A statute can clear every constitutional hurdle and still govern nothing. It can be passed by both Houses of Parliament, receive the assent of the President, be printed in the Gazette of India with an Act number of its own, be summarised in textbooks and taught in examination halls, and yet impose no obligation on any employer and confer no enforceable right on any worker, because one further step remains untaken: the executive has not appointed the day on which it is to come into force. The labour codes implementation delay is the largest modern demonstration of that gap in Indian lawmaking. Four codes, enacted in 2019 and 2020 to replace twenty-nine central labour statutes, waited more than five years for their appointed day, and when the day came on 21 November 2025, it arrived before the rules that make the codes operable, so that the long wait for commencement was followed immediately by a second wait of a different kind.

That sequence is what makes this the most instructive commencement story in the series. Other clusters in this series record statutes that were never notified at all, statutes notified in fragments, and statutes in force on paper while the institutions they created stood unbuilt. The labour codes supply the inverse case and complete the taxonomy: a reform whose stated obstacle for five years was that the subordinate legislation was not ready, commenced on a single day with most of that subordinate legislation still not ready, and held together in the interval by the device of applying rules framed under repealed statutes to the statutes that replaced them. A reader who understands why that happened, and what law actually governed an employer and a worker at each stage, understands something general about how Indian legislation behaves between enactment and operation, which is the stretch of a statute’s life that most commentary skips.

The argument of this article is contained in its title. Commencement and operability are different things, and the interval between them is not a technical footnote but a period with real legal content: rights accrue, prosecutions are launched, salary structures are designed, and disputes are filed inside it, all under a hybrid law that belongs fully to neither the old regime nor the new one. The record also resists the two easy verdicts. The delay was not simply sabotage or drift, because the federal explanation for it is genuine: labour sits on the Concurrent List, the codes delegate their operative detail to two levels of government, and a staggered commencement would have produced a patchwork that penalised exactly the uniformity the reform promised. Nor was the delay simply prudence, because five years is far longer than that explanation can carry, because the wait coincided with sustained political opposition that the government had every incentive to outlast quietly, and because commencement, when it came, did not wait for the uniformity that was supposed to justify it. Both readings are true, and the article holds them together rather than choosing the more comfortable one.

One more feature of the episode belongs in the opening frame, because it shaped what the delay meant for the people who had to act inside it. Commencement uncertainty is an information problem before it is a legal one, and the market for answers behaved accordingly. Across the first interregnum, and at flood tide in November 2025, employers and workers were served a large volume of confident and inaccurate guidance: summaries that gave commencement dates which had been rumoured and then deferred as though they had been law, accounts that described the codes as fully operative when their rules did not exist, and accounts in the opposite direction that treated the repealed statutes as though nothing had changed at all. Each error has a cost attached to a real decision, a salary structure rebuilt on a wrong date, a settlement signed on a wrong threshold, a claim filed in a forum whose governing statute had changed. The discipline this article applies to itself is therefore part of its subject. Every date in what follows is a date something was enacted, notified, published or answered in Parliament, stated with its source, and the rules position is described structurally, as a distribution across governments and time, rather than as a snapshot pretending to be a status. A reader who finishes the article should be able to say not only what happened but how the happening is known, because in a field where the official record includes a press release that celebrates and concedes in consecutive sentences, the method of knowing is half the knowledge.

The commencement position, stated precisely

The facts of enactment are short and undisputed. The Code on Wages, 2019, Act 29 of 2019, was passed by the Lok Sabha on 30 July 2019 and by the Rajya Sabha on 2 August 2019, and received the assent of the President on 8 August 2019. The other three codes were taken through Parliament a year later, in the monsoon session of September 2020: the Industrial Relations Code, 2020, Act 35 of 2020, the Code on Social Security, 2020, Act 36 of 2020, and the Occupational Safety, Health and Working Conditions Code, 2020, Act 37 of 2020, were passed by the Lok Sabha on 22 September 2020 and by the Rajya Sabha on 23 September 2020, and each received presidential assent on 28 September 2020, the Acts being published in the Gazette the following day. Between them the four codes subsumed twenty-nine central labour statutes, four under the Code on Wages, three under the Industrial Relations Code, nine under the Code on Social Security and thirteen under the occupational safety code, a distribution the framework article in this cluster sets out in full.

None of the four codes commenced on assent. Each carries, in its first section, an appointed-day clause of the standard Indian form: the code is to come into force on such date as the Central Government may by notification in the Official Gazette appoint, and different dates may be appointed for different provisions. That clause is the hinge on which this entire article turns. It converts commencement from an automatic consequence of enactment into a discretionary executive act, exercisable provision by provision and repeatable over time. Parliament, having spent its authority on the text, hands the timing to the government of the day, and a government that does not wish to operate a statute need not repeal it, amend it or defend its dormancy in the legislature. It need only refrain from signing a notification.

For five years that is, with two small exceptions, what happened. The first exception came on 18 December 2020, when the Central Government brought into force a small set of provisions of the Code on Wages concerned principally with the Central Advisory Board, the tripartite body through which floor wages are advised upon, and followed it with the Code on Wages (Central Advisory Board) Rules, 2021. The second came on 3 May 2023, when a set of provisions of the Code on Social Security was notified. Neither fragment touched the obligations of an ordinary establishment. No minimum wage was fixed under the Code on Wages, no standing order was certified under the Industrial Relations Code, no factory licence was issued under the occupational safety code, and no gig worker scheme was framed under the Code on Social Security during the whole of that period. The codes existed as enacted text and as a standing promise, and the twenty-nine statutes they were designed to replace continued to govern Indian workplaces in the ordinary way, their repeal clauses dormant because repeal, in each code, is itself tied to commencement.

The appointed day, when it arrived, arrived for all four codes at once. On 21 November 2025 the Ministry of Labour and Employment issued four commencement notifications, numbered S.O. 5319(E) to S.O. 5322(E), appointing 21 November 2025 as the date on which the provisions of the codes would come into force, and announced the commencement the same day by press release. The occupational safety code and the industrial relations code were commenced generally, while the notifications under the Code on Wages and the Code on Social Security enumerated the provisions brought into force, carrying forward the fragments commenced in 2020 and 2023 within the larger whole. From that date the twenty-nine subsumed statutes stood repealed, subject to the savings provisions in each code. The commencement of the Social Security Code required one correction after the fact: a corrigendum dated 19 December 2025 corrected the scope of certain provisions and exclusions in the November notification, an episode worth pausing on, because a commencement notification that itself needs correction within a month is a small, exact emblem of a commencement executed faster than its paperwork.

What did the labour codes repeal on the day they commenced?

On 21 November 2025 the repeal provisions of the four codes took effect and twenty-nine central statutes stood repealed, four under the Code on Wages, three under the Industrial Relations Code, nine under the Code on Social Security and thirteen under the occupational safety code, subject to savings for accrued rights, pending proceedings and rules made under the repealed law.

The savings were not a courtesy extended to the old regime. They were the mechanism that made a one-day repeal of twenty-nine statutes administrable at all, and the sections that follow return to them from the litigant’s side.

The detail of what replaced each repealed statute, provision by provision, belongs to the owner of the substantive comparison in this cluster, and this article does not redo it. What belongs here is the commencement mechanics that surrounded the repeal, because they determine what a person with a pending claim, a subsisting licence or an accrued liability actually holds. Three features matter. First, repeal and commencement are fused in the codes’ design: the repeal provisions operate because and when the codes commence, so the five-year postponement of commencement was also a five-year postponement of repeal, and the old law’s long survival was not an oversight but the direct legal consequence of the appointed-day clauses. Second, the repeal is not a bonfire. Each code carries a repeal and savings section, Section 69 of the Code on Wages, Section 104 of the Industrial Relations Code, Section 164 of the Code on Social Security and Section 143 of the occupational safety code, and behind those sections stands the general savings doctrine of the General Clauses Act, 1897, which preserves rights accrued and liabilities incurred under a repealed enactment and carries forward subordinate legislation made under a repealed statute into the corresponding provisions of its replacement so far as the two are not inconsistent. Third, the government’s own statements after commencement leaned explicitly on that continuity: the Ministry of Labour and Employment told the Lok Sabha, in an answer given by the Minister of State, Shobha Karandlaje, on 27 July 2026, that those provisions of the codes which do not require any rule to be notified are already in force, and that in the states, during the transition period, existing rules, regulations, notifications, standards and schemes under the repealed laws will continue to operate in so far as they are not inconsistent with the new codes, ensuring continuity and avoiding a legal vacuum. That sentence, delivered eight months after commencement, is the official description of the second interregnum, and it concedes the article’s central point in the government’s own words: the codes were in force, and the working law of the workplace remained, in large part, the law made under statutes that no longer existed.

The labour codes implementation delay, counted: how long each code waited

The delay is usually quoted as five years, and the figure is right for the three codes of 2020 and wrong for the first. The Code on Wages received assent on 8 August 2019 and commenced on 21 November 2025, a wait of six years and three and a half months, longer than the full term of the Parliament that enacted it and spanning two general elections. The Industrial Relations Code, the Code on Social Security and the occupational safety code received assent on 28 September 2020 and waited five years and just under two months. Measured not to commencement but to operability, the waits lengthen again: from assent to the Central Rules that gave the central sphere its rulebook, the Code on Wages took six years and nine months, and the codes of 2020 five years and seven months, with the state rulebooks that govern most establishments trailing behind those dates by months more. No single number therefore describes the delay, and the article uses the figures in their places: five years for the enactment-to-commencement interval of the 2020 codes, which is the interval the public argument is about, and the longer figures where the operative question is when an establishment’s governing law was complete.

The counting matters for a reason beyond precision. A wait of five years crosses thresholds that a wait of eighteen months does not. It exceeds the life of most of the governments that began it: the labour ministers, secretaries and state administrations of 2020 were substantially not the office-holders of 2025, so the reform was inherited, twice over in some offices, by officials who had not drafted it and whose institutional memory of its purposes was documentary. It exceeds the horizon of the wage settlements, standing orders and compliance systems it governed in prospect, so that the interregnum generation of each of those instruments was drafted in the codes’ shadow and had to be redrafted at their arrival. And it exceeds, by a wide margin, the period for which the explanation of preparation retains its natural force, since preparation is a process with products, and after the draft rules of 2020 and 2022 the products stopped appearing. A reader who keeps the arithmetic in view can locate, with unusual exactness, the point at which the delay changed its nature: not a date on which anything was decided, but the stretch, across 2023 and 2024, in which the documents stop accumulating and the story passes from the phases of pandemic and preparation into the phase of attribution. Legislative delay is usually a fog. This one has a chronology, and the chronology is the evidence.

The appointed day: a small clause with a long shadow

The default rule of Indian statute law is the opposite of what happened to the labour codes. Section 5 of the General Clauses Act, 1897 provides that where an Act of Parliament makes no provision for its commencement, it comes into operation on the day on which it receives the assent of the President. Enactment and operation, under the default, are simultaneous, and a statute that is to wait must say so. The appointed-day clause is the standard device by which it says so, and its rationale is genuine. Modern statutes depend on machinery that cannot be conjured on the day of assent: rules must be drafted, authorities staffed, forms prescribed, systems built, and a commencement tied to assent would either force that work into the period before passage, when the text may still change, or produce a statute in force and unusable from its first morning. The clause exists so that a government can finish building before it switches on, and in the ordinary case it works as designed, with commencement following assent by weeks or months and nobody outside the gazette noticing the seam.

