The standard account of Indian liberalisation is told in economic language: growth rates, trade shares, the price of goods in shops. Told in legal language, the same period looks different and stranger. The licence raj, a regime of control that had been built statute by statute over four decades, was dismantled between 1991 and 2011, yet very little of the dismantling was done by the method a law student would expect. Parliament did not assemble in July 1991 and repeal the architecture of control. Instead, a statement of industrial policy read out in Parliament on 24 July 1991, and a series of notifications issued under powers the control statutes themselves contained, ended the controls in practice. The statute book was then adjusted to match the new reality, slowly, unevenly, and in the most emblematic case not at all. The Industries (Development and Regulation) Act, 1951, the statute that created industrial licensing, has never been repealed. It stands in force, its licensing chapters largely disapplied by exemption, nearly three decades after licensing itself ended.

That gap, between the moment a law ceases to be applied and the moment it ceases to exist, is the subject of this article. The series of which this article forms part has repeatedly found gaps of this family: between enactment and commencement, between a judgment and its practical effect, between a rule and its parent Act. The liberalisation sequence is the largest instance of all, because it was not an accident of drafting or administration. It was the method. Every substantive liberalisation of the period was effected first by executive act, and the corresponding statutory change followed afterwards at intervals ranging from a few months to eighteen years. One statute, the capital issues control statute, breaks the pattern and was repealed in the same season its function was transferred to a new regulator. One statute, the licensing statute, completes the pattern by never being repealed at all.

The article proceeds in the order the evidence demands. It first sets out the control regime as what it legally was: not a mood or a metaphor but a list of statutes, each imposing a consent that a single industrial project might have to obtain from a different office under a different Act. It then examines 24 July 1991 as a legal event and asks what kind of instrument the Statement on Industrial Policy was, what it changed by its own force, and what it could not change. It follows each control through its ending, records the interval between the practical end and the statutory end in a single table, and then turns to the questions the sequence raises: why the courts mostly stood aside, whether a control ended by notification can be revived by notification, and how the two rival assessments of liberalisation, the favourable and the critical, each handle the legal facts. The thesis defended at the end is measurable rather than rhetorical. Of the principal controls, one was ended by prompt repeal, one by early amendment and late repeal, one by replacement after eight years, and the central one by executive withdrawal under a statute that survives. Liberalisation, in statutory terms, was mostly not repeal. It was the executive ceasing to use powers that Parliament had granted and, in the leading case, has never taken back.

Born of war and scarcity: where the control statutes came from

The control regime was not designed in one place or at one time, and its statutes carry the marks of the emergencies that produced them. Two of the principal instruments began as wartime measures. Control over capital issues entered Indian law through the Defence of India framework during the Second World War, when the state rationed access to capital so that private fundraising would not compete with war finance, and it was placed on a permanent footing by the Capital Issues (Continuance of Control) Act and then the Capital Issues (Control) Act of 1947. Control over imports and exports likewise began as a wartime and immediate post-war measure, continued after independence by the Imports and Exports (Control) Act, 1947, initially as a temporary statute whose life Parliament extended again and again until temporariness became a fiction and the control simply formed part of the permanent landscape. The pattern of the era is visible in miniature in those two histories: an emergency power, taken for a defined crisis, retained because the next crisis, scarcity of capital, scarcity of foreign exchange, always seemed to be arriving.

The Industries (Development and Regulation) Act, 1951 belonged to a different register. It was not an emergency measure but a planning measure, the statutory expression of the industrial policy resolutions through which the new republic announced that the commanding heights of the economy would be publicly held and that private industry would develop within a framework of state direction. Licensing was its chosen instrument because licensing converts every significant investment decision into a decision the state participates in: what to produce, how much, where, and by whom. The first schedule of industries under the Act was correspondingly broad. The statute also armed the government with powers that went beyond licensing and that matter later in this story: powers to investigate the affairs of an undertaking whose production had fallen or whose management was failing the public interest, powers in extreme cases to take over management, powers to control the distribution and price of specified articles, and, running quietly through the scheme, the exemption power that would eventually switch the licensing system off.

The monopolies statute of 1969 answered a political anxiety of the following decade, that licensing, designed to disperse industrial power, was being used by the largest houses to pre-empt capacity and entrench themselves, obtaining licences to keep rivals out rather than to produce. Parliament’s answer was to add the second gate described below, keyed to size. The foreign exchange statute of 1973 answered the anxiety of its own decade, that foreign exchange was chronically scarce and foreign ownership carried political as well as economic risk; it replaced the first foreign exchange regulation statute of 1947 with a markedly more stringent code, one whose restrictions on foreign shareholding produced the celebrated departures and dilutions of multinational companies in the late 1970s. The land ceiling statute of 1976 answered the urban counterpart of land reform sentiment. The Essential Commodities Act, 1955, the oldest temperament of the set, answered scarcity itself, and its control order machinery was the everyday face of the regime for traders and consumers who never saw an industrial licence.

Two conclusions about origins matter for the ending. The controls were heterogeneous in purpose, planning, anti-concentration, conservation, equity, scarcity management, and they were enacted across thirty years by Parliaments pursuing different projects; there was never a single control code that a reforming Parliament could repeal in a single bill. And each statute was drafted in the confidence of its own era that state direction was permanent, which is why the exemption and relaxation powers scattered through them were framed as safety valves for deserving cases. Nobody drafted Section 29B of the Industries Act, the exemption provision, as the eventual instrument of liberalisation. It became that instrument because it was there, wide enough, and in the hands of an executive in a hurry. The history of the control regime’s end begins in the drafting choices of its beginning.

The licence raj was a list of statutes

The phrase licence raj invites an image of a single office stamping a single form. The legal reality was a bundle of separate regimes, enacted at different times for different stated purposes, whose combined effect on one project was cumulative. A promoter planning a factory in, say, 1985 faced a sequence of consents no one of which was granted by the authority that granted the others, and the failure of any one consent could end the project regardless of the rest. Understanding liberalisation requires holding that bundle in view, because liberalisation did not repeal the bundle. It untied the bundle string by string, at different dates, by different instruments, and two of the strings were never cut at all but merely loosened and left in place.

The first string governed entry and capacity. The Industries (Development and Regulation) Act, 1951 (Act 65 of 1951) required a licence to establish a new industrial undertaking in a scheduled industry, to manufacture a new article, and to expand substantially an existing undertaking. The schedule of industries covered most of organised manufacturing. The licence specified the product, the capacity and the location, so the state controlled not only whether a factory existed but how much it might produce and where. That statute is examined in full in the pillar article on the Industries Act 1951, which owns its text; here it matters as the foundation stone of the bundle.

The second string governed size. The Monopolies and Restrictive Trade Practices Act, 1969 (Act 54 of 1969), in its original form, subjected large undertakings and dominant undertakings to a further layer of prior approval: approval to establish a new undertaking, to expand substantially, to amalgamate with or take over another undertaking. The trigger was asset size and market share rather than the nature of the product. A project could hold an industrial licence and still be stopped at this gate because its promoter was already large. The third string governed foreign participation and foreign exchange. The Foreign Exchange Regulation Act, 1973 (Act 46 of 1973) treated foreign exchange as a scarce public resource to be conserved: most transactions in foreign exchange required permission, dealings were channelled through authorised dealers, and foreign shareholding in Indian companies was restricted, with companies above the prescribed level of foreign equity expected to dilute or to justify their structure under guidelines administered by the Reserve Bank. The fourth string governed access to the savings of the public. The Capital Issues (Control) Act, 1947 (Act 29 of 1947) required the consent of the Controller of Capital Issues for a public issue of securities, and the consent extended to the pricing of the issue, so the state decided not only whether a company might raise capital but at what premium its shares might be sold.

The fifth and sixth strings governed what crossed the border and what could be sold at home. The Imports and Exports (Control) Act, 1947 (Act 18 of 1947) made imports subject to licence, so the machinery and raw materials a licensed factory needed might themselves need a separate permission. The Essential Commodities Act, 1955 (Act 10 of 1955) empowered control orders over the production, distribution and pricing of scheduled commodities, a power exercised through a large body of orders covering everything from foodgrains to fertiliser and drugs. The seventh string governed the site itself: the Urban Land (Ceiling and Regulation) Act, 1976 (Act 33 of 1976) imposed ceilings on vacant urban land holdings, so assembling a factory site in an urban agglomeration could require exemption or surrender of excess land. Around these principal statutes sat further controls of the same family: reservation of products for the small scale sector, effected by notification and at its height covering more than eight hundred items which larger units might not manufacture; price control orders for particular industries; and the phased manufacturing programmes, administrative schemes under which a foreign collaboration was approved on condition that imported components were progressively replaced by domestic manufacture on a timetable fixed by the government.

FERA in operation: the permission economy for foreign exchange

The foreign exchange statute governed a different scarcity from the licensing statute, and its texture in operation explains why its replacement took a full new code rather than an exemption scheme. The licensing Act asked whether a thing should exist. The Foreign Exchange Regulation Act, 1973 asked, of nearly every transaction with a foreign element, whether it should be permitted. Its prohibitions were cast generally and its permissions specifically: dealings in foreign exchange were confined to authorised dealers and money changers; payments to persons resident outside India, transfers of securities to non-residents, the establishment of branches and places of business by foreign companies, and the holding of foreign exchange itself were regulated heads, each administered by the Reserve Bank through directions, manuals and case by case approvals. Contravention was a criminal offence, and the enforcement directorate created under the Act investigated and prosecuted in the manner of a revenue police. The citizen’s default position before the statute was that a foreign exchange transaction was unlawful until permitted, and the burden of bringing the transaction within a permission lay on the citizen.

Foreign shareholding was the head of control that shaped corporate India most visibly. Companies in which the non-resident interest exceeded 40 per cent were, for the Act’s central purposes, placed in a distinct category and subjected to a regime of expectations: that they confine themselves to specified high technology or export oriented activities, or dilute. The dilution exercises of the late 1970s, in which household name subsidiaries of foreign parents reduced their foreign shareholding to Indian majorities by offering shares to the Indian public, were the direct product of this statutory threshold, and the ownership pattern of a generation of Indian companies, foreign branded, Indian owned in majority, professionally managed by the minority foreign partner, is unintelligible without it. The 1991 statement’s automatic route, approval as a matter of course for foreign equity up to 51 per cent in the listed high priority industries, reversed the polarity for the covered sectors: permission ceased to be negotiated and became registered. But the reversal happened inside the old Act’s machinery, by changing the practice and the guidelines under which permissions issued, and the Act’s criminal architecture stood behind the new liberality for nine more years, a permission based liberalisation layered over a prohibition based code, until the management Act of 1999 rebuilt the field on the new premise and repealed the old.

What did a single industrial project need consent for under the licence raj?

A typical large project needed an industrial licence for entry, capacity and location under the 1951 Act; prior approval under the monopolies statute if the promoter crossed its size thresholds; exchange permission for foreign collaboration and equity; capital issues consent to raise money publicly; import licences for machinery; and land ceiling clearance for its site.

