime it publicly opposed. That doubleness explains a feature of the later politics that puzzles outside observers, the relative quiet of Indian industry in the compliance argument compared with the research multinationals and the patient groups. The settlement of 2005 threatened one Indian business model and endowed another, and the mailbox’s own composition had announced the split a decade before the argument began. The examination was conducted from the Patent Office at Kolkata, where the mailbox had been held. The box had also been black in the literal sense: during the transition the applications were neither examined nor published, so the industry that would have to live under the eventual grants could not read the claims it might one day be held to infringe, and planned its own launches in ignorance of the register that was accumulating against it. The opening of the box on the amended Act’s commencement was therefore a disclosure event as much as an examination event, and the restored pre-grant opposition acquired its caseload in the same moment, as competitors and patient groups read the published claims and selected their targets. From 2005 the Controller began referring them for examination in filing order, against the patentability standards of the amended Act, including the new Section 3(d). A mailbox application that succeeded produced a patent whose twenty-year term ran from its filing date in the 1990s or early 2000s, so part of the term had always expired before the right existed. Two transitional rules then governed what the patentee could do with it.

The first is in Section 11A(7): for a patent granted on a mailbox application, the patentee’s rights accrue only from the date of grant. Nothing done before grant can be an infringement of a right that did not yet exist, so the years of generic manufacture during the transition were not retrospectively converted into wrongdoing. The second is the proviso to the same sub-section, the grandfather clause. Where an enterprise had made significant investment and was producing and marketing the product before 1 January 2005 and continues to produce it, the patent holder cannot restrain the enterprise from continuing; the patentee is entitled only to receive a reasonable royalty from it. The Supreme Court, describing the scheme in the Novartis litigation, treated the clause in just these terms: a patentee who prevailed could not drive from the market the generic producers who had built their position before 2005, but could require them to pay for the use.

When did a mailbox patent begin to bite?

Only on grant. The term ran from the filing date, so the clock had been consuming the twenty years while the application waited, but no right existed to enforce until the patent was sealed, and against an enterprise that had invested and produced before 2005 the right never extended beyond a reasonable royalty.

Two features of the proviso remain genuinely open, and honesty requires naming them rather than filling them in. The Act nowhere defines significant investment, and it nowhere quantifies a reasonable royalty; contemporary commentary on the Ordinance already noted the absence of interpretive guidance, and the clause’s key terms have never acquired an authoritative judicial measure, because where the filter of Section 3(d) and opposition refused the prominent mailbox claims, the royalty question rarely had a granted patent to attach to. The fate of exclusive marketing rights at the same moment was simpler. Rights already granted continued until the underlying application was decided, the applications behind them were examined at once for grant of a patent, and suits concerning a granted marketing right were dealt with as if they concerned a patent.

Glivec worked through every stage of this machinery in sequence, which is why it structures the memory of the transition. A mailbox application filed in 1998; an exclusive marketing right in November 2003; injunctions against generic producers; pre-grant oppositions by generic companies and by the Cancer Patients Aid Association once examination began; and, on 25 January 2006, orders of the Assistant Controller refusing the application, substantially on the ground that the crystalline form claimed was a new form of a known substance whose efficacy over the known compound had not been shown. Novartis’s challenge to the refusal, and its separate attack on Section 3(d) itself, produced the judgment analysed in the article on Section 3(d) and the Novartis judgment. For the amendment trail the episode’s significance is institutional: the settlement of 2005 was not a paper settlement. Its filters were used, against the most prominent application in the mailbox, within thirteen months of the box being opened.

The grant side of the ledger opened within weeks, and its first entry tested the other half of the redesigned machinery. In March 2006 the Controller General granted F. Hoffmann-La Roche a patent on pegylated interferon alfa-2a, sold as Pegasys for hepatitis C, the first pharmaceutical product patent of the new regime, on an application that had waited in the mailbox. The grant was opposed after grant by the Sankalp Rehabilitation Trust, a patients’ organisation, and by a manufacturer, and the Patent Office in 2009 rejected the opposition and upheld the patent. On appeal, the Intellectual Property Appellate Board revoked it in November 2012, holding both that a patients’ group could qualify as a person interested entitled to challenge a granted patent, and that the claims failed on the evidence, including the efficacy burden that Section 3(d) places on a new form of a known substance. The episode is the cleanest worked example the trail affords of the 2005 design operating as designed: a mailbox grant, a public-interest opponent with standing, an appellate tribunal created by the earlier amendments, and a patent that did not survive the scrutiny the passage had restored. Between the Glivec refusal and the Pegasys revocation, the register acquired both of its poles, and the hundreds of unreported mailbox examinations in between were decided in the space those two outcomes defined.

