For over seven decades, the Industrial Disputes Act 1947 decided a question that sounds simple and is not: when a worker and an employer fall out, who gets to decide the quarrel, and by what rules? The statute built a complete alternative world of labour justice, with its own forums, its own vocabulary and its own arithmetic of compensation, and it placed the government of the day at the gate of that world, deciding which disputes would be adjudicated at all. Almost everything else in Indian labour law, from the size of factories to the survival of sick mills, was shaped by how this one Act answered its opening questions.

One fact about the Act is stranger than anything in its text, and it organises everything this article has to say. The single most important word in the statute is “industry”, because the whole machinery switches on only where an industry exists. In 1978 the Supreme Court gave that word a very wide meaning. In 1982 Parliament enacted a narrower meaning and voted the wider one down. The executive then declined, for forty-four years, to bring Parliament’s narrower meaning into force, and the courts went on applying the meaning Parliament had rejected. A statute whose central term was governed by a definition the legislature had disowned is a fair emblem of the Act itself: powerful, contested, and administered for most of its life in a state of suspension between what Parliament said and what governments were willing to do.

This article works through the Act as a pillar treatment should: the problem it was built to solve, the design it inherited, the definitions that carry its weight, the adjudication machinery, the protections it builds around a workman while a dispute is pending, the rules on strikes and lockouts, the compensation architecture for layoff, retrenchment and closure, the permission regime of Chapter VB that became the most argued-over provision in Indian labour law, the state variations that quietly rewrote it, and the succession by the Industrial Relations Code 2020, which inherited its machinery and raised its thresholds. It closes with the definitional litigation that began in 1978 and reached a bench of nine judges in 2026, which is where the Act’s strangest story ends.

The problem the Act was built to solve

Industrial conflict in the India of the 1940s was not an abstraction. Mills struck, workers were locked out, and the state’s inherited tools for dealing with a stoppage were the tools of public order rather than of industrial justice: the police, the magistrate and the criminal law of unlawful assembly. A strike in a cotton mill or a jute works could idle thousands, interrupt supplies the economy could not spare, and turn violent in ways that made every stoppage a problem of government and not merely of contract. The question the statute had to answer was whether the state would treat an industrial dispute as a breakdown to be policed or as a disagreement to be investigated, conciliated and, if necessary, adjudicated by a tribunal applying the statute’s own standards.

The choice embedded in the Act was for supervision over free bargaining. In the mature collective bargaining model, the state sets background rules and leaves the parties to fight within them, with the strike and the lockout as lawful weapons whose economic pressure produces settlement. The Industrial Disputes Act took a different path. It built machinery through which the state could step into a dispute, attempt conciliation, and then send the dispute to a labour court or tribunal whose award would bind the parties whether they liked it or not. The strike weapon was not abolished, but it was fenced: hedged with notice requirements in public utility services, prohibited while proceedings were pending, and liable to be displaced altogether when the government prohibited its continuance on referring the dispute. The design premise, formed in wartime and carried into independence, was that industrial peace was too important to be left to the balance of economic strength between a mill owner and a union.

That premise explains both the Act’s durability and the resentment it accumulated. Employers experienced the design as a system in which the merits of a dispute mattered less than access to the government that controlled the reference. Unions experienced it as a system that blunted their one real weapon and substituted the uncertain charity of adjudication. Workers outside the organised sector, the overwhelming majority of the Indian workforce, experienced it hardly at all, because the Act’s protections attached to the status of workman in an industry and most Indian workers had neither. A complete account of the statute has to hold all three experiences together, because each of them is a true description of the same design.

What existed before, and the wartime inheritance

The Act did not invent its machinery. The Trade Disputes Act 1929 had already created courts of inquiry and boards of conciliation and had restricted strikes in public utility services, though it stopped short of compulsory adjudication of the merits. The decisive acceleration came with the Second World War. Rule 81A of the Defence of India Rules empowered the government to refer industrial disputes to adjudication, prohibit strikes and lockouts, and enforce the resulting awards, on the ground that wartime production could not be interrupted by industrial conflict. Provincial legislation, most elaborately the Bombay Industrial Relations Act 1946, built parallel systems of recognised unions and standing orders in the great textile centres.

Between the 1929 Act and the war stands the deliberative groundwork that gave the settlement its shape. The Royal Commission on Labour, chaired by John Whitley, which reported in 1931, had examined Indian industrial conditions in the aftermath of the strikes of the 1920s and recommended a standing machinery of conciliation and a more deliberate law of disputes than the 1929 Act’s improvisations; the trade union movement it surveyed had carried legal status only since the Trade Unions Act of 1926, and the Commission’s preference for organised and responsible bargaining left a mark on every later design. The war then built the institutions the peace inherited: from 1942, the Indian Labour Conference and the Standing Labour Committee organised centre, provinces, employers and workers into a tripartite habit of consultation that outlived the war, and Rule 81A supplied the working model of compulsory reference and restrained strikes which the 1947 Act converted, with remarkably little alteration of spirit, into permanent peacetime law. The choice confronting the legislators of 1947 was therefore not between adjudication and its absence; it was between keeping the wartime machine as an ordinary statute and dismantling it in favour of unfettered collective bargaining, and the machine was kept.

When the war ended, the emergency machinery was about to lapse, and the interim government chose to keep it in permanent statutory form. The Industrial Disputes Act, enacted as Act 14 of 1947 and brought into force on 1 April 1947, converted the wartime power to refer disputes and enforce awards into the ordinary law of the land, months before independence. The genealogy matters because it explains the Act’s instincts. A statute descended from emergency production rules naturally treats the stoppage as the problem and the settlement as the solution, and naturally gives the executive the keys. The later argument about the Act, conducted for seventy years in the language of workers’ rights and managerial freedom, was at bottom an argument about whether a wartime design should have become the permanent constitution of Indian industrial relations. Parliament in 1947, legislating in the shadow of partition and scarcity, did not pause long over that question. Its successors, as this article will show, repeatedly answered the question in one direction and then declined to act on the answer.

How the Industrial Disputes Act 1947 is structured

The Act is a short statute by modern standards, organised around a definitions section of unusual importance, a machinery Part, and a set of chapters that regulate specific employer actions. Its forty sections carry schedules that do much of the real work. The Second Schedule lists the matters within the jurisdiction of labour courts, which include the propriety or legality of an employer’s order, the application and interpretation of standing orders, discharge and dismissal of workmen and the grant of relief on reinstatement, the withdrawal of customary concessions, the illegality or otherwise of a strike or lockout, and residual matters other than those in the Third Schedule. The Third Schedule lists the wider matters reserved to industrial tribunals: wages and allowances, hours of work and rest intervals, leave and holidays, bonus and provident fund and gratuity, shift working, classification by grades, rules of discipline, rationalisation, retrenchment of workmen and closure of undertakings, and other prescribed matters. The Fourth Schedule lists the conditions of service for which an employer must give notice of change. The Fifth Schedule, added in 1982, catalogues unfair labour practices on the part of employers, unions and workmen.

The chapters follow a deliberate logic. The opening provisions define the cast: appropriate government, employer, industry, industrial dispute, workman, strike, lockout, layoff, retrenchment, settlement, award. A second block creates the authorities: works committees, conciliation officers, boards of conciliation, courts of inquiry, labour courts, industrial tribunals and national industrial tribunals, and regulates reference, proceedings, reports and awards. A third block protects the status quo while disputes are pending and regulates strikes and lockouts. Chapter VA, added in 1953, lays down the compensation code for layoff and retrenchment in the generality of establishments. Chapter VB, added in 1976, superimposes a permission regime on larger establishments. A final block covers recovery of money due, penalties, and the protection of persons acting under the Act. The architecture therefore runs from definition to forum to conduct to money, and the definitions come first because everything downstream depends on them. Whoever controls the meaning of “industry” and “workman” controls the reach of the entire statute, which is why those two words generated more litigation than all the operative provisions combined.

Two governments, one statute: the appropriate government

Almost every operative provision of the Act runs through a defined term that is easy to skip and impossible to avoid: the appropriate government. It is the appropriate government that appoints conciliation officers and constitutes labour courts and tribunals, that receives the notice of retrenchment and the application for closure permission, and that decides whether a dispute is referred. Section 2(a) divides that role between the Centre and the states on a functional map of the economy. The Central Government is the appropriate government for disputes concerning industries carried on by or under its authority, the railways, mines and oilfields, major ports, banking and insurance companies, and corporations and companies in which the Centre holds the controlling interest or which are constituted under central legislation. For everything else, the state government is the appropriate government, and since the economy’s everyday mass, factories, shops of industry, hospitals, universities, municipal services, lies in the state column, the Act as administered was predominantly a state-administered statute wearing central drafting.

That division had three consequences that shaped the Act’s life. The first is divergence of practice: reference policy, the speed of conciliation, the constitution and staffing of labour courts, and the temper of Chapter VB decisions varied by state and by government, so that the same statute protected differently in different industrial geographies, a variation the parties priced into their strategies. The second is divergence of text: because labour sits in the Concurrent List, a state legislature may amend the central Act for its state, and where the amendment is repugnant to the central text it requires the President’s assent to prevail. States used the power across the Act’s life, on definitions, machinery, penalties and thresholds, and the 1982 central amendment itself spent part of its energy clarifying the appropriate government for multi-state and centrally controlled establishments whose disputes had generated jurisdictional skirmishing. The third consequence is the one the threshold story illustrates: on a concurrent subject, the centre’s failure to move is never the whole legal picture, because a state with the will and the assent can move for itself, and the Act’s history is unintelligible if the state amendments are treated as footnotes to the central text rather than as part of the statute in force.

The definitions that decide everything

Section 2(k) defines an industrial dispute as any dispute or difference between employers and employers, between employers and workmen, or between workmen and workmen, which is connected with the employment or non-employment, or the terms of employment, or the conditions of labour of any person. Two features of that text did heavy work. The first is its breadth on subject matter: non-employment is inside the definition, so a dismissal is not the end of a dispute but one of its standard subjects, and the phrase “any person” allows a dispute to be fought over the treatment of persons who are not themselves parties to it. The second is the requirement, built by the courts on the statutory scheme, of espousal: historically, an individual workman’s grievance became an industrial dispute only when taken up by a union or by a substantial body of fellow workmen, on the theory that the Act addressed collective conflict and not private grievances. That doctrine left the lone dismissed workman, often the person most in need of the forum, dependent on finding a sponsor.

Parliament corrected that gap in 1965 by inserting Section 2A, which provides that where an employer discharges, dismisses, retrenches or otherwise terminates the services of an individual workman, any dispute or difference between that workman and his employer connected with or arising out of the termination is deemed to be an industrial dispute, even though no other workman or union is a party. The provision converted the commonest industrial conflict of all, the contested sacking, into a dispute the individual could carry into the machinery without a sponsor, subject to a limitation discipline added much later: an application concerning a dismissal must be made within three years of the date of discharge, dismissal or retrenchment, with a power in the forum to condone delay for sufficient cause. Section 2A is the quiet workhorse of the Act. The celebrated litigation concerns industry and Chapter VB, but the daily traffic of the labour courts, for decades, was the individual dismissal dispute that Section 2A made possible.

Can a single workman raise an industrial dispute alone?

Yes, in the case that matters most. Since 1965, Section 2A has deemed a dispute between an individual workman and the employer over that workman’s discharge, dismissal, retrenchment or other termination to be an industrial dispute by itself, without any union or group of fellow workmen having to espouse the cause.

The second definition is Section 2(s), workman: a person, including an apprentice, employed in any industry to do manual, unskilled, skilled, technical, operational, clerical or supervisory work for hire or reward, whether the terms of employment are express or implied, and including, for the purposes of a dispute proceeding, a person who has been dismissed, discharged or retrenched in connection with or in consequence of that dispute, or whose dismissal led to it. The exclusions mark the boundary. The definition does not cover a person subject to the Army Act, the Air Force Act or the Navy Act, a person in the police service or an officer or employee of a prison, a person employed mainly in a managerial or administrative capacity, or a person who, being employed in a supervisory capacity, draws wages exceeding a statutory ceiling and exercises functions mainly of a managerial nature. The ceiling was one thousand six hundred rupees a month for decades, a figure that inflation converted into an absurdity, and the Industrial Disputes (Amendment) Act 2010, effective from 15 September 2010, raised it to ten thousand rupees a month and made it adjustable by notification. The operative test the courts apply is the dominant nature of the workman’s duties: designation is not decisive, and a supervisor who spends the substance of the day on skilled or clerical work may be a workman while a handsomely titled employee doing routine work may not. Almost every coverage fight under the Act begins here, because a person who is not a workman has no industrial dispute to take anywhere.