Two features of this bundle matter for everything that follows. The first is that the controls were statutory. The licence raj was not executive habit or bureaucratic excess layered over a free legal order. Each consent traced to an Act of Parliament or to subordinate legislation made under one, which is why the phrase licence raj describes a legal regime and not merely an administrative culture. The second feature is the one that made 1991 possible. Most of these statutes also contained, within themselves, powers of exemption, relaxation and notification by which the executive could disapply their own requirements to classes of cases. The licensing Act empowered the Central Government to exempt undertakings from its licensing provisions. The policy revolution of 1991 was carried out largely through those internal doors, left in the walls by the original builders.

The regime in operation: what the consents did to a project

To see why the manner of the regime’s ending mattered, it helps to watch the regime work on a single project, because its practical character explains both the grievances that accumulated against it and the relief that its removal brought, on which the two assessments later in this article divide. Take a promoter in the mid 1980s proposing a new plant to manufacture a scheduled product, with a foreign collaborator, public capital and imported machinery, an entirely ordinary large project of the period. The first application was for a letter of intent and then an industrial licence under the 1951 Act. The licence was not a formality of registration. It determined the product, the installed capacity and the location, and the deciding process weighed the claims of competing applicants, the regional balance of industrialisation, the claims of the small scale sector in the same product line, and the availability of the foreign exchange the project would consume. Capacity was a rationed good. A licence for less capacity than the technology made efficient was a common outcome, and producing beyond licensed capacity was not enterprise but contravention, however strong the demand.

With the licence in hand, the project met the second gate if the promoter’s group crossed the asset thresholds of the monopolies statute: prior approval for the new undertaking itself, distinct from the industrial licence and administered under a different Act by reference to the size of the group rather than the merits of the plant. The third gate was the foreign collaboration approval and the exchange permissions behind it, where the terms of the collaboration, the royalty, the duration and the foreign equity proportion were negotiated against guidelines; equity above the prescribed ceiling required the dilution arrangements described in the answers below. The fourth gate was the capital issue: the promoter company could not simply decide the size and price of its public issue, because the Controller of Capital Issues consented to both, and the official pricing conventions of the period, which valued issues by formula rather than by demand, made the premium a matter of arithmetic performed in an office in Delhi. The fifth gate was the import licence for the machinery, and the sixth, running in parallel, was the land: an urban site could engage the ceiling statute, and a rural site the state level land and power and water clearances that sat outside the central bundle. Each gate had its own file, its own queue and its own discretion.

The systemic consequences were the ones the reformers of 1991 recited. Time itself became a cost allocated by the state, and delay operated as a hidden tax that fell most heavily on the entrant without an established Delhi presence. The licence, once obtained, was an asset in itself, a barrier behind which the incumbent could produce below capacity, with indifferent quality, in a market from which the licence system excluded challengers; the queue for permission became a moat. Information about the state’s intentions was worth money, and the class that specialised in obtaining permissions flourished in proportion. None of this required any individual official to behave badly. It followed from a legal design in which the right to produce was a grant rather than a liberty, allocated case by case under statutes that made the exceptional grant the normal path. That design is what the phrase licence raj names when it is used precisely, and it is why the abolition of licensing, even by notification, was experienced by industry as a constitutional change in the economy’s working arrangements rather than as an administrative relaxation.

The monopolies gate deserves a paragraph of its own because its career is the strangest in the bundle. Enacted to prevent concentration, it operated in practice as a tax on size and a brake on the expansion of the most efficient producers, since capacity growth by a large house required approval that a small entrant did not. Committees examining the working of industrial policy through the 1970s and 1980s recorded the inversion: the licensing system and the monopolies system between them protected incumbents from entrants and large incumbents from the discipline of expansion alike. The gate’s abolition in 1991 was therefore among the least contested elements of the reforms, and its statutory deletion came within weeks. What survived the deletion, the conduct jurisdiction over restrictive and unfair trade practices, was the part of the Act that had always resembled a modern regulatory function, policing behaviour rather than vetting entrants, and its survival until the competition statute was ready is the clearest illustration in this article of repeal waiting upon a successor.

Each lock had its own key, and its own ending

Because the bundle was statutory, ending it presented a technical problem with several distinct solutions, and the government of 1991 used nearly all of them across the following two decades. It is worth separating the techniques, because public discussion routinely merges them into a single event called liberalisation and thereby misdates almost everything. There were five techniques. The first was exemption: leave the statute in force, and notify that specified classes of industry or transaction are exempt from the requirement. The second was amendment: persuade Parliament to delete the offending provisions from a living Act. The third was repeal and replacement: remove the old statute and enact a new one on a different philosophy. The fourth was plain repeal: remove the statute and put nothing in its place, or place its function with a regulator created by a companion Act. The fifth was administrative withdrawal without any legal instrument at all, used for the purely administrative layers of control, such as phased manufacturing programmes, which had never been statutory and so needed no statutory funeral.

Industrial licensing was ended by the first technique, exemption, and it remains ended by that technique. The licensing Act conferred on the Central Government a power to exempt any industrial undertaking or class of undertakings from its licensing provisions, and the government exercised that power through notifications in 1991 and afterwards until the licensing requirement survived only for a short list of industries, principally those connected with defence, security and specified hazardous or strategic concerns. The list has been revised since, again by notification. The licence as an institution died in July 1991. The licensing Act did not die at all, and the detailed reasons why, together with the text of the surviving provisions, are set out in the article on the MRTP Act repeal and its replacement for the monopolies statute and in the pillar article already referenced for the Industries Act itself.

The monopolies control was ended by the second technique, amendment, remarkably quickly, and by the fourth technique, repeal, remarkably slowly. Within weeks of the policy statement, Parliament enacted the Monopolies and Restrictive Trade Practices (Amendment) Act, 1991 (Act 58 of 1991), which omitted Part A of Chapter III of the principal Act, the sections that had required prior approval for expansion, new undertakings, mergers and takeovers by large undertakings, with effect from 27 September 1991. The size gate that had stood beside the licensing gate was therefore removed by legislation within about two months. What survived was the conduct half of the Act, the provisions on monopolistic, restrictive and unfair trade practices, administered by the Commission the Act had created. That surviving half had work to do and a constituency, and the Act as a whole was not repealed until the Competition Act settlement matured: the repeal took effect on 1 September 2009, eighteen years after the core was deleted, when the pending cases were transferred under the scheme the repeal provision itself prescribed.

Foreign exchange control was ended by the third technique, repeal and replacement, on a delay of eight years. The policy direction changed in 1991 and the administration of the Foreign Exchange Regulation Act softened in stages; amendments in 1993 removed several of the restrictions that had borne on companies with substantial foreign shareholding, and the automatic route for foreign equity widened by administrative decision. But the statute itself, with its architecture of prohibition backed by criminal penalty and its premise that foreign exchange was a resource to be conserved, stood until Parliament enacted the Foreign Exchange Management Act, 1999 (Act 42 of 1999), which came into force on 1 June 2000 and repealed the 1973 Act. The replacement changed the legal character of the field, from offences criminally punishable to contraventions visited with monetary penalty, and from a regime in which a transaction needed permission unless allowed to one in which transactions were regulated by category. That transition is owned by the article tracing FERA to FEMA, and this article takes from it only the interval: nine years between the change of policy and the change of statute, during which the economic world operated under a law written for the opposite policy.

Capital issues control was ended by the fourth technique, plain repeal coupled with a regulator, and it is the solitary case where statute kept pace with policy. The Securities and Exchange Board of India, which had existed as a non-statutory body since 1988, received statutory status under the Securities and Exchange Board of India Act, 1992 (Act 15 of 1992). The Capital Issues (Control) Act was repealed in May 1992, the office of the Controller of Capital Issues was abolished, and free pricing of issues replaced official pricing within the same season. Control of the capital market moved from a government officer exercising a licensing discretion over each issue to a statutory regulator prescribing disclosure, on the principle that the investor, given the facts, prices the risk. The details of that settlement belong to the article on the SEBI Act 1992; its place in this article is as the exception that tests the rule.

The remaining strings of the bundle ended piecemeal. Import licensing withered as trade policy moved from licensing to a negative list and then to tariff-based regulation, and the statutory frame was replaced when the Foreign Trade (Development and Regulation) Act, 1992 (Act 22 of 1992) superseded the 1947 control Act, substituting a development-and-regulation scheme for the wartime-born control scheme. The land ceiling statute followed a two-stage path: several states diluted or abandoned its enforcement, and Parliament enacted the Urban Land (Ceiling and Regulation) Repeal Act, 1999 (Act 15 of 1999), which took effect in January 1999 in the states and territories to which it first applied and extended to other states only as each adopted it by resolution under the constitutional provision for such adoption, a process the land cluster of this series examines in the article on the urban land ceiling repeal. The Essential Commodities Act was never repealed and never fully dismantled: control orders were pruned, many commodities were removed from control, and licensing requirements for dealers in several commodities were withdrawn by order in the early 2000s, but the enabling Act survives entire, a standing power that governments reach for in price spikes. Small scale reservation, created and maintained entirely by notification, was ended the same way, by progressive dereservation across two decades until the last twenty reserved items were dereserved in April 2015.

Was the Industries Act 1951 repealed after liberalisation?

No. It was never repealed. Industrial licensing was ended by exemption notifications issued under a power the Act itself conferred, leaving the licensing requirement alive on paper for a short notified list and the statute in force with its remaining chapters on registration, investigation and control of undertakings still operative.

The controls nobody photographed: trade, distribution and exit

Beyond the celebrated core of licensing stood three further layers of the regime that complete the statutory picture and whose endings were the quietest. The first was trade. The import control statute of 1947 supplied the legal basis for the import licensing system, under which the right to import was allocated by licence against categories of importer and priorities of use, with the annual import policy, published as a handbook rather than as law, carrying the operative detail. Capital goods for a licensed project, raw materials for an existing unit and goods for the open market each travelled by this route, and the canalisation of specified imports through public sector agencies added a state trading layer on top. The system’s reform after 1991 proceeded by shrinking the licensed list, moving item after item to free import under what became the negative list method, and by the statutory replacement of 1992 already described; quantitative restriction as a governing idea survived longest in agriculture and consumer goods, defended on balance of payments grounds until the turn of the century. The trade layer thus ended by erosion over a decade, its statute replaced early, its practice dismantled late: an interval that runs in the opposite direction from licensing and warns against treating any single row of this article’s table as the whole pattern.

The second layer was distribution and price. The Essential Commodities Act empowered the Central and state governments to issue control orders regulating the production, supply, distribution and pricing of commodities declared essential, and the orders made under it, licensing dealers, fixing stock limits, controlling movement between states and prescribing prices, constituted the most widely experienced face of the control regime. The Act did not allocate investment; it policed the aftermath, and for that reason liberalisation bore on it oddly. A government freeing entry and production had no need to fix the price of what entered freely, and order after order was withdrawn as supply normalised, particularly in the early 2000s when dealer licensing for several commodities was removed by central order. But the enabling power was never surrendered, because the political demand that produces control orders, a price spike in a necessity, recurs in every generation and no government will disarm itself against it in advance. The Act’s survival is therefore the most rational of the survivals in this article: it is retained not by oversight but by a standing calculation, renewed each time it is used, that the option of control is worth more than the tidiness of repeal.