The mailbox grants also carried a temporal peculiarity that the 2005 Act built into Section 11A and that distinguishes them from every later patent. For an ordinary application, the patentee’s rights are reckoned from the date the application is published, although no infringement proceeding can be instituted until the patent is granted. For a mailbox application, Section 11A(7) displaces that reckoning: the patentee acquires no rights against anyone for the period before grant, however long the application waited, and the proviso then protects the pre-2005 producer even against the granted patent, on the royalty terms already described. The decade of lawful generic production was thus not merely unremedied as a matter of prosecutorial grace; it was placed outside the right by the statute itself. An applicant reading the trail in 1995 could have predicted the filing queue. Only the 2005 Parliament decided that the queue would also be a safe harbour, and that decision, taken in the proviso and in the main rule of the subsection together, is the transition’s most consequential act of protection after Section 3(d) itself.

Deadlines as drafting: how the treaty calendar wrote the Indian one

Read against the calendar, the trail resolves into four dates, each of a different legal species. The mailbox duty ran from the Agreement’s entry into force on 1 January 1995, with no transition at all, which is why it alone generated litigation. The general obligations, the twenty-year term, the exclusions discipline and the enforcement chapters, fell due on 1 January 2000 at the end of the developing-country transition, and were met by the Act of 2002, late on paper and unchallenged in the event, because no private right turned on the interval. The product patent obligation fell due on 1 January 2005 under Article 65.4, and was met on the day, by an Ordinance a week old. And the one obligation performed ahead of its legislative season, the mailbox’s statutory repair in 1999, was performed under the compulsion of a dispute ruling and a compliance period that ran to April of that year. Nothing in the trail was enacted early for goodwill, and nothing was enacted late enough to be sanctioned. Deadline discipline of this precision is not procrastination; it is a negotiating position maintained for a decade, in which time itself was the concession India had won at Marrakesh and declined to give back.

The term arithmetic of the mailbox shows what the discipline cost and bought. Article 70.8 required that patentability be judged by the criteria of the Agreement by reference to the filing or priority date, and that protection run from grant for the remainder of a term counted from filing. A mailbox application filed in 1996 and granted in 2007 on the twenty-year term therefore expired in 2016, carrying nine years of effective exclusivity for a decade of waiting, during much of which Indian firms had lawfully sold the product. The compression was the treaty’s own design, and it explains two otherwise puzzling features of the settlement: the research industry’s complaint that Indian mailbox patents were worth less than their foreign counterparts, which was true and was the transition working as written; and the proviso to Section 11A(7), which exists precisely because the waiting decade had created lawful markets that a grant would otherwise have destroyed overnight. Time, in this trail, was not the background of the negotiation. It was the substance of it.

One more feature of the drafting deserves its place in the audit, because it recurs at every stage of the trail: the systematic use of retrospective commencement to make the record continuous. The 1999 Act was deemed in force from 1 January 1995, four years before its enactment, so that filings made under administrative instruction and under a lapsed Ordinance stood validated as filings under the Act. The 2005 Act was deemed in force from 1 January 2005, so that the Ordinance period and the Act period form a single legal surface and nothing done under the Ordinance required separate authority. Retrospection of this kind, validating acts taken in anticipation of a statute, sits comfortably in Indian constitutional practice for civil and regulatory matters, and no challenge disturbed either exercise of it. But its repetition is diagnostic: a legislature that must repeatedly legislate backwards is a legislature working against deadlines set elsewhere, and the commencement clauses of this trail are the fingerprints of the treaty calendar on the Indian statute book.

The amendment that was studied and not made: the Mashelkar questions

A trail is defined by its abandoned branches as well as its enacted ones, and the largest abandoned branch of this trail opened within months of the 2005 Act. In the passage debates and after them, a recurring proposal sought to confine the restored product patents to new chemical entities, denying patents to other categories of pharmaceutical invention, and a cognate question asked whether micro-organisms needed to be patentable at all. Both proposals would have required a fourth amendment, and both turned on the same interpretive question that had governed the trail throughout: how much room the Agreement in fact left. The government referred the questions to a Technical Expert Group chaired by R. A. Mashelkar, constituted in 2005 and reporting in December 2006.

The Group answered both questions against the proposed restrictions. To limit patent protection for pharmaceutical substances to new chemical entities, it concluded, would not sit with the obligation that patents be available without discrimination as to the field of technology; and to exclude micro-organisms from patentability would contradict the express terms of Article 27.3(b), which permits the exclusion of plants and animals other than micro-organisms. A revised report in 2009, issued after the first had attracted controversy, confirmed both conclusions. The significance for the trail is twofold. Institutionally, Parliament was advised, on the record, that the 2005 text already stood at the frontier of what the treaty allowed, so the restrictionist amendment was examined and abandoned rather than merely outvoted. Analytically, the episode marks the moment the argument changed venue for good: after 2006, the serious contest over Indian patent law concerned the application of the enacted words by controllers and courts, and the legislative trail effectively ends where the expert group reported, with the statute of 2005 standing as its final entry.