The remaining definitions can be taken shortly, but each conceals a decision. A strike, under Section 2(q), is a cessation of work by a body of persons employed in any industry acting in combination, a concerted refusal or a refusal under a common understanding to continue to work or accept employment; the courts held that the definition does not require the cessation to be a breach of contract, so a strike on a holiday or in the notice period can still be a strike. A lockout, under Section 2(l), is the temporary closing of a place of employment, the suspension of work, or the refusal by an employer to continue to employ any number of persons. A layoff, under Section 2(kkk), is the failure, refusal or inability of an employer, on account of shortage of coal, power or raw materials, accumulation of stocks, breakdown of machinery, natural calamity or any other connected reason, to give employment to a workman whose name is borne on the muster rolls and who has not been retrenched. Retrenchment, under Section 2(oo), inserted in 1984, means the termination by the employer of the service of a workman for any reason whatsoever otherwise than as a punishment inflicted by way of disciplinary action, but excludes voluntary retirement, retirement on reaching the age of superannuation where the contract so stipulates, termination resulting from non-renewal of a contract on its expiry or under a stipulation in that behalf, and termination on the ground of continued ill-health. The Supreme Court read that language at its width in Punjab Land Development and Reclamation Corporation in 1990, holding that retrenchment covers termination for any reason other than the stated exclusions, a reading that pulled ordinary contractual terminations inside the compensation code and has been debated ever since.

And then there is Section 2(j), industry. As enacted, and as it stood throughout the Act’s operative life, the clause reads: “industry” means any business, trade, undertaking, manufacture or calling of employers and includes any calling, service, employment, handicraft, or industrial occupation or avocation of workmen. It is a definition of two halves that do not obviously cohere, the first half describing the employer’s side of the enterprise and the second the workman’s, and it nowhere says what the organising idea is. Is a hospital an industry? A university? A municipal corporation? A solicitor’s office? A temple? The text does not answer, and for thirty years the courts answered case by case, before a bench of seven attempted to answer for all of them at once. That litigation is the doctrinal heart of this article, and it is taken up after the machinery and the money, because its significance is clearest once the consequences of being inside the definition are visible. One preliminary point belongs here, because the brief for this article insists on it and because half the commentary gets it wrong: the meaning of “industry” that governed India for nearly five decades was not the 1947 text read plainly, and it was not the 1982 text at all. It was a judicial construction of 1978, and the 1982 text never governed anything.

The machinery: conciliation first, adjudication by leave of government

The Act’s first instinct is settlement without adjudication. Section 3 requires a works committee in any industrial establishment in which one hundred or more workmen are employed, composed of representatives of the employer and the workmen, charged with promoting measures for securing amity and good relations and with commenting on matters of common interest. The committees were the weakest part of the design in practice, routinely bypassed by unions and managements alike, who preferred to negotiate directly or fight directly. The 2010 amendment added a second in-house forum, the Grievance Redressal Committee under Section 9C, for establishments employing twenty or more workmen, to which an individual workman’s dispute must first be carried, with a thirty-day horizon for its decision before the dispute moves outward to conciliation. Between the works committee and the grievance committee, the statute expresses a persistent hope that disputes can be metabolised inside the establishment. The record is that most serious disputes were not, and that the real machinery begins with the conciliation officer.

Conciliation officers, appointed under Section 4 by the appropriate government for a specified area or industry, are charged with mediating in and promoting the settlement of industrial disputes. In a public utility service, conciliation is not optional: where a notice of strike or lockout has been given, the conciliation officer is bound to hold conciliation proceedings. The officer investigates the dispute, induces the parties toward a fair settlement, and reports to the government. If a settlement is reached, it is recorded in a memorandum signed by the parties; if not, the officer sends a failure report setting out the steps taken and the reasons settlement failed. The Act fixes tight timelines on paper, a report within fourteen days of the commencement of conciliation proceedings in the standard case, extendable only by written agreement of the parties. Boards of conciliation under Section 5, an ad hoc body with a chairman and two or four members representing the parties, and courts of inquiry under Section 6, which inquire into matters connected with a dispute and report, complete the non-adjudicatory tier. Both were used sparingly; the system’s centre of gravity lay in the choice between a conciliated settlement and a reference to adjudication.

That choice belonged, historically, to the executive. Under Section 10, where the appropriate government is of the opinion that an industrial dispute exists or is apprehended, it may refer the dispute, or specified matters connected with it, to a board, a court of inquiry, a labour court or a tribunal, and where the dispute relates to a matter in the Third Schedule it goes to an industrial tribunal. Where both parties apply jointly or separately asking for a reference, and the government is satisfied that the applicants represent the majority of each party, the reference must be made. Where the Central Government is the appropriate government, or considers a dispute to be of national importance or to affect establishments in more than one state, it may refer the dispute to a national industrial tribunal under Section 7B. Section 10 also arms the government to prohibit the continuance of a strike or lockout in connection with a dispute it has referred, a power that makes the reference itself a cooling device: the dispute moves into the forum and the weapon must be put down.

The gatekeeping character of Section 10 is the design decision from which the Act’s politics flowed. A workman or a union could not, for most of the Act’s life, walk into a labour court and file a dispute the way a plaintiff files a suit. The government decided, and its refusal to refer was reviewable only on narrow administrative law grounds: the courts would correct a refusal that ignored relevant material or rested on irrelevant considerations, but they would not substitute their own view of whether adjudication was wise. Governments of every complexion used the power, referring disputes that embarrassed them to refer, sitting on disputes that embarrassed them more, and occasionally referring disputes precisely to trigger the prohibition on strikes. Defenders of the design reply, correctly, that the reference power filtered out disputes that conciliation could still settle and let the state sequence adjudication according to industrial peace. Critics reply, also correctly, that a right whose exercise depends on a ministerial opinion is a privilege dressed as a right. The Industrial Relations Code eventually resolved the argument by abolishing the reference for most disputes and letting a worker apply to the tribunal directly after conciliation fails, a change described in the succession section of this article. For the decades the Act governed, the gate was real, and no account of the statute is honest that treats adjudication as something the parties commanded.

What is the difference between a labour court and an industrial tribunal?

Both are adjudicating authorities under the Act, divided by subject. A labour court, under Section 7, tries the Second Schedule matters: dismissal and discharge, the legality of an employer’s order, withdrawal of a customary concession, and whether a strike is illegal. An industrial tribunal, under Section 7A, tries Second and Third Schedule matters, including wages, hours, bonus, retrenchment and closure.

Once a dispute is inside, the forums work with civil court powers for compelling evidence and attendance, and the Act gives them a distinctive discretion that ordinary courts lack. Section 11A, inserted with effect from 15 December 1971, provides that where an industrial dispute relating to the discharge or dismissal of a workman has been referred, and the labour court, tribunal or national tribunal is satisfied that the order was not justified, it may set aside the dismissal and direct reinstatement on such terms as it thinks fit, or give such other relief, including a lesser punishment, as the circumstances require. Before 1971, on the authority of the Supreme Court’s reading of the tribunal’s powers in Indian Iron and Steel, a tribunal could interfere with a dismissal only on narrow grounds: mala fides, victimisation, unfair labour practice, a basic error of fact, or perversity. Section 11A replaced that supervisory posture with a power to reappraise the evidence and the punishment, and the Supreme Court confirmed in Firestone Tyre that the forum could take fresh evidence and substitute its own view of the appropriate penalty. Dismissal litigation was transformed: the question stopped being whether the employer had behaved outrageously and became whether the punishment fitted the misconduct on the whole record, which is why so much Indian dismissal law reads like a jurisprudence of proportionality.

The forums themselves are judicial in composition. A labour court consists of one person only, its presiding officer, who must be or have been a judge of a High Court, or have been a district judge or an additional district judge for not less than three years, or have held judicial office in India for not less than seven years, with a further route for persons who have presided over other tribunals for a prescribed period. An industrial tribunal’s presiding officer must likewise be or have been a High Court judge, or a district or additional district judge of three years’ standing; a tribunal may be assisted by two assessors, appointed to advise it, whose advice it is free to disregard. A national industrial tribunal is presided over by a person who is or has been a judge of a High Court, a deliberate elevation that marks the class of disputes thought to exceed ordinary industrial adjudication. Section 15 obliges every board, court and tribunal to hold its proceedings expeditiously and to submit its report or award within the period specified in the reference or the Act, and Section 17 requires the government to publish every report and award within thirty days of receipt. The qualifications scheme matters beyond ceremony: it made the presiding officer a judge doing industrial work rather than an administrator doing quasi-judicial work, and much of the forums’ authority, and of the doctrine they built, flows from that choice.

Representation before these forums is deliberately restricted. Section 36 bars legal practitioners from conciliation proceedings altogether, and permits them before a labour court or tribunal only with the consent of the other party and the leave of the forum, a rule meant to keep the process in the hands of unions and employers’ associations rather than counsel. The parties to a dispute may be represented by officers of registered trade unions, federations, or employers’ associations, or by officers of the establishment itself. Settlements and awards then do the binding. A settlement arrived at in the course of conciliation binds all parties, all workmen employed in the establishment, and their successors; a settlement arrived at otherwise binds only its parties, a distinction that generated its own case law on who is bound by a bargain struck with one union among several. An award, once published by the government, becomes enforceable after thirty days, subject to the government’s power, where it is a party, to declare on public grounds that giving effect to the award would be inexpedient, laying the matter before the legislature. An award ordinarily remains in operation for one year, extendable by the government, and binds the parties, their heirs and successors, and all workmen in the establishment, including those who join later. The output of the machinery, in other words, is not a judgment between two litigants but a small code for an establishment, and that is exactly what the 1947 design intended.

The reference in practice: refusals, arbitration and the gate’s discontents

Because the reference is the gate, the law of the gate deserves a section of its own. The government’s discretion under Section 10 is wide but not a prerogative: the courts settled that the government must form its opinion on relevant material, that a refusal which rests on an adjudication of the merits is an excess, because deciding who is right is the forum’s job and not the ministry’s, and that a refusal may be corrected by mandamus directing the government to reconsider, though not, in the ordinary case, by the court directing the reference itself. The practical texture that grew around those principles was of delay as a decision: a dispute that is neither referred nor formally refused sits in a limbo in which the strike weapon rusts, the conciliated record stales, and the parties learn to treat the secretariat, rather than the forum, as the first court of appeal. Governments also acquired the habit of referring disputes in terms so wide or so narrow that the framing itself predetermined the litigation, and the generated case law on whether a tribunal may travel beyond the terms of its reference, it may decide matters incidental to the points referred but not enlarge the reference, is the fossil record of that habit.

The Act offered one road around the gate, and the parties largely declined to take it. Section 10A, inserted in 1956, permits the employer and the workmen, where a dispute exists or is apprehended and has not yet been referred, to agree in writing to refer it to arbitration, naming the arbitrator, who may be a presiding officer of a labour court or tribunal; where the reference is made by an even number of arbitrators an umpire is provided for, the agreement must be published, and the government may then prohibit the continuance of any strike or lockout connected with the dispute. The arbitrator investigates and submits an award which binds as an award under the Act. Voluntary arbitration had distinguished advocates in Indian labour policy and was written into national labour rhetoric for decades, yet it remained a minor road, for a structural reason the statute could not cure: arbitration requires both parties to surrender the chance of a favourable gate decision and to accept an outcome neither controls, and a party that expects the government to refuse a reference, or expects delay to work in its favour, has no incentive to consent. The road around the gate stayed open and largely untravelled, which is itself a finding about the design: where the state offers adjudication as a controlled service, the parties organise their behaviour around the controller.

Protecting the status quo while a dispute is pending

Adjudication takes time, and an employer who can alter the facts while the forum deliberates can win by changing the subject. The Act answers with two freezes. The first is Section 9A: an employer who proposes to effect any change in the conditions of service applicable to a workman in respect of any matter specified in the Fourth Schedule must give the workman notice of the proposed change, and the change cannot take effect until twenty-one days after the notice. The Fourth Schedule covers the territory of everyday industrial life: wages and the mode of payment, contributions to provident fund and pension, compensatory and other allowances, hours of work and rest intervals, leave and holidays, starting and closing hours, classification by grades, withdrawal of customary concessions, rationalisation and changes in technology likely to end jobs, and increases or reductions in the number of persons employed. The notice does not forbid the change; it forces a pause in which the union can respond and a dispute can crystallise before the fact, rather than after.