The labour codes show what the device becomes when the seam is allowed to stretch. Three features of the clause, each innocuous in the ordinary case, combine badly at scale. The first is that the discretion it confers is unstructured: the clause names no outer date, requires no reasons, and imposes no duty to report the delay to Parliament, so a government that never commences a provision breaks no law and files no explanation. The second is that it is divisible: different dates may be appointed for different provisions, so commencement can be granted to the harmless fragments of a statute, an advisory board here, a definition there, while the operative core waits, and the statute book acquires the appearance of motion without its substance. The third is that the discretion is, for practical purposes, unreviewable: the courts have declined to direct the executive to exercise it, for reasons examined later in this article, so the only forum that can correct an indefinite postponement is the political one, in which an uncommenced reform is, as this article argues below, an asset rather than an embarrassment to the government holding it. A clause with these three features is not a scheduling device. It is a veto that Parliament builds into its own work, exercisable by silence.

It is worth setting beside this the way the codes’ own drafters handled the repeal side of the same design, because the contrast shows the choice was understood. The repeal of the twenty-nine statutes was not left to a separate act of executive will; it was written into the codes themselves, to operate on commencement, so that no government could take the credit of the new law while keeping the old law alive beside it. Whoever commenced the codes commenced the repeal in the same signature. The Indian drafting tradition, in other words, knew how to fuse two events when fusion mattered, and the one event it left unfused was commencement itself. Reform proposals for curing the pathology recur in the commentary and were rehearsed again across the codes’ five years: statutory outer limits on appointed days, commencement by default after a fixed period unless Parliament resolves otherwise, and a duty to lay a statement of reasons before Parliament for each year of non-commencement. None has been adopted as general law, and the labour codes were enacted, waited and commenced entirely within the traditional design. The design is therefore not background to this story. It is one of its authors.

Why a code cannot operate on its enacted text alone

To see why rules mattered so much in this reform, a reader has to see how the codes are built. A nineteenth-century statute typically carried its operative detail in its own sections: rates, forms, procedures and penalties on the face of the Act. The labour codes are built the other way. They state principles, definitions, entitlements and prohibitions at a level of generality, and then delegate the operative machinery downwards, hundreds of times over, to rules to be made by the appropriate government, to regulations to be made by the boards and corporations the codes create or continue, and to schemes to be framed for the benefits that are the codes’ main promise to the unorganised workforce. The floor wage under the Code on Wages is to be fixed by the Central Government through the advisory board process the rules structure. The manner of fixing minimum wages, the forms of registers and returns, the timelines for claims, the composition and procedure of the authorities who hear them, all sit in rules. The social security code’s celebrated extension to gig and platform workers operates through schemes: the code defines the worker, the platform and the aggregator, fixes the band within which aggregator contributions fall, and then leaves the benefit itself, its quantum, its eligibility conditions and its delivery machinery, to schemes the government must frame. The occupational safety code’s licensing, registration and inspection architecture likewise runs through rules made by the government appropriate to the establishment. A code of this design, commenced without its subordinate legislation, is a skeleton that has been told to stand.

The second structural fact compounds the first. Labour is not a Union subject. Entries 22, 23 and 24 of the Concurrent List in the Seventh Schedule cover trade unions and industrial disputes, social security and employment, and the welfare of labour including conditions of work, and the division of the field between the Centre and the states, which the article on the Seventh Schedule lists explains in its own right, means that for the great majority of establishments the appropriate government is the state government, not the Central Government. The codes respect that division on their face: rule-making power is vested in the appropriate government for each establishment, so the Central Government makes rules for the central sphere, railways, mines, major ports, banking and insurance and central public sector undertakings among them, and each state government must make its own rules for the factories, shops, plantations and transport undertakings that fill its territory. There is no single rulebook for the codes. There are, in prospect, one central rulebook per code and over thirty state rulebooks per code, and the codes operate fully in a state only when that state’s rules exist. A reform of this architecture cannot be switched on from Delhi alone, and every actor in the story, government, opposition, unions and industry, understood that from the beginning.

The delegation runs at every level of the codes’ design, and a short inventory shows how little of the operative law Parliament itself enacted. Under the Code on Wages, the floor wage, the manner of fixing and revising minimum wages, the constitution and procedure of the advisory boards, the form of the wage slip and the register, the timelines and forms for claims, and the compounding of offences all sit below the Act. Under the Industrial Relations Code, the model standing orders are an executive product, the procedure of works committees and grievance redressal committees is prescribed, the recognition mechanics of negotiating unions and councils run through rules, and the contribution and administration of the worker re-skilling fund is scheme and rule work. Under the Code on Social Security, the schemes are the substance, as the following sections show, and the rates, ceilings and machinery of the funds move by notification. Under the occupational safety code, registration and licensing, the standards for hazardous processes, the duties around contract labour and inter-state migrant workers, and the inspection scheme are delegated in the same way. None of this delegation is improper; much of it is unavoidable in a field whose detail changes faster than Parliament sits. But its consequence for commencement is severe and was visible from the day of enactment: the codes could not be commenced responsibly by signature alone, and everyone in a position to know it knew it, which is why the story of the delay is properly told as a story about rule-making and not about the codes’ text at all.

This is also why the familiar comparison with a statute like the Goods and Services Tax legislation, often invoked in argument about the delay, misleads. The GST Acts commenced on dates fixed with the subordinate legislation, the council machinery and the network built in parallel and in public, over a prepared runway. The labour codes were drafted on the assumption that the runway would be built after enactment, by thirty-seven governments acting at their own pace on a politically contested subject, with tripartite consultation at each stage in a field where consultation is not a courtesy but the operating system. The drafting choice to delegate so much, to so many, is the deep cause of everything that followed. It does not excuse five years. It explains why the unit of delay in this story was never going to be weeks.

How long did the second interregnum last?

For the central sphere it ran from 21 November 2025 to 8 May 2026, when the Central Rules under the four codes were notified: just under six months in which the codes were the governing law and the central rulebooks did not exist. For establishments governed by state rules the interval is longer and varies by state.

The state interval has no single closing date, because it closes state by state and code by code: state rule-making continued well past May 2026, and final rules under all four codes stood in place in only a minority of states long after the Central Rules appeared, with the carried-forward rules of the repealed statutes filling the remaining ground in the rest.

The government’s handling of the transition acknowledged the problem in terms. The press release of 21 November 2025 itself stated that during the transition the relevant provisions of the existing labour Acts and their rules, regulations, notifications, standards and schemes would continue to remain in force. Practitioners’ advisories in the days after commencement, from firms including JSA and Khaitan and Co, converged on the same practical instruction to employers: the substantive standards of the codes had arrived, day-to-day compliance continued to run on the forms, registers and procedures of the old regime, and the sensible posture was to prepare for the codes’ machinery without dismantling the machinery that still, in law, governed the file in front of them. When a reform’s first professional commentary is a set of instructions for living in two regimes at once, the reform has commenced in the gazette sense and not yet in the administrative one. The remainder of this article reconstructs how India arrived at that posture twice over: once by waiting five years, and once by not waiting for the rules.

The first interregnum: five years between assent and commencement

The five years divide into phases, and the phases matter because the explanation that fits one does not fit the others. The first phase ran from enactment to the middle of 2021 and is the phase the pandemic genuinely explains. The three codes of 2020 were passed in a Parliament working under pandemic conditions, in a session dominated by the farm legislation and by the opposition’s boycott of the Rajya Sabha on the day the labour codes passed that House. The administrative state that would have had to frame rules was, in those same months, managing a lockdown, a migrant worker exodus that the codes’ own inter-state migrant provisions were written in response to, and a collapse and recovery of employment that made any immediate disturbance of wage structures politically and humanly hazardous. Several states moved in the opposite direction in 2020, suspending or relaxing labour law application to restart economic activity, which is a fair index of the climate: the question in 2020 was not how fast the codes could be commenced but whether labour standards would hold at all. A government that had commenced the codes in the winter of 2020 would have been restructuring provident fund contributions and retrenchment thresholds in the middle of the second pandemic year. The delay of the first phase needs no conspiracy to explain it.

The second phase, from 2021 into 2023, is the phase of the first draft rules, and it is the phase that gives the delay its most respectable documentary record. The Central Government pre-published draft rules under each of the four codes in the second half of 2020, the draft Code on Wages (Central) Rules appearing in July 2020 and the set complete by November of that year, inviting stakeholder comment in the ordinary way. States followed at their own pace. By 25 July 2022 the Ministry of Labour and Employment could tell Parliament, in a published answer, that the Central Government had pre-published its draft rules and that, as per available information, thirty-one states and union territories had pre-published draft rules under the Code on Wages, twenty-six under the Industrial Relations Code, twenty-five under the Code on Social Security and twenty-five under the occupational safety code. The two commencement fragments of December 2020 and May 2023 belong to this phase and show the machinery being assembled piece by piece. Read on its own, this record describes a reform in genuine, slow preparation: drafts published, comments invited, states at different stages of a heavy rule-making burden, and a Centre repeating that commencement would follow readiness.

The third phase, from 2023 to November 2025, is where the respectable explanation thins. Draft rules aged without being finalised. The general election of April to June 2024 intervened, and with it the ordinary paralysis of a model code of conduct period and the reconstitution of the government, but the election does not explain the years on either side of it. Assembly elections cycled through the large states across the whole period, so that at almost any moment some state whose rules were needed was approaching a poll, and the argument from elections, if accepted without limit, becomes an argument for never commencing anything. The government’s own attribution of the delay in this phase is on record and is worth quoting in substance: in a statement in Parliament during the monsoon session of 2025, the position reported was that thirty-two states and union territories had pre-published draft rules under the codes, that West Bengal and Lakshadweep had not pre-published draft rules under any of the codes, that Delhi had pre-published drafts only under the Code on Wages, and that the delay in implementation was largely due to the lag in the states. A reader should notice what that statement does. It records real progress in draft-making and simultaneously locates the obstacle outside the Centre, in the states, at the very moment when the decision to commence, drafts or no drafts, rested with the Centre alone. Four months later the Centre commenced the codes over the top of the incomplete state position it had described, which does not make the July 2025 statement false, but does reveal that state readiness was a reason the Centre could weigh rather than a wall it could not pass.

The fourth phase is the commencement itself, and it belongs to the next section. What the three phases together establish is that the five-year delay is not one fact but three: a pandemic phase that explains itself, a preparation phase that is genuinely documented, and a long final phase in which the documents stop accumulating and the explanation shifts from preparation to attribution. Any verdict on the delay that does not distinguish the phases will be wrong about at least two of them.