The third layer was exit, the control regime’s least examined wall. The same legal order that vetted entry made departure nearly impossible: a failing industrial company could not readily close, retrench or be wound up, and the Sick Industrial Companies (Special Provisions) Act, 1985 institutionalised the protection of the failing unit through the Board for Industrial and Financial Reconstruction, before which proceedings suspended the remedies of creditors. The 1991 statement acknowledged the problem by promising an exit policy and referring sick public enterprises to the Board, but the legal reconstruction of exit took the longest of all the reforms: the repeal of the sick companies statute was enacted in 2003 and not brought into force until December 2016, when the Insolvency and Bankruptcy Code supplied the successor regime. Entry was freed in a notification in 1991. Exit was not freed in law for a quarter of a century. Between those two dates stands the measure of how partial the legal liberalisation was at the points where it touched workers, creditors and the state’s promise to protect the industrial sick, and any verdict on the sequence that ignores the exit layer mistakes the easiest third of the reform for the whole.

Twenty four July 1991 as a legal event

The Statement on Industrial Policy of 24 July 1991 is the hinge of the whole story, and it deserves to be read as a lawyer reads it, by asking what kind of instrument it was and what legal force each of its announcements carried. It was a policy statement laid before Parliament by the government of the day, in the middle of a balance of payments crisis in which the country’s foreign exchange reserves had fallen to a level measured in weeks of import cover and an external financing programme was being negotiated. It was not a bill. It amended no statute by its own words. Its announcements took legal effect, where they took effect at all, through instruments made under existing statutory powers: notifications exempting industries from licensing, amendments to the lists and schedules administered under those powers, Reserve Bank circulars and approval routes under the exchange control framework, and in one major case a bill brought to Parliament within weeks.

The statement’s industrial announcements can be sorted by their legal mechanics. Industrial licensing was abolished for all industries except a specified list of eighteen, later reduced; that announcement was executed by exemption notification under the licensing Act, and the Act was not amended in the process. The requirement of prior approval under the monopolies statute for large undertakings was removed; that required legislation, because the requirement was embedded in the text of the 1969 Act, and the amending Act of 1991 followed in September. Automatic approval of foreign direct investment up to 51 per cent of equity was announced for a specified list of high priority industries; that was executed administratively through the exchange control machinery, which had always operated by permission and could widen the permission by changing its own practice and guidelines. The phased manufacturing programmes, being administrative conditions rather than statutory requirements, were simply abolished. The statement also announced the disinvestment of part of the government’s holding in selected public enterprises and the referral of chronically sick public enterprises to the board that handled industrial and financial reconstruction, matters of executive action within existing law.

Two conclusions follow from reading the statement this way, and both are load-bearing for the argument of this article. The first is that the speed of 1991 was possible only because the control statutes had been drafted with internal executive doors. A licensing Act that had admitted no exemption power would have required amendment before a single industry could be delicensed, and amendment would have meant a parliamentary timetable in the middle of a crisis. The breadth of the exemption power, conferred decades earlier so that the executive could soften the regime at the margins, became the means of switching the regime off at the centre. The second conclusion is the counterpart. What the executive could do by notification, a later executive can undo by notification, subject to the constraints examined below. The settlement of 1991 was therefore legally provisional in a way that a settlement by repeal would not have been. The licensing Act still contains its licensing chapters. They apply, by force of the notifications, to very little. The chapters and the notifications are both law, and the difference between the India of 1990 and the India that followed is located almost entirely in the second of those two layers.

Why could licensing be abolished without amending the licensing Act?

Because the Act itself empowered the Central Government to exempt any undertaking or class of undertakings from its licensing provisions by notification. The 1991 notifications used that power for all but a short list of industries, so the requirement ceased to apply without a single section of the Act being altered or repealed.

The road to 1991: the regime criticised in its own terms

Liberalisation did not arrive without intellectual preparation, and the preparation matters legally because it determined which instruments the reformers trusted. The first authoritative criticisms came from inside the system. R. K. Hazari’s study of the licensing system in the mid 1960s documented, from the licensing record itself, how the largest industrial houses had used applications and licences to pre-empt capacity in industry after industry, securing positions they did not always proceed to build, so that the instrument of dispersal had become an instrument of foreclosure. The Industrial Licensing Policy Inquiry Committee under Subimal Dutt, reporting in 1969, confirmed the pattern from the files: licensing had failed to disperse ownership, had failed to prevent concentration, and had conferred its benefits disproportionately on the houses best placed to work it. Parliament’s legislative answer of that year, characteristically, was more law: the monopolies statute added the second gate rather than removing the first. The episode fixed a lesson the reformers of 1991 drew on: adding controls to correct the failures of controls compounds the original error, and the correction of a permission regime is its abolition, not its supervision.

The decade before 1991 then rehearsed the method in miniature. The governments of the 1980s, professing liberalising intent within the system’s premises, proceeded exactly as 1991 would, only less far: the asset threshold that triggered the monopolies gate was raised substantially, to one hundred crore rupees in 1985, exempting a generation of mid sized expansion from the second gate by a monetary revision; broad banding was introduced, permitting a licensed unit to vary its product mix within a band of related articles without a fresh licence, a relaxation effected through the administration of the licensing system rather than by amending it; and a first tranche of industries was delicensed outright in the mid 1980s, again by the executive instruments the statutes supplied. These rehearsals established both the feasibility and the limit of reform by instrument. Feasibility, because each relaxation took effect at once and the sky did not fall. Limit, because a reform conducted inside the permissions of the old statutes remained what the statutes said it was: a series of exemptions under a licensing Act, a threshold revision under a monopolies Act, administration all the way down. When the crisis of 1991 forced the question of principle, whether permission would remain the organising idea of industrial law, the answer given was no in policy and, as this article has shown, only partly and belatedly in law. But the continuity of technique is the point: 1991 was not a departure from the legal method of the 1980s. It was that method driven to its conclusion by an emergency, which is why the reform’s legal character, executive, notification borne, reversible in form, was visible in embryo a decade before the reform itself.

The crisis that dictated the method

The choice of instrument in 1991 cannot be separated from the calendar of the crisis. Through the first half of that year, India’s foreign exchange reserves fell to a level that covered a few weeks of imports; gold was pledged abroad to secure emergency finance; an arrangement with the International Monetary Fund was under negotiation; and a new minority government had taken office in June with weeks, not years, in which to demonstrate that the external position would be corrected. The finance minister’s budget speech and the industrial policy statement of 24 July were delivered within this compression. In such a setting, the order of operations was decided by the instruments available. What could be done by notification was done within days. What required a bill was begun at once and completed in September. What required a new institution, or a replacement code, was announced as direction and pursued across the following decade. The method that this article has described as executive first was not chosen from a menu of equally available methods by a government philosophically committed to governing without Parliament. It was the only sequence that fit the timetable of a balance of payments emergency, and its long tail, the statutes lagging by years, is in part the tail of that emergency working itself out through drafting offices and parliamentary sessions that had other business.

This matters to the fairness of the later debate. The critical assessment of liberalisation treats the notification method as a democratic defect, and on the legal facts the description is accurate: the most consequential economic decisions of the period received no committee scrutiny and no clause by clause vote. But the counterfactual the criticism implies, a Parliament leisurely repealing the control statutes through 1991 and 1992 while the reserves ran out, was not on offer, and the governments that followed, which had the leisure the crisis government lacked, are the proper address for the complaint that the statute book was left unreformed for eighteen years in one case and permanently in another. The emergency explains the method of July 1991. It does not explain the unrepealed Act of the decades afterwards, which is a separate choice, made by omission, by every administration that found the exemption notifications convenient and repeal unnecessary. Holding those two judgments apart is essential to an accurate legal history, because they concern different actors at different times making different kinds of decision.

Delicensing in detail: what the notifications actually did

The mechanics of delicensing repay attention because they show the exemption power doing work far beyond anything its drafters contemplated. The Industries Act required a licence for a new undertaking, for a new article and for substantial expansion in the scheduled industries. The exemption power allowed the Central Government, by notification in the official gazette, to exempt any industrial undertaking or class of undertakings from the licensing provisions where it thought the exemption necessary in the discipline of the Act’s purposes. In July and August 1991 the government issued notifications exempting all industries other than the eighteen on the retained list. The legal architecture of the change is worth pausing over. Nothing was repealed, nothing was amended, and the schedules of the Act continued to enumerate the industries of India as they had since the 1950s. What changed was a layer of subordinate legislation standing between the citizen and the Act. A factory established in a delicensed industry after 1991 was not unlicensed because Parliament had freed it; it was unlicensed because a notification said the licensing chapter did not apply to it, and it remained, in strict legal form, an exempted subject of a licensing statute.

The retained list itself became an instrument of continuing policy. Industries remained licensed for reasons of security and defence, of public health and hazard, and, in the early years, of reservation for the small scale sector in particular lines. Successive notifications cut the list from eighteen to a small residue, and the grounds of retention narrowed towards defence, atomic energy related activity and specified hazardous manufacture. Because the list lived in notifications, each reduction was an executive act of the same character as the original delicensing, and the boundary of the licence raj in its afterlife has been redrawn repeatedly without Parliament’s involvement in either direction. The same technique governed the small scale reservation, whose list shrank by notification across two decades to emptiness in 2015, and the geography of the reform is inscribed in the gazette rather than in the statute book. A researcher who reads only the Acts of Parliament for these years will find the SEBI Act, the trade Act, the insurance Act, the FEMA and the Competition Act, and will miss the single most consequential legal event of the period, which is recorded in a series of exemption notifications under a 1951 statute that the statute book presents as unrepealed and, on its face, unremarkable.

Policy first, statute later: the intervals measured

The table below is the argument of this article in tabular form. Each row records one control, the date it ceased in practice, the instrument that ended it at that point, the date of the corresponding statutory change, and the interval between the practical end and the statutory end. Read down the interval column and the pattern asserts itself without commentary: months in one case, eight years in another, eighteen years in a third, a quarter of a century for the reservation that lived and died by notification, and, in the case of the statute at the centre of the whole regime, no interval at all because the second event never occurred.

