Every account of Indian competition law begins with a contrast, and most accounts state the contrast correctly and then fail to use it. The Monopolies and Restrictive Trade Practices Act 1969 asked how large an undertaking was. The Competition Act 2002 asks what an undertaking has done to a market. The MRTP vs Competition Act comparison is therefore not a comparison between an old statute and its amendment, or between a weak regulator and a strong one. It is a comparison between two different theories of what the state should fear in private economic power: accumulation itself, or conduct and transactions that damage the competitive process. Everything else in this article, the thresholds, the remedies, the merger rules, the leniency programme, the fate of consumer complaints, follows from that single difference, and the verdict at the end is derived by counting how the difference plays out across nine dimensions rather than by asserting it at the start.

A second caution belongs at the front, because it changes what an honest comparison can claim. The two regimes never operated in their full forms one after the other. The older statute lost its concentration provisions in 1991, eighteen years before it was repealed. The newer statute was enacted in 2002 but its conduct provisions took effect only in 2009 and its merger provisions only in 2011. Between 1991 and 2011 India had neither the old permission regime nor the new merger control, and between 1970 and 2009 it had a competition authority that could not impose a fine. Any verdict that compares the two statutes as though the second simply succeeded the first on a single date is comparing two things that never stood in that relation, and this article therefore assesses both regimes against a third state, the interval in which neither operated, because naming that interval converts an easy comparison into an informative one. The wider field in which both statutes sit is mapped in the guide to industrial and competition law in India, and this article takes its texts from the two owning accounts rather than restating them.

The first regime: the state as gatekeeper of size

The Monopolies and Restrictive Trade Practices Act 1969, Act 54 of 1969, in force from 1 June 1970, was built on a proposition drawn from the Directive Principles and from the inquiries that preceded it: that the concentration of economic power in private hands was itself an evil to be prevented, whether or not any particular exercise of that power had harmed anyone. The full statutory history, the committees behind it and the mechanics of its repeal are set out in the account of the MRTP Act 1969 and its end, and only the features that decide the comparison are taken up here.

The organising measure of the statute was the asset value of an undertaking. An undertaking whose assets, together with those of interconnected undertakings, crossed the prescribed threshold, fixed at twenty crore rupees at the start and raised to one hundred crore rupees in 1985, attracted the concentration provisions of Part A of Chapter III, sections 20 to 26. A dominant undertaking was identified by a market share test of one third, later one fourth, combined with a far lower asset floor. Once an undertaking stood on the register, its growth became a matter of prior permission. Section 21 required approval before a substantial expansion. Section 22 required approval before establishing a new undertaking. Section 23 required approval before an amalgamation, merger or takeover. The trigger was not conduct and not harm. The trigger was size, crossed in advance of the act that would increase it, and the deciding authority in the permission decisions was the Central Government acting on advice, which made the regime an instrument of industrial policy quite as much as a branch of law.

Around that core the statute arranged its conduct provisions. Restrictive trade practices were defined and then enumerated, with a list of registrable agreements in section 33, and a practice that fell within an enumerated head could still escape through the gateways in section 38, a set of justifications that included such matters as the public interest in negotiating fair terms and the absence of a substantial lessening of competition. Monopolistic trade practices formed a separate chapter and were tested against the public interest. Unfair trade practices were added in 1984, in section 36A, giving individual consumers and traders a forum for complaints about misleading representations and false promises at a time when no dedicated consumer statute existed. The MRTP Commission inquired, and its remedies were characteristic of the design: it could direct that a practice be discontinued and not repeated, it could pass interim injunctions under section 12A, and under section 12B it could award compensation to a person who had suffered loss or damage and applied for it. What it could not do, at any point in its forty years, was fine anyone. Appeals from the Commission lay directly to the Supreme Court under section 55.

The design had a coherence that later commentary often misses. If the harm is accumulation, then the remedy is permission before accumulation, and a commission that inquires and directs is a sufficient instrument, because the heavy work is done at the gate. The regime failed, where it failed, in the premises rather than in the drafting: asset value is a poor proxy for market power, a firm can be large in assets and harmless in every market it serves, and a firm can be small in assets and hold a stranglehold over a narrow market that the threshold never sees. The 1991 amendment, Act 58 of 1991 with effect from 27 September 1991, deleted Part A of Chapter III altogether, which was Parliament conceding the premise while keeping the building. What remained from 1991 to 2009 was a conduct statute without its core, adjudicating restrictive and unfair practices with remedies that had been designed as the lesser half of a larger scheme.

Why did the older regime fall out of favour with policymakers?

The short answer is that its premise was abandoned before its machinery was. The Raghavan Committee, whose report of 22 May 2000 recommended repeal and replacement, treated concentration as a fact to be assessed for its effects rather than a wrong to be licensed.

The economic policy of the decade after 1991 had already removed the permission system the statute existed to operate. A gatekeeping statute cannot survive the decision to open the gate.

That answer needs its other half stated at full strength, because the stance of this series is that contested questions get both sides in their own voice. Defenders of the older design, and the Raghavan Committee itself recorded the argument before departing from it, held that in an economy of scarcity and licensed capacity, size was not a neutral fact: the house that accumulated assets across unrelated markets accumulated political influence, preferential access to credit and a capacity to pre-empt entrants that no single market inquiry would capture. On that view the asset threshold was crude because the phenomenon it targeted was crude, and the effects based test that replaced it is blind to a real species of power. The counter argument, which prevailed, is that a test blind to some power is still better than a test that punishes size without regard to power at all, and that the conglomerate concern, if it is to be addressed, needs an instrument designed for it rather than a threshold designed for something else. This article returns to the conglomerate question below, because it is the one dimension on which the older statute did something the newer one does not attempt.