The dates that were announced and passed

A distinctive feature of the first interregnum, and the one that did most to damage the credibility of every subsequent assurance, is the procession of expected dates that entered public discussion and then passed without event. Across 2021 and 2022, reporting based on official statements successively placed the intended rollout at the start of a financial year, then at a mid-year date, then at the start of the following financial year, each expectation expiring into the formula that commencement would follow state readiness. Ministers continued to speak of implementation as imminent across the whole period, and the imminence was not fabricated: the draft rules existed, the conferences met, and at any given moment a commencement within the following two quarters was a defensible forecast. What no actor in the system supplied was a date the government was bound to, and the distinction matters for how the period should be judged. A missed deadline is a failure of execution that can be located and priced. A permanently receding horizon is a technique of management, in which the reform is always arriving and never arrives, and in which employers are kept in a posture of preparation that costs money while unions are kept in a posture of protest that wins nothing. The professional literature of the period bears the marks of the technique: advisories from the major law and accounting firms run in an unbroken series from 2020 to 2025, each explaining the codes’ provisions to clients who had to decide, quarter after quarter, whether to restructure pay in anticipation of a commencement that each advisory correctly said could not be predicted.

The procession had a second effect, inside the state governments whose lag was the official explanation. A state labour department deciding how much of its scarce drafting and administrative capacity to commit to finalising code rules had to weigh the Centre’s horizon like everyone else. Draft rules were the rational product of that calculation: they demonstrated progress, satisfied the Centre’s reporting, and committed the state to nothing it would regret if commencement receded again, whereas final rules drafted against a moving central design risked rework. The near-universality of drafts and the rarity of finals, the distribution recorded in the parliamentary answers, is exactly what a system of rational actors produces when the principal’s deadline is understood to be soft. The Centre’s complaint about state lag and the states’ inference from the Centre’s behaviour are the same fact read from opposite ends, and any account of the delay that adopts only one end of it has been captured by one of the parties. When commencement finally came without warning, it broke the equilibrium at a stroke, and the speed of everything that followed, drafts in five weeks, central rules in six months, state finals multiplying through 2026, is the best available evidence that the constraint all along had been the credibility of the deadline and not the capacity of the system.

The old law’s afterlife: what the wait did to the twenty-nine statutes

An interregnum is usually imagined as a pause, a period in which nothing happens because the new thing has not started. The first interregnum was not like that for the twenty-nine statutes, which lived a full and in places energetic afterlife while their replacements waited. They were administered, amended, litigated and, in several states, reformed in the direction of the codes that were supposed to supersede them, so that by November 2025 the governing law was not the law of 2019 preserved in amber but a moving body of doctrine whose centre of gravity had begun to shift towards a regime that had no legal existence. Understanding the delay requires holding that picture, because it dissolves the simple image of an old order frozen and a new order waiting, and replaces it with two orders influencing each other across a boundary that existed in politics and not yet in law.

The clearest channel of influence was state legislation. The codes’ signature thresholds had a state prehistory: Rajasthan’s amendments of 2014, made with the assent of the President under Article 254(2) of the Constitution, had already raised the prior-permission threshold in the Industrial Disputes Act to three hundred in that state and had tested the factory and contract labour thresholds at higher levels, and the codes’ national figures were drawn in part from that experiment. During the interregnum other states worked the same seam within the old statutes, adjusting thresholds, digitising inspections and consolidating state-level licences, each change lawful under the old law and legible as preparation for the new. An employer in Jaipur or Gandhinagar in 2023 could be operating under thresholds closer to the codes than to the central Acts the codes were to repeal, which is one reason the state patchwork argument in the next sections cuts both ways: the patchwork was not a risk the delay avoided, because variation was accumulating under the old law too. It was a condition the delay managed without curing.

The administrative layer moved as well, and its movement is the part of the interregnum with the strongest claim to have been genuine preparation. The ministry’s digital architecture for the codes, a revamped Shram Suvidha portal intended to carry single registration, common licences, a common annual return and web-based randomised inspection under the new regime, was under construction during the wait, and states were asked to connect their own labour portals to the central platforms so that registrations, returns, inspection records and grievances could move between them. This is unglamorous work and it is also the work that decides whether a code, once commenced, is experienced as a simpler regime or merely a differently labelled one. A law that prescribes electronic filing cannot be operated by a paper office, and the states whose portals were weakest were predictably the states whose rules were slowest, administrative capacity and rule-making pace being two faces of the same department. When the government attributed the delay to the lag in the states, this is the lag it could most plausibly have meant, and the attribution deserves the respect of being partly true.

Registration infrastructure tells the same story in numbers of a durable kind. The e-Shram portal, launched in August 2021 under the old unorganised workers framework to build a national register of unorganised workers, accumulated registrations in the tens of crores during the interregnum and became, in practice, the delivery spine on which the Code on Social Security’s promises to the unorganised and platform workforce would have to run. A register built under a statute the code was to repeal, populated while the code waited, and indispensable to the code once it commenced, is the afterlife problem in miniature: the transition the savings provisions govern is not between two sealed regimes but between two accumulations of administrative fact, and the second regime inherits the first’s data, staff, habits and arrears along with its rules.

The third feature of the afterlife is doctrinal, and it is the subtlest. Litigation under the old statutes did not pause for the codes. Courts went on construing the definition of industry, the conditions of retrenchment and the scope of the wage and bonus statutes through the whole of the interregnum, producing precedent that the savings architecture would carry into the codes’ era as the received reading of corresponding provisions. Employers and unions, for their part, drafted around the waiting law: settlements, standing orders and salary structures from the middle of the period routinely took positions on the codes’ wage definition and thresholds in advance, pricing a statute that might commence in a year or in a decade. The interregnum therefore had a jurisprudence and a practice of its own, organised around a text that governed nothing directly and influenced nearly everything. When commencement came, it did not interrupt a stable old regime. It landed on a field already tilted towards it, which is part of why the landing, for all its abruptness, did not produce the chaos the unions predicted or the clean break the government advertised.

The reasons given and the reasons visible

The stated reasons, taken from the government’s own pronouncements across the period, form a coherent case and deserve to be put at full strength before they are weighed. The first stated reason was federal. Labour is a Concurrent List subject; the codes cannot operate fully in a state without that state’s rules; and a commencement that ran ahead of state readiness would produce the patchwork the reform existed to abolish, with a multi-state employer facing the new regime in one plant and the old regime in another. The second stated reason was uniformity: the government’s declared preference was for a simultaneous rollout across the country, a single appointed day on which the twenty-nine old statutes would give way everywhere at once, rather than a staggered commencement state by state and provision by provision. The third was the pandemic, which consumed the first two years and made the disturbance of wage and separation arrangements a cost no government would voluntarily schedule. The fourth was consultation: the codes were framed, the government maintained, after wide consultation during drafting, and the framing of rules would likewise be done with stakeholders, a process that in the tripartite tradition of Indian labour administration is slow because it is real. Put together, the stated case is that commencement was delayed by the structure of the reform itself, and that a government which had commenced earlier would have commenced worse.

The visible reasons sit alongside the stated ones and are also documented rather than conjectured. The first is organised opposition. The joint platform of ten central trade unions resisted the codes from enactment onwards: protests followed the Code on Wages in 2019 and culminated in a general strike in January 2020; a general strike on 26 November 2020 accompanied the farmers’ movement to Delhi; a two-day strike followed in March 2022; joint actions with the Samyukt Kisan Morcha continued through February 2024; and a general strike on 9 July 2025, months before commencement, restated the demand that the codes be scrapped. A government does not need to concede anything to such a campaign for the campaign to work on its timing: commencement converts a dormant text that unions protest against in principle into an operative regime that disturbs specific arrangements in specific workplaces, and every year of delay was a year in which that conversion, and its political cost, was postponed. The pandemic also left a specific deposit in the politics of the middle phase that has to be reckoned with before the visible reasons are weighed. In 2020, in the name of reviving economic activity, several state governments moved to suspend the application of large parts of the labour statute book or to raise thresholds and extend working hours within it, and the Union government was criticised by the unions for acquiescing in a competitive dilution of standards between states. The episode taught every participant a lesson that outlasted it. The unions learned that labour standards in a crisis would be treated as adjustable costs, and read the codes thereafter as the permanent form of the emergency’s temporary dilutions. The states learned that labour flexibility was an investment-attraction instrument on which they would be compared. And the Centre learned that the politics of commencing a contested labour reform into a fragile recovery would be conducted in the shadow of an exodus of migrant workers that the country had watched on its highways. None of this appears in the stated reasons, and all of it is part of why the stated reasons sufficed for as long as they did.

The second visible reason is state reluctance of a partisan as well as administrative kind. Some state governments had political reasons of their own to move slowly on a central reform their parties opposed, and the draft-rule record shows the spread: near-universal draft-making by the middle of the period, but finalisation concentrated in very few states, which is the signature of governments willing to be seen preparing and unwilling to be first to operate. The third visible reason is the electoral calendar already described, operating on the Centre as much as on the states: a reform whose most discussed consequence was a restructuring of take-home pay through the wage definition was never going to be commenced in the approach to a general election if it could be commenced after one.

Weighing the two sets of reasons against each other, the honest conclusion is that the stated reasons explain the shape of the delay and the visible reasons explain its length. Federalism explains why commencement could not be instant and why the exercise was measured in years rather than months. It does not explain why, by 2024, final state rules existed almost nowhere while draft rules existed almost everywhere, a distribution that suggests the constraint was not drafting capacity but the absence of a deadline. Deadlines are what governments supply when they intend a thing to happen, and for five years no deadline was supplied, announced and missed dates being replaced by the formula that commencement would follow readiness. The single most revealing document of the whole period may be the remark of the Union Labour and Employment Minister, Mansukh Mandaviya, in December 2025, explaining why fresh draft Central Rules were being published after commencement: the central government and the states had pre-published draft rules earlier, but that was a long time ago, and there was a need to bring the draft rules again in sync with present times. The remark is candid and, in its candour, conclusive about the second and third phases. The drafts of 2020 had not been a runway under construction. They had been allowed to go stale, like milk rather than like wine, and the work of the second interregnum had to begin by redoing the work of the first.

The political economy underneath both sets of reasons deserves one further observation, because it generalises beyond this cluster. An enacted but uncommenced statute is, for a government, a nearly costless possession. It can be claimed as reform in every speech, cited as intent in every negotiation, and held in reserve against every demand for action, while none of its costs in compliance, contribution or political friction has yet been incurred by anyone. Commencement is the moment the possession becomes expensive. This asymmetry explains why appointed-day clauses, drafted as flexibility devices, so often become instruments of indefinite postponement in Indian practice, and why the labour codes, the largest reform held in that posture in a generation, were always likely to be held there until some consideration stronger than readiness supplied a reason to let go. What that consideration was in November 2025, the government did not say, and this article does not invent one. The visible facts are that a national conference of state labour and industry ministers met in New Delhi on 11 and 12 November 2025, that the draft Shram Shakti Niti, the proposed national labour and employment policy, was placed in circulation in the same period, and that nine days after the conference closed, commencement was announced. Sequence is not motive, and the record is left as a sequence.