Three transitions: what other large markets did with the same obligation

The Indian route was one of three by which large developing markets entered the product patent world, and the comparison isolates what the trail actually decided. Brazil, in its industrial property law of 1996, chose importation: a pipeline mechanism under which subject matter already patented abroad and not yet marketed could receive a Brazilian patent for the remainder of the foreign term, without the application being examined on its merits under Brazilian standards. China chose anticipation: pharmaceutical product patents had been accepted in its 1992 amendment, before the WTO existed, under bilateral pressure, so the system was joined with the central question already closed and no transition to manage. India chose deferral: a queue in which applications waited a decade and then faced examination under Indian standards, including a Section 3(d) that no other major system possessed and an opposition procedure open to any person.

The consequences of the choice are visible in this article’s own record. Under a pipeline, the Glivec application would have entered as a granted right and the argument would have been about revoking it. Under anticipation, there would have been no mailbox, no proviso and no transition generation of generic producers to protect. Deferral kept the state’s examiner, and behind the examiner the state’s Parliament, between the foreign filing and the Indian right, and the two emblematic mailbox outcomes, a refusal at Chennai and a revocation at the Appellate Board, are products of that interposition. Critics of the settlement who argue that product patents were inevitable once the treaty was signed are correct about the destination; the trail’s subject is the road, and the road determined which institutions would hold the power to say no.

The constitutional register: health, trade and the treaty power

One further frame surrounded the passage and belongs in a complete trail, because it explains the register in which the debate was conducted. Legislative competence was never in issue: patents fall within the Union’s field under the Constitution’s Seventh Schedule, and Article 253 empowers Parliament to legislate for implementing an international agreement across the whole territory regardless of the ordinary distribution of subjects, which is also why the treaty deadlines of this trail could be met by Ordinance when Parliament was not sitting. The contest was conducted instead in the language of rights. Opponents of the product patent argued from the right to life in Article 21 and from the Directive Principle that charges the state with raising the level of nutrition and public health, contending that a Constitution read in that light could not license a regime that priced the dying out of treatment. The government’s answer was not to deny the premise but to locate the discharge of the duty inside the statute: the flexibilities, the licensing doors and the efficacy threshold were presented, in debate, as the constitutional content of the settlement, the means by which a treaty obligation was performed without surrendering the health obligation.

Notably, the constitutional argument was won and lost politically rather than judicially during the passage itself. No court was asked in 2005 to measure the amendment against Article 21, and the constitutional testing of the settlement came later and in the courts’ own time, attached to individual provisions in individual cases, a division of the question that belongs to the articles on those provisions. For the trail, the significance of the constitutional register is that it fixed the burden of justification: after March 2005, a defender of the Indian settlement did not have to argue that the treaty permitted the flexibilities, but could argue that the Constitution commended them. That inversion of the burden is one of the settlement’s most durable achievements, and it was accomplished in debate rather than in litigation.

The compliance argument, stated both ways

The contention that India’s patent law breaches the Agreement has been maintained by the originator industry and reflected in the trade instruments of the United States for as long as the settlement has existed, and it deserves to be stated at the strength its proponents give it. The Pharmaceutical Research and Manufacturers of America, in submissions to the Special 301 process, has argued that Section 3(d) imposes a patentability criterion beyond the three the Agreement names, novelty, inventive step and industrial application, and so discriminates against pharmaceutical inventions; that the twin oppositions, and particularly an open pre-grant opposition usable by commercial rivals, subject applicants to delay and uncertainty unknown in comparable systems; and that the working requirement, read with the compulsory licensing chapter, treats importation as suspect and the patent as a manufacturing licence. Successive Special 301 Reports of the United States Trade Representative have placed India on the Priority Watch List in report after report reaching back across decades, and have described India in terms such as one of the world’s most challenging major economies for the protection and enforcement of intellectual property.

The Indian answer, maintained by governments of both coalitions and developed by commentators such as Carlos Correa of the South Centre, is that every element of that critique mislocates the obligation. The Agreement requires that patents be available for inventions and names three criteria without defining any of them; defining the boundary between an invention and a new form of a known substance is an exercise of the very discretion the text leaves open. Compulsory licensing is addressed in Article 31 of the Agreement through procedural conditions, and the Doha Declaration records that each member has the freedom to determine the grounds on which licences are granted. Exhaustion is expressly reserved by Article 6. And the compliance charge has never been adjudicated: the only patent rulings the WTO has given against India are the 1990s decisions on the missing transition machinery, decided before the 1999 Act repaired it. A Special 301 listing, on this side of the argument, is a unilateral assessment by one trading partner under its own statute, issued through a process that receives industry submissions; it is pressure, not a finding.