The second regime: the state as policeman of conduct

The Competition Act 2002 is, despite its short title, Act 12 of 2003, and its text is owned by the account of the Competition Act 2002 explained, from which the comparison takes its anchors. The statute prohibits, in section 3, agreements which cause or are likely to cause an appreciable adverse effect on competition within India, and it prohibits, in section 4, the abuse of a dominant position. It regulates, in sections 5 and 6, combinations above prescribed thresholds, which must be notified before they are closed and may not be given effect until the Commission has decided or the statutory period has run. Holding a dominant position is not an offence under it. Being large is not an offence under it. Size enters only as one factor among the enumerated factors in section 19(4) by which dominance in a relevant market is assessed, and dominance enters only as the precondition for asking whether specified conduct amounts to abuse.

The analytical apparatus is written into the statute in a way the older Act never attempted. Relevant market, relevant product market and relevant geographic market are defined in sections 2(r), 2(s) and 2(t). The factors for assessing appreciable adverse effect are enumerated in section 19(3), those for dominance in section 19(4), and those for combinations in section 20(4). Horizontal agreements of the kinds listed in section 3(3), price fixing, output limitation, market sharing and bid rigging among them, are presumed to cause an appreciable adverse effect, and the burden moves to the parties to displace the presumption. Section 32 gives the Commission express jurisdiction over acts taking place outside India which have or are likely to have an appreciable adverse effect within it. Section 46 provides for a lesser penalty where a cartel participant makes a full and true disclosure, the leniency provision, and the Competition (Amendment) Act 2023, Act 9 of 2023, layered on a leniency plus mechanism in force from 20 February 2024 and settlement and commitment mechanisms for non cartel conduct in force from 6 March 2024, together with a deal value threshold for combinations in force from 10 September 2024.

Enforcement was staged. The Competition Commission of India was established on 14 October 2003. Sections 3 and 4 took effect on 20 May 2009. Section 66, which repealed the older statute and transferred its pending business, took effect on 1 September 2009. The combination provisions took effect on 1 June 2011. Institutional design changed with the substance: the Commission investigates and adjudicates, a Director General conducts the investigation under its direction, and appeals lie to a separate appellate body, the Competition Appellate Tribunal until 2017 and the National Company Law Appellate Tribunal thereafter, rather than to a single court of final appeal as under the older Act. The remedies changed most of all, and that change is taken up in the dimensions below.

The nine dimensions, taken one at a time

A comparison table can record a verdict, but it cannot earn one. This section works through the nine dimensions in prose, and the table that follows it does no more than collect the conclusions reached here, so that the verdict column can be checked against the reasoning rather than taken on trust.

The organising concept: assets against effect

The older statute organised itself around a stock: the value of assets an undertaking held. The newer statute organises itself around a flow: the effect of conduct or of a transaction on competition in a market. The difference sounds abstract and is anything but. An asset test can be applied by an accountant from a balance sheet, without knowing what the undertaking sells, to whom, or against whom it competes. An effects test cannot be applied at all until the market has been defined, the undertaking’s position in it measured, and the conduct characterised, which is why the newer statute needs definitions of relevant market and enumerated assessment factors and the older one needed none. The asset test has one genuine virtue, certainty of application, and one genuine vice, that it measures something other than the thing the law cares about. A steel plant and a software firm with identical asset values present entirely different competitive questions, and a statute that treats them identically at the threshold has decided to regulate a proxy. The effects test has the opposite profile: it measures the right thing and is correspondingly expensive, slow and contestable. Every later dimension in this comparison is a working out of that trade, and any reader who understands this first dimension has the key to the other eight.

The trigger: crossing a line against doing an act

Under Part A of Chapter III the trigger was a state of affairs, the crossing of an asset line, and the legal consequence attached before any arguably harmful act occurred, at the point of expansion, establishment, amalgamation or takeover. Under the Competition Act the conduct provisions are triggered by an act, an agreement or an abuse, and the combination provisions by a transaction that meets the statutory description in section 5 and is prohibited from closing by section 6 until cleared. The practical consequence is a reversal of who must move first. In the permission regime the undertaking had to approach the state and justify its growth in advance, and silence or delay on the state’s part was itself a decision against it. In the prohibition regime the undertaking acts, and the state must approach it, build a case on evidence, and justify intervention after the fact, except in merger control, where the posture of prior notification was deliberately retained. That exception is instructive: Parliament in 2002 did not reject prior control as such. It rejected prior control of growth by existing firms, and kept prior control of structural change through combination, which suggests the draftsman had learned to distinguish between supervising what a firm is and supervising what a transaction does.

The treatment of size: suspect against neutral

The older statute treated size as presumptively suspect and dominance as a status carrying obligations of prior restraint. The newer statute is formally neutral about size: section 4 prohibits abuse of a dominant position and says nothing against holding one, and the Competition Commission has repeated in its advocacy and its orders that dominance is not prohibited, only its abuse. The neutrality is real but should not be overstated, and the honest statement of the position requires both halves. A dominant firm under the newer regime is subject to a set of prohibitions, on unfair or discriminatory pricing, on limiting production or technical development, on denial of market access, on tying and on leveraging into another market, that do not bind its smaller rivals at all, so the regime is not indifferent to size; it attaches special duties to market power while refusing to attach them to asset value. The older regime attached duties to asset value whether or not market power existed. Between a duty that follows power and a duty that follows wealth, the first is better aimed, and the second is easier to administer, and the whole history of the two statutes is contained in that sentence.