What does a notice of change under Section 9A require?

It requires the employer to tell the affected workmen, in the prescribed manner, of any proposed change in conditions of service on a Fourth Schedule matter, and to wait twenty-one days before the change takes effect. Wages, allowances, hours, leave, grades and reductions in employment are all Fourth Schedule matters, so most unilateral changes pass through this pause.

The second freeze is Section 33, and it is the sharper of the two. While any proceeding under the Act is pending before a conciliation officer, board, labour court, tribunal or national tribunal, the employer may not, in regard to any matter connected with the dispute, alter the conditions of service of the workmen concerned to their prejudice, and may not, for any misconduct connected with the dispute, discharge or punish a workman save with the express permission in writing of the authority before whom the proceeding is pending. For matters not connected with the dispute, the employer may act in accordance with the standing orders or the terms of the contract, but if it discharges or dismisses a workman, it must pay one month’s wages and apply to the authority for approval of the action taken. The architecture is precise: connected matters need prior permission, unconnected dismissals need prompt subsequent approval, and in neither case can the employer present the forum with a fait accompli. Section 33A gives the workman a direct complaint to the forum where the employer contravenes Section 33, and the forum adjudicates the complaint as if it were a referred dispute. Section 33C then provides the recovery route for money due under a settlement, award or the layoff and retrenchment chapters, first by certificate to the Collector as arrears of land revenue, and, under Section 33C(2), by application to the labour court to compute money due or the money value of a benefit, the route through which generations of workmen converted paper entitlements into computed sums.

Section 33 singles out one category for reinforced protection. A protected workman, under Section 33(3), is a workman who, being a member of the executive or an office bearer of a registered trade union connected with the establishment, is recognised as such in accordance with the rules; the number recognised is one per cent of the total number of workmen employed in the establishment, subject to a minimum of five and a maximum of one hundred. In the case of a protected workman, the employer cannot alter conditions of service or discharge or punish at all during pendency, even on matters unconnected with the dispute, without the express written permission of the authority. The rationale is structural rather than sentimental: if union office bearers can be picked off while a dispute is being fought, the dispute itself is being rigged, and the forum’s eventual award will describe an establishment whose union has been beheaded in the interval.

Who is a protected workman, and why does the status matter?

A protected workman is a workman who is an executive member or office bearer of a registered trade union connected with the establishment and is recognised as protected under the rules, up to one per cent of the establishment’s workmen, with a floor of five and a ceiling of one hundred.

The status matters because during the pendency of proceedings the employer cannot dismiss, punish or worsen the conditions of a protected workman on any matter without the forum’s express written permission, a bar stricter than the one protecting the general body of workmen.

The last piece of the pendency scheme pays the workman while the employer litigates. Section 17B, inserted by the 1982 amendment and in force from 21 August 1984, provides that where a labour court, tribunal or national tribunal has made an award directing the reinstatement of a workman, and the employer prefers proceedings against the award in the High Court or the Supreme Court, the employer is liable to pay the workman full wages last drawn, inclusive of allowances, during the pendency of those proceedings, if the workman has not been employed in any establishment during that period and files an affidavit to that effect. The provision attacks the oldest attrition strategy in dismissal litigation, in which the employer’s appeal costs the workman his livelihood whether or not the award was right. Its condition, that the workman be unemployed, and the affidavit requirement, are the employer’s safeguards, and the courts police them; but the direction of the provision is unmistakable, and it redistributed the economics of every reinstatement appeal after 1984.

Strikes and lockouts: a fenced weapon

The Act regulates the strike by distinguishing where it happens and when. Section 22 governs public utility services, a defined category in Section 2(n) that includes railways, major port and dock transport, sections of industrial establishments on whose working the safety of the establishment or its workmen depends, postal, telegraph and telephone services, public conservancy and sanitation, power, light and water undertakings, and any industry the appropriate government declares to be a public utility service for six months at a time on account of public emergency or public interest. In a public utility service, a workman may not go on strike in breach of contract without giving the employer notice of strike within six weeks before striking, or within fourteen days of giving such notice, or before the date specified in the notice, or during the pendency of conciliation proceedings before a conciliation officer and seven days after their conclusion. The lockout provisions mirror the strike provisions. The notice is not a request for permission; the employer cannot refuse it. Its functions are cooling and signalling: it starts the conciliation clock, warns the public that an essential service may stop, and gives the government the chance to refer the dispute, whereupon Section 10(3) can prohibit the stoppage.

Outside public utility services, the Act does not require notice at all, a point that contradicts the common belief that Indian law demands warning before every strike. What it does instead is prohibit the weapon during the fight’s formal phases. Section 23 bars strikes and lockouts in breach of contract during the pendency of conciliation proceedings before a board and seven days after, during the pendency of proceedings before a labour court, tribunal or national tribunal and two months after, during the pendency of arbitration proceedings where a notification has issued and two months after, and during the operation of a settlement or award in respect of matters covered by it. A strike or lockout commenced or declared in contravention of Section 22 or Section 23, or continued in contravention of a prohibitory order under Section 10(3) or Section 10A(4A), is illegal under Section 24; and a strike that is illegal at its inception is not rescued by being otherwise justified, though the courts developed a doctrine that an illegal strike may still be justified in fact and a legal strike may be unjustified, with consequences for wages and discipline rather than for legality itself. The penalty scheme in Sections 26 to 29 is famously antique in its figures: a workman who commences or continues an illegal strike is punishable with imprisonment up to one month, or a fine that the statute still expresses in tens of rupees, or both; instigation and the giving of financial aid to an illegal strike attract heavier terms. The real sanctions for an illegal strike were never the fines. They were the loss of wages for the strike period on the principle the courts summarised as no work, no pay, exposure to disciplinary proceedings in which justification mitigates but illegality aggravates, and the risk that the employer answers with a lawful lockout.

The balance the Act strikes here is frequently misstated from both directions. It is not true that the Act made strikes practically impossible: the notice condition applied only to public utilities, the general prohibitions bit only when proceedings or a settlement were running, and Indian industrial history is full of long, lawful strikes. It is also not true that the strike under the Act is a free weapon of the kind contemplated by pure collective bargaining theory: the reference power, the power to prohibit a strike on reference, and the pendency prohibitions together mean that the state can usually convert a threatened stoppage into an adjudication, at which point striking becomes a contempt of the design rather than an incident of the bargain. The successor Code changed the balance in one conspicuous way, by requiring fourteen days’ notice of strike and lockout in every industrial establishment and confining the strike to the window following that notice, thereby generalising the public utility discipline to the whole economy; that change, and the argument it continues to generate, is taken up in the succession section.

Public utility services: the Act’s emergency lane

Woven through the strike provisions is a second legal map of the economy, the definition of public utility service in Section 2(n), which separates employments whose interruption the Act treats as a public danger from those it treats as private quarrels. The statutory list covers railway and transport services for passengers and goods, ports and docks, the sections of an industrial establishment on the working of which the safety of the establishment or its workmen depends, postal, telegraph and telephone services, industries supplying power, light or water to the public, and systems of public conservancy or sanitation. To that list the appropriate government may add, by notification, any industry specified in the First Schedule, the coal, textile, steel, banking, transport and like industries, where it is satisfied that public emergency or public interest so requires, for periods not exceeding six months at a time, renewable within the same limit. The First Schedule device is the Act’s emergency lane in statutory form: an ordinary industry may be converted into a public utility for the emergency, with the consequence that its strikes and lockouts attract the stricter notice regime of Section 22, six weeks’ notice, no strike within fourteen days of it, no strike before the date the notice itself specifies, and no strike while conciliation is pending and for seven days after its conclusion, in place of the Section 23 general bars that govern ordinary establishments. The declaration power, repeatedly used in the decades of scarcity for industries such as banking and transport, is the point at which the Act’s wartime ancestry shows most plainly: the state reserves the right to move any significant industry into the emergency lane by notification, and the boundary between the emergency lane and ordinary law is drawn by executive satisfaction rather than by the statute’s own categories. Readers comparing the successor will find the Industrial Relations Code generalising the stricter notice to every establishment, which is the emergency lane made universal by a quieter route.

Standing orders: the rulebook the Act polices

No pillar account of the Act is complete without its sibling, because the substantive law of discipline inside most establishments is not in this Act at all. The Industrial Employment (Standing Orders) Act 1946 requires establishments of the prescribed size to have certified standing orders, rules framed with the certifying officer’s approval that define the classification of workmen, shifts and attendance, leave, termination, and the acts that constitute misconduct. The Industrial Disputes Act takes that rulebook as given and polices it at three points. The Second Schedule makes the application and interpretation of standing orders a labour court matter, and Section 13A, inserted in 1982, allows any employer, workman, union or government to refer a question as to the application or interpretation of a standing order to a labour court, whose decision is final and binding, a small provision that quietly converted standing orders from house rules into adjudicated law. Section 33’s freezes operate, in the unconnected-matters branch, expressly by reference to the standing orders: the employer may act in accordance with them, and the forum’s approval inquiry under Section 33 asks whether the action is prima facie justified under them and not a device of victimisation. And Section 11A’s assessment of punishment proceeds against the background of the penalties the standing orders prescribe. Suspension pending enquiry sits inside the same architecture: the power to suspend comes from the standing orders or the contract, the workman’s claim to subsistence during suspension from the standing orders and state law, and the forum’s power to treat a suspension that is really a punishment as a dismissal in disguise. The two statutes were repealed together into the Industrial Relations Code, which carries the standing orders regime inside itself, and they are best understood as one instrument in two volumes for their entire joint life.

Wages for the strike period: justification as doctrine

Legality is only half the strike story; the other half is pay. A strike that is not illegal may still be unjustified, and the forums developed a jurisprudence of justification that decided who bears the wage cost of the stoppage. The governing propositions were settled in stages. The wage rule begins with the contract: for days not worked, the maxim of no work no pay states the default, and in Bank of India v T. S. Kelawala in 1990 the Supreme Court applied it to a partial stoppage in which bank employees had struck for part of the day, upholding the deduction of wages for the hours not worked and rejecting the argument that a day partly worked must be paid in full. But adjudication under the Act is not confined to the contract, and the forums long exercised a discretion to award wages for a strike period where the strike was both legal and justified, justified meaning provoked by the employer’s own unfair conduct, a demand that was reasonable, or a refusal to use the machinery the Act provides. Gujarat Steel Tubes v Its Mazdoor Sabha, decided in 1980 on the ruins of a long strike and a mass termination, is the doctrine’s great statement, treating justification as a question of industrial fairness to be weighed with the legality, the provocation and the conduct of both sides. Syndicate Bank v K. Umesh Nayak in 1994 organised the learning into its modern form: wages for the strike period are not an entitlement and not a forfeit, but a discretionary award turning on whether the strike was legal and justified, with the burden lying on those who claim the wages to show both. Lockouts run in mirror image, with a lockout declared in consequence of an illegal strike standing on a different footing from one used as an offensive weapon. The government’s own employees stand apart: in T. K. Rangarajan v Government of Tamil Nadu in 2003, in the context of a mass strike by state employees and teachers, the Supreme Court held that government employees have no fundamental right to strike, a holding whose breadth drew sustained academic and judicial unease and which remains part of the subject’s contested margin rather than its settled centre. For the student of the Act the significance of the wages doctrine is structural: it is the clearest example of the forums legislating in the gaps, because the Act itself says nothing at all about wages for a justified strike, and the entire law of the subject, from provocation to apportionment, is judge-made.

The domestic enquiry: the trial inside the factory

Most dismissal disputes begin long before any forum is seized of them, in the domestic enquiry, the internal disciplinary proceeding the employer conducts before punishing. The Act does not prescribe the enquiry; it grew from the standing orders and from the common law of the employment relation, and the forums then built the standards by which it is judged. The settled principles are these. The workman must be told the charges with enough precision to answer them, ordinarily by a charge sheet; the enquiry must be held by an officer who is not himself a witness to the incident and not otherwise biased; the workman must have a real opportunity to present evidence, to cross-examine the management’s witnesses and to be assisted, commonly by a co-worker or union representative rather than by counsel; the findings must rest on some evidence, the standard being preponderance of probabilities and not proof beyond reasonable doubt, since the enquiry is not a criminal trial; and the punishment must follow the findings rather than precede them. A failure of natural justice in the enquiry, a biased officer, a denial of the chance to defend, a finding with no evidence behind it, vitiates the proceeding, and the forum’s response to a vitiated enquiry itself evolved: the employer may be permitted to prove the misconduct afresh before the forum, a doctrine the Supreme Court consolidated when it held that the forum’s power under Section 11A includes receiving evidence on the merits where the enquiry is defective or absent, so that the litigation decides the misconduct rather than only the paperwork.