A vocabulary point clarifies what the central charge is really about. The industry’s word for Section 3(d) is evergreening, the practice of extending a franchise by patenting successive forms of one medicine as the original patent runs down. India’s answer is that the section does not police the calendar at all: a new form that genuinely enhances efficacy may be patented and will hold its own twenty-year term from its own filing, and a new form that does not is refused, whenever it is filed. The objection, on the Indian account, is therefore not to a rule against term extension but to a rule that examines merit at the level of the molecule’s known family, and the disagreement is about the unit of invention rather than about time. The pressure, meanwhile, has been calibrated rather than constant: the United States conducted an out-of-cycle review of India’s listing in 2014, after the research industry had sought the graver designation of Priority Foreign Country, and retained the Priority Watch List placement at its close, an outcome each side reported as vindication.

Two exhibits dominate the mature form of the argument. The first is regulatory data. The research industry’s standing complaint, repeated in the Special 301 record, is that India maintains no effective system to protect the test data submitted for marketing approval against reliance by generic applicants, and treats the absence of data exclusivity as a TRIPS shortfall. India’s answer is textual: Article 39.3 of the Agreement obliges members to protect such data against unfair commercial use and disclosure, and prescribes no period of exclusivity; a system that approves generics on their own or on referenced data under its drugs law, without misappropriation, honours the obligation as written. The second exhibit is the licensing chapter’s record. In the years since 2005 the Controller has entertained few Section 84 applications and refused most at the threshold, including applications by BDR Pharmaceuticals and by Lee Pharma, while granting one, to Natco in 2012 over Bayer’s sorafenib, which survived appeal. The industry’s reading is that the single grant proves the chapter is a standing expropriation risk priced into every launch. India’s reading is that a chapter used once in two decades, under judicial supervision, on a medicine whose price and supply were the record against it, is evidence of restraint. Both readings are really arguments about what the 2005 Parliament intended, which is why the trail, and not the rhetoric, is the evidence that matters.

The single grant repays a closer reading, because it is the licensing chapter’s proof of operation and both sides cite it for opposite conclusions. The patent covered sorafenib tosylate, Bayer’s medicine for kidney and liver cancer, and had been granted in 2008 on an application of the mailbox generation. Natco applied under Section 84 in 2011, the three-year bar having run, after seeking a voluntary licence and failing to obtain one on terms it could accept. The Controller’s order of March 2012 found the statutory grounds made out in combination: the reasonable requirements of the public were not satisfied, the medicine reaching only a small fraction of the patients estimated to need it; the patented article was not available at a reasonably affordable price, Bayer’s price standing at about 280,000 rupees for a month of treatment against the 8,800 rupees a month at which Natco proposed to sell; and the invention was not worked in India by importation alone. The licence issued on payment of a royalty of six per cent of net sales, which the Appellate Board raised to seven per cent while upholding the grant, and the Supreme Court declined to interfere in December 2014. Note what the episode does and does not show. It shows the chapter operating exactly as the 2005 settlement designed it: an applicant-led process, evidence on price and supply, a reasoned order, appellate supervision, and a royalty that kept the patentee paid. It also shows the state-initiated doors standing unused beside it: no notification under Section 92 has been issued, and no export licence under Section 92A granted, so the whole of the chapter’s practical history runs through the single applicant door of Section 84. The research industry’s systemic reading, that one grant in a decade proves the risk is structural, and India’s reading, that one grant in a decade proves the restraint, are both honestly drawn from the same record; the trail can only insist that the record, and not the apprehension, is what compliance should be judged on.

After 2005 the argument migrated into a second channel, and the trail’s later entries are negotiating texts rather than statutes. In the free trade agreement negotiations between India and the European Union that ran from 2007, the European side sought commitments beyond the TRIPS floor, including a period of data exclusivity for pharmaceutical test data and extensions of the patent term, the two devices known in the debate as TRIPS-plus. India refused both, under sustained pressure from the same access coalition that had shaped the 2005 passage, and European representatives in due course disclaimed any intention to press demands that would damage the Indian generic industry. The episode completes the pattern this article has traced. What the multilateral system had failed to obtain from India in 1995, and what the dispute system could not obtain in 1998, the bilateral channel attempted after 2005; and the defence in every channel was the same, that the 2005 settlement marked the boundary of the bargain. A compliance argument that began as a question about a mailbox thus ends as a standing instruction to Indian trade negotiators, which is perhaps the clearest measure of what the amendment trail built.

What is the legal status of a Special 301 listing?

It is an act of United States domestic trade administration, not an adjudication by any international body. Listing initiates engagement and signals possible escalation under United States law; it determines no state’s rights under the TRIPS Agreement, which only the WTO dispute system can do, and that system has never been asked to rule on the 2005 settlement.