Merger control: permission, then nothing, then notification

This dimension contains the most interesting fact in the entire comparison, and the one most accounts omit. Under the older statute in its full form, amalgamations, mergers and takeovers involving undertakings above the threshold required the prior approval of the Central Government under section 23. The 1991 amendment deleted that requirement along with the rest of Part A. The Competition Act restored merger control in the form of mandatory pre notification to a specialist regulator with a standstill obligation, but the combination provisions took effect only on 1 June 2011. It follows that from 1991 to 2011, a period of twenty years covering the most intense restructuring of Indian industry after liberalisation, no statute in India controlled mergers at all. Acquisitions that would have required permission before 1991 and notification after 2011 required neither in between. The comparison on this dimension is therefore not between two systems but between a system, an absence, and a system, and the absence lasted longer than the full operation of the regime that preceded it. Assessed on design, the notification model is superior to the permission model in two respects: the deciding body is a specialist regulator applying published criteria in section 20(4) rather than a government applying industrial policy, and the standstill is time bound, so that the state’s silence approves rather than blocks. Assessed on history, the dimension must be scored against both regimes together, because the gap was created by repealing the old control twenty years before switching on the new one, and that sequencing decision belongs to neither statute and damages the record of both.

The analytical apparatus: enumeration against definition

The MRTP Act worked by enumeration. It listed the agreements that had to be registered, listed the practices that counted as restrictive, and listed the gateways through which a practice could be justified. It contained no definition of a relevant market, no statutory concept of dominance built on market analysis, and no definition of a cartel. Its central defined term, restrictive trade practice in section 2(o), was framed around the manipulation of prices or conditions of delivery and the prevention, distortion or restriction of competition in general terms, and the adjudication that followed was largely an exercise in classification: did the conduct fall within an enumerated head, and if it did, did a gateway save it. The Competition Act works by definition and factor. It defines the market, enumerates the factors for adverse effect and for dominance, presumes the effect of listed horizontal agreements, and requires the Commission to reason from evidence about a defined market to a conclusion about effect. The difference in apparatus is the difference between a statute that asks its adjudicator to sort and a statute that asks its adjudicator to analyse. Sorting is faster, more predictable and less dependent on the quality of the individual decision maker. Analysis is more accurate, more demanding, and capable of reaching conduct that no list anticipated, including conduct in markets that did not exist when the statute was drafted. Digital markets make the point concretely: a list of registrable agreements written in 1969 could not have contemplated a platform’s self preferencing, whereas an effects test written in 2002 reaches it without amendment to the prohibition, whatever difficulties the assessment itself presents.

Enforcement powers: direction against penalty

Here the difference is not of degree but of kind, and it is the dimension a practitioner would name first. The MRTP Commission could inquire, could direct that a practice be discontinued and not repeated, could injunct in the interim under section 12A, and could award compensation under section 12B to an applicant who proved loss. It could not impose a penalty, and a direction to stop a practice that had already yielded its profit was, for a calculating firm, a cost free instruction to keep the proceeds and change the method. Section 27 of the Competition Act empowers the Commission to impose a penalty of up to ten per cent of the average turnover of the preceding three financial years, and for a cartel the proviso permits the greater of three times the profit made or ten per cent of turnover for each year the agreement continued. Section 28 empowers it to order the division of a dominant enterprise. Section 33 empowers interim relief. The Supreme Court’s construction in Excel Crop Care v Competition Commission of India confined the turnover base to the turnover relevant to the infringement rather than the whole of a multi product enterprise’s turnover, a discipline on the power that the older regime never needed because it had no such power to discipline. The enforcement record that these powers produced, and the controversies over quantum and recovery that came with them, are examined in the account of CCI antitrust enforcement in India, which this comparison does not duplicate. For the verdict column the point is narrower: deterrence requires that infringement be capable of costing more than it earns, and exactly one of the two statutes was capable of that.

Incentives to disclose: silence against leniency

A cartel is a secret, and a statute that offers nothing to the first participant who breaks the secret must detect cartels from outside, by complaint or by accident. The MRTP Act offered nothing. It had no leniency provision, no settlement mechanism and no reward for cooperation, and its remedial scheme gave a participant every reason to deny and none to disclose. Section 46 of the Competition Act introduced the lesser penalty for a participant that makes a full, true and vital disclosure, and the Competition (Amendment) Act 2023 added leniency plus, in force from 20 February 2024, under which a participant already cooperating in one cartel can earn an additional reduction by disclosing a second and separate cartel, and settlement and commitment mechanisms, in force from 6 March 2024, under which an enterprise in a non cartel case can offer commitments or settle rather than litigate to the end. The significance for the comparison is that disclosure incentives change the epistemics of enforcement: the authority stops depending on the visibility of the offence and starts depending on the instability of the conspiracy, which is a far more reliable foundation. On this dimension there is no counter case for the older statute to make, and the verdict column records it accordingly.

Reach: the territory of the offence

The MRTP Act had no extraterritorial provision. Its concern was with undertakings and practices within the Indian economy as the statute conceived it, and conduct arranged abroad that produced its effects inside India sat awkwardly with its scheme. Section 32 of the Competition Act is express: the Commission may inquire into an agreement, an abuse or a combination taking place outside India if it has or is likely to have an appreciable adverse effect on competition in a relevant market in India. In an economy in which the parties to a cartel may sit in three jurisdictions and the market harmed may be a fourth, an effects based reach is not an embellishment but a condition of the statute working at all. It also completes the logic of the first dimension: a law organised around effects within a market has no reason to stop at the border where the conduct occurred, whereas a law organised around the assets of undertakings registered within a permission system naturally stops at the boundary of that system.