Control	Ceased in practice	Instrument ending it in practice	Statutory change	Instrument of statutory change	Interval
Industrial licensing	July 1991	Exemption notifications under the Industries Act 1951	None; the Act stands unrepealed	None	Never
Monopolies prior approval	September 1991	Policy statement followed by amendment	Core deleted 1991; Act repealed 1 September 2009	Amendment Act 58 of 1991; repeal under the Competition Act 2002	Eighteen years to repeal
Capital issues consent	May 1992	Repeal and transfer to a statutory regulator	May 1992	Capital Issues (Control) Repeal Act 1992 and the SEBI Act 1992	Months
Foreign exchange permission regime	From 1991 in stages	Administrative widening of permissions and automatic routes	1 June 2000	Foreign Exchange Management Act 1999 repealing the 1973 Act	About eight years
Import licensing	From 1991 in stages	Trade policy moved to a negative list	1992	Foreign Trade (Development and Regulation) Act 1992	About a year
Urban land ceiling	Through the 1990s by state dilution	State level abandonment of enforcement	January 1999 in the first states and territories	Urban Land (Ceiling and Regulation) Repeal Act 1999	Years, and state by state
Small scale reservation	Progressive from the 1990s	Dereservation notifications	None; reservation was never statutory	None	Ended April 2015 by notification
Essential commodities control	Pruned from the early 2000s	Control orders withdrawn for many commodities	None; the 1955 Act survives entire	None	Never

Three observations discipline the reading of the table. The first is that the intervals are not random. Where the control rested on a discretionary permission that the executive itself administered, liberalisation could be immediate and its legal form could stay administrative indefinitely: licensing, reservation, import lists. Where the control was hard coded into statutory text that an administrator could not disapply, as with the monopolies approval sections, legislation had to follow quickly or the policy statement would have been an empty promise. Where the control embodied a philosophy the government wished to repudiate formally, replacement legislation eventually came, but at a pace set by drafting capacity and parliamentary time rather than by the policy itself, which is the FERA story. The second observation is that repeal, when it came at all, usually came attached to a successor institution: the capital issues control died when the securities regulator was ready to receive its function; the monopolies Act died when the competition commission was operational; the land ceiling died when the political constituency for ceiling had dissolved. Statutes in this period were not so much abolished as retired, on the day their replacement reported for duty or their purpose was conceded to be gone. The third observation concerns the two rows marked never. An unrepealed licensing Act and an unrepealed essential commodities Act are not museum pieces. Both are live law. The licensing Act’s surviving chapters have been used in later decades, and the commodities Act’s control powers are exercised whenever prices provoke intervention. The statute book of the control era is therefore not an archive of the control era. Parts of it remain the working law of the successor regime.

The exception: capital issues, and what it proves

The capital market row repays a closer look because it is the one case that ran the other way, legislation first or at least legislation at once, and the reasons for the exception illuminate the rule. The Controller of Capital Issues exercised a discretion that could not simply be widened into its opposite. Consent to an issue and approval of its price were individual acts of administration; there was no exemption power broad enough to convert the office into a disclosure regulator, because the office’s function was the control itself. If issues were to be priced by the market, the consent requirement had to be abolished, and abolishing a statutory requirement takes a statute. At the same time, the receiving institution already existed in embryo: the securities board had been functioning without statutory power since 1988 and could be clothed with authority by a short Act. The repeal ordinance of May 1992, followed by the repeal Act, and the SEBI Act of the same year therefore formed a single legislative package, conceived together. The episode supports the general finding rather than disturbing it. The executive route was used where a statute already conferred the power to take it. Where no such power existed, because the control was the statute’s whole content, the legislative route was taken, promptly and completely. The determinant of method was not ideology about Parliament. It was the drafting history of each control statute: what doors its authors had left in it.

Which control statute was repealed in the season its function moved?

The Capital Issues (Control) Act, 1947. It was repealed in May 1992 as the statutory securities regulator took over the capital market under the SEBI Act, 1992, and the office of the Controller of Capital Issues was abolished, making it the one control whose statutory end matched its practical end.

The regulators: the part of the story Parliament did write

Against the pattern of executive withdrawal stands a counter pattern that any honest account must weigh. The successor institutions of the liberalised economy were created by legislation, one by one, and this was not incidental. A regulator needs legal personality, coercive powers of inquiry and penalty, a budget and appointment machinery; none of that can be conjured from a notification under an existing Act whose subject matter lies elsewhere. The securities regulator received its statute in 1992. The insurance market was opened and its regulator created by the Insurance Regulatory and Development Authority Act, 1999 (Act 41 of 1999), which amended the insurance statutes and admitted private and foreign participation within limits the Act itself fixed. The telecom regulator had been created by statute in 1997, in a reform that predates the main sequence’s second half. The electricity sector was reorganised by the Electricity Act, 2003 (Act 36 of 2003), which delicensed generation, restructured the state electricity boards’ statutory successors and built the sector’s regulatory commissions on a statutory base. The competition commission, the institution that finally allowed the monopolies statute to be repealed, was created by the Competition Act, 2002 (Act 12 of 2003), whose substantive chapters commenced in stages, agreements and abuse of dominance in May 2009 and combination regulation in June 2011, as the owning article in this series records.

The contrast is instructive and cuts in a direction neither camp in the liberalisation debate emphasises. Removal was executive; construction was legislative. Where the task was to stop doing something the state had done under powers already granted, the executive acted alone and instantly. Where the task was to build something that had to wield power over others, the full legislative process was used, with its delays: the competition statute took the better part of a decade between enactment and full commencement. The legal history of the period is therefore not a simple story of Parliament sidelined. It is a story of the two branches dividing the labour along the line drawn by the pre-existing statute book, the executive clearing ground with the tools the old Acts gave it, and Parliament legislating the new structures that no old Act could supply.

Ownership without amendment: the public sector reservation

One further control of the era belongs in the statutory anatomy because its legal form is the most surprising of all: the reservation of whole industries for the public sector rested on no statute at all. The Industrial Policy Resolution of 1956 divided industries into schedules, reserving seventeen for exclusive state development and a further group for progressive state expansion, and for thirty five years that division governed the boundary between public and private enterprise across the commanding heights of the economy, steel, mining, heavy machinery, telecommunications, power among them. Yet the resolution was a policy instrument. Private entry into a reserved industry was prevented in practice because the licensing Act, administered in conformity with the resolution, would not license it, and because the sectoral statutes, in insurance, banking, coal and elsewhere, enacted public monopolies for particular fields. The reservation was thus doubly supported: by policy at the general level and by statute at the points where nationalisation had been legislated. The 1991 statement accordingly dismantled it by the appropriate instruments at each level: the reserved list was cut by the statement itself from seventeen heads to eight, and reduced by later policy to three, atomic energy, defence and railway operations, while the statutory monopolies required, and in insurance received, amending legislation before private entry could occur. Where a monopoly existed only as policy, it ended by policy. Where it existed as an Act, it waited for an Act. The whole liberalisation sequence is compressed into that contrast.

Disinvestment, the sale of government shareholding in public enterprises, travelled the same executive road. The shares were the government’s property, and their sale in tranches, first of minority stakes in bundles to financial institutions and later of strategic stakes to purchasers who took management with them, required no legislation, being an exercise of the ordinary rights of a shareholder conducted under policy guidelines the government made and remade. The challenges to disinvestment were accordingly fought on the doctrines described in the courts section, and the courts’ refusal to convert the policy into a justiciable question completed the legal design of the reform era: the commanding heights could be vacated by the same informality with which they had been occupied, because occupation had itself been a matter of policy executed through corporations the state owned. Nationalisation had taken Acts, nineteen banks and whole industries transferred by statute. Denationalisation, which never bore that name, took decisions.

Location, freight and the geography the licence drew

The licence was never only about whether and how much. It was also about where, and the locational dimension of the regime left marks on the industrial map that outlasted the regime itself. Licences for industries in congested metropolitan areas were refused or conditioned in favour of dispersal; incentives, concessional finance, tax concessions and subsidised industrial estates, were assembled to pull investment towards districts classified as backward; and the freight equalisation arrangements for key inputs such as steel and coal, under which delivered prices were equalised across distances by a pool mechanism, removed the natural advantage of the mineral bearing east so that industry might spread. None of these instruments was a prohibition in the licensing sense, and none rested on a single statute: they were a tissue of policy, fiscal concession and administered pricing laid over the licensing skeleton. Their withdrawal in the reform years followed the same administrative path as their creation, freight equalisation being abolished by decision in the early 1990s as pricing was freed, and the locational conditions fell away with delicensing itself, since a state that no longer licenses entry cannot direct where entry occurs.

The consequences are part of the legal history because they show how thoroughly a regime of permissions shapes facts that no subsequent repeal can redistribute. Industrial concentration in the western and southern states, which the reform decades deepened, was in part the free working of advantages the control era had deliberately muted; the eastern states’ complaint that freight equalisation had subsidised their competitors at the cost of their own industrialisation was a complaint about a policy instrument, litigated in the court of politics rather than of law, and withdrawn without any legislature ever voting on the scheme that bore its name. A reader tracing the geography of Indian industry through the statute book will find the Industries Act and the taxing statutes and almost nothing of the instruments that actually drew the map, because the map was drawn in the same layer of policy and notification in which, forty years on, it was redrawn. The point generalises the article’s finding beyond the celebrated statutes: the operative economic law of the control era lived substantially below the level of the Act, and it is that depth, more than any single provision, that the word raj accurately describes.

The afterlife of licensing in the retained sectors

The retained licensing list, reduced to a residue, has had an active afterlife that displays the 1991 method in continuing operation. Defence production is the standing example. It remained licensed, for reasons of security that no reform government questioned, and the terms of private and foreign participation in it have been liberalised in stages ever since, each stage executed by the characteristic instruments of the era: a policy announcement, a press note in the consolidated foreign investment policy, and consequential changes in the exchange regulations, raising the ceiling for foreign investment from 26 per cent to 49 per cent and in time to 74 per cent under the automatic route, with the licensing requirement itself retained and administered under the Industries Act throughout. The pattern of 1991 thus became the permanent technique of economic government in the retained sectors: the statute stands, the list stands, and the substance moves by instrument beneath them. An observer who asks whether India still has industrial licensing receives two true answers, which this article exists to reconcile: as a general system of entry control, licensing ended in 1991; as a statutory power held in reserve and exercised over a short list, it never ended at all, and the distance between those answers is the distance between the economy and the statute book with which this article began.

Banking supplies the same lesson from the financial side. The nationalisation of the major commercial banks was effected by statute, in 1969 and again in 1980, and the licensing of bank branches was administered by the Reserve Bank under the Banking Regulation Act, 1949 (Act 10 of 1949), a permission regime that determined the geography of credit as the industrial licence determined the geography of production. Liberalisation in banking proceeded, characteristically, through the regulator’s instruments rather than through repeal: guidelines issued by the Reserve Bank in 1993 admitted new private banks for the first time in a generation, branch licensing was progressively relaxed towards authorisation regimes, and the nationalisation statutes were left in force with their public sector banks operating under them in a market the guidelines had transformed. Entry by permission gave way to entry by prudential guideline, a change in the temperature of the regime accomplished without touching the Acts that had built its predecessor. Finance, industry, exchange and trade all tell the one story with local variations, and it is the story this article has documented at length: the Indian state liberalised its economy largely by deciding, under powers it already held, to stop doing what its statutes entitled it to do, and the statutes, largely, are still there.