On the narrow question of what the record establishes, the position can be stated without taking either side’s brief. India performs every obligation the Agreement’s text specifies for a member in its position: availability, term, rights, enforcement, transition. The contested provisions occupy ground the text does not specify. Until a panel rules otherwise, the statement that India’s patent law is non-compliant is an assertion by interested governments and industries, and the statement that it is plainly compliant is a prediction about litigation that has never been brought. Both are regularly printed as findings. Neither is one.

Obligation, choice and flexibility: the 2005 settlement at a glance

The whole argument of this article fits in one table. Each element of the regime produced by the three amendments is marked as something the TRIPS Agreement required, a choice India made within the latitude the Agreement leaves, or a flexibility the Agreement expressly preserves. Readers preparing teaching material from the trail can lay the sequence out slide by slide with the ReportMedic presentation viewer, which suits a table whose columns carry the analysis.

Element of the regime	Required, choice or preserved flexibility	Provision
Product patents in all fields, including food, drugs and chemicals	Required by the Agreement	Section 5 deleted by the 2005 amendment
Twenty-year term from the filing date	Required by the Agreement	Section 53, as recast by the 2002 amendment
A means of filing product applications from 1995	Required by the Agreement	Section 5(2) and the mailbox scheme of 1999
Exclusive marketing rights during the transition	Required by the Agreement	Sections 24A to 24F, inserted 1999 and deleted 2005
Reversal of the burden of proof in process cases	Required by the Agreement	Section 104A, inserted by the 2002 amendment
New forms of a known substance excluded without enhanced efficacy	Indian choice within the Agreement’s latitude	Section 3(d), substituted by the 2005 amendment
Opposition before and after grant, open to any person before grant	Indian choice	Section 25(1) and Section 25(2)
Pre-2005 producers protected against injunction, royalty only	Indian choice, transitional	Proviso to Section 11A(7)
Compulsory licences on grounds including non-working	Flexibility the Agreement expressly preserves	Sections 84 and 92
Compulsory licence for export to countries without manufacturing capacity	Flexibility the Agreement expressly preserves	Section 92A, inserted by the 2005 amendment
Research and regulatory use outside infringement	Flexibility the Agreement expressly preserves	Section 107A(a), inserted 2002
Parallel importation from a person authorised under the law	Flexibility the Agreement expressly preserves	Section 107A(b), widened 2005

Read the rows in the first group and a pattern emerges: they are few, they are mechanical, and every one of them is a matter of availability, duration or procedure. Read the remaining rows and a different pattern emerges: everything in Indian patent law that generates argument, litigation and trade friction sits in the second and third groups, in the territory the Agreement left to members or expressly preserved to them. The columns are the article’s thesis in compressed form. India’s compliance and India’s restrictiveness are not in tension, because they describe different rows of the same table.

The table’s silences are as informative as its rows. There is no row for price control, because the settlement deliberately kept the pricing of medicines in the separate machinery of the Essential Commodities Act and its control orders, refusing to fuse the patent grant with a price condition. There is no row for data exclusivity, because no Indian instrument has ever created it, the demand having been refused in the passage and in the trade channel alike. And there is no row for a patent linkage between the drug regulator and the patent register, because the two systems were kept apart, leaving the marketing approval of a generic medicine to turn on safety and quality rather than on patent status. An audit that reads only what the amendments enacted will miss the settlement’s negative space, the list of devices, familiar in other jurisdictions, that Indian law was invited to adopt at every stage of the trail and declined at every stage. The declines were decisions, and they bind the future in the same way the enactments do: a negotiator or a minister proposing any of them must first explain why the settlement of 2005 should be reopened.

What the trail predicts

An amendment trail is worth writing only if its direction tells the reader something about what comes next, and this trail has a consistent direction. Each amendment widened the rights the statute confers and widened the safeguards around them in the same instrument: 1999 added the filing queue and the marketing right but kept Section 5; 2002 added the twenty-year term and treaty-grade enforcement and added the exceptions and the revised licensing chapter in the same Act; 2005 added the product patent itself and added the efficacy threshold, the restored opposition, the export licence and the widened parallel import in the same Act. The Indian method across a decade was never to trade safeguards for compliance but to legislate both simultaneously, and there is no point in the trail at which that pattern reverses. A reader predicting the shape of any future amendment should expect the same grammar: movement on the right and movement on the safeguard, together, with the safeguard carrying the parliamentary negotiation.