The day commencement came

Nothing about 21 November 2025 resembled the simultaneous, prepared rollout that five years of explanation had described. There was no announced commencement date published weeks ahead, no countdown in which employers were told to ready their payroll systems for a named Monday, no final set of rules laid on the table alongside the notifications. The commencement was announced by a press release of the Ministry of Labour and Employment on the day itself, with the four gazette notifications of the same date, and the professional reaction records the surprise in terms a historian can use. The law firm JSA, in its client analysis of 27 November 2025, described the government as having acted through an unforeseen announcement by way of a press release. DLA Piper’s employment practice recorded the same characterisation. Employers who had been told, in substance, for five years that commencement was imminent and would follow readiness received, when it came, no lead time at all: the codes were in force from the Friday of the announcement, and the salary structures, standing orders, registrations and contribution arrangements of every establishment in the country changed their governing statute over a weekend.

The substance of what commenced also deserves precision, because commencement day reporting flattened it. The Industrial Relations Code and the occupational safety code were commenced in general terms. The Code on Wages and the Code on Social Security were commenced by enumeration, their notifications listing the provisions brought into force, incorporating the fragments of 2020 and 2023 and leaving specified matter, including parts of the repeal and miscellaneous provisions, to be handled with a care that produced, within a month, the corrigendum of 19 December 2025 to the Social Security Code notification. The repeal of the twenty-nine statutes took effect with commencement, subject to savings. The Central Rules under the codes did not exist. The state rules, with exceptions under individual codes in a small number of states, did not exist in final form. What existed was the codes’ text, the old rules carried forward by the savings and continuity arrangements, the advisory machinery of the old regime continuing in office, and a press release.

The press release that announced the commencement deserves a paragraph of its own, because its two halves state the whole episode. Its first half was celebration: the decision was described as historic, the codes as laying the foundation for a future-ready workforce and stronger, resilient industries driving labour reform for a self-reliant India, and the twenty-nine laws they replaced were characterised as products of the pre-Independence and early post-Independence era that had failed to keep pace with the world of work. Its second half was the transition admission: that in line with the consultations carried out during drafting, the government would engage with the public and stakeholders in framing the corresponding rules, regulations and schemes, and that during the transition the relevant provisions of the existing labour Acts and their rules, regulations, notifications, standards and schemes would continue to remain in force. Read together, the halves announce a completed revolution and a continuing old regime in consecutive sentences, and both were true. The release is also the only official narrative of why the day had come when it came, and it is silent on the question: it gives the codes’ virtues at length and the timing no explanation at all beyond the decision’s announcement of itself. In the documentary record of a five-year wait, the culminating document is a celebration with a continuity clause, and a reader who wants the reason for November 2025 will find, in the official record, that there isn’t one.

The reactions of the day map the politics exactly. The joint platform of ten central trade unions issued its statement on 21 November 2025 itself, condemning what it called the blatantly unilateral implementation of anti-worker, pro-employer labour codes, describing the notification in its headline as a deceptive fraud on the working people of the nation, demanding a complete rollback, and calling for combative resistance and defiance on 26 November 2025 in mobilisation alongside the Samyukt Kisan Morcha. The Bharatiya Mazdoor Sangh, the labour federation outside the joint platform, took the differentiated position it had prepared: it welcomed the implementation, saying the reforms would ensure dignity, protection and progressive welfare for the workforce, and then, after a delegation met the Labour and Employment Minister, issued a further statement strongly opposing certain provisions of the Industrial Relations Code and the occupational safety code while endorsing the Code on Wages and the Code on Social Security. Industry spoke in the opposite register and just as promptly. Chandrajit Banerjee, Director General of the Confederation of Indian Industry, called the commencement a historic milestone for India’s labour landscape and the consolidation of twenty-nine laws into four codes a transformative leap towards a modern, simplified and future-ready labour ecosystem, promising better wages, stronger social security, enhanced workplace safety and a more predictable regulatory environment. The same industry bodies would be writing to the ministry within weeks seeking clarifications and protection against retrospective cost, which is the next section’s story; on the day, the welcome was unqualified, and the contrast between the welcome and the resistance is the political fact of the commencement. A reform commenced after five years of consultation claims was received by one side as a milestone and by the other as a fraud, and the disagreement was not about the contents of the codes, which had been fixed since 2020, but about the manner and the moment of their arrival.

There is a final detail of the day that a commencement record should not lose. The notifications were issued under the commencement clauses of each code by the ministry that had held the codes for five years, and they were, in form, four short instruments of a few lines each, appointing a date. Six years of legislation, five years of postponement, the repeal of twenty-nine statutes and the re-governing of every workplace in India were effected by documents shorter than the press release that announced them. The disproportion is not an irony to be savoured but the mechanism itself, made visible: in a system of appointed days, the largest legal events are accomplished by the smallest instruments, and the entire politics of implementation reduces, at the last, to the politics of a signature. Everything in this article before this section is an account of why the signature took five years. Everything after it is an account of what the signature could not do by itself.

The two interregnums: what law governed, stage by stage

The table below is the article’s findable artifact, and it is worth reading down the final column slowly, because that column is where the two interregnums become visible as a single phenomenon. At every stage the question is not which statute Parliament had enacted but which law an employer and a worker could actually invoke, and the answer changes its character three times: the old law fully, then the old law by default, then the new law’s text operating through the old law’s machinery.

Stage	Dates	Law in force	What actually governed an employer and a worker
Enactment	8 August 2019 and 28 September 2020	The twenty-nine central labour statutes; the four codes enacted but uncommenced	The old statutes governed everything; the codes were text without legal operation, citable in argument about intent and nothing more
The first interregnum	29 September 2020 to 20 November 2025	The twenty-nine statutes, with two fragments of the codes commenced on 18 December 2020 and 3 May 2023	The old statutes governed in full; the commenced fragments concerned advisory machinery and altered no obligation of any establishment
Commencement day	21 November 2025	The four codes, with the twenty-nine statutes repealed subject to savings	The codes’ text governed rights and prohibitions; pending matters and accrued rights continued under the repealed statutes by force of the savings provisions
The second interregnum	21 November 2025 to 7 May 2026	The four codes, without Central Rules and almost without final state rules	The codes governed in principle; forms, registers, procedures and day-to-day compliance continued under rules framed under the repealed statutes, so far as not inconsistent with the codes
Central Rules	8 May 2026	The four codes with the Central Rules notified	Establishments in the central sphere received their rulebook, including the model standing orders; establishments under state governments awaited state rules of equivalent content
State rule phase	From 8 May 2026 onwards	The four codes with Central Rules and a growing, incomplete set of state rules	The governing law became a function of geography: final rules under all four codes in a minority of states, partial rules elsewhere, and the carried-forward old rules still filling the gaps

Two features of the table reward attention. The first is the asymmetry of the two interregnums. The first lasted five years and two months from the assent of the three codes of 2020, and over six years from the assent of the Code on Wages, and during it the legal position was at least coherent: everyone knew which law governed, because it was the old one. The second interregnum is shorter but stranger, because during it the governing law is a composite that has to be assembled case by case, provision by provision, by asking of each rule of the old regime whether it survives, and of each provision of the codes whether it needs a rule to bite. Coherence was traded for movement. The second feature is the identity of the device that spans both interregnums: the savings and continuity arrangements, which in the first interregnum preserved the old statutes by leaving the codes uncommenced, and in the second preserved the old rules by carrying them into the codes. The Indian state’s deepest habit in statutory transition is to ensure that the outgoing law never quite leaves the room, and the labour codes exhibit the habit at both ends of their journey.

In force without a rulebook: the record after commencement

The months after 21 November 2025 followed a sequence that can be stated as dates, and the dates are the criticism and the defence rolled into one. On 30 December 2025, five and a half weeks after commencement, the Ministry of Labour and Employment pre-published fresh draft Central Rules under all four codes, allowing thirty days for comment on the Industrial Relations Code draft and forty-five days on the other three. The Labour Secretary, Vandana Gurnani, had told a newspaper on 27 November 2025 that it would take about ninety days from the publication of draft rules for the codes to become operational. The Press Trust of India reported at the end of December 2025 that the government intended to fully operationalise the four codes from 1 April 2026 across the country in one go. The final Central Rules were in fact notified on 8 May 2026, the Code on Wages (Central) Rules, the Industrial Relations (Central) Rules, the Social Security (Central) Rules and the occupational safety (Central) Rules, together with the model standing orders under the Industrial Relations Code. An official told the Press Trust of India on the publication of the rules that the process of full implementation of the labour codes was complete. That claim, like the claim of readiness that preceded commencement, requires the geographic qualification: the Central Rules bind where the Central Government is the appropriate government, a sphere the Ministry’s own frequently-asked-questions document of 16 March 2026 described by reference to railways, mines, banking and insurance and central public sector undertakings, and the ordinary private establishment in a state awaited its own government’s rules for the corresponding machinery.

The state record after commencement continued the pattern of the first interregnum, preparation outpacing finalisation. The most careful contemporary count, published in the November 2025 employment bulletin of Khaitan and Co from the position at commencement, recorded that only two of the thirty-six states and union territories had yet to publish draft rules under the Code on Wages and the occupational safety code, three under the Industrial Relations Code and one under the Code on Social Security, while final rules had been released only under some of the codes and only in a few states, Gujarat, Karnataka, Uttar Pradesh and Mizoram among them. Draft-making, in other words, was nearly universal and final-making was rare, which is the distribution the first interregnum had already produced and which commencement did nothing by itself to change. The later record shows movement without completion. A tracking report by Grant Thornton Bharat published in September 2026 counted Andhra Pradesh, Arunachal Pradesh, Bihar, Dadra and Nagar Haveli, Gujarat, Lakshadweep and Puducherry as having final rules under all four codes, with Uttar Pradesh, Rajasthan, the Andaman and Nicobar Islands and Punjab having finalised rules under one or more codes, and large industrial states still at the draft stage for much of their rulebook. The schemes layer of the social security code advanced on its own track: on 29 June 2026 the Central Government notified the Employees’ Provident Fund Scheme, 2026, the Employees’ Pension Scheme, 2026 and the Employees’ Deposit Linked Insurance Scheme, 2026 under the Code on Social Security, replacing the schemes of 1952 vintage through which provident fund law had actually been lived. And in 2026 the government introduced an Industrial Relations Code (Amendment) Bill whose object was to put beyond argument the repeal and savings position under Section 104 of that code, a legislative footnote that tells its own story: two years of transition had generated enough doubt about the interaction of repeal and savings to make a clarifying amendment worth Parliament’s time.

The texture of the second interregnum at ground level was clarification. A commenced code without rules generates questions faster than any ministry can answer them, and the months after November 2025 produced the predictable instruments: a frequently-asked-questions document from the Ministry of Labour and Employment dated 16 March 2026 addressing the reach of the coming Central Rules, circulars from the insurance and provident fund administrations on the wage definitions their collections would apply, and, recorded in the industry representations of December 2025, conflicting circulars on the revised insurance wage position that employers were left to reconcile for themselves. Each clarification settled a point and advertised the unsettled remainder. This is the administrative signature of the fifth mode of the taxonomy, and it distinguishes the second interregnum from the first in kind: the first was a silence, in which the governing law was old and known, while the second was a noise, in which the governing law was new, partially specified, and expounded in instalments by the authorities charged with operating it. Noise is progress of a sort, since it is produced by a system in motion. It is also a cost, borne in professional fees, payroll rework and deferred decisions by every establishment that had to act before the instalments finished arriving.