The instruments compared: what a notification can and cannot do

The legal craftsman reading this history will want the taxonomy stated plainly, because the entire sequence turns on the capacities of subordinate instruments. A notification under an exemption power does three things well. It is fast, taking effect on publication in the gazette without a parliamentary session. It is granular, capable of covering exactly the industries, transactions or regions intended and no others, which is why the retained licensing list could be trimmed industry by industry for twenty years. And it is reversible by the same hand that made it, which the reversibility section treats as the settlement’s defining legal feature. Against these stand three incapacities. A notification cannot contradict its parent Act: the delicensing of 1991 was possible only because the exemption power, fairly read, authorised exemption without numerical limit, and a power drawn more narrowly would have confined the government to case by case relief. A notification cannot create an institution, a tax, a criminal offence beyond what the parent Act authorises, or a monopoly’s abolition where the monopoly is itself enacted; each of those requires a bill. And a notification carries a lighter democratic warrant, subject to laying procedures before Parliament that are, in practice, formalities of tabling rather than occasions of decision. The reformers of 1991 used the instrument up to the edge of its capacities and stopped, perforce, at that edge: the monopolies deletion, the capital issues repeal, the insurance opening and every regulator in the new landscape mark the points where the edge was reached and the bill had to follow. A map of where notifications sufficed and where they failed is therefore also a map of the pre-1991 statute book’s internal distribution of power between Parliament and the executive, drawn by the drafters of the control era for purposes of control, and used after 1991, with perfect legality, for the opposite purpose. That is perhaps the deepest irony the legal record holds: the licence raj carried the means of its own abolition in its pocket from the beginning, in exemption clauses inserted so that the state could be merciful in hard cases.

The second wave: Parliament catches up in part, 1999 to 2011

The years from 1999 to 2011 are the period in which the legislative branch of the sequence did its work, and the character of that work confirms the division of labour described earlier. Parliament did not, in these years, revisit the decision to liberalise; no repeal of a delicensing notification was debated, no control was restored by statute. What Parliament did was replacement and construction. The Foreign Exchange Management Act, 1999 replaced a prohibition code with a management code nine years after the policy it codified had taken effect, and the debates on it were the debates of consolidation: the philosophy had been settled in the administrative practice of the intervening years, and the Act’s task was to give that practice a coherent statutory home, decriminalising contraventions and sorting transactions into the free and the regulated. The Insurance Regulatory and Development Authority Act, 1999 opened a sector that the executive could not have opened by notification, because the insurance statutes themselves confined the business to public sector corporations; here legislation had to lead, since no exemption power in the old Acts reached the monopoly they enacted, and the sequence accordingly ran statute first for once, with the market following the Act. The Electricity Act, 2003 repeated that pattern in a sector where the state boards’ statutory structure had to be dismantled by law. The Competition Act, 2002, enacted at the turn of the period and commenced in stages through 2009 and 2011, closed the longest interval in the table by supplying the institution whose readiness finally permitted the monopolies statute to be repealed.

Read together, the second wave shows Parliament legislating at the points where the executive method ran out: where a statutory monopoly had to be ended, where a new coercive institution had to be created, where an entire code had to change its philosophy in text and not merely in administration. Everywhere else, the notifications of 1991 were simply left to stand, increasingly venerable, gathering the authority of long usage. By the end of the second wave the legal map of the economy had a curious double character that persists. The visible peaks are modern regulatory statutes, enacted with deliberation. The plain beneath them is crossed by the older enactments, the Industries Act of 1951 and the Essential Commodities Act of 1955 chief among them, unrepealed, their operative content inverted or pruned by subordinate instruments, legible only to a reader who knows to look below the Act to the gazette. The statute book records the destination. The journey is recorded elsewhere.

The courts: deference on policy, insistence on purpose

A change of regime executed through exemption notifications and shifting approval practice generated litigation at predictable points, and the courts’ response forms the third leg of the legal account. Two lines of doctrine did most of the work. The first is judicial restraint in economic policy. The Supreme Court held repeatedly, in challenges to price fixation, industrial policy and later to disinvestment, that the framing of economic policy is committed to the executive and the legislature, that the court does not sit in appeal over the wisdom of such policy, and that intervention requires a showing of constitutional violation, mala fides or arbitrariness of a kind that negates reasoned decision. The proposition runs from the price fixation cases, including Shri Sitaram Sugar Company v Union of India, decided in 1990 on the eve of the reforms, through the disinvestment litigation of the following decade, including BALCO Employees Union v Union of India, where the Court declined to arrest a disinvestment decision taken as policy. Applied to liberalisation, the doctrine meant that the decision to delicense, to widen foreign equity routes or to withdraw a control order was, as a policy choice, effectively unreviewable.

The second line of doctrine cuts the other way and is the one practitioners relied on. A statutory power must be exercised for the purpose for which it was conferred, and subordinate action that travels beyond the parent Act falls. Exemption notifications under the control statutes were therefore always open to the challenge that they were not exemptions at all but an abdication or a disguised amendment, and the courts approached such challenges by asking whether the power, read fairly, bore the use made of it. In the event, the exemption powers in the licensing and kindred statutes were drawn widely enough to bear the weight placed on them, and the great delicensing notifications survived. But the doctrine set the boundary of the executive method: the government could go as far as its statutory powers reached and no further, which is precisely why the monopolies approval requirement, embedded in operative sections rather than in an exemption scheme, had to wait the few weeks for an amending Act.

The third doctrine, legitimate expectation and its senior relative promissory estoppel, is where disappointed beneficiaries of the old and new regimes actually fought. The courts recognised, following Motilal Padampat Sugar Mills v State of Uttar Pradesh, that a clear governmental promise, acted upon to a party’s detriment, could bind the state in equity. Against that stood the limit stated in later cases, of which Kasinka Trading v Union of India is representative: an exemption granted in public interest may be withdrawn in public interest, no estoppel arises against the exercise of a statutory power for the purpose it exists to serve, and a trader who takes the benefit of a notification takes it subject to the power of withdrawal the statute preserves. The combined effect for the liberalisation sequence was a ratchet with a legal spring. Firms that had invested on the faith of the new policies acquired expectations the courts would weigh, political facts that made reversal costly; but no firm acquired a legal right that any control, once lifted, must stay lifted, and no incumbent acquired a right that a protection, once granted, must stay granted. The law protected reliance on specific promises. It did not protect the permanence of regimes, in either direction.

Reversibility: the spring inside the settlement

The feature of the post-1991 legal position that commentators in both camps return to is its reversibility, and it deserves exact statement rather than metaphor. Industrial licensing could be restored for an industry, or in principle for industries generally, by the same instrument that removed it: a notification narrowing or withdrawing the exemption, issued under the same section of the same unrepealed Act. No bill, no debate, no presidential assent would be required for the legal act itself, whatever the political cost. Essential commodities controls can be and are reimposed by order. Dereserved products could be reserved again by notification. By contrast, the elements of the settlement that were enacted cannot be undone except by Parliament: the securities regulator cannot be disestablished by circular, the competition prohibitions cannot be switched off by the ministry that administers them, and the management philosophy of the foreign exchange statute cannot revert to the regulation philosophy without an Act, because the 1973 Act no longer exists to be revived; it would have to be re-enacted in substance.

The two camps read this asymmetry from opposite ends, and the article’s stance charter requires each reading in its own voice. Supporters of the reforms, among them economists who had advised the reform governments and commentators who chronicled the control regime’s failures, treat the unfinished statutory work as a defect to be repaired: a settlement left in notifications invites incremental recontrol, each step too small to provoke resistance, and the periodic official proposals to repeal obsolete and redundant statutes, which have repeatedly listed enactments of the control era, acknowledge the untidiness without curing the central case, since the licensing Act is retained deliberately as an enabling shell. Critics of the reforms read the same asymmetry as the record of a choice: a transformation of the economy’s legal basis carried out under external financial pressure, in a crisis, by instruments that bypassed the deliberative scrutiny a bill receives, left the transformation permanently open to the charge that it lacked a legislative mandate, and preserved for future governments the legal capacity to reverse course without persuading Parliament, a capacity the critics regard variously as a safeguard or as proof that the settlement was never meant to be constitutional in character. Both readings agree on the legal fact. They disagree about whether the fact is the settlement’s weakness or its honesty.

One refinement belongs to neither camp and is this article’s own. Reversibility is not uniform across the field, because two decades of operation under the new policies built legal facts that a notification cannot erase: regulators with statutory duties and decided cases, contracts and investments made under the foreign exchange management regime, a generation of undertakings whose existence traces to exemption notifications on which third parties have relied. A notification can change the rules for the next factory. Unwinding the accumulated position of the existing ones would engage the doctrines of reliance and non-retroactivity and, at scale, the legislative process itself. The settlement is reversible at its edges by executive act. At its centre it has set, the way cooling metal sets, through the accretion of rights and institutions around it. The unrepealed Act gives a future government a door. The room beyond the door is no longer the room of 1990.

Can industrial licensing be restored without a new Act?

In form, yes. Because licensing was ended by exemption notifications under the Industries Act 1951, which was never repealed, a notification withdrawing or narrowing the exemption would revive the licensing requirement for the industries affected. The political and reliance obstacles would be large, but the legal instrument required is executive, not legislative.

Why the shell is kept: obsolete law and the politics of the unrepealed Act

A question remains that neither the crisis nor the division of labour answers: why, in the decades after 1991, did no government repeal the Industries Act? The question is sharpened by the fact that repeal of obsolete law became, in those same decades, an avowed programme. Successive official reviews catalogued enactments that had outlived their purpose; the Law Commission had recommended the repeal of spent statutes across the life of the republic; and Parliament passed a series of Repealing and Amending Acts that struck large numbers of dead enactments from the book. Control era statutes of minor consequence were swept up in such exercises. The central licensing statute was not, and the reasons are more instructive than oversight.

The first reason is utility. The Industries Act is not wholly spent. Its chapters on the investigation of failing undertakings, on the control of supply and price of specified articles in defined circumstances, and on the registration framework continue to supply legal authority that departments use, and a repeal bill would have to decide, provision by provision, what to preserve and how to rehouse it, which converts a ceremonial repeal into a drafting project with opponents. The second reason is option value. A government that repeals its dormant licensing power surrenders it; a government that retains the Act retains the legal capacity to reimpose licensing by notification should a future policy demand it, and executives of every complexion have preferred to keep that door unlocked. The third reason is the absence of a constituency for repeal. The costs of the unrepealed Act are borne by legal clarity itself, which has no minister; the benefits of retention are felt by administrators, who do. The fourth reason is risk asymmetry in legislative management: a repeal bill invites debate on the whole settlement of 1991, its gains and its distribution, and no government schedules a debate it does not need on a question it considers closed.