The interaction between enquiry doctrine and dismissal outcomes explains a large slice of Indian dismissal practice. Because the forum can reappraise, the enquiry’s real function is evidentiary: it assembles the record on which the employer will stand or fall, and managements learned to conduct enquiries as if a presiding officer were reading over the officer’s shoulder, because one eventually would be. Because the workman’s protections in the enquiry are procedural rather than substantive, unions learned to fight on procedure, where the employer’s errors are findable, rather than only on the facts. And because reinstatement with full back wages was the classical remedy for a dismissal held unjustified, the stakes of procedural error were for decades out of proportion to the error, a disproportion the courts themselves later moderated by withholding back wages where the workman was gainfully employed, by substituting compensation where the relationship of trust was destroyed, and by refusing reinstatement where the original appointment was itself made in defiance of recruitment rules. None of that moderation appears in the statute; it is the adjudication system correcting its own incentives in the open, and it belongs in a pillar account because a reader who knows only the text of the Act does not know how dismissal actually worked under it.

Chapter VA: the compensation code for layoff, retrenchment and closure

Chapter VA, inserted in 1953, is where the Act stops regulating process and starts pricing exits. Its provisions apply to the run of industrial establishments, with the layoff compensation provision pitched at factories, mines and plantations in which fifty or more workmen are employed, and they rest on a unit of measurement the whole chapter shares: continuous service. Section 25B provides that a workman is in continuous service for a period if he has rendered uninterrupted service for that period, counting service interrupted by sickness, authorised leave, an accident, a strike which is not illegal, a lockout, or a cessation of work not due to any fault of the workman. A workman who has actually worked under the employer for not less than two hundred and forty days in the preceding twelve calendar months, one hundred and ninety days in the case of a workman employed below ground in a mine, is deemed to have completed one year of continuous service. The two hundred and forty day count is the hinge of the chapter: it converts casual and interrupted engagement into statutory seniority, and disputes about whether the count is met, on muster rolls and attendance records, are a staple of the case law.

Layoff is the temporary exit. Where a workman in a covered establishment, whose name is borne on the muster rolls and who has completed not less than one year of continuous service, is laid off, Section 25C entitles him to compensation from the employer for all days of layoff, except weekly holidays, equal to fifty per cent of the total of basic wages and dearness allowance that would have been payable had he not been laid off. If the layoff extends beyond forty-five days in any period of twelve months, the compensation obligation for the further period falls away unless an agreement provides otherwise, and the employer may instead retrench the workman in accordance with Section 25F, setting off the layoff compensation already paid against the retrenchment compensation. The design concedes the employer’s predicament, an establishment without power, materials or orders cannot pay full wages indefinitely, while refusing to let the cost of the stoppage fall wholly on the workman who did not cause it.

Retrenchment is the permanent reduction, and Section 25F states its conditions as conditions precedent, meaning that a retrenchment done without them is not merely irregular but void, with reinstatement as the standard consequence. The conditions are three, and the article states them exactly because the brief for it identifies the formula as one of the things readers come to get right. No workman who has been in continuous service for not less than one year may be retrenched until, first, the workman has been given one month’s notice in writing indicating the reasons for retrenchment and the period of notice has expired, or has been paid wages in lieu of the notice for its period; second, the workman has been paid, at the time of retrenchment, compensation equivalent to fifteen days’ average pay for every completed year of continuous service or any part thereof in excess of six months; and third, notice in the prescribed manner is served on the appropriate government or the authority it specifies. The arithmetic repays a worked example. A workman with eleven years and eight months of continuous service counts twelve units, eleven completed years plus a part exceeding six months, and is owed fifteen days’ average pay twelve times over, which is one hundred and eighty days’ average pay, six months’ wages in substance, in addition to the month of notice or its wages. Two rounding rules inside that arithmetic are constantly missed: the part-year counts only if it exceeds six months, and the service count runs on the Section 25B measure, deemed years included, not on the calendar alone. Around Section 25F the chapter builds its supporting rules: Section 25FF gives transfer compensation where an undertaking changes hands, unless service is uninterrupted, terms are no less favourable, and the new employer is liable as if service had been continuous; Section 25G prescribes the order of retrenchment, ordinarily the last person employed in a category goes first, unless the employer records reasons to depart; Section 25H obliges the employer who later re-employs to give retrenched workmen the first opportunity, in the prescribed manner. Section 25FFF prices the closing of an undertaking: notice and compensation as if the workmen had been retrenched, with a proviso that where the undertaking is closed on account of unavoidable circumstances beyond the employer’s control, the compensation is capped at three months’ average pay.

A practitioner reading this section will notice what it does not attempt, and the omission is deliberate: the general position stated here is information about the statute, not advice on any individual case, and a reader facing an actual retrenchment should take advice on the facts, because the consequences turn on records, categories and notices that a general account cannot weigh. What the general account can do is locate Chapter VA’s real significance, which is not the size of its cheques. Fifteen days’ pay a year is modest severance by any standard. The chapter’s bite lies in the voidness rule and its interaction with Section 11A adjudication: an employer who retrenches without the notice, the compensation or the government intimation has not committed a fineable irregularity but has, in the eyes of the doctrine, not retrenched at all, and the workman remains in service in law with back wages at the forum’s discretion. Reformers attacked that consequence as disproportionate, and courts themselves, over the decades, moderated the back-wages reflex toward compensation in lieu in many classes of case, particularly where the appointment itself was irregular. But the structure Parliament built stands as the chapter’s logic: because India offered the dismissed workman no unemployment insurance and no welfare floor worth the name, the statute put the price of exit on the party making the exit decision, and enforced the price with the strongest remedy labour law knows.

The wider wage jurisdiction: dearness allowance, bonus and the tribunal as pay commission

The pillars of the Act that lay readers notice are dismissals and retrenchment, but the forums’ largest historical jurisdiction was pay. The Third Schedule put wages, allowances, hours, leave, bonus, provident fund and gratuity into the tribunal’s hands, and for the first three decades the industrial tribunal functioned in industry after industry as a standing pay commission, fixing wage scales and dearness allowance systems for establishments and whole industries on principles the forums themselves developed. Dearness allowance doctrine, the extent to which the allowance should neutralise the rise in the cost of living, and for which wage bands full neutralisation was appropriate, was tribunal-made law long before it was anyone’s statute. Bonus was the largest creation of all: workers claimed a share in profits as a matter of industrial right, employers answered that profits belonged to capital after a fair return, and the Labour Appellate Tribunal’s Full Bench, in 1950, reduced the contest to an accounting formula, the available surplus computed after prior charges including depreciation, reserves and a fair return on capital and working capital, from which bonus could be awarded. The formula governed bonus adjudication until the Payment of Bonus Act 1965 converted the doctrine into a statutory scheme of minimum and maximum bonus, and gratuity travelled the same road from tribunal awards, which ordered gratuity schemes as a terminal benefit on retirement and long service, to the Payment of Gratuity Act 1972. The pattern is worth naming in a pillar treatment, because it reverses the usual direction of Indian labour legislation: in pay matters, adjudication came first and codification followed the doctrine the forums had built, whereas in the definition of industry codification was attempted and adjudication was left in possession. A reader who wants to understand why Indian employers historically dreaded the tribunal less for its dismissal jurisprudence than for its wage jurisdiction, and why industry-wide bargaining in sectors such as banking, steel, cement and ports developed the shapes it did, should start from this pay commission character of the Act’s forums.

Chapter VB: the permission regime and the forty-year argument

Chapter VA prices exits in the generality of establishments. Chapter VB, added by the Industrial Disputes (Amendment) Act, Act 32 of 1976, with effect from 5 March 1976, forbids them outright in the largest establishments until the government agrees. As inserted in 1976, the chapter applied to an industrial establishment, not being of a seasonal character or one in which work is performed only intermittently, in which not less than three hundred workmen were employed on an average per working day for the preceding twelve months; the industrial establishment for this purpose means a factory, a mine or a plantation. The 1982 amendment substituted one hundred for three hundred in Section 25K, with effect from 21 August 1984, and it is the figure of one hundred, not the original three hundred and not the figure in many summaries, that governed from 1984 until the Code. Within a covered establishment, Section 25M prohibits layoff of any workman, other than a badli or casual workman, without the prior permission of the appropriate government or the specified authority, unless the layoff is due to shortage of power or natural calamity, or, in a mine, fire, flood or excess of inflammable gas. Section 25N imposes the corresponding discipline on retrenchment: the workman must be given three months’ notice in writing indicating the reasons, or wages in lieu; the employer must apply for prior permission; and the government or authority, after inquiry and hearing both sides, may grant or refuse permission by a reasoned order, with the permission deemed granted if no order is communicated within sixty days of the application; where permission is granted, the workman is entitled at retrenchment to compensation at the same fifteen days’ average pay per completed year of service, or part exceeding six months. Section 25O imposes the discipline on closure: an employer intending to close down an undertaking must apply for prior permission at least ninety days before the intended date of closure, stating the reasons; the government decides on stated criteria, including the genuineness and adequacy of the reasons, the interests of the general public and all other relevant factors; permission uncommunicated within sixty days is deemed granted; and a closure carried out without permission, or after refusal, is illegal, the workmen being entitled to all the benefits under any law for the time being in force as if the undertaking had not been closed down.

The chapter’s constitutional life began in failure. The Supreme Court in Excel Wear in 1978 struck down Section 25O as originally enacted, holding that a power to refuse closure indefinitely, without guidance, hearing discipline or time limit, was an unreasonable restriction on the right to carry on business, which includes the right to close it. Parliament recast Section 25O in 1982 with the hearing, criteria and deemed-permission features described here, and when Section 25N as substituted by Act 49 of 1984 came under attack the Constitution Bench in Workmen of Meenakshi Mills in 1992 upheld it, treating the permission requirement as a reasonable restriction in the interest of the general public and reading the sixty-day default as the answer to the charge of indefinite executive veto. The chapter thus stood, for the rest of the Act’s life, as constitutionally valid law whose operation was the controversy. Governments deciding applications weighed the genuineness of the employer’s reasons against the public interest in continued employment, and both sides of the resulting record are part of the statute’s real history. Permission was refused far more often than it was granted in the experience reported to commissions and recorded by employers’ organisations and unions alike, and litigation over refusals, over the adequacy of reasons, and over establishments that sickened while applications aged became a standard sequel. The Second National Commission on Labour, whose 2002 report proposed gathering the labour statutes into a small number of codes, treated the permission threshold and the compensation level as a pair to be renegotiated together, higher protection pay in exchange for a higher gate, which is the bargain the Industrial Relations Code eventually struck at three hundred.

The threshold architecture only makes sense inside the whole labour statute landscape, and the companion guide to the full framework of Indian labour law maps how the Act’s gates sit among the other statutes’ gates. The reader who wants the architecture should go there; this article’s business is what happened at this particular gate, and what happened is the liveliest economics argument in Indian legislation.

The employer case against Chapter VB, in its strongest form, is not that workers deserve no protection. It is that a permission regime converts a business decision into an administrative favour, with four documented consequences. First, exit becomes discretionary: a unit that cannot cover its costs cannot lawfully die, so it lingers, consuming capital and management attention that a functioning market would redeploy. Second, entry and expansion are deterred at the threshold: an establishment that grows past ninety-nine workmen crosses into a legal category in which its workforce becomes, for practical purposes, permanent, so rational firms organise production to stay below the line, splitting units, capping headcounts and substituting machinery and contractors for employees. Third, the margin of adjustment shifts to the unprotected: if the organised core cannot be retrenched, employers hire through contractors and casual arrangements, so the statute protects a shrinking insider class by pushing risk onto outsiders with no protection at all. Fourth, the empirical literature, most prominently the study by Timothy Besley and Robin Burgess published in the Quarterly Journal of Economics in 2004, which coded state amendments to the Act between 1958 and 1992 as pro-worker, pro-employer or neutral, found that states which amended in a pro-worker direction experienced measurably lower output, employment and investment in registered manufacturing, and higher urban poverty, than the coding predicted otherwise.