The professional record of the second interregnum is the employer-side mirror of the table above. The Institute of Chartered Accountants of India advised auditors that any increase in gratuity liability arising from the codes’ wage definition had to be recognised in the profit and loss account from the quarter of commencement, even though the detailed rules were pending, which is accounting’s way of saying that the codes’ consequences were real before their machinery existed. The Confederation of Indian Industry, having welcomed commencement in November 2025, was by the end of December 2025 urging the Labour Ministry to avoid retrospective application of the new provisions, flagging confusion over revised insurance wage definitions and conflicting circulars issued after commencement, and seeking clarity on gratuity, on leave encashment and on the uniformity of central and state rules. Anshul Jain, national leader for regulatory affairs at PwC India, put the practitioner position in one sentence in the same period: firms still lacked clarity on gratuity calculations, on the treatment of service before 21 November 2025, and on the inclusion of stock options and variable pay, because the key central and state rules remained unnotified. Between the government’s announcement that full implementation was complete and the profession’s statement that the operative questions were unanswered lies the whole of the second interregnum, and it is measured not in rhetoric but in the distance between 21 November 2025 and the day a given establishment’s rulebook, in its state, under its code, finally closed. For establishments in the central sphere that distance was just under six months. For the rest, it is counted in the states’ gazettes, which continued to fill through 2026.

What did industry ask for after the welcome?

Within weeks, industry bodies were asking the Labour Ministry for what commencement had not supplied: assurance against retrospective application of the provisions whose cost turned on the wage definition, clarity on gratuity and leave encashment for service already rendered, resolution of conflicting circulars on insurance wage ceilings, and uniformity between central and state rules.

The Confederation of Indian Industry’s December 2025 representations and the Institute of Chartered Accountants’ advice that gratuity liabilities be recognised from the quarter of commencement together show a regulated community pricing a statute whose text it had and whose machinery it awaited.

The sequence, welcome followed by clarification-seeking, is sometimes reported as inconsistency. It is better read as the two halves of the employer’s real position, which the commencement separated in time. The codes as policy, consolidation, thresholds, a single wage definition, digitised compliance, were what industry had sought for years and welcomed sincerely. The codes as an operative event, arriving on a Friday without rules, with payroll consequences that accountants had to provide for in the same quarter and lawyers could not fully map, were a different experience. Both experiences were produced by the same decision: to commence first and complete afterwards. A commencement that had followed the Central Rules by even the five months that separated them in the event would have denied industry its complaint and the unions their strongest procedural argument, at the price of a delay the government had, by November 2025, evidently decided it would no longer pay.

The schemes layer and the missing conference table

One stratum of the codes’ machinery deserves separate treatment because it is the stratum on which the codes’ largest promise rests, and because it is the stratum where the second interregnum runs deepest. The Code on Social Security is built around schemes. For the provident fund, insurance and pension architecture, the code continues the great statutory corporations and replaces their founding statutes, and the operative content, contribution mechanics, benefit scales, eligibility and administration, lives in schemes framed under the code. For the unorganised, gig and platform workforce, the code goes further: it promises benefits that did not exist before, to workers it defines for the first time, financed partly by aggregator contributions it fixes as a band, and the whole of that promise is scheme-dependent. Nothing in the code itself pays anyone anything. The 29 June 2026 notification of the Employees’ Provident Fund Scheme, 2026, the Employees’ Pension Scheme, 2026 and the Employees’ Deposit Linked Insurance Scheme, 2026 therefore mattered more, in lived terms, than the commencement of the parent code seven months earlier: it was the point at which the largest single block of social security law acquired operative content under the new dispensation. The schemes for gig and platform workers stand on a slower track still, entangled with the state-level gig worker statutes that several states enacted during the first interregnum, so that the newest category of worker in Indian law is governed by a double wait, for a central scheme and for the reconciliation of state schemes with it. A reader measuring the implementation of the labour codes by their commencement date will miss all of this. The codes are being implemented in layers, text first, rules second, schemes third, and the layers are separated by months and years rather than days.

The consultation record belongs in the same section, because the argument about it was, at bottom, an argument about which layer counted. The government’s case across both interregnums was that consultation was extensive: at the drafting of the codes, through the draft rules exercises of 2020 and 2025, and in the conferences of state labour ministers that punctuated the period, including the two-day national conference of labour and industry ministers and secretaries held in New Delhi on 11 and 12 November 2025. The unions’ counter-case fastened on the institution that was missing. The Indian Labour Conference, the apex tripartite body of Indian labour administration, has not been convened since 2015, and the joint platform’s demand, repeated across the period and restated after commencement, was that the codes and their rules be placed before that conference before implementation, a demand the government never met. Both statements are accurate, and the space between them is the real finding. Consultation in the form of circulated drafts, invited comments and ministerial conferences occurred repeatedly and on the record. Consultation in the form the tripartite tradition regards as authoritative, the standing conference at which government, employers and workers confront the same text in the same room, did not occur at all across the entire life of the codes from drafting to commencement. A government is entitled to prefer the first form. It is not entitled to have the preference pass unnoticed, and the unions’ procedural complaint, that commencement was unilateral, draws its force from that absence rather than from any shortage of paperwork. The complaint should be recorded at that strength: not that nobody was consulted about anything, but that the one forum whose consent would have constituted the settlement the codes needed was never called into session.

The stakes: why commencement could not be treated as a formality

It might be asked why any of this mattered enough to fight over, given that the codes’ text had been public since 2019 and 2020 and every consequence had been debated for years. The answer is that commencement is the moment abstract consequences acquire dates, and several of the codes’ consequences were large enough that their dating was itself a major economic event. The substantive comparison between the codes and the twenty-nine statutes they replaced belongs to the comparison of the labour codes with the old laws, which owns it; what belongs here is the subset of consequences whose weight explains both the five years of hesitation and the shock of the single day.

The heaviest is the definition of wages. All four codes share a single definition under which the core components, basic pay and dearness allowance and retaining allowance, are measured against a ceiling on exclusions: where the excluded components, house rent allowance, employer contributions, overtime, commission and the rest, exceed one half of total remuneration, the excess is treated as wages. Because provident fund contributions, gratuity, bonus and compensation are computed on wages, the definition moves money in every payroll in the country. Establishments that had structured salaries with a small basic component and a large allowance component faced higher provident fund outgo, higher gratuity accruals measured over the whole of an employee’s service, and a lower take-home figure for the same gross cost, unless gross pay itself was restructured. That is why the question of the day was never whether the definition was wise, it had been enacted for six years, but when it would begin to bite, on which pay cycles, with what treatment of past service, and under rules that existed only as drafts. The Institute of Chartered Accountants’ instruction to recognise the gratuity effect from the quarter of commencement shows the bite arriving through accounting standards before it arrived through any rule.

The thresholds carried the second weight. The Industrial Relations Code raised the threshold for certified standing orders and for prior government permission for layoff, retrenchment and closure from one hundred workers to three hundred. The occupational safety code raised the factory threshold from ten and twenty workers to twenty and forty, depending on the use of power, and raised the contract labour threshold to fifty. Each number redraws the map of which establishments carry which obligations, and each had been frozen in argument for five years: unions treated the three hundred figure as the codes’ central act of deregulation, since establishments below it would separate workers without prior permission, and industry treated the old figures as the central absurdity of the old regime, thresholds set for the economy of the 1970s policed into the economy of platform work. Commencement turned both arguments from positions into facts on a single date. The same is true, in the other direction, of the codes’ extensions: the statutory right to an appointment letter, the universalisation of minimum wages beyond scheduled employments, the recognition of fixed term employment with gratuity after one year, and the first statutory definitions of gig work and platform work with a fund and schemes to follow. Extensions of that kind are only promises until the machinery exists, and their commencement without schemes is precisely the gap the second interregnum names. A gig worker’s statutory existence dates from 21 November 2025. The benefit that existence was to carry awaited the scheme, and the distance between those two events is the distance this article exists to measure.

The payroll month that straddled two regimes

The abstraction of commencement becomes concrete in a single administrative fact: salaries in India are computed by the month, and the codes commenced on the twenty-first day of one. November 2025 was therefore paid, in every establishment in the country, across a seam, with part of the month governed in expectation by the old statutes and the whole of it governed in law, from the twenty-first, by the codes and their new definition of wages. The questions that reached advisers in the days after commencement were correspondingly granular. Did the new wage definition govern the November payroll or the December one? Were provident fund contributions for the month to be computed on the old basic or the restructured one? What was to be done about appointment letters, which the codes require for every employee, for a workforce that had never been issued them? The occupational safety code and the commencement statements supplied the transitional answers in general form: existing arrangements continued, appointment letters for existing staff were to follow within the period the transition allowed, and contribution practice ran on until the schemes and rules displaced it. But the granularity is the point. A commencement that lands mid-cycle on a Friday does not merely change the law; it changes the law underneath transactions already half-completed, and the administrative cost of the manner of commencement, separate from the cost of the reform itself, was paid in that month in every payroll office in the country. A government is entitled to decide that the cost is worth paying to end a five-year wait. The record of November 2025 does not show the cost being counted in advance, and the five years of preparation that preceded the day could have produced, at trivial expense, a commencement timed to a month-end and a quarter-end with the payroll profession briefed to receive it. That it did not is the smallest and most telling evidence that the decision to commence was taken as a political act first and an administrative act second.

The floor wage that waited with everything else

The floor wage is the single provision whose biography best exposes the structure of the delay, because it was the one piece of the codes the government actually began to operate during the first interregnum and then stopped. The Code on Wages replaces the patchwork of scheduled employments with a universal minimum wage and sets beneath it a national floor wage, to be fixed by the Central Government on the advice of the Central Advisory Board, below which no state minimum may fall. The fragment of the code commenced on 18 December 2020 was chosen for a reason: it brought into force the provisions constituting that advisory machinery, and the Code on Wages (Central Advisory Board) Rules, 2021 followed, so that the board through which the first floor wage would be advised existed in law from the first year of the interregnum. There the motion stopped. No floor wage was fixed under the code during the five years, because the operative wage provisions of the code were not in force and a floor fixed under them would have had nothing to govern. In the meantime the floor level under the old machinery continued its separate existence: the government told Parliament in February 2024, in answer to a question on the subject, of a floor level of Rs 176 a day effective from 1 June 2017 under the pre-code arrangements, a figure from a different statutory world doing duty while the code’s own floor waited for the code.

The episode is small and it generalises. Commencing the advisory fragment allowed the government to say, truthfully, that parts of the Code on Wages were in force, and the statement appeared in official accounts of the codes’ progress across the period. But the fragment was chosen precisely because it obligated nobody: a board can be constituted, can meet and can advise without any employer’s wage bill moving by a rupee. The pattern is worth naming because it recurs in Indian implementation practice and because readers of official progress reports need a defence against it. Partial commencement is a real technique with real uses, and it is also the readiest instrument for converting a stalled statute into a statistic of motion. The corrective question is never whether provisions have been commenced but which provisions, and what obligation in the world changed when they were. Asked of the labour codes at any point between 2021 and 2025, that question returns the same answer: none, and the floor wage, the codes’ most universal promise, waited with everything else for a November day five years on.