The open edges of the settlement are also identifiable from the trail, because they are the points where the drafters chose standards over rules. Significant investment and reasonable royalty in the transitional proviso were left undefined in 2005 and remain undefined. Working in India, the hinge of the compulsory licensing chapter, was left as a principle in Section 83 and a ground in Section 84 rather than a definition, and has been argued case by case ever since. Enhancement of known efficacy was left as a phrase for the courts. Standards of that kind do not settle policy; they relocate it, from the legislature to the Controller and the courtroom, which is where the subsequent history of Indian patent law has in fact been written. The settlement’s durability is itself a finding of the trail. No later amending Act has reopened the core bargain: product patents stand, Section 3(d) stands in its 2005 words, and the licensing and opposition chapters operate as the passage left them. What changed after 2005 was the venue of the argument, which moved from Parliament to the Patent Office, the Appellate Board and the courts, and from the existence of the flexibilities to their application in individual cases. That migration was the passage’s own prediction. A legislature that chooses case-by-case filters over categorical rules is choosing future litigation as its instrument, and the Indian state, having spent a decade negotiating the right to make that choice, has defended it in every forum since, against complainants in Geneva it never had and critics in Washington it answers annually. For the reader of the trail, the stability of the text is the strongest evidence of what the text was: not an interim concession awaiting a stricter round, but a considered settlement that its authors regarded as final.

The trail also predicts the procedure of any future change. Twice in this record a contested amendment was preceded by an attempt to fix the treaty’s meaning before the drafting began, once by dispute in Geneva and once by expert reference to the Mashelkar group; and the one amendment attempted without such a preliminary, the Ordinance of 2004, was rewritten in passage. A reader watching the open margins of the settlement, data protection in the trade channel, the working requirement in the licensing chapter, the efficacy threshold in the examination room, should therefore expect movement to announce itself first as a reference, a report or a ruling rather than as a Bill, and should expect the Bill, when it comes, to move the right and the safeguard together as every amendment in this trail has done.

The trail from the Patents Act’s amendment file is therefore best read as a record of two decisions taken ten years apart in form but once in substance: to accept the product patent, and to decide, domestically and in detail, what kind of product patent regime India would run. The first decision was made in Marrakesh. The second was made in New Delhi, and it is the second that the rest of this series keeps encountering. The account above is a legislative history offered for general information on the public record; it is not legal advice for any application, product or dispute.

Frequently Asked Questions

Q: What did India have to change to comply with TRIPS?

At the foundation, one thing: India had to make product patents available in the fields its Patents Act had kept closed since 1970, namely food, drugs, medicines and substances produced by chemical processes, where only processes could be patented. Around that core, the Agreement required a twenty-year term, defined exclusive rights, enforcement machinery including a reversible burden of proof in process cases, and, during the transition, a mailbox for filing applications and exclusive marketing rights for qualifying products. It did not require India to alter its patentability thresholds, opposition design, licensing grounds or exceptions, which is why the compliance legislation and the restrictive features of Indian law arrived in the same Acts.

Q: What were the three patent amendments?

The Patents (Amendment) Act, 1999, deemed in force from 1 January 1995, created the mailbox and exclusive marketing rights after India lost the WTO dispute over their absence. The Patents (Amendment) Act, 2002, in force from 20 May 2003, completed general compliance outside the closed fields: a uniform twenty-year term, reversal of the burden of proof in process infringement suits, expanded exclusions, a revised compulsory licensing chapter and the Section 107A exceptions. The Patents (Amendment) Act, 2005, deemed in force from 1 January 2005, deleted Section 5 to admit product patents in all fields, deleted the marketing rights chapter, substituted Section 3(d), rebuilt opposition, added the export licence in Section 92A and protected pre-2005 producers.

Q: What was the mailbox and the exclusive marketing right?

The mailbox was the filing facility the TRIPS Agreement required of a member using the extra transition period for product patents: from 1995, applications for medicines and agrochemicals could be filed and date-stamped in India, preserving novelty and priority, but they were not examined until the period expired at the end of 2004. The exclusive marketing right was the interim reward attached to a mailbox application. Where the applicant had obtained a patent and marketing approval abroad and marketing approval in India, the Controller could grant the exclusive right to sell or distribute the product in India for five years or until the application was granted or rejected, whichever came first.

Q: What did the 2005 patent amendment introduce?

The amendment introduced the product patent itself, by deleting Section 5 of the Patents Act, and with it the full set of rights over substances in every field of technology. It also deleted the exclusive marketing rights chapter, substituted Section 3(d) so that new forms of known substances require enhanced efficacy, added an Explanation treating salts, polymorphs and similar derivatives as the same substance, rebuilt pre-grant and post-grant opposition, inserted Section 92A for export compulsory licences, widened parallel importation, inserted the transitional proviso protecting enterprises that had invested and produced before 2005, and strengthened the national security provisions on filing inventions abroad.

Q: Why was the 2005 amendment brought by ordinance?