The worker case, also in its strongest form, begins by refusing the premise that the alternative to Chapter VB was a flexible labour market with a safety net. India, for the whole of the chapter’s life, had no unemployment insurance system, no general income floor for the dismissed, and a retrenchment cheque under Chapter VA of fifteen days’ pay a year, which for a workman of twenty years’ service is ten months of wages against the loss of a livelihood in an economy where the next formal job may never come. In that setting, the economists’ flexibility is a transfer of the entire risk of demand fluctuation, technological change and managerial failure onto the party least able to bear it, and the permission requirement is better understood as a crude substitute for insurance: the state, unwilling to pay the dismissed, forbade the dismissal. The empirical case is contested on its own terms as well. Aditya Bhattacharjea’s re-examinations of the Besley and Burgess coding showed that several amendments were miscoded or coded in the wrong direction, and that the headline results weaken considerably when the coding is corrected; and the threshold story itself is double-edged, since the Economic Survey of 2018-19, in its discussion of firms that remain small throughout their lives, treated dwarfism as a product of a family of size-based policies, of which Chapter VB was one member and not the largest. The honest verdict the evidence supports is narrower than either side’s slogan. Chapter VB did not make it impossible to retrench in India: it made retrenchment, layoff and closure in establishments above a modest size a matter of state discretion, exercised unevenly, litigated expensively and evaded predictably, and the evasions, contract labour above all, are where the regime did its least defensible damage, because they unprotected the workforce without ever repealing the protection. It is in precisely that gap, between the protected core the chapter imagined and the contract workforce employers built to avoid it, that the companion statute on contract labour does its work, and the two statutes are best read together, as the account of the contract labour regime sets out. The same gap explains the Act’s most melancholy by-product: units that could neither obtain permission to close nor earn their keep were allowed to sicken in place, their workmen unpaid in fact while employed in law, until they arrived at the sickness and revival machinery by the back door, a pathology the series traces in its account of the SICA and BIFR regime.

Permission in practice: how governments decided, and the Bombay warning

The permission sections generated their own practice, distinct from their text. An application under Section 25N or Section 25O required the establishment to state its case on the statutory factors, the genuineness and adequacy of the reasons, the interests of the workmen, the financial condition and efforts made, and the deciding authority, ordinarily the labour department of the state acting for the appropriate government, heard the parties and passed a speaking order, which the aggrieved side could seek to have reviewed, or referred to a tribunal for adjudication, within the thirty days Section 25N allows. The litigation that followed taught governments to write fuller orders, because the reviewing courts examined whether relevant factors were considered and irrelevant ones excluded, and taught employers to build their record in advance, since a permission refused on an adequate record was difficult to move. The deemed permission clauses did real work at the margin: the expiry of sixty days under Section 25N, or of the two months under Section 25O, converted administrative silence into authority to act, and departments learned to decide within the clock or lose the decision.

Behind the doctrine stands the era’s defining industrial trauma, which the permission regime predated by a few years and could not in the event prevent. In January 1982, Bombay’s textile workers, some quarter of a million of them in the common telling of the episode, struck under the leadership of Datta Samant in one of the largest and longest strikes in Indian industrial history. The strike ran for about eighteen months; the mills, already weakened, largely never recovered their place, and the closures that followed worked themselves out through the Act’s machinery, through sickness proceedings and through simple abandonment over the following decades, transferring the city’s industrial land question to the centre of its politics. The episode is the standing warning attached to the permission debate from both directions: for those who defend permission, it shows what unmediated industrial collapse does to a workforce; for those who attack it, it shows that a permission regime did not save the industry either, and that the decisive variables were the economics of the mills and the choices of the parties, on which the statute’s conditions had only marginal purchase. The permission sections outlasted the episode by a generation, which suggests the debate was never only about Bombay: it was about who, in the Indian system, is trusted with the decision to end employment at scale, the market, the employer, or the state, and the Act’s answer, the state, held until the codes replaced it.

State variation: the Concurrent List in action

Labour is a Concurrent List subject, and the Act was always two laws in practice: the central text, and the central text as amended for each state by its legislature with the President’s assent where the amendment touched repugnancy. States amended definitions, machinery, penalties and, in Maharashtra’s case under its own recognition statute, the whole bargaining structure around the Act. The threshold of Chapter VB became the favourite lever. Rajasthan moved first in the modern phase, amending the Act in its application to Rajasthan in 2014, with the President’s assent, to raise the Section 25K figure from one hundred back to three hundred, and pairing the change with amendments to the Factories Act and the Contract Labour Act thresholds in the same direction. Gujarat and Uttar Pradesh were among the states that later made the same change to the retrenchment threshold, and by the time Parliament enacted three hundred as the national figure in the Industrial Relations Code, it was ratifying a direction a line of state legislatures had already taken. The sequence matters for the series’ thesis: the most contested number in Indian labour law was changed, lawfully and without national legislation, by the mechanism the Constitution provides for exactly this, state amendment on a concurrent subject, and the states that changed it generated the nearest thing the long argument ever had to evidence about what a higher threshold does. The evidence was argued over, as state-level evidence always is, but the constitutional point is clean: Chapter VB’s rigidity was never a fact of nature. It was a setting, and settings can be changed by any legislature with the competence and the assent, which is why the forty-year failure to change it belongs to politics and not to law.

Unfair labour practices: the Fifth Schedule

The 1982 amendment also completed the Act’s moral vocabulary by inserting the Fifth Schedule and Sections 25T and 25U with effect from 21 August 1984, prohibiting unfair labour practices and making their commission punishable. The Schedule lists practices on the part of employers and their organisations, including interfering with or coercing workmen in the exercise of their right to organise, threatening workmen with discharge for joining a union, discharging or punishing a workman for urging others to join, abolishing regular work and giving it to contractors as a measure of breaking a strike, transferring a workman mala fide under the guise of management policy, showing favouritism to one of several unions, employing workmen as badlis or casuals for years with the object of depriving them of the status of permanents, and refusing to bargain collectively in good faith with a recognised union. It lists practices on the part of workmen and their unions in parallel: advising or actively supporting an illegal strike, coercing workmen in the exercise of the right not to join a union, picketing in such a manner that non-striking workmen are physically barred from entering, indulging in acts of force or violence or holding out threats of intimidation against non-striking workmen or managerial staff, and refusing to bargain in good faith. The reciprocity was the point. For thirty-five years the Act had regulated the parties’ weapons without naming the fouls; the Fifth Schedule named them on both sides, and although prosecutions under Section 25U were rare, the Schedule’s real career has been as the standard by which forums characterise conduct inside other proceedings, from Section 33 permission applications to Section 11A assessments of punishment.

The word that swallowed the Act: industry before 1978

Return to Section 2(j), with the machinery and the money in view, because the consequences of the definition are the consequences of everything just described: a finding that an establishment is an industry switches on the conciliation machinery, the forums, the status quo freezes, the compensation chapters and the permission regime for everyone it employs. The pre-1978 courts read the clause with a mixture of hospitality and embarrassment. In D. N. Banerji v P. R. Mukherjee in 1953, the Supreme Court held a municipal corporation to be an industry, reading the clause functionally. In State of Bombay v Hospital Mazdoor Sabha in 1960, the Court held that a group of hospitals run by the state was an industry, reasoning that the second half of the definition, the calling and service of workmen, could not be read down by the first, and that an undertaking producing services for the community in an organised form fell within the clause even without a profit motive. The Corporation of the City of Nagpur decision of the same period brought municipal services generally within the Act. Then the line wobbled. University of Delhi v Ram Nath in 1963 held that a university is not an industry, on the ground that education is not an undertaking of the kind the clause contemplates and that teachers are not workmen, a conclusion in open tension with the hospital case’s logic. Safdarjung Hospital in 1970 held that a hospital, on the facts, was not an industry, distinguishing and narrowing Hospital Mazdoor Sabha on the basis that the hospital was a public welfare institution without the attributes of an undertaking. Gymkhana clubs were held outside; a research institute’s fate depended on how commercial its operations looked. By the mid-1970s the law was a patchwork in which the same institution could be an industry in one state and not in another, and the question in every case, what the Act applied to at all, had to be litigated before the merits could be reached. It was a failure of the definition, not of the courts: the clause had been drafted in 1947 for a world of factories, mines and docks, and was being applied to a welfare state.

Two subsidiary doctrines of the pre-1978 decades deserve recording, because they survived the great restatement and still organise borderline cases. Where an establishment carries on several activities, some industrial and some not, the courts asked which activity is predominant, the dominant nature test: if the predominant activity is an industry and the others are inseverable adjuncts, the whole establishment is an industry; if a unit is severable, it is judged on its own character. Parliament later wrote that severability logic into the 1982 package’s definition of industrial establishment, which was among the clauses that did commence, so the doctrine has a statutory echo even though the definition it served never arrived. And the professions stood at the boundary from early on: in National Union of Commercial Employees v M. R. Meher in 1962, the Supreme Court held that a solicitor’s firm was not an industry, on reasoning that leaned on the liberal professions as callings in which the cooperation the definition imagines is absent, a conclusion Bangalore Water Supply treated with reserve without formally interring, leaving professional firms of scale in a argued-over margin that the 1982 exclusions, exempting small professional practices, were plainly drafted to settle. The margin mattered in practice precisely because the modern service economy grows at it: hospitals, universities, firms of accountants and lawyers, research companies and clubs all live near the line the cases drew, and each new organisational form, from the private hospital chain to the technology services company, arrived at the forums as a fresh argument about a sentence drafted in 1947.

Bangalore Water Supply, 1978: the triple test

The reconsideration came in Bangalore Water Supply and Sewerage Board v A. Rajappa, decided on 21 February 1978 by a bench of seven judges. The immediate dispute was small: whether the Bangalore water board was an industry, and so whether its employees could raise industrial disputes. The bench used the occasion to restate the whole subject. The principal opinion of Krishna Iyer J. held that the clause, read as a whole, yields a functional test directed at the activity and not at the employer’s motives or legal form, and distilled it into the triple test that has organised the subject since: there must be systematic activity; the activity must be organised by cooperation between employer and employee; and its object must be the production or distribution of goods or services calculated to satisfy human wants and wishes, not being wants or wishes that are merely spiritual or religious in nature. Capital investment is unnecessary; a profit motive is unnecessary; the identity of the employer, state, municipal, charitable or private, is not decisive. What matters is the organised cooperation of labour toward the satisfaction of human wants, and where that is found, an industry is found. Two carve-outs bounded the test: activities that are sovereign functions of the state in the strict sense, of the kind that are inalienably governmental, lie outside, as do activities so casual, isolated or small in scale that the employer-employee cooperation the test requires is absent, the judgment instancing the tiny office of an individual professional and domestic service in a household. The consequences were stated without flinching: hospitals are industries, educational institutions are industries, research institutions and charitable bodies organised on employer-employee lines are industries, government departments performing non-sovereign services are industries, and clubs and professional firms of real scale are industries. Earlier authority in the way, Safdarjung and the University of Delhi line included, was overruled or recast. The bench was equally explicit that a definition of this breadth was for Parliament to settle if Parliament wanted it narrower: the opinions record the Court’s awareness that it was laying down a provisional settlement, and the invitation to legislate was part of the judgment’s structure.

What is the triple test for industry in Bangalore Water Supply?

The triple test asks three things: that the activity be systematic; that it be organised by cooperation between an employer and employees; and that it aim at producing or distributing goods or services calculated to satisfy human wants or wishes, other than wants that are merely spiritual or religious. Profit and capital are irrelevant; sovereign functions fall outside.

The judgment’s internal architecture explains some of its later vulnerability. Seven judges produced a sheaf of separate opinions rather than a single voice: the leading judgment of Krishna Iyer J. carried four colleagues with it in substance, Chandrachud J. concurred on his own reasoning, Chief Justice Beg dissented, and Bhagwati J. differed in part. The majority was candid about the exercise it was performing, describing its construction as reaching the widest amplitude the words would bear and the scheme would discipline, and it expressly contemplated that Parliament might legislate the boundaries differently, as Parliament then attempted to do. Two features of that candour mattered later. The first is that a test announced by a fractured bench, with its author acknowledging its provisional, policy-laden character, invited a later bench to treat reconsideration as a continuation of the original conversation rather than an overruling of settled law, which is exactly the use the five-judge bench made of the 1978 papers in 2005. The second is that the triple test’s very breadth displaced the reasoning into the exceptions: because almost everything systematic and cooperative qualified, the real law of industry after 1978 is found in the margins, in sovereign functions, in the dominant nature of mixed establishments, in the treatment of minuscule and purely spiritual activities, and the thousands of reported decisions of the following decades are overwhelmingly margin cases, hospitals and research institutes, clubs and trusts, departments and missions, each asking the forums to run the same three questions over a new organisational form. When the nine-judge bench closed the reference in 2026 by keeping the test for the repealed Act’s docket while freeing the successor Code’s definition from it, it was in substance ratifying that division of labour: the 1978 test as the law of the old world’s margins, and a fresh start, unanchored, for the new.