The documentary habits of the period tell the same story from the archive side. Because the operative law of liberalisation lives in notifications, circulars, approval guidelines and files rather than in Acts, its preservation depends on the record keeping of the departments and regulators that issued them, and researchers reconstructing the sequence have learned to treat the gazette and the departmental file with the care that practices for protecting sensitive source documents apply to papers that powerful institutions would rather lose. The statute book, the public and permanent record, is precisely the layer of the story that changed least, and a reader confined to it would misdate the end of the licence raj by decades or miss it altogether. That is a finding about where Indian law keeps its history, and it generalises beyond this period: in a system where subordinate legislation and exemption notifications carry the operative content of policy, the archive of change is the archive of the executive.

The two assessments, stated at full strength

The favourable assessment of the sequence begins from the record of the control regime itself. Entry licensing, capacity licensing and size approval together produced an economy of shortage and queuing, in which the scarce asset was the permission rather than the product; incumbents were protected from entry by the very machinery meant to plan investment; quality stagnated in sheltered markets; and the discretionary allocation of licences, import entitlements and foreign exchange generated the rents and the corruption that the word raj came to denote. On this account the method of 1991 was not a constitutional shortcut but the only route proportionate to the emergency: reserves covered weeks of imports, a parliamentary programme of repeals would have consumed years the balance of payments did not contain, and the executive powers used had been conferred by Parliament itself, so the democratic objection proves too much, since a Parliament that disliked the delicensing could have legislated against it at any point in the decades that followed and instead legislated in the same direction whenever it acted at all, in 1992, in 1999, in 2002 and in 2003. The results are then entered as evidence: sustained higher growth in the decades after reform, an expansion of consumer choice and industrial capability, the emergence of sectors, software services being the standard example, that the licensing regime had never contemplated and could not have licensed intelligently. Defenders of the sequence concede the unfinished statute book and treat it as an agenda item, the repeal of spent and redundant enactments being a housekeeping exercise that official committees have periodically recommended.

The critical assessment begins from the mandate. The reforms of 1991 were not in the election manifesto of the government that enacted them in any detail that would have warned the electorate, they were undertaken alongside an external financing programme whose conditions pressed in the same direction, and the method, government by notification under old powers, insulated the most consequential economic decisions of the republic’s history from the committee scrutiny, amendment and recorded division that attend a bill. On this account the eighteen year survival of the monopolies statute after its core was deleted, and the permanent survival of the licensing Act, are not quirks but symptoms: a transformation that could not command legislative time because it had never commanded legislative deliberation. The distributional record is then entered in reply: growth concentrated in services and in regions already advantaged, manufacturing employment expanding more slowly than the reform’s architects projected, agriculture left largely outside the liberalised settlement, and the gains of delicensing accruing in the first instance to those with the capital and connections to use the new freedoms, the same class the old regime’s rents had enriched by a different mechanism. Critics divide among themselves on the remedy: some would have preferred the reforms legislated openly and owned politically; others regard the preserved reversibility of the notification based settlement as the one redeeming feature, the retained legal capacity of a future Parliament or government to reassert control without first repealing a wall of successor statutes.

The article takes no side in that contest, which is about economics and democratic theory, because the evidence available to a legal history cannot settle it. It does take a side on the prior legal question, which is separable and which the record does settle. As a matter of legal technique, the liberalisation of 1991 onwards was an executive event with legislative consequences, not a legislative event. The proportions can be stated with some precision from the table above. Of the eight controls tabulated, one was ended by legislation within the year, capital issues; one was cut by legislation within months but not repealed for eighteen years, monopolies; two were replaced by legislation after intervals of one and eight years, imports and foreign exchange; one was repealed state by state under a 1999 Act after years of administrative decay, land ceiling; and three were never the subject of repealing legislation at all, because they lived by notification, ended by notification, or survive as standing powers, licensing, small scale reservation and essential commodities control. A reader who wishes to argue that liberalisation was Parliament’s work must explain away six of the eight rows. A reader who wishes to argue it was executive usurpation must explain the rows where Parliament acted, promptly or eventually, always in the same direction and never once to reverse the executive. The sequence shows something more specific than either slogan: a ratification pattern, in which the executive moved first under colour of powers Parliament had granted, and Parliament, given repeated opportunities across four decades and governments of different parties, chose each time to build on the new position rather than restore the old. The statutes were the last thing to change. When they changed, they changed to confess what had already happened.

What 1991 did not touch, and the view from the states

A legal history of liberalisation that stopped at the Centre’s economic statutes would overstate the transformation, because large fields of economic law were never liberalised at all, and others were reformed, if at all, decades later. The labour statutes of the control era, with their permissions for retrenchment and closure in larger establishments, stood essentially unreformed through the whole period this article covers, consolidated rather than liberalised when Parliament eventually enacted the labour codes a generation later. Agricultural marketing remained governed by state statutes channelling the sale of produce through regulated markets, reformed piecemeal by states and never nationally. Land, the factor on which every project ultimately sits, remained under state law entirely: ceiling apart, the acquisition, tenancy and conversion regimes that governed industrial land were creatures of state legislation and state administration, untouched by the notifications of 1991, which may explain why land assembly displaced licensing as the binding constraint on new projects in the decades that followed. The licence was abolished in Delhi. The queue moved to the state capital and the district office.

The states also supply a footnote that complicates any simple story of executive government. Several states, competing for the investment the Centre had freed, enacted their own single window clearance statutes, promising a single point of contact and deemed approvals on the expiry of stated periods. The gesture is the licensing idea inverted, the state binding itself to decide quickly rather than reserving the right to decide at leisure, and it was done by legislation, since a promise enforceable against departments needs statutory form. Even in the age of the notification, then, the durable settlement of a new position repeatedly turned out to require an Act: a regulator, a promise, a successor code. The executive instrument excelled at subtraction. Everything that had to be built, guaranteed or institutionalised still went the long way through the legislature, which is the final refinement of this article’s thesis. Liberalisation by notification was real, extensive and effective, and it was also confined, by the nature of legal instruments, to the negative half of reform.

The jurisprudence that made deference: the courts and the control regime before 1991

The courts’ restraint during the reforms had deep roots, because the judiciary had spent the life of the control regime declining, with rare exceptions, to dismantle it. The control statutes were challenged as they were enacted, on the ground that licensing and control orders violated the constitutional freedom to practise a trade or carry on a business, and the Supreme Court’s answer, settled early and repeated for decades, was that the freedom was subject to reasonable restrictions in the public interest and that a scheme of licensing directed to the allocation of scarce resources or the planned development of industry was such a restriction. In Harishankar Bagla v State of Madhya Pradesh, decided in 1954 on a control order under the essential supplies legislation, the Court upheld the control framework against the challenge that it delegated unguided power, finding the policy of the statute a sufficient guide. A line of decisions on licensing, price fixation and distribution control followed the same course. The consequence for 1991 is direct. A judiciary that had held for forty years that the existence of the controls was constitutionally permissible policy was never going to hold that their removal was constitutionally suspect policy; the doctrine was symmetrical, and both the imposition and the withdrawal of the regime were placed, by the same reasoning, beyond ordinary judicial review. The few points at which courts intervened in the liberalisation period were the technical points the doctrine had always reserved: whether a particular notification travelled beyond its parent Act, whether a specific assurance bound the state in equity, whether a retrospective withdrawal of a benefit offended the rules against arbitrary state action. The architecture stood; the paperwork was checked.

That inheritance also explains a feature of the reforms that foreign observers found odd: the absence of any great constitutional case about liberalisation. There is no Indian judgment that plays, for economic reform, the role that the great cases play for reservation or federalism, because the doctrinal settlement of the 1950s had removed the questions such a case would have asked. The legal battles of the era were fought instead in the interstices: over the transfer of pending monopolies cases, over the survival of prosecutions under the repealed exchange Act for conduct predating its repeal, over tax exemptions promised in one policy and withdrawn in the next. Transitional law, the unglamorous discipline of savings clauses and accrued rights, did the work that constitutional law did not. The repeals, when they came, were drafted with the standard savings for things done and liabilities incurred, and the courts applied the savings in the ordinary way, so that the repealed FERA continued to govern the prosecutions of its own era long after the management Act had taken over the living field, one more interval, inside the courtroom, between a statute’s death and its effects.

The proportions, restated as a finding

The thesis of this article can be stated with the precision the evidence allows, and it is worth doing so in a single place, since the claim is of a kind that can be checked against the table. Eight controls constitute the regime as this article has surveyed it. Of these, one, capital issues, was terminated by repeal in the same season as its function was transferred to a regulator created by Act of Parliament: liberalisation there was a legislative act, complete and contemporaneous. Two, import control and foreign exchange, were terminated by replacement statutes after intervals of about one year and about nine years respectively, the replacement in the exchange case transforming the character of the law from criminal prohibition to civil regulation. One, the monopolies approval regime, was cut out of its Act by amendment within two months of the policy statement, while the Act that housed it survived eighteen further years and was repealed only when its surviving conduct business could be transferred under the competition settlement. One, the urban land ceiling, decayed administratively through the 1990s and was repealed in 1999 by an Act whose own commencement then proceeded state by state, some states retaining the ceiling in force for years after the central repeal. And three, industrial licensing, small scale reservation and essential commodities control, were never the subject of repealing legislation at all: the first lives on as an exempted statute of full formal validity, the second lived and died entirely in notifications between 1967 and 2015, and the third stands entire, its use pruned by order and revived by order as prices command.

Counting generously, then, legislation initiated the change in none of the eight cases at the moment the change took effect in the world; it ratified one change within the year, two within the decade, one within two decades, and three it has not ratified to this day. The complementary half of the reform, the building of the successor order, shows the opposite distribution: the securities regulator, the insurance regulator, the electricity commissions and the competition commission were every one of them created by Act, and the exchange code was rebuilt by Act, so that the architecture of regulation is parliamentary work throughout. The Indian liberalisation, in statutory terms, was executive in its subtractions and legislative in its additions, and the two halves proceeded at different speeds, which is why the statute book of the period reads as it does: swift silence where the old law was concerned, slow construction where the new was being built, and between them the long intervals that this article set out to measure. The finding is offered not as an epitaph on the reforms, about whose economic merits this article holds its peace by design, but as the answer to the One Test the brief for this article prescribed: the reader who can name the statutes, date the removal of each control, and state which instrument removed it, understands the legal shape of Indian economic reform better than the standard economic narrative, unaided, can teach.

The tariff: protection moves to the one instrument Parliament votes every year

There is a final displacement in the story that ties the legal analysis together, because the function the licence performed did not vanish; part of it migrated. The industrial licence had been, among its other characters, an instrument of protection: by refusing entry and rationing imports of competing goods, it sheltered the domestic producer from foreign competition more completely than any tariff could. When licensing and import licensing were withdrawn in the 1990s, protection did not disappear from Indian economic law. It relocated to the customs tariff, the schedule of duties enacted in the Customs Tariff Act and varied by the annual Finance Act, where peak rates, among the highest charged by any major economy at the start of the 1990s, were reduced budget by budget across the decade under the reform programme. The relocation changed the politics and the law of protection together. A licence was an individual, discretionary, invisible act of protection, its beneficiary and its cost both hidden in a file. A tariff is a general, published, legislative act of protection, voted in a Finance Bill, visible to every importer and trading partner, and disciplined by the international commitments India undertook in the same decade as a member of the World Trade Organization, whose bound rates capped the tariff from outside. Protection survived liberalisation, as it survives every liberalisation, but it was obliged to move from the gazette notification to the statute, from the particular to the general, and from discretion to rule. If the argument of this article needed a single emblem for its distinction between the executive layer and the legislative layer of Indian law, it would be this migration: what could remain discretionary remained executive; what had to be defended in public had to become law.