Consumer protection: a forum that moved rather than improved

Unfair trade practices entered the MRTP Act in 1984, in section 36A, and for five years the MRTP Commission was the only specialised forum in India in which a consumer could complain of a false representation, a misleading advertisement or a bargain price that was not intended to be offered at all. The Consumer Protection Act 1986 then created the consumer fora, and the two jurisdictions overlapped until the repeal, when section 66 of the Competition Act sent the pending unfair trade practice business to the consumer side under the transfer scheme whose mechanics the owning article sets out. The Competition Act itself contains no consumer complaint jurisdiction over misleading advertising: its prohibitions are addressed to effects on competition, and a false claim that harms buyers without harming the competitive process is not its business. On this dimension the change cannot be scored as an improvement in the competition statute, because the function was not improved but relocated, and whether the relocation served consumers depends on the performance of the consumer fora and of the Central Consumer Protection Authority under the Consumer Protection Act 2019, which is a question about a different statute. What can be said for the newer design is that it separated two functions that the older statute had combined by historical accident, the protection of the competitive process and the redress of individual deception, and that a specialist competition authority deciding advertising complaints was always an anomaly. What can be said against it is that something was genuinely lost in the move, namely a single forum in which a small trader injured by a rival’s false claims could frame the complaint as a trade practice and obtain a direction quickly, and the verdict column records this dimension as a relocation with losses on both sides.

Nine dimensions, two regimes, one verdict column

Dimension	The MRTP regime	The competition regime	Verdict
What the law measures	The asset value of an undertaking and its interconnected undertakings	The effect of conduct or of a transaction on competition in a defined relevant market	Improved
What triggers intervention	Crossing an asset threshold, before expansion, establishment, merger or takeover	An agreement, an abuse of dominance, or a notifiable combination	Improved
Treatment of size	Size suspect in itself, carrying prior restraints until 1991	Size neutral; dominance lawful, only its abuse prohibited	Improved
Merger control	Prior approval of the Central Government until 1991	Mandatory pre notification to a specialist regulator with a standstill, from 2011	Improved in design; absent from 1991 to 2011
Analytical apparatus	Enumeration of practices and gateways; no definition of relevant market or cartel	Statutory definitions and enumerated factors for effect, dominance and combinations	Improved
Enforcement powers	Direction to discontinue, interim injunction, compensation on application; no penalty	Penalty on turnover, division of an enterprise, interim relief	Improved
Incentives to disclose	None; no leniency, no settlement, no reward for cooperation	Lesser penalty for disclosure, leniency plus, settlement and commitment	Improved
Reach beyond India	No extraterritorial provision	Express jurisdiction over foreign conduct with effects in India	Improved
Consumer complaints	Unfair trade practices adjudicated by the Commission from 1984	No such jurisdiction; the function sits with the consumer fora	Relocated, not improved

The count is the conclusion in embryo. Seven of the nine dimensions improved, and none of the seven is marginal: each represents a capability the older statute did not possess at all, a penalty power, a disclosure incentive, a market definition, a reach beyond the border. One dimension, merger control, improved in design and failed in history. One dimension, consumer complaints, was not improved but moved. A comparison that ended there would be accurate and incomplete, because it would treat the two statutes as contemporaries, and the sections that follow restore the dimension of time and then test the design verdict against the three constituencies the law is supposed to serve.

Institutional design: one body against a separated structure

The dimension the table cannot hold is institutional. The MRTP Commission combined the functions of inquiry and decision within a single body served by a Director General of Investigation and Registration, and its orders were appealable directly to the Supreme Court under section 55, which made the Supreme Court the first and only appellate forum for trade practice adjudication. The Competition Act separates the functions in stages: a Director General investigates under the Commission’s direction and reports, the Commission adjudicates, and an appeal lies to a specialist appellate tribunal, the Competition Appellate Tribunal from 2009 until its merger into the National Company Law Appellate Tribunal in 2017, with a further appeal to the Supreme Court on questions of law. Each design has a characteristic failure. The combined body risks deciding on the strength of its own inquiry, and the direct appeal to the Supreme Court meant that a small number of trade practice questions received the attention of the highest court while the great mass of the Commission’s output was never tested on appeal at all, because the cost and gravity of a Supreme Court appeal filtered out all but the largest disputants. The separated structure distributes scrutiny more evenly and builds a body of appellate precedent at a level where volume is possible, at the price of an additional tier of litigation and the delays that attend it. On balance the separated structure is the better design for a regime whose decisions impose penalties in crores, because the gravity of the sanction is what makes intermediate appellate scrutiny necessary, and it is one more instance of the same pattern: the older institution was adequate to the older remedies, and would not have been adequate to the newer ones.

How did each regime treat a dominant firm that charged very high prices?

Under the older statute, unreasonably increasing the price of goods or services could itself be inquired into as a monopolistic trade practice and tested against the public interest, so the price charged by a powerful firm was a direct object of the law.

Under section 4 of the Competition Act, the imposition of an unfair or discriminatory price, including a predatory price, is an abuse where the firm is dominant in a relevant market, so the price is reachable only through the gateway of dominance in a defined market and a characterisation of unfairness.

The contrast is the whole comparison in miniature. The older route was shorter: no market definition, no dominance inquiry of the statutory kind, and a direct question about reasonableness. It was also correspondingly more exposed to the objection that a tribunal deciding what a reasonable price is has become a price controller, which was an office the Indian state of that era was willing to occupy and the Indian state of the later era was not. The newer route is longer and demands economic evidence, but it confines intervention to cases where market power has first been established, and it leaves the prices of firms without dominance alone however high they may be. Which of those is the better rule depends on a judgment about the competence and proper role of the state that the statutes embody but cannot prove, and the honest comparison records the difference rather than pretending that one of the two answers is a drafting error.

What did each regime do about public sector undertakings?