Because the treaty deadline and the parliamentary calendar collided. The additional transition period for product patents expired on 31 December 2004, the Bill intended for the winter session had not been taken up, an earlier Bill of 2003 had lapsed when the Lok Sabha was dissolved, and Parliament was not in session when the deadline fell due. The President therefore promulgated the Patents (Amendment) Ordinance, 2004 on 26 December 2004, effective from 1 January 2005. An Ordinance is temporary by design, and Parliament replaced it with the Act in March 2005, amending its text in respects that define the final settlement.

Q: Which TRIPS flexibilities did India retain?

India retained the full set the Agreement and the Doha Declaration preserve. Patentability criteria were left to domestic definition, producing the Section 3 exclusions including the efficacy threshold for new forms of known substances. Opposition before and after grant was kept and widened. Compulsory licensing was retained on grounds including non-working, unreasonable pricing and unmet public requirements, with an emergency route and an export licence for countries lacking manufacturing capacity. The research and regulatory use exception and international exhaustion, permitting parallel imports, were written into Section 107A. And the transition itself was used to the last day, with pre-2005 investment protected by the grandfather proviso in Section 11A(7).

Q: What happened to applications filed in the mailbox?

They waited, and then they were judged under the new law. Examination of the 8,926 mailbox applications began after the transition closed, in filing order, against the patentability standards of the amended Act, including Section 3(d). Applications that succeeded produced patents whose twenty-year term ran from the original filing date, so part of the term had already run. Rights under such a patent accrued only from the date of grant, and against an enterprise that had made significant investment and was producing and marketing the product before 1 January 2005, the patentee could not obtain an injunction, only a reasonable royalty under the proviso to Section 11A(7).

Q: Is India’s patent law TRIPS compliant?

On every obligation the Agreement’s text specifies, India performs: patents are available in all fields, the term is twenty years, the prescribed rights and enforcement procedures exist, and the transition duties were discharged. The contested provisions, including Section 3(d), the opposition design, the working requirement and the licensing grounds, occupy ground the Agreement leaves undefined or expressly reserves, and the Doha Declaration supports reading that ground in favour of public health. No WTO panel has held the post-2005 settlement inconsistent with the Agreement; the only adverse patent rulings concerned the missing transition machinery of the 1990s. Compliance in the contested zone is therefore asserted by India and denied by some trading partners, without an adjudication either way.

Q: What did the Patents (Amendment) Act, 1999 change?

It changed procedure, not rights. The Act, deemed in force from 1 January 1995, inserted Section 5(2) so that a claim for a medicine or drug itself could be filed and held, and inserted Chapter IV-A, Sections 24A to 24F, directing the Controller to keep such applications unexamined until the end of 2004 and creating exclusive marketing rights for qualifying applicants. It validated filings made under the lapsed Ordinance of 1994, and it answered the WTO rulings that had found India’s informal arrangements legally unsound. Section 5(1), the bar on product patents in the field, remained untouched, so the process-only regime survived the amendment that built its replacement queue.

Q: What did the Patents (Amendment) Act, 2002 change?

It completed the compliance that was due by 2000 and modernised the system’s working parts. Section 53 was recast to give every patent a uniform term of twenty years from filing, ending the shorter term for process patents in food and drugs. Section 104A reversed the burden of proof in process infringement suits in defined circumstances. The Section 3 exclusions were expanded, including a clause against patenting traditional knowledge. The compulsory licensing chapter was revised and the emergency licence in Section 92 recast. Provision was made for Patent Cooperation Treaty applications, and Section 107A created the research and regulatory use exception and permitted parallel importation.

Q: What is the Doha Declaration, and why does it matter for Indian patent law?

The Declaration on the TRIPS Agreement and Public Health was adopted by the WTO Ministerial Conference at Doha on 14 November 2001. Its fourth paragraph affirms that the Agreement can and should be interpreted and implemented in a manner supportive of members’ right to protect public health and to promote access to medicines for all, and its fifth paragraph confirms the flexibilities, including freedom to determine the grounds of compulsory licences and freedom on exhaustion of rights. It matters because it is an interpretive text endorsed by the entire membership: the reading of TRIPS on which India’s 2005 settlement rests is not an Indian gloss but a position the membership itself recorded.

Q: What is the Section 92A export compulsory licence?

Section 92A, inserted by the 2005 amendment, empowers the Controller to grant a compulsory licence solely for the manufacture and export of a patented pharmaceutical product to a country that has insufficient or no manufacturing capacity in the pharmaceutical sector, to the extent necessary to address the public health problem there. The importing country must have granted a compulsory licence for the product or have allowed its importation from India by notification or otherwise, wording widened during the Bill’s passage. The provision implements the mechanism built under paragraph 6 of the Doha Declaration for members whose need for a licence exceeded their ability to manufacture under one.