The permission that migrated: environment, land and the surviving prior controls

An accurate epitaph for the licence raj must record that prior permission, as a technique of Indian economic government, did not die in 1991; it changed its subject. The industrial licence had vetted a project for its conformity with the plan: its product, its size, its promoter, its place. As that vetting fell away, another vetting rose to practical dominance over the same project, the environmental clearance, administered under the Environment (Protection) Act, 1986 through the impact assessment notifications of 1994 and their successor of 2006, under which specified categories of project may not begin construction without a prior environmental clearance granted after an assessment process. Around it stand the forest clearance under the forest conservation statute for projects on forest land, the coastal regulation zone notifications governing construction near the shore, and the state level consents to establish and operate under the water and air pollution statutes, administered by the state pollution control boards. A promoter of the reform era therefore did not escape the file and the queue. The questions asked in the file changed, from whether the economy needed the plant to what the plant would do to its air, water and neighbourhood, and the asking moved from the industry ministry to the environment ministry and the state boards.

The migration is instructive for the argument of this article in two ways. First, it confirms that the Indian state’s withdrawal in 1991 was a withdrawal from a purpose, planning the composition of industry, and not from the technique of prior permission as such, which was redeployed to purposes, environmental protection chief among them, whose legitimacy no camp in the liberalisation debate contests. The permission state survives; it has been re-founded on new grounds, and its legal instruments display the family’s familiar traits: an enabling Act of general powers, operative content in notifications, exemption categories revised by executive instrument, and a practical regime that a reader of the parent Act alone would never reconstruct. Second, the migration explains the texture of business complaint in the decades after reform, in which licensing as a word survived the licence as an institution, applied by promoters to the environmental and land clearances that had become the binding constraints on a project. The complaint was loose in its law and accurate in its experience. Entry had been freed at the Centre in a notification of 1991. Beginning, at the site, remained a matter of permissions, and the distance between entry and beginning is where the licence raj’s ghost has lived ever since.

One industry through the whole sequence: cement and steel

The pattern of this article can be tested against a single industry, and cement serves because it met every layer of the regime. Capacity in cement was licensed under the Industries Act, so the map of Indian cement production was drawn by the licence: how much, and where, in a product whose economics are dominated by the location of limestone and the cost of carriage. Price and distribution were controlled under the essential commodities framework, with a levy system obliging producers to surrender a share of output at a controlled price for public works and priority distribution, the free sale residue carrying the economics of the plant. Freight pooling softened the geography the licence had drawn. The industry’s reform accordingly arrived in layers matching the bundle. Price and distribution control was dismantled first, in stages through the 1980s and completely by 1989, an executive withdrawal under the control order machinery, before the crisis and before the word liberalisation attached to the process. Licensing of capacity ended with the general delicensing of 1991. The industry that resulted, consolidated, capacity rich and priced by the market, was built inside the shell of the two statutes that had governed its controlled era, neither of which was repealed for the purpose.

Steel repeats the sequence with the public sector at its centre. The great steel plants were public enterprises under the policy reservation; private capacity was licensed; price and distribution of steel were administered through a joint plant committee structure that operated the industry’s pricing as a matter of arrangement rather than of market; and imports of steel were canalised and licensed. Reform unpicked the bundle string by string on the familiar pattern: administered pricing and distribution were withdrawn in the early 1990s as the economy opened, entry was freed by delicensing, and the sector’s subsequent expansion was executed substantially by private producers operating under the same Industries Act that had licensed their predecessors, in plants the licensing system would have rationed. Neither industry required, or received, an industry specific statute of liberalisation. The general instruments, exemption notification, control order withdrawal, pricing decision, did the work, which is the strongest practical evidence for this article’s central claim: by 1991 the legal technique of liberalisation was already standardised, because it consisted of running the control regime’s own instruments in reverse.

A note on the evidence, for the reader who checks

The dates and instruments in this article have been stated at the level at which they can be verified against the public legal record: the texts of the Acts and their amending and repealing Acts, the commencement and exemption notifications under them, the policy statement of 24 July 1991, and the decided cases named in the courts section, each described by its proposition rather than by pinpoint citation. Where the record itself is untidy, the article has preferred the durable statement to the precise one: the reserved list for the public sector is given in its settled heads rather than by every intermediate revision, the foreign investment ceilings in the retained sectors are traced in their stages rather than by the date of every press note, and the rundown of the small scale list is dated by its end in April 2015 rather than by the score of notifications that composed it. That discipline costs the narrative some granularity and buys it the property this series values above narrative pace, that a reader who checks any load bearing statement in it, the Act numbers, the repeal dates, the intervals in the table, will find the record as the article left it. The errors this period attracts in casual writing, the licensing Act described as repealed, the monopolies repeal dated to 1991, the policy statement treated as a statute, are precisely the errors the discipline exists to prevent, and they are worth naming once more at the end, because a reader who has absorbed nothing else should leave knowing that the most famous repeal of the Indian twentieth century is one that never took place.

Three dates in the life of a law

The liberalisation sequence suggests a general lesson with which this article closes, about the biography of a statute. A law in India has at least three dates, and they need not coincide. The first is the date of enactment, when the text enters the statute book. The second is the date of commencement, which this series has repeatedly found trailing the first by months or years, entire Acts awaiting the appointed day. The third, which standard accounts rarely record, is the date on which the law ceases to be applied, the date its requirements stop governing conduct in the world. For the repealed control statutes the third date preceded the legal death of the text by years. For the Industries Act, the third date arrived in July 1991 for nearly the whole of its licensing scheme, and the law has lived ever since in the interval between application and existence, a condition that has no name in the standard vocabulary of legislation and that this period made normal.

For the student of Indian law, the practical instruction of the sequence is a habit of reading. A statute cited in a textbook is the beginning of an inquiry and not its end: the reader must ask whether the Act has been commenced in the relevant part, whether its requirements have been exempted, relaxed or superseded by notification, whether its functions have been transferred in practice to a successor body whose own statute stands elsewhere, and whether its repeal, if it has been repealed, saved the liabilities of its operative years. The control era’s statutes reward exactly that discipline, because every one of those questions has, somewhere in this article, carried the whole answer. The licence raj ended, in the world, on dates this article has given: July 1991 for licensing, September 1991 for the size gate, May 1992 for capital issues, the turn of the century for exchange, 2009 for the monopolies statute, 2015 for reservation. It ended, in the books, on fewer and later dates, and for its central statute it has not ended at all. Both statements are true, and the space between them is where the legal history of modern India is written.

A reader of the series will recognise the family resemblance to its other findings: commencement notifications that never issued, judgments whose directions awaited machinery, rules struck down for travelling beyond their Acts. In each case the formal record and the operative record diverge, and the truth of the legal position lies in the second. Liberalisation is the grand instance because it was deliberate and successful: a government that knew the difference between changing the law and changing the legal position chose the second as its instrument, and a legal system whose statutes had been drafted with internal executive doors let it. The statutes were the last thing to change. In the case that symbolises the whole era, they have not changed yet, and the licence raj, which ended for the citizen in the monsoon of 1991, continues on paper as an available future, filed between the Acts of 1951 and everything that has happened since. That double existence, dead in the world and alive in the book, is not a defect in the record of the reforms. It is the record of the reforms, the truest single statement of how they were made: by a state that had learned to govern its economy through the instruments its statutes placed in its own hands, and that changed the economy first and its law afterwards, in its own time, and in one celebrated case not at all.

Frequently Asked Questions

Q: What was the licence raj in statutory terms?

It was a bundle of consent requirements, each imposed by a separate statute, that together governed whether a business could enter an industry, how much it could produce, how large it could grow, whose money it could take, what it could import and where it could build. The principal statutes were the Industries (Development and Regulation) Act, 1951 for entry and capacity; the Monopolies and Restrictive Trade Practices Act, 1969 for size and expansion; the Foreign Exchange Regulation Act, 1973 for foreign exchange and foreign shareholding; the Capital Issues (Control) Act, 1947 for public issues and their pricing; the imports control statute of 1947; the Essential Commodities Act, 1955; and the urban land ceiling statute of 1976. Around them sat reservation of products for small scale units and phased manufacturing programmes imposed as conditions of approval.

Q: What did the 1991 industrial policy statement change?

The Statement on Industrial Policy of 24 July 1991 abolished industrial licensing for all industries except a specified list, removed the requirement of prior approval under the monopolies statute for substantial expansion, new undertakings, mergers and takeovers by large undertakings, opened an automatic approval route for foreign equity up to 51 per cent in a specified list of high priority industries, abolished the phased manufacturing programmes, and announced disinvestment in selected public enterprises and the referral of chronically sick public enterprises to the reconstruction board. Its legal force was indirect. Licensing ended through exemption notifications under the licensing Act, the monopolies change through an amending Act passed that September, and the foreign equity route through the administrative machinery of exchange control. The statement itself amended no statute.

Q: Which licence raj controls were removed by law and which by notification?

Industrial licensing was removed by notification, through exemptions issued under a power in the Industries Act itself, and small scale reservation was created and ended by notification as well. The monopolies prior approval requirement was removed by law, through the amending Act of 1991 that deleted the relevant sections. Capital issues control was removed by law, through repeal in 1992. Foreign exchange control was changed by law only in 2000, when the management Act replaced the regulation Act, having been loosened administratively for the intervening years. Import licensing faded through trade policy before its statutory frame was replaced in 1992. The land ceiling was repealed by law in 1999, state by state. Essential commodities control was pruned by order and never removed by law at all.

Q: Which control statutes were repealed after liberalisation?

Three of the central control statutes were repealed outright. The Capital Issues (Control) Act, 1947 was repealed in 1992 as the statutory securities regulator took over the capital market. The Foreign Exchange Regulation Act, 1973 was repealed by the Foreign Exchange Management Act, 1999, which replaced it from 1 June 2000 on a philosophy of management rather than conservation. The Monopolies and Restrictive Trade Practices Act, 1969 was repealed with effect from 1 September 2009 under the Competition Act, eighteen years after its core approval provisions had been deleted. The urban land ceiling statute was repealed by an Act of 1999 that took effect state by state. The Industries Act 1951 and the Essential Commodities Act 1955 were not repealed, and both remain in force.

Q: Why did the statutes lag behind liberalisation policy?