The older statute was extended to public sector undertakings by the amendment of 1991, the same amendment that removed its concentration core, so that for the last eighteen years of its life the state enterprise stood inside the conduct provisions while the private house had been released from the permission provisions. Under the Competition Act an enterprise includes a government department or undertaking engaged in an economic activity, and the Commission has proceeded against public sector entities, most prominently in the Coal India litigation over the conditions of fuel supply agreements, in which the argument that a statutory monopoly stood outside competition law was rejected. Sovereign functions of government are not an enterprise activity, and section 54 lets the Central Government exempt a sector or enterprise in the interests of state security or public interest, a power the older statute did not confer in that form. The trajectory is one of widening formal coverage and continuing argument about its edges, and it is common ground between the regimes’ defenders that neither statute solved the problem of the state as market participant, since the state’s advantages are conferred by instruments other than the competition statute and can be withdrawn only by changing those instruments.

Can the two regimes be compared at all if they never ran in full together?

They can, provided the comparison is between designs and between the periods in which each design actually operated, and provided the long interval between them is treated as evidence rather than as an embarrassment to be passed over in polite silence.

What cannot be done is to compare outcomes across the whole span and attribute the difference to the statutes, because the span also contains liberalisation, trade opening, the growth of services and a technological transformation, and no method exists for isolating the statutes’ contribution from that company.

The interval: the third state against which both must be judged

Between the deletion of Part A in 1991 and the commencement of the combination provisions in 2011, India ran an experiment that no committee recommended and no statute described: a large, liberalising economy with no merger control, no permission regime for the growth of large firms, and a conduct authority whose remedial ceiling was a direction to stop. The surviving MRTP Act continued to adjudicate restrictive and unfair trade practices, and after 2002 a Competition Act existed on the statute book with a Commission established in 2003 whose principal prohibitions were not in force. The sequencing is the subject of the narrative account of the licence raj and its dismantling, and the comparison needs only its consequences.

Three consequences matter. The first is that the consolidation of Indian industry in the two decades after liberalisation, the assembly of the large diversified groups in their present form, the consolidation of cement, of telecommunications, of aviation and of organised retail, took place under no structural supervision whatever, and the market structures that the Competition Commission inherited in 2009 and 2011 were formed in that interval and could be examined by it only prospectively. Merger control is inherently prospective: it can prevent a concentration, and it cannot readily unscramble one that closed lawfully twenty years earlier. The second consequence is that a generation of practitioners, officials and judges formed its habits in the interval, in which competition law was a minor branch of practice concerned with trade practice complaints, and the analytical apparatus of the new Act had to be learned by the institution charged with applying it at the same time as it was applied. The third is that the interval is the strongest available evidence about what each regime was for. When the permission regime was abolished, nothing in the political system demanded its replacement for a decade, which suggests that the constituency for concentration control had always been the state itself rather than any group that suffered from its absence. When cartel penalties began to be imposed after 2009, by contrast, the complaints came from purchasers and rivals who had lacked a remedy, which suggests that the second regime answered a demand the first had not been designed to meet.

Counting the interval against the record changes the verdict in one respect and leaves it standing in another. It leaves standing the judgment about design, because the comparison of the two statutes’ capacities does not depend on when they were switched on. It changes the judgment about the transition, because a reform that removes a control in 1991 and replaces the function in 2011 has, for the intervening period, not reformed the control but abolished it, and any assessment of the change that stops at the two statutes’ texts will miss the largest single fact in the story.

What the older statute did that the newer one does not

The counter reading required by an honest comparison can be put as a list of functions performed by the MRTP Act for which the Competition Act contains no equivalent, and each deserves to be stated without the softening that usually attends it.

It addressed concentration of economic power as such. An undertaking could attract the statute’s attention by accumulating assets across markets that had nothing to do with each other, and the state’s objection was to the accumulation. The Competition Act has no provision addressed to conglomerate accumulation that produces dominance in no defined relevant market. Commentators divide on whether this is a gap or a correct narrowing. The narrowing case, associated with the Raghavan Committee’s reasoning, is that power which cannot be exercised against anyone in any market is not a competitive harm and that the state’s objection to it was an objection to private wealth in a particular form, properly a matter for taxation and company law rather than for a competition statute. The gap case, pressed in the digital markets debate by those who argue that a platform group’s presence across many adjacent markets gives it advantages of data, attention and bundling that a market by market inquiry understates, is that the definition of the relevant market does the excluding before the analysis begins, so that cross market power is not found because the instrument cannot look at it. Both positions are held by serious people, and the comparison should record that the older statute, whatever its crudity, was at least capable of posing the question.

It provided a forum in which an individual could complain of a trade practice and obtain a direction, including a trader complaining of a rival’s conduct and a consumer complaining of a deception, without first establishing an effect on competition in a market. The Competition Act receives information from any person, but its inquiry is into the statutory prohibitions, and a grievance that does not translate into an agreement, an abuse of dominance or a combination is outside its remit however meritorious. The individual redress function moved to the consumer fora, and the small trader’s complaint about a rival’s false claim must travel by that road or by the law of passing off and trade marks, neither of which was designed for it.

It also kept a register. The registration of registrable agreements under section 33, and of undertakings under the concentration provisions, meant that the state possessed a standing record of the arrangements and the accumulations it was charged with watching. A prohibition regime keeps no such record, because it has no occasion to: it learns of arrangements when a complaint or a leniency application or its own research brings them to notice. The loss is easy to exaggerate, since registers of that kind decay into formalism and were widely evaded, but the change from a state that required notification of arrangements as a matter of routine to a state that discovers them by enforcement is a real change in the relation between the state and business, and it belongs on the older statute’s side of the ledger.