What the wide definition did to public employment

The widest consequences of Bangalore Water Supply were felt not in the private factory but inside the state itself. Once government activity of a systematic, service-providing kind counted as industry, the daily-wage, muster-roll and work-charged employees of public works, irrigation, forest, water supply and municipal departments acquired access to a forum their service rules had never contemplated, and they used it. The litigation that followed ran on two tracks that eventually collided. On the first track, the forums entertained the termination disputes of departmental daily-wagers as workmen, ordering reinstatement and continuity where the engagement had ripened into long service, and departments found their casual labour practices audited by tribunals applying Chapter VA to the state as employer. On the second track, the constitutional discipline of public employment asserted itself: engagement to public posts must follow the recruitment rules and the equality guarantee, appointments made through the back door could not be converted into permanence by length of service alone, and in Secretary, State of Karnataka v Umadevi in 2006 a Constitution Bench held that courts and tribunals could not direct the regularisation of illegally appointed employees merely because they had served long, permitting only a narrow, one-time regularisation of irregular appointments against sanctioned posts. The collision was real and never fully harmonised: the labour forums, applying a welfare statute, continued to protect long-serving workmen against termination without the Chapter VA conditions, while the service jurisprudence refused the further step of permanence the protection seemed to promise, and a generation of government daily-wagers lived in the gap, secure against arbitrary termination and insecure in everything else. The sovereign functions exception was fought over in the same cases, department by department: irrigation and social forestry were held industries in one line of authority and state functions in another, and it was precisely that unresolved department-by-department warfare, visible in the Jagannath Maruti Kondhare and Pratamsingh Parmar conflict, that the five-judge bench used in 2005 to justify reopening the definition itself.

The amendment that governs by not existing

Parliament accepted the invitation with unusual speed, and then the executive declined to deliver what Parliament enacted. The Industrial Disputes (Amendment) Act 1982, Act 46 of 1982, was a large amending statute: it inserted the Fifth Schedule and the unfair labour practice provisions, inserted Section 17B on wages during appeals, substituted Section 25O on closure permission, lowered the Chapter VB threshold from three hundred to one hundred, and, in clause (c) of its Section 2, substituted a new definition of industry. The new definition deserves to be read for what it was: not a repudiation of the triple test but a codification of it with a fence. It enacted the test’s own language, speaking of systematic activity carried on by cooperation between an employer and his workmen for the production, supply or distribution of goods or services with a view to satisfying human wants or wishes not merely spiritual or religious, whether or not capital is invested and whether or not the activity is carried on with a motive of gain or profit, and it expressly included the activities of the Dock Labour Board and activity relating to the promotion of sales or business. It then excluded nine categories: agricultural operations, except integrated operations in which the other activity predominates; hospitals or dispensaries; educational, scientific, research or training institutions; institutions owned or managed by organisations wholly or substantially engaged in charitable, social or philanthropic service; khadi or village industries; activity of the government relatable to its sovereign functions, including the departments of the Centre dealing with defence research, atomic energy and space; domestic service; a profession practised by an individual or a body of individuals employing fewer than ten persons; and an activity of a co-operative society, club or similar body employing fewer than ten persons. A reader comparing the two definitions side by side, as anyone working through a downloaded copy of the amending Act in the ReportMedic office file viewer can readily do, sees the design at once: the Court’s principle kept, the Court’s conclusions about hospitals and universities reversed by name.

The amending Act’s commencement clause, Section 1(2), left commencement to notification by the Central Government, with power to appoint different dates for different provisions. On 21 August 1984 the government brought almost the whole Act into force by notification, including the threshold reduction, the unfair labour practices and Section 17B. Clause (c) of Section 2, the new definition, was not in the notification. It has never been in any notification since. The legal consequence is exact and is the point most commentary blurs: the 1982 definition never became law in force, the 1947 definition never ceased to be the text, and the courts, asked to apply the text, applied it as construed in 1978, since a definition enacted but not commenced has no legal operation and cannot displace a judicial construction of the definition actually in force. The episode is the centre of this series’ account of non-commencement, and it has a dedicated treatment in the article on the labour codes’ stalled commencement and the wider pattern of non-commencement, where this instance belongs as the pattern’s most consequential specimen.

Why was it withheld? The explanation given to the courts, and repeated by the Union across the decades of litigation, was that the excluded categories could not be left without any forum: the employees of hospitals, educational institutions and charitable bodies, excluded from the Act by the new definition, would lose the dispute machinery without acquiring a substitute, and the new definition was therefore to be commenced together with a separate law providing grievance machinery for the excluded institutions. The companion promise failed on its own terms: a Bill for the redressal of grievances of employees of hospitals and other institutions was introduced in 1988 and was dropped in the face of union and worker resistance, and no successor machinery was enacted. The brief for this article directs that the government’s stated reason be taken at face value, and it should be: leaving lakhs of hospital and university employees forum-less by a notification was a real harm, and a government that declined to inflict it was not acting absurdly. What cannot be defended is the sequel, in which the interim arrangement hardened into permanence. For forty-four years, through governments of every persuasion, the commencement decision was simply never taken, the separate law never built, and the litigation burden the 2005 bench described fell on forums and parties who had no say in the matter. Non-commencement here did not delay a legislative decision. It reversed one, silently, for two generations, and it did so in the area of law where the reversed decision concerned the jurisdiction of the deciding courts themselves.

The long reference: 1998 to 2026

The courts, left to apply a definition Parliament had voted to narrow and governments refused to narrow, did not sit quietly. In Coir Board v Indira Devi in 1998, a bench of two judges observed that the Bangalore Water Supply ruling had, in their assessment, done more damage than good, not merely to organisations but to employees whose employment opportunities it curtailed, and referred the question to a larger bench; a bench of three judges declined the reference, holding that a bench of two could not seek reconsideration of a decision of seven. The immediate conflict that finally moved the machinery arose over social forestry: a bench of three in Chief Conservator of Forests v Jagannath Maruti Kondhare had held a state social forestry department to be an industry under the 1978 test, while a bench of two in State of Gujarat v Pratamsingh Narsinh Parmar had taken a different view of a similar department. In State of Uttar Pradesh v Jai Bir Singh, a Constitution Bench of five judges, by judgment of 5 May 2005, confronted the accumulation: the 1978 decision had not been unanimous in its reasoning; the majority in 1978 had itself treated its construction as provisional pending legislative intervention; the legislative intervention had been enacted and then withheld for more than two decades; and the expansive reading had, in the bench’s phrase as the case is summarised in the Court’s own later proceedings, produced a docket explosion in the labour forums, as every government scheme and welfare activity became arguable industry. The five judges referred the interpretation of Section 2(j) to a larger bench for reconsideration. In 2017 a bench of seven directed that the reference be placed before a bench of nine, as the reconsideration of a seven-judge decision required. There the matter rested for a further nine years.

The bench of nine, headed by the Chief Justice of India, heard the reference over three days in March 2026 and reserved judgment on 19 March 2026. Judgment was delivered on 20 August 2026. The opinions disagreed, in ways the article must report carefully, on whether the reference had been necessary at all and on whether the triple test should be recast: the Chief Justice’s opinion proposed a reformulated version of the test, in terms that presented it expressly as a hypothesis and left the existing law undisturbed, while the balance of the opinions that addressed the merits held that the 1978 interpretation called for no reconsideration and remained the correct construction of Section 2(j). On the two questions that decide what happens in practice, the Court spoke with one voice. Pending disputes arising under the 1947 Act, which by the date of judgment had been repealed, continue to be decided under the Bangalore Water Supply test, undisturbed. And the definition of industry enacted in Section 2(p) of the Industrial Relations Code 2020 is to be interpreted independently, on its own text and scheme; the 1978 judgment is not the sheet anchor of the new Code. The reference thus closed the Act’s doctrinal story with a paradox that fits it: the great reconsideration, five decades in the making, changed nothing for the statute it was about, because by the time it was answered the statute itself had been replaced, and its meaning had become a legacy question for pending cases. The table that follows is the definitional timeline in one view.

Stage	What happened	The operative definition at that stage
1947	The Act is enacted with Section 2(j) in its original spare form	The 1947 text: any business, trade, undertaking, manufacture or calling of employers, and any calling, service, employment, handicraft or industrial occupation or avocation of workmen
1953 to 1970	The first readings: hospitals and municipalities held in, a university and one hospital line held out	The courts divided case by case, with no organising test
21 February 1978	Bangalore Water Supply decided by seven judges	The triple test governs; profit irrelevant; only sovereign functions and activities too small for organised cooperation fall outside
1982	Parliament enacts Act 46 of 1982, substituting a narrower definition with nine exclusions	Unchanged in practice: the substituted clause awaits a commencement notification
21 August 1984	The rest of the 1982 Act is brought into force; clause (c) of Section 2 is withheld	The 1947 text as construed in 1978 continues to govern
5 May 2005	A bench of five in Jai Bir Singh doubts the 1978 reading and refers it to a larger bench	The 1978 test continues, under a formal reference
2017	A bench of seven directs the reference to a bench of nine	The 1978 test continues; the bench of nine is not heard for nine more years
2020	The Industrial Relations Code enacts its own definition in Section 2(p)	The Code’s triple-test core with named exclusions exists on paper beside the operative 1978 test
21 November 2025	The labour codes are brought into force and the 1947 Act is repealed, with savings	Pending disputes remain under the 1978 test; new disputes fall to the Code
20 August 2026	The bench of nine delivers judgment in Jai Bir Singh	The 1978 test is left undisturbed for the repealed Act; the Code’s definition is to be read independently

After the award: the writ courts and the absence of an appeal

The Act provides no appeal from an award, and that omission, deliberate in a scheme designed for speed, generated the system’s characteristic second tier: the writ jurisdiction of the High Courts under Article 226, and of the Supreme Court under Article 136, through which awards have been reviewed for two-thirds of a century. The grounds were settled early and held with reasonable consistency. A High Court does not sit in appeal over a tribunal; it corrects errors of law apparent on the face of the record, findings of fact so unsupported by evidence as to be perverse, violations of natural justice, and exercises of jurisdiction the tribunal does not possess, including travelling beyond the terms of its reference, and it leaves findings of fact supported by some evidence alone even where it would itself have weighed the evidence differently. Dismissal jurisprudence under Section 11A was repeatedly disciplined by that boundary, the Supreme Court reminding forums and High Courts in turn that reappraisal is a power to be exercised on the record’s defects rather than a licence to substitute sentencing instincts. The practical consequence was a two-speed system: the tribunal decided, and then the real litigation frequently began, with awards of reinstatement suspended by stay orders while writs aged, so that a workman who had won everything below might wait out the remaining years of his service for the enforcement of it, and the final outcomes were shaped as much by interim orders as by final judgments.

The writ tier also decided the system’s boundary questions. Whether a dispute existed at all, whether the claimant was a workman, whether the establishment was an industry, whether the appropriate government had acted on relevant material, whether a closure was genuine or a device, the constitutional attacks on Chapters VA and VB, were all ultimately pronounced upon not by the industrial forums but by the constitutional courts, and the treatises on the Act are, in consequence, treatises on Supreme Court and High Court case law with the statute as their skeleton. For the pillar reader the point is diagnostic: a statute that makes its forums final invites its superior courts to become its legislature of detail, and the Industrial Disputes Act’s enormous case law, thousands of reported decisions elaborating every phrase of Section 2, is the predictable product of final forums, no appeal, and definitions drafted short. The Industrial Relations Code preserves the design in this respect, tribunals with judicial and administrative members, awards without a statutory appeal, and review left to the writ courts, so that the successor’s case law, when it accumulates, will grow in the same soil.