Pending matters, accrued rights and the savings bridge

For the advocate and the litigant, the practical question of the two interregnums is narrower and harder than the politics: what happens to a matter that began under a repealed statute? The question arises in every statutory replacement, and the codes answer it the way Indian law standardly answers it, in three layers. The first layer is the codes’ own savings provisions, the repeal and savings sections already identified, which provide in substance that the repeal does not affect rights, privileges, obligations or liabilities acquired or incurred under the repealed enactments, that proceedings and remedies in respect of them may be continued and enforced as if the codes had not been enacted, and that things done under the old law, registrations effected, licences granted, standing orders certified, settlements and awards in force, are deemed to have been done under the corresponding provisions of the codes. The second layer is Section 6 of the General Clauses Act, 1897, the standing savings rule of Indian statutory interpretation, which protects accrued rights and pending proceedings against repeal unless a contrary intention appears. The third layer is Section 24 of the same Act, the provision on which the second interregnum’s strangeness legally rests: where an enactment is repealed and re-enacted, rules and orders made under the repealed enactment continue in force, so far as not inconsistent with the re-enacted provisions, and are deemed to have been made under them. The codes’ commencement statements and the ministry’s parliamentary answers are applications of that section in modern dress.

Two consequences follow for a person holding a position under the old law. A dispute raised, a claim filed, an inspection conducted or a prosecution launched under a repealed statute does not fall with the statute; it continues under the savings bridge, decided by the law under which it accrued, though the forum and procedure may migrate to the codes’ machinery as that machinery is built. And a right that had not yet accrued, an expectation rather than an entitlement, receives no such protection, which is why the litigation after commencement was always going to concern the boundary between the two: service rendered before 21 November 2025 counted towards gratuity under a definition of wages that commenced on that date is the paradigm hard case, and the clarification-seeking of December 2025 shows employers identifying it within weeks. The general position stated here is information about how the savings architecture works, not advice on any individual matter, and a person with a pending claim should take advice on the specific savings provision and forum that govern it, because limitation periods and transitional rules are unforgiving in detail.

The architecture is easier to hold with three worked instances, kept at the level of principle their facts deserve. Take first a retrenchment dispute raised before a labour court in October 2025 under the Industrial Disputes Act, with pleadings complete and evidence under way when commencement arrived in November. The savings provisions preserve the proceeding: the dispute continues and is decided under the Act that created it, because the rights in contest accrued under that Act, and the repeal of the Act does not reach backwards to dissolve a lis pending under it. Take second a factory licence granted under the Factories Act for a term running past November 2025. The licence is a thing done under the repealed statute and is deemed, under the savings and corresponding-provisions doctrine, to have been granted under the occupational safety code, holding good for its term while renewals and fresh applications migrate to the code’s machinery as the applicable rules arrive. Take third the hardest of the three, a gratuity claim by an employee who retires in 2026 with twenty years of service behind her, the first nineteen rendered under the Payment of Gratuity Act and the last under the Code on Social Security, with the wage definition changed between the two. Her entitlement accrued year by year, but its quantification turns on the definition of wages in force at the legally relevant point, and the interaction of accrual, repeal and a redefined measure is exactly the question employers were putting to the ministry in December 2025 and the question on which rule-makers and courts, and not commencement statements, would have the last word. The three instances together mark the boundary of what commencement settles. It settles which statute book is open. It does not settle, and cannot settle, the cases that straddle the closing of the old one.

On the one question of commencement law that the brief for this article flags, the position is settled and short. There is no direct litigation that can force a commencement, because the appointed day is committed to the executive. When the Supreme Court was asked, in A. K. Roy v Union of India, to direct the government to bring into force a provision whose commencement had been withheld, the Court declined to issue such a direction, treating the timing of commencement as a matter for the government under the clause Parliament had enacted. That holding, arising from the withheld provision of the Forty-fourth Amendment discussed in the next section, is the reason commentary asking whether the codes’ commencement, or their earlier non-commencement, could be challenged in court receives the honest answer that commencement itself is not readily justiciable. What courts can and will police is the operation of the codes once in force, the consistency of carried-forward rules with the codes’ text, and the treatment of accrued rights at the boundary. The delay itself was never going to be ended by a judge. It was ended, as it was always going to be ended, by a signature.

Which commencement wait in this series was the longer one?

The withheld definition of industry in the Industrial Disputes (Amendment) Act, 1982 waited forty-four years and was never commenced at all, its parent Act being repealed first. The labour codes waited five years and two months from the assent of the three codes of 2020: shorter in duration, but far larger in the scale of what was held in postponement.

The comparison also separates two kinds of patience. The 1982 clause waited because a specific piece of machinery, a grievance forum for excluded institutions, was missing and identifiable. The codes waited although the missing machinery was the entire operative apparatus of four statutes, a difference of kind disguised as a difference of degree.

A taxonomy of legislative inertia

This article completes a classification that the series has been assembling cluster by cluster, and it is worth setting out whole, because the labour codes take their full meaning only inside it. Indian statutes fail to operate in distinguishable ways. The first mode is enactment without notification: the statute is passed and assented to and the appointed day never comes at all, as with the Whistle Blowers Protection Act, 2014, enacted after a decade of campaigning and never brought into force, and as with Section 3 of the Forty-fourth Amendment Act, 1978, which was to have rewritten the preventive detention safeguards of Article 22 and remains uncommenced across four and a half decades. The second mode is notification in fragments: commencement exercised provision by provision so that a statute exists in instalments, a mode the labour codes themselves exhibited in miniature in December 2020 and May 2023 and which staged commencement clauses make permanently available. The third mode is delay by institutional dependency: the statute commences, but the body it creates is constituted late, staffed thinly or left without its head, so the law operates at a fraction of its design. The fourth mode is commencement without machinery of the administrative kind: rules unmade, forms unprescribed, funds unconstituted, the statute in force and its promise undeliverable in practice.

The cluster’s own archive supplies a fifth illustration inside the first mode, and it is the closest relative of the codes’ story. When the Industrial Disputes (Amendment) Act, 1982 was brought into force on 21 August 1984, one clause was withheld: clause (c) of its second section, which substituted a new definition of industry. The explanation given by the government, and recorded in the article on the Industrial Disputes Act, was that the categories the new definition excluded, hospitals, educational and charitable institutions among them, could not be left without dispute machinery, and that the definition would be commenced together with a separate law providing grievance machinery for the excluded institutions. The companion Bill was introduced in 1988 and dropped, the machinery never arrived, and the definition was never notified in the forty-four years that followed, so that the courts went on applying the earlier definition as construed in the Bangalore Water Supply case while the narrower enacted text sat in the statute book without legal operation. The parallel with the labour codes is exact in structure and opposite in outcome. In 1982 a provision waited for machinery that was never built, and the wait consumed the provision. In 2025 four codes waited for machinery that was largely unbuilt, and the wait was ended by commencing anyway, machinery to follow. The labour codes are the fifth mode of the taxonomy, and its inversion of the first: notified before the machinery was ready, producing a period in which the statute is law and the law is not yet usable.

The third and fourth modes repay a closer look inside the codes’ own record, because the codes exhibit both in subordinate parts even though their headline story is the fifth mode. Institutional dependency, the third mode, was largely avoided in this reform for a reason worth recording: the codes did not create their delivery institutions from nothing. The provident fund and insurance corporations continued from the old regime under new statutory authority, the inspectorates and labour departments of the states carried over, and the advisory boards were continuations of bodies the old law knew. Where a code did depend on a body to be built, the gig and platform worker boards and funds of the social security code, the dependency reappeared in its classic form and the corresponding promise is the slowest in the entire reform to acquire machinery. The fourth mode, commencement without administrative machinery, is the second interregnum itself described from the administrator’s side: registers that did not exist in the new forms, returns that could not yet be filed in the new formats, thresholds that applied to establishments whose registration category the new rules had not yet defined. The modes are not alternatives between which a statute chooses. They are layers, and a statute of sufficient size will exhibit several at once in different provisions, which is why single-cause accounts of implementation failure, blaming drafting, or administration, or politics, each capture one layer and misdescribe the rest.

The drafting lesson the taxonomy teaches is addressed to the legislator and the staffer, and it is sharper than the usual moral that governments should hurry. Every mode in the classification is produced by the same initial decision, taken at drafting, about where a statute’s operative content will live. Content placed in the Act itself commences with the Act or not at all, and its postponement is visible and politically priced. Content delegated to rules, schemes and a body to be constituted can be postponed invisibly, in fragments, by omission, at no parliamentary cost, which is why the modern Indian statute, drafted for flexibility, is structurally prone to every mode in the taxonomy at once. The labour codes are not an aberration of implementation. They are the predictable biography of a statute that delegated its operative life to two levels of government and thirty-seven rule-makers, and the predictability is the lesson. A Parliament that wishes a reform to operate on a date should write the operative detail into the Act, fix commencement in the Act itself or subject the appointed day to a statutory outer limit, and treat every delegation as a decision about timing and not merely about convenience. Nothing in the Indian drafting tradition requires any of that, which is why the taxonomy has five modes and counting.

Was commencing before the rules were ready the right call?

The strongest version of the case against the November 2025 commencement is the one the joint platform of trade unions made and the one the practitioners’ record supports from the other side of the table. A government that had justified five years of delay by the need for rules, consultation and uniformity commenced the codes on a day’s notice, without the rules, over the stated opposition of the organised workforce, and with state readiness at roughly the position it had occupied when readiness was the reason for waiting. On that record, the delay purchased nothing: not preparedness, which the December 2025 draft rules exercise admitted had to be redone, not consent, which the November protests recorded as absent, and not uniformity, which the state rule position of 2026 continued to withhold. If the codes could be commenced in that condition in November 2025, they could have been commenced in that condition in 2022, and three years of the delay stand unexplained by any reason the government itself offered.

The strongest version of the case for the commencement is not a defence of the five years, and it should not be confused with one. It is a claim about 21 November 2025 itself. By that date the pattern of the first interregnum was established beyond argument: drafts aged and went stale, consultation produced positions rather than convergence, states finalised rules only under the pressure of an actual commencement to prepare for, and every passing year reset the readiness argument without advancing it. On that pattern, making commencement wait for completed rules was not a plan but a formula for the Whistle Blowers outcome, enactment as a permanent condition. Commencement changed the incentives of every actor at once. States that had drafted at leisure acquired a repealed statute book and a live code to operate; industry that had priced the codes as a contingency had to price them as law; the ministry itself was driven to produce draft rules in five weeks and final Central Rules in under six months, a pace the previous five years had never approached. And the feared legal vacuum did not occur, because the savings and continuity arrangements held the floor: old rules governed where new ones were absent, accrued rights survived, and the composite law of the second interregnum, untidy as it is, is a functioning law and not an absence of law. On this reading, commencing with rules pending was not a blunder interrupting a preparation that would otherwise have completed. It was the only step in the entire sequence that caused preparation to complete.