The function only one statute had: advocacy

Section 49 of the Competition Act obliges the Central Government, when formulating a policy on competition or a law on a competition issue, to be able to seek the Commission’s opinion, and empowers the Commission to take suitable measures for the promotion of competition advocacy and the creation of awareness about competition issues. The MRTP Act contained nothing of the kind, and the omission is characteristic. A permission regime has no need to persuade anyone of anything: its instrument is the application and the order, and the regulated population’s understanding of the policy is relevant only to the extent of securing compliance with a queue. An effects based prohibition regime is in a different position. Its prohibitions are general, its tests are evaluative, and its success depends on thousands of firms, procurement officers, trade associations and lawyers understanding, without being told individually, that a price fixing understanding is a serious offence and that a leniency route exists. Advocacy is the mechanism by which such a regime manufactures its own deterrent climate, and the Commission’s first years, when its prohibitions were not yet in force, were spent largely on it, which turned the accident of staged commencement into something resembling a preparation period. The Commission’s advocacy output, its papers on public procurement, its engagement with state governments whose purchasing practices had competition consequences, and its market studies are part of its statutory identity in a way that has no parallel in the older body, whose public face was its orders. For the comparison the point is that the newer institution was designed to change a culture as well as to decide cases, and that this ambition, whatever its uneven fulfilment, belongs on the list of capabilities the older design never attempted.

The tempo of intervention

Both regimes could act before a final order, the older through the interim injunction in section 12A and the newer through the interim relief in section 33, which permits the Commission to restrain a party temporarily where an act in contravention has taken place or is about to take place. The verbal similarity hides a difference in use that follows from the rest of each design. Interim relief in a permission regime is a holding operation inside a system whose default position is that growth awaits approval, so that restraining a party pending inquiry disturbs no settled expectation. Interim relief in a prohibition regime cuts the other way: the default position is that a firm may act unless the state proves its case, and restraining it before that proof reverses the regime’s own presumption for the duration of the inquiry, which is why the power has been exercised sparingly and argued about intensely, and why appellate courts have scrutinised its use closely. The comparison here is not between a strong power and a weak one but between two systems in which the same words carry different weights because of what surrounds them, and it is a useful reminder, for anyone reading statutes comparatively, that the operative unit of comparison is never the section but the section in its system. Much of the confusion in popular accounts of this transition comes from comparing remedies section by section, as though a direction, an injunction and a penalty were items on a single scale of severity, when in fact each remedy derives its meaning from the trigger that precedes it and the appeal that follows it.

Two commissions as employers of expertise

The older Commission’s work required lawyers, accountants and administrators, and its inquiry reports read accordingly. The newer Commission requires, in addition, industrial economists in numbers, a professional investigation directorate with powers of search and seizure exercised through the Director General, and a cadre capable of reviewing combination filings against filing deadlines measured in days. This difference in establishment is easy to overlook in a comparison of texts, and it explains a part of the newer regime’s history that its text does not: the institution had to be built while it was operating, its economic capability had to be recruited against the competition of private practice, and several of its early difficulties, in the drafting of orders, in the handling of confidential material, in the recovery of penalties, were difficulties of an institution growing into a design that assumed capacities it was still acquiring. The older institution never faced that problem, because its design assumed little. Judged on the fit between institution and task, the older design was the more realistic piece of drafting for the state that enacted it, and the newer design the more ambitious for the state that enacted it, and the history of the second is in part the history of the state deciding, over two decades, to become the kind of state its own statute presupposed.

What each regime knew about the economy it governed

A statute is also an information system, and the two regimes were built to know different things. The permission regime knew the intentions of large firms in advance, because they had to declare them: expansions, new undertakings, amalgamations and the agreements they entered were notified or registered, and the state’s picture of the economy, however imperfectly the registers were maintained, was a picture assembled from declarations made before the event. The prohibition regime knows the economy retrospectively and adversarially. It learns of an arrangement from an informant, from a leniency applicant, from a filing made because a threshold was crossed, or from a market study it commissions after a suspicion has formed, and its picture is assembled from evidence gathered against resistance. Each information posture has a characteristic blindness. The declaratory system is blind to whatever its categories do not require to be declared, and its categories were drawn around the formal acts of large undertakings, so that the informal understandings among small and medium firms, and the practices of firms below the threshold, lay outside its vision almost by definition. The adversarial system sees what it looks for with great clarity and has no general vision at all: it has no register of agreements, no standing census of arrangements, and its knowledge of a sector between cases may be thinner than the knowledge the older machinery possessed as a by product of routine. Competition advocacy and market studies are the newer regime’s attempt to rebuild a general vision without rebuilding the register, and the attempt is instructive about the trade the transition made. The state gave up knowing a moderate amount about everything as a matter of routine in exchange for the ability to know a great deal about one thing when it mattered, and to act on that knowledge with a penalty. Whether the exchange was well made depends on the frequency of the occasions that matter against the value of routine knowledge, and the honest answer is that India has had enough cartel and abuse cases of consequence to justify the exchange, and has also had sectors drift into structures nobody was watching, which is the price of it.

Procedure grew with the penalty

A direction to stop a practice can be issued on a procedure that would be intolerable if the order took a tenth of a company’s turnover, and the procedural history of the two regimes follows their remedies with precision. The MRTP Commission’s proceedings were comparatively summary, its inquiry was its own, and the discipline imposed on it by the appellate court was the ordinary discipline of administrative law. The Competition Commission’s proceedings have been rebuilt by appellate scrutiny into something much closer to a trial: the Director General’s report must be disclosed and answered, the material relied on must be shared, a party must be heard on the penalty as a distinct question, confidential material must be ring fenced by order rather than by practice, and orders have been set aside or remanded for failures of each of these kinds, in litigation of which the steel and cement cases are the large examples. Critics of the newer regime count this as evidence of its dysfunction, and defenders count the same facts as evidence of its maturation. The comparison suggests a third reading. Procedural rigour is a cost a legal system pays in proportion to the severity of what it authorises, and a regime that fines on turnover must be tried to a standard a regime that directs discontinuance need not meet. The older statute’s procedural economy was therefore not a virtue it chose but a consequence of its weakness, and it could not have been retained if the penalty power had been added to it, as reform proposals periodically suggested. Adding a fine to the MRTP Act, a suggestion made more than once in its lifetime, would not have produced a shorter route to the newer regime. It would have produced the newer regime’s procedure wearing the older statute’s definitions, which is to say it would have reproduced the difficult half of the transition while keeping the weak half of the design.