The enforcement tail: penalties, prosecution and the fine print

The Act’s penal scheme, in Sections 26 to 31, is the clearest surviving evidence of its 1947 vintage, and an honest pillar treatment should show the numbers rather than paraphrase them away. A workman who commences, continues or acts in furtherance of an illegal strike is punishable with imprisonment up to one month, or with a fine up to fifty rupees, or both; an employer who commences or continues an illegal lockout faces imprisonment up to one month, or a fine up to one thousand rupees, or both; financial support of an illegal strike or lockout draws up to six months or one thousand rupees. Breach of a settlement or award binding under the Act is punishable with imprisonment up to six months, or fine, or both, with an additional daily fine for a continuing breach, and commits the further civil consequence that the breach itself becomes a fresh industrial dispute. An unfair labour practice carries imprisonment up to six months, or a fine up to one thousand rupees, or both, under Section 30; contravention of Section 33 carries the same scale under Section 31; and a residuary offence, any contravention for which no other penalty is provided, is punishable with a fine up to one hundred rupees. Prosecutions are filtered through Section 34: no court may take cognizance of an offence under the Act except on a complaint made by or under the authority of the appropriate government, and no court inferior to a metropolitan magistrate or a judicial magistrate of the first class may try one.

The enforcement design that emerges is characteristic of the whole statute. The real sanctions were never the fines, fossilised at 1947 price levels and never revised upward with any seriousness; they were the civil consequences the Act attaches automatically, the illegality of the strike, the approval jurisdiction over dismissals, the recovery of money as arrears of land revenue, the liability for compensation when permission procedures are ignored, and the prospect of the dispute machinery itself being turned against the defaulter. Section 29’s conversion of a breached settlement into a fresh dispute, and Section 33C’s executive recovery, did more work in a month than the penal code of the Act did in a decade. A reader of the Act who stops at its sanctions chapter would misunderstand it entirely: this is a statute that enforces itself chiefly by controlling the parties’ next moves, not by punishing their last ones, and its fines are best read as archaeology, the price list of industrial misconduct as a colonial-era draftsman expected it in 1947.

The Act in operation: delay, recognition and the unprotected majority

Three operational facts complete the picture, each of them a standing criticism that any honest pillar treatment must record. The first is delay. The Act’s paper timelines, fourteen days for a conciliation report, expeditious disposal by the forums, were not the lived ones: references aged in secretariats, tribunal proceedings ran for years and sometimes for decades, awards were challenged in the High Courts under the writ jurisdiction and in the Supreme Court beyond them, and the parties’ real remedy was frequently exhaustion. Delay was not a neutral defect. It systematically favoured the party who benefited from the status quo, which in a dismissal dispute is the employer holding the job and in a wage dispute is usually the employer holding the money, and Section 17B was Parliament’s partial admission of that asymmetry. The forums’ dockets, dominated numerically by individual dismissal and allied disputes rather than by the great collective questions, filled with the small change of industrial discipline, and the system’s capacity to decide a major industry-wide reference was correspondingly thin.

The second fact is the recognition gap. The Act regulates disputes involving unions at every turn, presupposing unions as the workman’s collective voice, yet the central Act contains no machinery for recognising a union as the bargaining agent of an establishment, no certification, no ballot, no statutory duty to bargain with the representative union. Recognition was left to voluntary codes and to state legislation, of which the Maharashtra Recognition of Trade Unions and Prevention of Unfair Labour Practices Act 1971 is the most elaborate, so that in most of the country the preliminary question, who speaks for the workmen, had to be fought before the substantive question could be reached, and rival unionism became an industrial condition rather than an accident. The Industrial Relations Code’s negotiating union and negotiating council provisions, recognising a union with fifty-one per cent support, or a council of unions where none reaches it, are the successor’s answer to a gap the 1947 Act left open for its entire life.

The third fact is coverage. The Act’s world, for all the breadth Bangalore Water Supply gave it, was the organised sector: establishments with identifiable employers, muster rolls, standing orders and unions. The Indian workforce’s great majority, in agriculture, in tiny enterprises, in casual and home-based and self-employed work, stood outside the definitions in practice, unable to point to an employer against whom a dispute could be raised or a forum that would take their case. The Act therefore governed the most protected stratum of Indian labour with an intensity that coexisted, without contradiction, with the near-total absence of statutory protection for the strata beneath. Every debate about the Act’s rigidity is silently a debate about that stratum’s experience generalised to the economy, and every debate about its necessity is a debate about the same stratum as the thin edge of a protection that ought to have spread. The codes’ expansion of definitions to gig and platform work belongs to the successor’s story; the 1947 Act’s story is that it built a cathedral of labour justice on a narrow foundation, and argued for seventy years about the architecture while the foundation stayed narrow.

The succession: the Industrial Relations Code 2020

The Industrial Relations Code 2020, Act 35 of 2020, which received the President’s assent on 28 September 2020, subsumed three statutes: this Act, the Trade Unions Act 1926 and the Industrial Employment (Standing Orders) Act 1946. It is the successor statute, and its treatment of the 1947 design is selective inheritance rather than rejection. From the Act it takes the definitions architecture, the conciliation-first sequence, the layoff and retrenchment compensation formulas at unchanged rates of fifteen days’ average pay per completed year, and the permission principle itself, which it retains for layoff, retrenchment and closure. What it changes is the settings and the doors. The permission threshold rises from one hundred to three hundred workers in a factory, mine or plantation, with power in the appropriate government to raise it further by notification, and the retrenchment bargain adds a worker re-skilling fund, to which the employer contributes an amount equal to fifteen days’ wages last drawn by each retrenched worker, credited to the worker’s account within forty-five days. The standing orders regime, under which establishments must have certified rules of conduct and classification, is pitched at three hundred workers as well. The reference gate is dismantled for the generality of disputes: after conciliation fails, a party may apply to the industrial tribunal directly, and the multiplicity of labour courts, boards and courts of inquiry gives way to a single tribunal system, with a national tribunal for disputes of national importance or spanning states. Notice of strike and lockout, fourteen days, given in advance and effective within a sixty-day window, is extended to every industrial establishment, ending the public utility distinction on which Sections 22 and 23 of the 1947 Act turned. Fixed term employment is recognised expressly, with parity of benefits and gratuity, and its expiry is placed outside retrenchment.

On the definition that consumed this Act, the Code did what the executive of 1982 would not: it commenced a fenced definition. Section 2(p) of the Code enacts the triple test’s core, speaking of systematic activity carried on by cooperation between an employer and workers for the production, supply or distribution of goods or services with a view to satisfying human wants or wishes not merely spiritual or religious, and then excludes institutions owned or managed by organisations wholly or substantially engaged in charitable, social or philanthropic service, activity of the government relatable to sovereign functions including the departments dealing with defence, atomic energy and space, domestic service, and such other activities as the Central Government may notify. The quarrel of 1982 is visible in the text: Parliament in 2020 adopted the substance of the withheld amendment, trimmed to fewer exclusions, and this time the commencement followed. The four labour codes were brought into force on 21 November 2025, and the Industrial Disputes Act stood repealed from that date by Section 104 of the Code, with the standard savings: proceedings, references and awards under the repealed Act continue as if it had not been repealed, and acts done under it are carried into the corresponding provisions of the Code. Commencement arrived, as this series records elsewhere, ahead of the complete body of central and state rules, so the succession runs for a period on savings clauses, continuing instruments and the forums’ accumulated practice. The full analysis of the successor, its negotiating union machinery, its tribunal design and its own unsettled questions, belongs to the pillar treatment of the Industrial Relations Code itself, and this article does not trespass on it.

Two further Code changes complete the comparison, because they recalibrate the 1947 Act’s two other pressure points. The Code’s definition of worker, the successor to workman, raises the supervisory wage ceiling that marks the upper edge of protection to eighteen thousand rupees a month, where the 1947 Act’s figure stood at ten thousand, and recasts the exclusions in the same recognisable shape, managerial and administrative capacity out, higher-paid supervision out, the substance of the employment deciding. And the Code absorbs the standing orders regime into itself, applying it to establishments with three hundred or more workers, where the 1946 sibling statute had applied on a much lower threshold, so that the rulebook jurisdiction described earlier in this account survives, attached to larger establishments only. Both changes repeat the Code’s characteristic gesture: the 1947 architecture retained, the thresholds lifted, the coverage narrowed at exactly the points where the older law had pressed hardest on the medium-sized employer.

The transition: what repeal does to a living docket

A statute that has run for seventy-eight years does not stop on a date; it unwinds. The repeal of the 1947 Act by the Code, with effect from the codes’ commencement on 21 November 2025, is governed by the savings in Section 104 of the Code and by the general savings doctrine, and its mechanics deserve a paragraph of precision in a pillar treatment because they decide live cases. Rights, privileges, obligations and liabilities acquired or incurred under the repealed Act are preserved; proceedings and remedies in respect of them may be instituted, continued and enforced as if the repealing Act had not been passed; and anything done under the repealed enactment, appointments of conciliation officers and forums, references made, settlements recorded, awards published, is carried into the corresponding provisions of the Code so far as it is consistent with them. The practical consequences sort themselves into three streams. Disputes and references already inside the old machinery continue there and are decided by the old law, which is why the nine-judge bench’s answer in 2026, confining itself to the repealed Act and its pending docket, has real work to do for years. Settlements and awards in operation on the commencement date continue for their terms, binding the parties under the law that made them. And disputes arising after commencement, including disputes about terminations that occur after it, are raised under the Code’s own doors, before its tribunals, on its definitions, with its notice rules, even where the establishment’s habits, standing orders and union arrangements were all formed under the Act. The boundary disputes that result, over which regime governs a cause of action that straddles the commencement date, over the fate of a reference made but not yet heard, over whether a settlement under the old Act can be reopened under the new, are the predictable first generation of Code litigation, and they are governed less by the Code’s text than by the savings architecture and the courts’ standard hostility to the retrospective disturbance of accrued rights. An Act’s afterlife, in short, is part of the Act, and the Industrial Disputes Act, which spent forty-four years haunted by an uncommenced amendment, has begun an afterlife haunted in the opposite direction: fully repealed, yet still deciding cases.

Assessment: what the Act did, and what it could not do

A pillar treatment owes the reader a verdict, and the verdict on this Act has to be given in parts because the Act was several instruments wearing one title. As a dispute settlement machine, it worked, in the plain sense that it processed industrial conflict through forums applying law for over seven decades, generated a body of doctrine on dismissal, wages, bonus and discipline that remains the grammar of Indian employment practice, and gave individual workmen, through Section 2A and Section 11A, a remedy against arbitrary dismissal that no other statute of its era attempted. As a definition statute, it failed at its centre and knew it: the word on which its jurisdiction depended was settled by judges in 1978, unsettled by Parliament in 1982, left unsettled by the executive for forty-four years, and finally mooted by replacement in 2025 and answered, for legacy purposes, in 2026. As an instrument of economic policy, its Chapter VB was the most consequential experiment in job security legislation in the democratic world, and the record supports neither the claim that it destroyed Indian manufacturing nor the claim that it cost nothing: it priced and policed exit in a segment of the formal economy, pushed employment growth toward forms the Act could not reach, and protected, genuinely and sometimes desperately, the workers inside the segment it could reach. As a piece of constitutional practice, it is the standing exhibit on commencement power: everything else in the Act was amended, litigated and administered in the open, while its central definition was governed by an omission.

The through-line, for this series, is the notification gap. The series has catalogued statutes never notified, provisions notified selectively, and codes notified before their machinery existed. The Industrial Disputes Act supplies the limiting case: a provision whose non-notification did not merely postpone a policy but inverted one, maintaining in force, for the working life of two generations of workers, a legal rule that the legislature had considered, rejected and replaced on paper. That this happened without scandal, in the plain sight of the statute book, where the uncommenced clause sits printed in every edition of the Act with a footnote recording that it awaits its day, is the deepest fact in this article. Law students meet the footnote early and are rarely told what it cost. It cost the coherence of the definition of industry for forty-four years, and it is the reason a bench of nine judges of the Supreme Court was still, in 2026, answering a question that Parliament had answered in 1982.

Frequently Asked Questions

Q: What did the Industrial Disputes Act 1947 govern?

It governed how industrial disputes in India were investigated, conciliated and adjudicated, and it regulated the conduct surrounding them. Its core subject was the industrial dispute: any dispute between employers and workmen, or between workmen themselves, connected with employment, non-employment, terms of employment or conditions of labour. Around that core it built conciliation and adjudication machinery, rules on strikes and lockouts, protection of conditions of service while proceedings were pending, a compensation code for layoff, retrenchment and closure in Chapter VA, and a permission regime for larger establishments in Chapter VB. It applied wherever the contested definitions of industry and workman were satisfied, which is why those two words decided its practical reach.

Q: What counts as an industry under the Industrial Disputes Act?