Q: What happened to the Patents (Amendment) Bill of 2003?

It lapsed. A Bill to make the third and final set of TRIPS amendments was introduced in 2003, was considered by a Group of Ministers amid sustained controversy over its treatment of patentability and opposition, and was still pending when the Lok Sabha was dissolved in 2004, on which pending Bills fall. The incoming government took up the subject again with a revised text, intended to legislate in the winter session of 2004 and, when the session ended without it, proceeded by Ordinance on 26 December 2004. The lapse is the reason the final amendment has a two-stage history, an Ordinance text and an Act text that differ in important respects.

Q: What did the 2005 amendment restore to pre-grant opposition?

It restored the opposition the Ordinance had taken away. The Ordinance of December 2004 had cut the grounds of pre-grant opposition from nine to two and removed the opponent’s entitlement to a hearing in person. The Act as passed reinstated the full list of grounds in Section 25(1), opened the representation to any person after publication of the application, and provided that the Controller shall hear the opponent if a hearing is requested, making the opponent a party to the proceedings rather than a source of information. The restoration proved consequential at once, because the mailbox applications were examined under it, Glivec’s among them.

Q: What protection did the 2005 Act give enterprises already making a mailbox product?

A right to continue, at a price. Under the proviso to Section 11A(7), where an enterprise had made significant investment and was producing and marketing a product that a mailbox application covered, before 1 January 2005, and continued to produce it, the patentee could not restrain the enterprise from carrying on. The patentee’s remedy was limited to a reasonable royalty. The clause protected investment made in lawful reliance on the process-only regime from being destroyed by a right that came into existence later, while recognising the patentee’s claim to be paid. Its two key terms, significant investment and reasonable royalty, were left undefined by the Act and have never been given an authoritative judicial measure.

Q: Did the Patents (Amendment) Act, 2005 operate retrospectively?

In a measured sense, yes. The Act received assent on 4 April 2005 but its main provisions were deemed to have come into force on 1 January 2005, so that no gap opened between the Ordinance it replaced and itself, and the product patent regime ran from the treaty deadline rather than the assent date. A small number of provisions were reserved for later commencement by notification. The retrospectivity was about continuity of the regime, not about past conduct: rights under mailbox patents accrued only from grant, and manufacture undertaken lawfully before 2005 was protected rather than penalised, through Section 11A(7) and its proviso.

Q: What role did the Left parties play in shaping the 2005 Act?

A decisive drafting role. The United Progressive Alliance governed with the outside support of the Left parties, and that support for a product patent law was conditional on safeguards being written into the statute. The amendments moved during the March 2005 passage, tightening the definitions in Section 2, adding the Explanation to Section 3(d), restoring the full grounds and hearing in pre-grant opposition, widening the export licence and the parallel import limb, and dropping the Ordinance’s widening of software patentability, were the product of that negotiation. The flexibility suite in the Act is therefore best understood as the recorded price of the majority that enacted product patents.

Q: Why does the United States keep India on the Special 301 Priority Watch List?

Because the Office of the United States Trade Representative, receiving submissions from American industry, maintains that features of the Indian settlement restrict patent rights beyond what trading partners should accept: Section 3(d)’s efficacy threshold, the breadth of opposition, the working requirement and compulsory licensing exposure, and the absence of regulatory data protection are recurring entries. Successive reports have kept India on the Priority Watch List across decades and describe it as among the most challenging major economies for intellectual property protection. The listing is a unilateral act of United States trade administration. It records a negotiating position, carries no adjudication of India’s treaty obligations, and has continued irrespective of the compliance record described in this article.

Q: What did the 2005 Act change in the definition of inventive step?

During its passage the Act put the definition in Section 2(1)(ja) into its durable form: inventive step means a feature of an invention that involves technical advance as compared to the existing knowledge or having economic significance or both and that makes the invention not obvious to a person skilled in the art. The disjunctive structure matters. An invention may qualify through technical advance, through economic significance, or through both, provided in every case that it clears the non-obviousness requirement. The same round of amendments added definitions of new invention and of pharmaceutical substances, the latter confined to a new entity involving one or more inventive steps, tying the medicines vocabulary of the Act to the inventive step threshold.

Q: What happened to exclusive marketing rights when the 2005 amendment took effect?

The scheme ended, and its remnants were wound into the patent system. The 2005 Act deleted Chapter IV-A, Sections 24A to 24F, so no new exclusive marketing rights could be granted after 1 January 2005. A right already granted did not vanish overnight: it continued until the patent application behind it was granted or rejected, the applications supporting granted rights were taken up for examination for the grant of a patent, and proceedings concerning a granted marketing right were dealt with as if they concerned a patent. The device had always been transitional, a five-year bridge between the mailbox and the product patent, and the amendment retired it on schedule.