The consumer statute that grew up in the older regime’s shadow

For twenty three years, from 1986 to 2009, India had two statutes under which a deceptive trade practice could be pursued, the MRTP Act before its Commission and the Consumer Protection Act 1986 before the district, state and national commissions, and the coexistence is a neglected part of the comparison because it explains what the repeal actually did to consumers. In the coexistence period the two forums divided the work by the character of the complainant and the remedy sought. A trader injured by a competitor’s false representation belonged naturally before the MRTP Commission, whose order could direct the practice to stop and whose compensation provision could be invoked by an applicant who proved loss. An individual buyer belonged naturally before the consumer forums, which were cheaper, local and empowered to order the ordinary consumer remedies of refund, replacement and compensation. The overlap was untidy and occasionally produced parallel proceedings, but it was not useless, because the two forums were good at different things and a complainant with advice could choose. The repeal of 2009 ended the trader’s route without replacing it, since the consumer forums are built around the definition of a consumer and a trader complaining of a rival is not one. What the consumer gained in the same period came from the strengthening of the consumer side, culminating in the Consumer Protection Act 2019 with its central authority empowered to act against misleading advertisements on its own motion, a power closer in spirit to the MRTP Commission’s than to the district forum’s. The fair conclusion is that the consumer protection function was never really part of competition law’s logic and was parked in the MRTP Act for want of a better address, that it has since found a better address, and that the trader’s grievance against a rival’s deception, which was genuinely competition adjacent, fell between the two addresses in the move and remains the one unrepaired casualty of an otherwise well sorted transfer.

Dominance in depth: a share of one fourth against a list of factors

No single provision shows the distance between the regimes as economically as the treatment of the leading firm. The older statute identified a dominant undertaking by arithmetic: an undertaking which, by itself or with interconnected undertakings, supplied not less than one third of the goods or services in India, later one fourth, and met a modest asset condition. The test could be applied from published figures, it made no inquiry into the structure of the market, the height of its barriers or the countervailing power of its buyers, and it treated a firm with a fourth of a fragmented national market and a firm with a fourth of a market fenced by licences as the same legal object. Section 19(4) of the Competition Act lists the considerations from which dominance is to be inferred: market share, the size and resources of the enterprise, the size and importance of its competitors, its economic power and commercial advantages, vertical integration, the dependence of consumers, monopoly or dominant position acquired by statute or by being a government company, entry barriers of every kind including regulatory and financial barriers and economies of scale, the countervailing buying power in the market, the market’s structure and size, and the social obligations and social costs of the enterprise, together with its relative advantage by way of contribution to economic development and any other factor the Commission considers relevant. The list is long because the phenomenon is complex, and its length is the point. Dominance under the newer statute is a conclusion argued from evidence about a market, and it can be contested on every element of the list, which makes the inquiry slower and the result less predictable and the conclusion, when reached, about the thing itself rather than about a fraction. The older test had the virtues of a rule and the defects of a rule applied to a matter that does not come in rule shaped units. It is worth adding that neither test is a test of conduct, and both regimes understood that a leading position is a condition and not a wrong; they differed in what followed from the condition, prior restraint of growth in the one case and a schedule of prohibited abuses in the other, and that difference of consequence is larger than the difference of measurement.

Cartels: the offence neither statute named alike

The older statute had no offence of cartelisation and no definition of a cartel. Collusive arrangements came before the Commission, when they came at all, as restrictive trade practices within the registered agreement framework or as monopolistic trade practices, and the analytical question was whether the arrangement fell within a head and survived the gateways, among which the gateway protecting agreements that prevented a substantial lessening of competition in a trade carried an irony the case law never quite resolved, since the respondent in a collusion inquiry was invited to justify the arrangement by the very criterion the arrangement offended. The newer statute names the conduct in section 3(3), presumes its effect, attaches the highest penalty in the Act to it through the proviso to section 27, and surrounds it with the lesser penalty machinery of section 46 and the leniency plus extension of the 2023 amendment, so that the cartel provisions form a connected system of prohibition, presumption, deterrence and induced defection. The contrast is the sharpest in the comparison after the penalty dimension itself, and it is not accidental that it is so. Bid rigging and price fixing flourished in the interval and under the older regime in sectors, construction inputs, pharmaceuticals distribution, transport and public procurement among them, where the purchasers were dispersed and the practice was invisible from outside. A regime that cannot see an offence and cannot punish it and cannot reward its disclosure has in substance decided not to pursue it, whatever its text says, and the older regime’s text did not even isolate the offence for pursuit. The newer regime’s cartel record is open to criticism on quantum, on delay and on recovery, and the enforcement account in this series makes those criticisms, but the criticism is of a working system, which is a different order of fault from the absence of one.