Under the test laid down in Bangalore Water Supply in 1978, an industry is any systematic activity organised by cooperation between an employer and employees for the production or distribution of goods or services calculated to satisfy human wants or wishes, other than wants that are merely spiritual or religious. Profit motive and capital investment are irrelevant, and the employer’s identity is not decisive, so hospitals, universities, research bodies, charities organised on employment lines, municipal services and non-sovereign government departments all qualified. Only activities that are sovereign functions of the state in the strict sense, and activities too small or casual to involve organised cooperation, fall outside. That judicial test, and not the narrower definition Parliament enacted in 1982, was the operative law throughout the Act’s life.

Q: Why was the 1982 amendment to the definition of industry never notified?

The Industrial Disputes (Amendment) Act 1982 substituted a narrower definition of industry, excluding hospitals, educational and research institutions, charitable bodies, khadi and village industries, domestic service and sovereign functions. Its commencement clause left the date to the Central Government. When the rest of the Act was brought into force on 21 August 1984, the definition clause was withheld, and it was never notified in the forty-four years that followed. The explanation the Union gave the courts was that the excluded institutions’ employees would be left with no forum for their grievances, and that the new definition was to commence together with a separate law providing alternative machinery. That separate law, attempted in a Bill in 1988, was dropped, and the commencement decision was simply never taken.

Q: What is Chapter VB and why was it so contested?

Chapter VB, inserted in 1976 and widened in 1984, is the part of the Act that required establishments employing one hundred or more workmen in a factory, mine or plantation to obtain the prior permission of the appropriate government before laying off workmen, retrenching them, or closing the undertaking. It was contested because it made exit from a business a matter of executive discretion. Employers argued that it trapped capital in unviable units, deterred establishments from growing past the threshold, and pushed hiring into contract and casual forms outside the Act’s protection. Unions and worker advocates answered that, in an economy without unemployment insurance, permission was the only real check on closure of viable units and on retrenchment used to break organisation. Both arguments are about the same provision, and both are grounded in its actual operation.

Q: What conditions must an employer meet to retrench a workman?

Under Section 25F, a workman who has completed not less than one year of continuous service cannot be retrenched until three conditions are met. First, the workman must be given one month’s notice in writing stating the reasons for retrenchment, and the notice period must expire, or wages for the notice period must be paid in lieu. Second, the workman must be paid, at the time of retrenchment, compensation equal to fifteen days’ average pay for every completed year of continuous service, or any part of a year exceeding six months. Third, notice in the prescribed manner must be served on the appropriate government or its specified authority. Retrenchment without these conditions is void, and in establishments covered by Chapter VB the additional requirement of prior government permission applies on top.

Q: Who decides whether an industrial dispute goes to a tribunal?

For most of the Act’s life, the appropriate government decided. Under Section 10, the government formed the opinion whether an industrial dispute existed or was apprehended and referred it to a labour court, a tribunal or a national tribunal, and its refusal to refer was open to challenge only on narrow administrative law grounds. The main exception was a joint or majority application by both parties, where the reference had to be made. Conciliation came first, and the conciliation officer’s failure report informed the government’s decision. Individual dismissal disputes could be carried by the workman through the conciliation route under Section 2A, but the adjudication itself began by government reference, not by the parties’ filing, until the Industrial Relations Code replaced the reference with direct application to the tribunal.

Q: Who is a workman under the Industrial Disputes Act?

A workman, under Section 2(s), is a person, including an apprentice, employed in any industry to do manual, unskilled, skilled, technical, operational, clerical or supervisory work for hire or reward, and it includes a person dismissed, discharged or retrenched in connection with a dispute for the purposes of that dispute. The definition excludes members of the armed forces, police and prison staff, persons employed mainly in a managerial or administrative capacity, and supervisors who draw wages above the statutory ceiling, ten thousand rupees a month since 2010, and exercise mainly managerial functions. The decisive test is the dominant nature of the duties actually performed, not the job title, so coverage is fought case by case on what the employee really does all day.

Q: When can a strike be declared illegal?

A strike is illegal under Section 24 if it is commenced or declared in contravention of the notice rules for public utility services in Section 22, or of the general prohibitions in Section 23, or if it is continued in contravention of an order prohibiting its continuance made when the dispute is referred. Section 23 prohibits strikes during the pendency of conciliation before a board, during proceedings before a labour court or tribunal and for two months after they conclude, during arbitration proceedings covered by a notification, and during the operation of a settlement or award on the matters it covers. Legality and justification are separate questions: a strike may be illegal yet provoked, and legal yet unjustified, and the distinction affects wages and discipline rather than the strike’s legal character.

Q: What is the difference between layoff, retrenchment and closure?

The three are different depths of separation. Layoff is temporary: the employer, for reasons such as shortage of power or raw materials, breakdown of machinery or accumulation of stocks, fails or is unable to give employment to workmen whose names remain on the muster rolls, and the employment relationship subsists while compensation of fifty per cent of basic wages and dearness allowance is paid for the layoff days. Retrenchment is the permanent termination of a workman’s service for reasons other than punishment, voluntary retirement, superannuation, expiry of contract or continued ill-health, and it attracts the notice and compensation conditions of Section 25F. Closure is the permanent shutting of the undertaking or a part of it, ending the employment of all its workmen, with compensation computed as if they had been retrenched.

Q: How is retrenchment compensation calculated under the Industrial Disputes Act?

The statutory formula is fifteen days’ average pay for every completed year of continuous service, plus fifteen days’ average pay for any remaining part of a year that exceeds six months. Continuous service is measured under Section 25B, under which a workman who has actually worked for at least two hundred and forty days in the preceding twelve months, one hundred and ninety days for work below ground in a mine, is deemed to have completed one year, and interruptions such as authorised leave, sickness, accident, a strike that is not illegal, a lockout or stoppage not due to the workman’s fault do not break continuity. So a workman with eleven years and eight months of qualifying service is paid for twelve units, one hundred and eighty days of average pay in all.

Q: Can one dismissed workman raise an industrial dispute on his own?

Yes. Before 1965, an individual grievance generally had to be espoused by a union or a substantial body of workmen before it counted as an industrial dispute, which left a lone dismissed workman dependent on finding a sponsor. Section 2A, inserted in 1965, reversed that position for the commonest case: where an employer discharges, dismisses, retrenches or otherwise terminates the services of an individual workman, the dispute between that workman and the employer connected with the termination is deemed to be an industrial dispute even if no other workman or union is a party to it. The workman can therefore carry a dismissal dispute into conciliation personally, subject to the limitation period for making the application, which is three years from the termination with a power to condone delay.

Q: Did the Supreme Court overrule the Bangalore Water Supply case?

No. A bench of nine judges, answering the reference made in State of Uttar Pradesh v Jai Bir Singh in 2005, delivered judgment on 20 August 2026 without displacing the 1978 ruling for the statute it interpreted. The opinions differed on whether the reference had been necessary and on whether the triple test should be recast; a reformulated version of the test was proposed in the Chief Justice’s opinion expressly as a hypothesis that left the existing law undisturbed. All the judges agreed that pending disputes under the Industrial Disputes Act, repealed by then, continue under the Bangalore Water Supply test, and that the definition of industry in the Industrial Relations Code is to be interpreted independently, not through the 1978 precedent.

Q: Can an employer dismiss a workman while a dispute is pending?

Only within the limits of Section 33. While proceedings are pending before a conciliation officer, board, labour court or tribunal, the employer cannot alter a workman’s conditions of service to his prejudice in regard to any matter connected with the dispute, and cannot discharge or punish him for misconduct connected with the dispute, without the express written permission of the authority. For matters unconnected with the dispute, the employer may act under the standing orders or contract, but a dismissal requires an application for approval and payment of one month’s wages. A protected workman, a recognised union office bearer, cannot be dismissed or punished on any matter during pendency without express permission. A dismissal in contravention can be challenged by complaint under Section 33A, and the forum decides it as a dispute.

Q: What does Section 11A of the Industrial Disputes Act allow a tribunal to do?

Section 11A, in force from 15 December 1971, empowers a labour court, tribunal or national tribunal, in a referred dispute over the discharge or dismissal of a workman, to set aside the dismissal if it is satisfied the order was not justified, and to direct reinstatement on terms it thinks fit, or to give other relief, including the award of a lesser punishment. It replaced the earlier narrow doctrine under which the forum could interfere only for mala fides, victimisation, unfair labour practice, basic error or perversity. Under Section 11A the forum reappraises the evidence itself, may receive fresh material, and judges whether the punishment is proportionate to the misconduct proved. Reinstatement with back wages is a discretionary outcome, not an automatic one, and forums increasingly substitute compensation where reinstatement is impractical or the appointment was irregular.

Q: What counts as an unfair labour practice under the Industrial Disputes Act?

The Fifth Schedule, read with Sections 25T and 25U, lists unfair labour practices on both sides of industry and prohibits them. On the employer’s side they include interfering with or coercing workmen in their right to organise, threatening discharge for joining a union, victimising a workman for urging others to organise, transferring a workman mala fide under the guise of policy, abolishing regular work and handing it to contractors as a strike-breaking measure, keeping workmen as casuals or temporaries for years to deny them permanent status, and refusing to bargain in good faith with a recognised union. On the union and workmen’s side they include supporting an illegal strike, coercive picketing that physically bars entry, acts of force or intimidation against non-striking workmen or managers, and refusal to bargain in good faith.

Q: Has the Industrial Disputes Act 1947 been repealed?

Yes. The Industrial Relations Code 2020, which received assent on 28 September 2020, subsumed the Industrial Disputes Act along with the Trade Unions Act 1926 and the Industrial Employment (Standing Orders) Act 1946. When the labour codes were brought into force on 21 November 2025, Section 104 of the Code repealed the 1947 Act. The repeal carries savings: disputes, references, proceedings and awards begun under the old Act continue and are decided under it, and the nine-judge bench confirmed in August 2026 that pending matters remain governed by the old law, including the Bangalore Water Supply test. References to the Act in older settlements, standing orders and awards therefore remain operative for the matters they govern, even though new disputes are raised under the Code.

Q: What does last come first go mean in retrenchment?

It is the statutory order of retrenchment in Section 25G. Where a workman in an establishment is to be retrenched and belongs to a particular category of workmen, the employer must ordinarily retrench the workman who was the last person employed in that category, unless the employer departs from the rule for reasons recorded. The rule converts seniority into a measure of protection: the longest-serving workmen are retrenched last, and a departure must be justified and recorded rather than unexplained, and it can be challenged before the forum if the reasons are a cover for victimisation. Section 25H completes the scheme from the other side: if the employer later proposes to take on employees again, retrenched workmen who offer themselves must be given preference over fresh applicants.

Q: How many days must a workman work to complete one year of continuous service?

Two hundred and forty days of actual work in the preceding twelve calendar months, under Section 25B, and one hundred and ninety days for a workman employed below ground in a mine. Reaching that count deems the workman to have completed one year of continuous service, which is the qualifying period for layoff compensation and for the retrenchment conditions in Section 25F. Days of interruption are not all lost: service interrupted by sickness, authorised leave, an accident, a strike that is not illegal, a lockout, or a cessation of work not due to the workman’s fault is still continuous service, so the count is of days actually worked combined with interruptions the statute forgives. Muster roll and attendance records decide the question in contested cases.

Q: What is the difference between a settlement and an award under the Act?

A settlement is the parties’ own bargain, recorded under the Act; an award is the decision of the adjudicating forum imposed on them. A settlement arrived at in the course of conciliation proceedings binds all parties to the dispute, all workmen employed in the establishment and those who subsequently join, and the parties’ successors; a settlement reached outside conciliation binds only the parties who made it. An award is the determination of a labour court, tribunal or national tribunal on a referred dispute, becomes enforceable thirty days after its publication by the government, and binds the parties, their heirs and successors and all workmen in the establishment. A settlement runs for the period it states, or six months if it states none; an award ordinarily operates for one year, which the government may extend.

Q: What happens if an employer closes a factory without government permission?

In an establishment covered by Chapter VB, closure without the prior permission required by Section 25O, or closure after permission has been refused, is illegal. The statutory consequence is that the workmen are entitled to all the benefits under any law for the time being in force as if the undertaking had not been closed down, which in practice means the closure is treated as not having happened for the purposes of their service and wages until a lawful closure occurs. The application itself must be made at least ninety days before the intended closure date, stating the reasons; the government must consider the genuineness and adequacy of those reasons, the interests of the general public and other relevant factors, and if it does not communicate its order within sixty days, permission is deemed to have been granted.