Four reasons recur in the record. The first is capacity: a crisis government in 1991 had weeks, and exemption notifications under existing powers could be issued in days, while bills required drafting, cabinet approval and parliamentary time. The second is design: the control statutes had been drafted with wide exemption and relaxation powers, so the executive rarely needed Parliament to switch a control off. The third is sequence: repeals were generally timed to the readiness of a successor, the securities regulator in 1992, the competition commission by 2009, so the statute waited on the institution. The fourth is indifference: once a control ceased to bite, its statute harmed no organised constituency, and legislative time went to measures with living supporters. A dead statute is politically weightless, which is why dead statutes accumulate.

Q: What replaced the licence raj with economic regulation?

The replacement was a set of statutory regulators and a conduct based competition law, built by legislation over the following decade. The securities market passed to the Securities and Exchange Board of India under its Act of 1992, applying disclosure requirements rather than consent and pricing control. Insurance was opened and regulated under the Insurance Regulatory and Development Authority Act, 1999. Electricity was reorganised by the Electricity Act, 2003, which delicensed generation and erected regulatory commissions. Telecommunications had received a statutory regulator in 1997. Competition, the successor discipline to monopolies control, was enacted in the Competition Act, 2002, with its prohibitions on agreements and abuse of dominance in force from 2009 and combination regulation from 2011. Regulation replaced permission: the state stopped deciding in advance who might act and reserved itself the power to police how they acted.

Q: Which licence raj controls were never formally removed?

Three stand out. Industrial licensing was never formally removed from the statute book: the Industries (Development and Regulation) Act, 1951 retains its licensing chapters, disapplied by exemption notification to all but a short list of industries, and the Act continues in force with its other chapters operative. The Essential Commodities Act, 1955 was never removed either; its control powers were pruned by the withdrawal of orders for many commodities, but the enabling Act survives complete, available whenever a government wishes to control production, distribution or price of a scheduled commodity. The third is subtler: the monopolies statute’s conduct provisions on restrictive and unfair trade practices were never separately repealed in 1991 and lived on inside the Act until the whole statute was repealed in 2009. Formally, the control era’s central statute and its commodity control statute outlived the control era without a closing date.

Q: What does the sequence show about Indian economic reform?

It shows that the operative legal instrument of reform was the executive act and that legislation played a confirming rather than an initiating role. Every major control was ended in practice before it was ended in law, the intervals ran from months to eighteen years, and in the leading case the statutory change never arrived. It also shows that Parliament, whenever it did legislate in this field across four decades and under governments of different parties, legislated in the direction the executive had already taken: it built regulators, replaced the exchange statute and repealed the monopolies statute, and it never once restored a control the executive had lifted. The pattern is neither parliamentary initiative nor executive usurpation. It is executive initiative under powers Parliament had conferred, followed by parliamentary ratification at leisure, with the statute book left permanently behind the economy it purports to govern.

Q: Did the 1991 liberalisation require an Act of Parliament to take effect?

Most of it did not, and the parts that did are identifiable precisely. Delicensing took effect through exemption notifications under the Industries Act, the foreign equity automatic route through administrative decisions under exchange control, and the abolition of phased manufacturing programmes by executive decision, none of which required legislation. Two elements required Acts and received them: the removal of prior approval for large undertakings, because the requirement stood in the operative sections of the monopolies statute and was deleted by the amending Act of 1991, and the abolition of capital issues control, because a consent requirement of that kind cannot be notified away, and it was repealed in 1992. The division between what needed Parliament and what did not was dictated by how each control had been drafted decades earlier.

Q: Who was the Controller of Capital Issues and what replaced the office?

The Controller of Capital Issues was the government officer who administered the Capital Issues (Control) Act, 1947. A company wishing to issue securities to the public needed the Controller’s consent, and the consent machinery extended to the amount, the type of security and the price, including the premium at which shares could be offered. The office therefore priced a large part of the capital raised in the economy by administrative decision. It was abolished in 1992 when the Act was repealed, and its function was not transferred to another officer but transformed: the Securities and Exchange Board of India, given statutory status the same year, regulates issues through disclosure requirements and eligibility conditions, leaving the price to be set by the issuer and tested by the market.

Q: How did FERA control foreign shareholding in Indian companies?

The Foreign Exchange Regulation Act, 1973 approached foreign shareholding as a claim on a scarce national resource. Companies with foreign equity above the prescribed level, set at 40 per cent for the general run of cases, were treated as foreign controlled for the purposes of the Act and were expected either to dilute their foreign holding to the permitted level or to justify their structure within guidelines administered by the Reserve Bank, which directed them towards export oriented or high technology activity in return for toleration. Below that level, entry of foreign equity was by permission, sector by sector and case by case. The consequence was a generation of ownership structures designed around a statutory threshold, including the famous dilution exercises by multinational subsidiaries in the late 1970s, before the guidelines, the threshold policy and finally the statute itself were successively abandoned.

Q: What changed in legal character when FEMA replaced FERA?

The change was one of kind, not merely of degree. The 1973 Act was built on prohibition: a transaction in foreign exchange was impermissible unless allowed, contravention was a criminal offence punishable with imprisonment, and the enforcing directorate worked in the manner of a criminal agency. The Foreign Exchange Management Act, 1999, in force from 1 June 2000, is built on regulation by category: current account transactions are free unless restricted, capital account transactions are permitted as the regulations specify, and contravention is a civil matter visited with monetary penalty, compounding being available for most defaults. The long title records the shift of purpose, from conservation of foreign exchange to facilitating external trade and payments and promoting the orderly development of the foreign exchange market. The citizen’s posture before the state changed from supplicant for permission to regulated actor.

Q: Which industries still require an industrial licence after 1991?

A short list, revised from time to time by notification, principally of industries connected with defence, security and specified strategic or hazardous activity. The 1991 notifications left licensing in place for eighteen industries; successive notifications reduced the list, and its settled content in the decades after reform comprised a handful of heads, arms and ammunition and allied defence equipment, specified defence and aerospace items, and certain hazardous chemicals and industrial explosives among them. The precise list is a matter of the notification in force rather than of the Act, which illustrates the article’s central point: the boundary of compulsory licensing in India is drawn by executive instrument under a 1951 statute, and it moves when the government notifies, without any amendment to the parent Act.

Q: Why did the MRTP Act survive until 2009 after losing its core in 1991?

Because only half of it had died. The amending Act of 1991 deleted the provisions on concentration of economic power, the prior approval regime for large undertakings, and left standing the provisions on monopolistic, restrictive and unfair trade practices, together with the Commission that administered them. That surviving half had live work: complaints of restrictive and unfair practices continued to be filed and decided, and no successor body existed to receive them. The succession was designed in the Competition Act, 2002, whose Section 66 both repealed the 1969 Act and distributed its pending business, cases of monopolistic and restrictive practice to the appellate tribunal, unfair practice matters towards the consumer fora, and investigations as the section directed. Repeal waited until that machinery was ready, taking effect on 1 September 2009.

Q: What were the phased manufacturing programmes that the 1991 statement abolished?

They were administrative schemes attached to approvals of foreign collaboration and import of technology. When a firm was permitted to manufacture a product with imported components or foreign technical collaboration, the approval could carry a phased manufacturing programme: a timetable, fixed by the government, by which the imported content of the product was to be progressively replaced with domestically manufactured content over a stated number of years. The programme was a condition of doing business rather than a provision of any statute, enforced through the approval machinery and the allocation of import licences for components. Its purpose was indigenisation by compulsion. Because it had never been statutory, its abolition in 1991 required no legal instrument beyond the policy decision itself, and it is the purest example in the sequence of a control that existed and ended entirely at the administrative layer.

Q: How did small scale industry reservation come to an end?

By the same instrument that created it: notification. Reservation of items for exclusive manufacture by small scale units was effected through notifications under the industrial policy framework administered with the Industries Act, beginning with a list of 47 items in 1967 and growing at its height to more than eight hundred items that larger units were barred from manufacturing. From the 1990s onwards, committees reviewing the policy recommended progressive dereservation, and the government dereserved in tranches, government by government, as the economic case for protecting each product weakened. The process took about two decades and ended in April 2015, when the final twenty items on the list were dereserved and the reservation list stood empty. No Act was passed at any point in the story, because reservation had never been enacted; it was policy with legal effect, and it died as it had lived.

Q: Could a later government bring back industrial licensing without a new Act?

For industries covered by the exemption notifications, the legal answer is yes, in form. Licensing was not repealed; it was disapplied by notifications issued under the exemption power in the Industries Act, 1951, and a notification withdrawing or narrowing an exemption would revive the requirement for the industries affected, without any bill being introduced. Three constraints would operate in practice. The doctrines of reliance weigh against abrupt reversals where specific assurances were given and acted upon. The political cost of reimposing a control that a generation of enterprise has never known would be substantial. And the rest of the liberalised settlement, the regulators, the competition statute, the exchange management regime, stands in enacted law that a notification cannot touch. Licensing could be revived by signature. The economy it would govern could not be restored by signature.

Q: Do firms have a legal right that an industrial policy will stay unchanged?

No. The courts have consistently held that economic policy is the province of the executive and the legislature, that a change of policy is not reviewable merely because it disappoints those who benefited from the earlier policy, and that no estoppel arises against the state for the exercise of a statutory power in public interest, so an exemption or concession granted under a power may be withdrawn under the same power. What the law protects is narrower: reliance on a specific, clear promise made to an identifiable party and acted upon to that party’s detriment, which the doctrine of promissory estoppel, recognised in the sugar mills case of 1979 and qualified in later cases on exemption notifications, may hold the state to. A regime is not a promise. A firm trading under a policy holds a position defeasible by a lawful change of that policy.

Q: Which control statute was repealed in the same phase as its function was transferred?

The Capital Issues (Control) Act, 1947. Its repeal in May 1992 formed part of a single legislative package with the Securities and Exchange Board of India Act, 1992, which gave statutory status to the regulator that received the capital market. The office of the Controller of Capital Issues was abolished, consent and official pricing of issues ended, and disclosure based regulation began, all within the same season. It is the exception that defines the pattern in this article. The repeal was prompt because it could not be otherwise: the control consisted entirely of a statutory consent requirement administered by a named office, there was no exemption power broad enough to convert it into market regulation, and the receiving institution already existed in non-statutory form waiting to be clothed with authority.

Q: What happened to import licensing and essential commodities controls after 1991?

They diverged. Import licensing withered as a system: quantitative restrictions administered through licences gave way to a negative list approach and then to regulation principally by tariff, and the statutory frame itself was recast when the Foreign Trade (Development and Regulation) Act, 1992 replaced the control Act of 1947, changing the governing idea from control of imports and exports to development and regulation of foreign trade, though licensing survives within it for specified goods. Essential commodities control took the opposite path in form while moving in the same direction in practice. The 1955 Act was never repealed or substantially rewritten. Instead, governments pruned its application by order, removing commodities from control, withdrawing dealer licensing for several commodities in the early 2000s and narrowing stock holding limits, leaving the enabling power intact and periodically reactivated when prices rise.