One further consideration belongs to the case for the exit, because it is the one the government’s defenders stated least clearly and it may be the strongest. Commencement is a disciplining event for the state itself. An uncommenced code imposes no duties on the administration that sponsors it: no authorities need be staffed for it, no portals built for it, no questions in Parliament answered about its working. A commenced code generates all of those pressures at once, and it generates them in a form, unanswered questions, unfiled returns, unbuilt machinery under a live statute, that a bureaucracy experiences as unfinished business rather than as policy. The acceleration of everything after 21 November 2025, the draft rules in five weeks against five years of drift, the Central Rules in under six months against drafts that had aged since 2020, the clarifying amendment introduced within the following year, is consistent with that discipline operating. It does not follow that commencing unprepared is a general technique to be recommended. It follows that where a statute has entered the condition of permanent preparation, the preparation has usually become the obstacle, and only an event that changes the costs of incompletion will end it. November 2025 was such an event, and the second interregnum’s untidiness is what the ending of a five-year postponement looks like when it is done by the only instrument that was ever going to end it.

Both cases are true, and they are true about different things, which is why the argument about the labour codes so often passes itself by. The case against is a case about the delay: five years in which the reasons given did not produce the thing they named, and in which the governing technique was postponement without deadline. The case for is a case about the exit from the delay: that once a reform has entered the posture of permanent preparation, commencement first and completion after is a rational, even a necessary, way out, and that the second interregnum, for all its strangeness, is a better legal condition than the first, being shorter, dated, and driven by machinery visibly under construction rather than by drift. The verdict this article defends joins the two. The five-year delay was longer than its federal explanation supports, and the manner of the exit, without notice after years of imminence, imposed avoidable costs on the employers and workers who had to absorb a new governing statute over a weekend. But the exit itself, commencement ahead of the rules under the protection of the savings bridge, was defensible and probably correct, because the alternative on offer was not a better-prepared commencement. It was, on the whole of the record, no commencement at all.

What the episode predicts for the next consolidation

A series that reads statutes as records should end an implementation account with the prediction the record supports, and this record supports a specific one. The consolidation technique, many statutes into few codes, is the Indian state’s preferred instrument of legal modernisation, and it will be used again in fields beyond labour. The labour codes’ biography predicts that any such consolidation which delegates its operative content to rules by two levels of government, on a subject either level can slow, will reproduce the two interregnums in some proportion: a wait between enactment and commencement whose length is set by politics, followed by a wait between commencement and operability whose length is set by rule-making capacity, with the savings architecture carrying the citizen across both. The proportions can be improved at the drafting stage, by writing operative detail into the code, by fixing a default commencement date in the text, and by building the schemes and portals before assent rather than after it. They cannot be abolished by exhortation, because they are produced by the structure of delegation and not by the failings of the individuals who administered this instance. Drafters of the next consolidation who read the labour codes’ record carefully will therefore spend less effort promising speed and more effort deciding, provision by provision, what must be law on the first day and what can honestly wait. That decision, taken in the drafting room, is where the implementation of a statute is in fact determined. Everything after it, including five years of ministerial statements and one November Friday, is the working out of consequences the drafting had already chosen.

The comparative check: a commencement that was planned in the open

The taxonomy gains from one comparison outside the labour field, because it isolates the variable that decided the codes’ fate. The three criminal law codes of 2023, which replaced the penal code, the procedure code and the evidence Act, were enacted with appointed-day clauses of the same family and commenced on 1 July 2024, roughly six months after assent, on a date announced in advance and prepared in public, with rules, training and systems work run against the deadline and the deadline kept. Whatever else is said of that reform, its commencement was an administrative campaign with a named day. The labour codes, with an identical legal device in their first sections, received five years and a surprise. The difference between the two records is not the clause. It is that the criminal codes’ machinery sat within a single level of government and a unified administrative command, the police, prosecution and courts of the states working to a date the Centre had publicly committed to, while the labour codes’ machinery sat in thirty-seven governments whose commitment was never publicly extracted. The comparison therefore supports the federal explanation of the labour delay at the level of structure, and undermines it at the level of excuse: intergovernmental machinery is a reason a commencement is hard, and the criminal codes show that hardness is overcome by the one input the labour codes never received until November 2025, a date somebody was willing to be held to.

What applies in practice: a reader’s map

For the employer and the adviser, the practical position through the second interregnum can be stated as four working rules, each anchored in the record set out above. The codes are the governing law: their definitions, thresholds, prohibitions and entitlements commenced on 21 November 2025, and any analysis that treats the old statutes as still governing in their own right is wrong. The procedure runs on carried-forward law: forms, registers, returns, wage-fixing machinery and inspection practice continue under the rules of the repealed statutes so far as those rules are not inconsistent with the codes, and the inconsistency question, where it arises, is answered by the codes’ text and not by preference. The rulebook that binds a given establishment is its appropriate government’s: the Central Rules of 8 May 2026 govern the central sphere, and an establishment in a state is governed by that state’s rules where they exist in final form, with the old rules filling the remainder. And preparation should be organised state by state and file by file, keeping the commencement notifications, the corrigendum, the savings provisions and the applicable rules in a single verifiable record for each establishment, the same discipline of one complete file of record that leads journalists to keep sensitive material in an encrypted workspace rather than in scattered notes. Multi-state employers face the patchwork in its full form and should expect it to persist: the state rule phase has no single end date, and each state’s final rules will arrive as a separate small commencement of their own.

For the worker and the union, the map is the mirror image. Rights conferred by the codes’ text, the appointment letter, the universal minimum wage entitlement, the definitions that widen coverage, are law and can be invoked; benefits that depend on schemes, above all the social security architecture for gig and platform workers, await the scheme and should be monitored in the gazettes rather than assumed. Claims and disputes that accrued under the repealed statutes survive on the savings bridge and are pursued under the law that created them. For the researcher, the file this article has assembled, assent dates, commencement notifications and their numbers, the corrigendum, the two rounds of draft rules, the parliamentary answers of July 2022, July 2025 and July 2026, is the skeleton of the commencement record, and the gaps in it are data: the absence of any announced reason for the timing of November 2025 is itself a finding about how commencement decisions are communicated. For the legislator, the lesson stands in the taxonomy section. Commencement is a stage of lawmaking, and a statute that does not plan for it has not finished being drafted.

The series’ thesis thread can be closed in one sentence. A statute is not its text alone but the whole record around it, and in the case of the labour codes that record includes five years in which the text was not law, a day on which it became law without machinery, and a continuing period in which the machinery is being built underneath it; the reader who holds all three periods in view at once knows what the labour codes are, which is something neither the bare text nor the announcement of their arrival can tell.

Frequently Asked Questions

Q: Why did the labour codes take five years to come into force?

The delay divides into phases with different explanations. The pandemic accounts for the first phase, in which the administrative system was managing a lockdown and a recovery and the disturbance of wage structures was a cost no government would schedule. The framing of rules accounts for the second phase, in which the Centre pre-published draft rules in 2020 and the states followed at their own pace. The final and longest phase is explained by the interaction of sustained trade union opposition, reluctance in several states to finalise rules, a general election in 2024 and assembly elections across the period, and the absence of any deadline the government was bound to meet. The federal structure explains why commencement could not be quick. It does not explain why it took five years, and the government’s own statements located the obstacle in the states’ rule-making lag.

Q: When did the labour codes actually take effect?

The four labour codes came into force on 21 November 2025. The Code on Wages, 2019 had received presidential assent on 8 August 2019 and the Industrial Relations Code, the Code on Social Security and the Occupational Safety, Health and Working Conditions Code on 28 September 2020, but none of those dates made the codes operative law. Commencement waited on notifications under each code’s appointed-day clause, and the Ministry of Labour and Employment issued the four commencement notifications, numbered S.O. 5319(E) to S.O. 5322(E), on 21 November 2025, announcing the commencement by press release the same day. Two small fragments had been commenced earlier, provisions of the Code on Wages on 18 December 2020 and provisions of the Code on Social Security on 3 May 2023. The Central Rules under the codes followed separately, on 8 May 2026.

Q: Why do the labour codes still not fully operate?

Because the codes delegate their operative detail to subordinate legislation that is incomplete. Each code leaves its machinery, forms, registers, fixation procedures, licensing, inspection schemes and benefit schemes, to rules, regulations and schemes made by the appropriate government. For most establishments the appropriate government is the state government, and state rules under the four codes remain incomplete, finalised under all four codes in only a minority of states and under individual codes in others. The Central Rules notified on 8 May 2026 complete only the central sphere, covering establishments such as railways, mines, banking, insurance and central public sector undertakings. Until a state’s rules exist, the codes operate in that state through their own text supplemented by the rules framed under the repealed statutes, which continue to apply so far as they are not inconsistent with the codes.

Q: Which states had labour code rules ready at commencement?

Almost none in final form. The careful contemporary count, published by Khaitan and Co in November 2025 from the position at commencement, recorded that draft rules had been pre-published nearly everywhere: only two of the thirty-six states and union territories had yet to publish draft rules under the Code on Wages and the occupational safety code, three under the Industrial Relations Code and one under the Code on Social Security. Final rules were a different matter. They had been released only under some of the codes and only in a small number of states, with Gujarat, Karnataka, Uttar Pradesh and Mizoram named among them. Everywhere else the position at commencement was drafts at various stages of consultation. The distribution, drafts nearly universal and finals rare, is the single most informative fact about the first five years, because it shows governments willing to be seen preparing and unwilling to be first to operate.

Q: What applies while the labour code rules are awaited?

A composite of the codes and the carried-forward law of the repealed statutes. The codes themselves govern: their definitions, thresholds, prohibitions and entitlements are in force, and provisions that need no rule to operate apply directly. Procedure runs on continuity: the rules, regulations, notifications, standards and schemes made under the repealed statutes continue to operate so far as they are not inconsistent with the codes, under the codes’ savings provisions and Section 24 of the General Clauses Act, 1897. The Ministry of Labour and Employment stated the position to the Lok Sabha on 27 July 2026 in those terms, adding that the arrangement ensures continuity and avoids a legal vacuum during the transition. Where a code provision and a carried-forward rule conflict, the code prevails. Where an establishment’s state has notified final rules, those rules displace the carried-forward ones to the extent they cover the same ground.

Q: Did trade unions oppose the labour codes?

Yes, in an organised and sustained campaign. The joint platform of ten central trade unions resisted the codes from enactment, with protests after the Code on Wages in 2019 culminating in a general strike in January 2020, a general strike on 26 November 2020 alongside the farmers’ movement, a two-day strike in March 2022, further joint actions in February 2024 and a general strike on 9 July 2025 demanding that the codes be scrapped. On commencement day, 21 November 2025, the platform condemned what it called the blatantly unilateral implementation of anti-worker, pro-employer codes, described the notification as a deceptive fraud on the working people of the nation, demanded a complete rollback and called for nationwide resistance on 26 November 2025. The Bharatiya Mazdoor Sangh stood outside the platform: it welcomed the Code on Wages and the Code on Social Security while strongly opposing provisions of the Industrial Relations Code and the occupational safety code.

Q: Does the Concurrent List explain the labour code delay?

Partly, and the partition matters. Labour sits in Entries 22, 23 and 24 of the Concurrent List, and the codes vest rule-making in the appropriate government for each establishment, which for most workplaces is the state. A reform of that design genuinely cannot be operated from the Centre alone, and the need to carry over thirty state rule-makers explains why commencement was a matter of years rather than months and why uniformity was a real objective rather than a slogan. What the Concurrent List does not explain is the distribution of the outcome: draft rules nearly everywhere and final rules almost nowhere for most of the period, and a commencement that, when it came, did not wait for the uniformity the federal argument required. The List explains the shape of the delay. Its length is explained by opposition, elections and the absence of a binding deadline, which are political facts, not constitutional ones.

Q: What does the labour code delay teach about Indian lawmaking?

Three lessons stand out. First, commencement and operability are different events, and a statute that delegates its operative content has only begun its journey at enactment: the appointed-day clause in a modern Indian Act is a discretion exercisable by silence, without an outer date, reasons or review, and it can hold a reform for years at no parliamentary cost. Second, delegation is a decision about timing and not merely about convenience, because content placed in rules, schemes and bodies to be constituted can be postponed invisibly and in fragments, while content placed in the Act itself cannot. Third, the savings and continuity arrangements are the system’s real transition machinery, the device that lets an old regime’s rules operate a new regime’s text, and any reader who wants to know what law governs at a moment of statutory change should read the savings provisions before the speeches. The labour codes exhibit all three lessons at the largest scale in the statute book.

Q: Was the commencement of the labour codes announced in advance?

No. The commencement was announced by a press release of the Ministry of Labour and Employment on 21 November 2025, the same day as the four gazette notifications that effected it, and employers received no lead time in which to prepare payroll systems, salary structures or registrations for a named date. The law firm JSA, in its client analysis of 27 November 2025, described the step as taken through an unforeseen announcement by way of a press release, and practitioner commentary converged on the same description. The absence of notice is the sharper because it followed five years in which commencement had repeatedly been described as imminent and conditional on readiness: the regulated community had been told to expect a prepared landing and received instead a same-day arrival, with the codes in force from the Friday of the announcement. No reason for the choice of date was given in the official record.

Q: What is the difference between a statute being enacted and a statute being in force?

Enactment is the completion of the legislative process: passage by both Houses of Parliament and the assent of the President, after which the text is an Act with a number and a place in the statute book. Being in force is a further condition, operation: the point from which the Act’s provisions impose obligations and confer rights that courts and authorities will enforce. Under Section 5 of the General Clauses Act, 1897, an Act that says nothing about commencement comes into operation on the day of assent, so the two events coincide by default. Most modern Indian Acts displace the default with an appointed-day clause, under which the government notifies the date, and may notify different dates for different provisions. The labour codes were enacted in 2019 and 2020 and in force only from 21 November 2025, and the repeal of the twenty-nine statutes they replaced was tied to the same event, so enactment alone changed no one’s legal position for over five years.

Q: Which parts of the labour codes were already in force before 21 November 2025?

Two small fragments. On 18 December 2020 the Central Government brought into force a set of provisions of the Code on Wages concerned principally with the Central Advisory Board, the tripartite body through which the floor wage is advised upon, and followed it with the Code on Wages (Central Advisory Board) Rules, 2021. On 3 May 2023 a set of provisions of the Code on Social Security was similarly notified. Neither fragment touched the obligations of an ordinary establishment: no minimum wage or floor wage was fixed under the codes during the five years, no standing order was certified under the Industrial Relations Code, no licence issued under the occupational safety code and no gig worker scheme framed under the Code on Social Security. The fragments are nevertheless part of the record, because official accounts of the codes’ progress cited them, and they illustrate how partial commencement can produce the appearance of motion without altering any duty in the world.

Q: Why were fresh draft Central Rules published in December 2025 when draft rules existed from 2020?

Because the government treated the earlier drafts as spent. The Central Government had pre-published draft rules under all four codes in the second half of 2020, and states had built their own draft exercises on that round. When fresh draft Central Rules were pre-published on 30 December 2025, thirty days being allowed for comment on the Industrial Relations Code draft and forty-five days on the other three, the Union Labour and Employment Minister, Mansukh Mandaviya, explained the repetition in terms: the central government and the states had pre-published draft rules earlier, but that was a long time ago, and there was a need to bring the draft rules again in sync with present times. The explanation is candid about the first five years. Drafts that are neither finalised nor refreshed become unusable as the surrounding law, administration and politics move, so the second interregnum began by redoing work the first interregnum had claimed as preparation.

Q: Why did the Central Rules notified on 8 May 2026 not finish the job of implementing the codes?

Because of the appropriate government division written into the codes. The Central Rules bind only where the Central Government is the appropriate government for the establishment, a sphere the Ministry of Labour and Employment’s frequently-asked-questions document of 16 March 2026 identified with railways, mines, banking and insurance and central public sector undertakings, among others. For the ordinary private establishment, a factory, shop, plantation or transport undertaking in a state, the operative rulebook is the one made by that state government, and state rules remained incomplete when the Central Rules appeared and continued to be notified piecemeal afterwards. The Central Rules also arrived with the model standing orders under the Industrial Relations Code, and an official described their publication as completing the process of full implementation. That description is accurate for the central sphere. Implementation for most workplaces is completed state by state, and the last of those completions is a separate event from 8 May 2026.

Q: What happens to registrations and licences granted under the repealed labour laws?

They survive. The codes’ repeal and savings provisions, supported by Sections 6 and 24 of the General Clauses Act, 1897, provide in substance that things done under the repealed statutes, registrations effected, licences granted, standing orders certified, notifications issued, continue to have effect and are deemed to have been done under the corresponding provisions of the codes, so far as they are not inconsistent with them. A factory licence granted under the Factories Act for a term running past 21 November 2025 therefore holds good for its term, and a registration under a repealed statute does not have to be taken afresh merely because the statute book changed. Renewals, modifications and fresh applications migrate to the codes’ machinery as the applicable rules are notified. Proceedings pending under the repealed statutes likewise continue and are decided under the law that created them, and rights and liabilities that had accrued before repeal are preserved against it.

Q: Can a court order the government to bring a statute into force?

In practice, no. The appointed-day clause commits the timing of commencement to the executive, and the Supreme Court has declined to convert that discretion into a duty a court can enforce. In A. K. Roy v Union of India, the Court was asked to direct the government to bring into force Section 3 of the Forty-fourth Amendment Act, 1978, whose commencement had been withheld, and it refused to issue such a direction, treating the appointed day as a matter for the government under the clause Parliament had enacted. That position is the reason informed commentary answers the question whether non-commencement, or commencement, of the labour codes could be challenged with the conclusion that commencement itself is not readily justiciable. What courts can police is the operation of a statute once in force: the consistency of carried-forward rules with the code’s text, the treatment of accrued rights at the boundary of repeal, and the legality of action taken under colour of the new law.

Q: Did employers have to restructure salaries on the day the labour codes commenced?

The legal obligation and the administrative machinery arrived on different schedules, and the honest answer separates them. The codes’ definition of wages, under which excluded components above one half of total remuneration are treated as wages, commenced on 21 November 2025, and its consequences for provident fund, gratuity and bonus computation date from commencement. The Institute of Chartered Accountants of India advised auditors that any increase in gratuity liability arising from the definition had to be recognised in the profit and loss account from the quarter of commencement, rules or no rules. The machinery for contribution and collection, however, ran on through the carried-forward schemes and rules until the new schemes and Central Rules displaced them, and the Confederation of Indian Industry spent December 2025 seeking assurance against retrospective application and clarity on gratuity for past service. Employers therefore had to provide for the new definition from commencement day, restructure prospectively as the rules arrived, and treat the interval as a period in which the liability was live and the paperwork transitional.

Q: Why did the government commence the codes before the rules were ready instead of waiting longer?

The government gave no reason for the timing in the official record, and any answer is a reading of the pattern rather than a report of a stated motive. The pattern supports one reading strongly. Across five years, waiting for readiness had not produced readiness: draft rules aged without being finalised, consultation restated positions without converging them, and states finalised rules only in a handful of places. Commencement reversed the incentives of every actor at once: the ministry produced draft Central Rules within five and a half weeks and final rules in under six months, states acquired a repealed statute book and a live code to administer, and the savings arrangements prevented the legal vacuum that waiting was supposed to avoid. On that reading, commencing first was the step that made completion happen, and the alternative on offer was not a better-prepared commencement at a later date but the indefinite continuation of the first interregnum. The costs of the manner, no notice and a mid-month arrival, stand separately and are not answered by the pattern.

Q: How many times were draft rules under the labour codes published for comment?

At the Centre, twice. The first round ran through the second half of 2020, the draft Code on Wages (Central) Rules appearing in July 2020 and drafts under all four codes complete by November of that year. The second round opened on 30 December 2025, after commencement, with thirty days allowed for comment on the Industrial Relations Code draft and forty-five days on the drafts under the other three codes, leading to the final Central Rules of 8 May 2026. Around those two central rounds ran the states’ own exercises: by 25 July 2022 the Ministry told Parliament that thirty-one states and union territories had pre-published draft rules under the Code on Wages, twenty-six under the Industrial Relations Code and twenty-five under each of the other two codes, and state draft rounds resumed after commencement as governments worked towards final rules. The count of rounds is itself an implementation finding: a reform that needed its central drafts published twice, five years apart, was not being prepared continuously in between.

Q: What did the corrigendum to the Social Security Code commencement notification correct?

The commencement of the Code on Social Security was effected by notification S.O. 5319(E) of 21 November 2025, which, unlike the general notifications for the industrial relations and occupational safety codes, enumerated the provisions brought into force and the parts withheld. A corrigendum dated 19 December 2025 corrected the scope of certain provisions and exclusions in that enumeration. The correction mattered because an enumerating notification is the operative text: which sections of a 164-section code are law, and which items of its repeal provision have taken effect, is determined by the list, and an error in the list is an error in the law in force until it is corrected. The episode is a small, exact emblem of the commencement as a whole. A step executed at speed, after five years of deliberation, produced an instrument that required correction within a month, in the code whose provisions most needed enumeration to be got right.

Q: What should an employer with establishments in several states do while state rules remain incomplete?

Work state by state, because the governing position is state by state. For each establishment, identify the appropriate government, apply the codes’ text as the governing law, and run procedure on the rules that bind that establishment: the Central Rules of 8 May 2026 in the central sphere, the state’s final rules where they have been notified, and the carried-forward rules of the repealed statutes for everything the new rules have not yet displaced. Keep a single verifiable file for each establishment containing the commencement notifications, the corrigendum, the savings provisions and the rules relied on, so that the basis of every payroll and separation decision can be shown. Track each state’s gazette as a separate commencement watch, since every state’s final rules arrive as a small commencement of their own and change the position in that state from the date they take effect. Expect the variation between states to persist rather than converge quickly, and structure pay, standing orders and registrations so that a change in one state’s rules does not force a rebuild in the others.