Why 2002 did not mean 2002

Students meeting this material in an examination regularly date the transition to 2002, and the error is worth correcting in a comparison because it conceals the comparison’s most important variable. The Competition Act received assent as Act 12 of 2003. Its Commission was established in October 2003 and for several years the institution’s effective work was advocacy, research and the building of a cadre, because the prohibitions it existed to enforce had not been commenced, and a constitutional challenge to the institutional design, in Brahm Dutt v Union of India, occupied the interval and produced amendments about the composition and selection of the Commission before the adjudicatory scheme settled. The conduct provisions commenced on 20 May 2009, the repeal of the older statute on 1 September 2009, and the combination provisions on 1 June 2011, so that the new regime arrived in three instalments across eight years, each instalment changing what the comparison can say. A comparison written in 2005 would have contrasted a functioning if weakened older regime with a new regime that existed as text and institution but not as law in action. A comparison written in 2010 would have had the conduct provisions and the repeal but no merger control. Only from 2011 do both regimes exist in forms complete enough to compare, and by then the older one had been dead for two years. The lesson generalises beyond this pair of statutes: the date of a reforming Act is the least informative date in its history, and a series organised by statutory dates must keep returning, as this article has, to the dates of commencement, because they are where the law begins to act on anyone.

Speed, cost and access: the comparison from the user’s side

Design comparisons are conducted from the statute book outward, and they flatter whichever statute reads better. From the user’s side the picture is more mixed, and the mix should be recorded. An MRTP proceeding was comparatively cheap to start and comparatively quick to a direction, the Commission’s procedure was inquisitorial and informal by the standards of a court, and an individual complainant could appear without constructing an economic case. Its weakness was that the direction obtained at the end of that accessible process was worth little against a respondent who had already taken the profit, and that compensation required a separate application and proof of loss to a standard many small applicants could not meet. A Competition Act proceeding is expensive in a different currency. The information filed by an informant costs little, and the Commission’s inquiry is conducted at the state’s expense, which is a real and underrated form of access, since the informant’s burden ends where the Director General’s begins. But the proceedings are long, the orders are long, the appeals are substantial, the penalty litigation has in prominent instances run for years and been reduced or remanded on questions of turnover and natural justice, and a small party’s practical position depends on the authority’s priorities rather than on its own persistence. Access in the older system was access to a weak remedy in person. Access in the newer system is access to a strong remedy by proxy. Which of those a given complainant prefers depends on whether the complainant’s need is to be heard or to prevail, and the honest comparison reports that the two regimes optimised for different needs because they were built by states with different views of their own role, the first willing to listen and unwilling to coerce, the second willing to coerce and structured to listen mainly through the filter of a market wide inquiry.

Three fact patterns run through both regimes

Abstract comparison hardens into understanding when the same facts are put to each statute in turn, and three standard patterns do most of the work.

Take first a merger between the second and third largest producers in a concentrated industry, each comfortably below any asset threshold that mattered by the standards of the time and neither dominant in the market. Under the older statute in its full form the transaction required the prior approval of the Central Government under section 23 if the parties fell within Part A, and the inquiry was into concentration and the public interest in the large, conducted within the machinery of industrial policy; after 1991 it required nothing at all, and between 1991 and 2011 this transaction, in the very decades when Indian industry consolidated, was nobody’s business. Under the newer statute it is a combination, notifiable if the section 5 thresholds of assets or turnover, or the deal value threshold, are met, incapable of being closed until the Commission has decided or the statutory period has run, and assessed on the factors in section 20(4), which ask about the level of combination in the market, the barriers to entry, the likelihood of a significant and sustained increase in price or profit margins, the countervailing power of buyers and the nature of vertical integration. The notification itself is a substantial documentary exercise, the annexures running to agreements, board papers and financial statements in spreadsheet and presentation form, commonly worked through with an office file viewer for Excel, Word and PowerPoint documents before the notice is settled. The same transaction is thus, across the three periods, a matter of permission, a matter of indifference and a matter of notification and effects review, and the middle period is the longest of the comparison’s surprises.

Take second a leading firm’s refusal to supply a distributor who has begun stocking a rival’s goods. Under the older statute the complaint would be framed as a restrictive trade practice, refusal to deal being among the registrable and examinable practices, and the inquiry would run to classification and then to the gateways, where the firm might plead the protection of its distribution standards or the absence of a substantial lessening of competition, with a direction to discontinue as the ceiling of the outcome. Under the newer statute the same facts must first clear two gates the older inquiry lacked: the firm must be dominant in a defined relevant market, and the refusal must amount to a denial of market access or another abuse within section 4. If those gates are cleared the outcome can include a penalty on turnover. If they are not cleared, because the firm is merely large in a market the Commission defines widely, the distributor has no competition remedy at all, whereas under the older statute it had one that did not depend on dominance. The pattern shows the newer regime at its most characteristic: harder to enter, far more consequential once entered, and deliberately indifferent to grievances that fall below the threshold of market power.

Take third a trade association that circulates a recommended price list to its members, who follow it. Under the older statute the arrangement was a registrable agreement and a restrictive practice, the association’s defence ran through the gateways, and the realistic outcome was a direction. Under the newer statute the arrangement is an agreement between enterprises at the same level that determines prices, it attracts the presumption of section 3(3), the association itself is reachable as a facilitator and after the amendment of 2023 a non competing participant acting in furtherance of it is caught by the hub and spoke provision, the participating members face the cartel penalty under the proviso to section 27, and any member that discloses the arrangement fully may earn a lesser penalty under section 46 and further credit under leniency plus if it discloses a second arrangement elsewhere. No single pattern separates the regimes more completely. The first statute saw a document to be registered and a practice to be directed to stop. The second sees the most serious offence in its book, presumes its effect, prices it by turnover and profit, and offers a bounty in reduced penalties to the first participant who betrays it.

The edges of each regime: exemptions and carve outs

A regime is also defined by what it leaves out, and the omissions differ instructively. The older statute as enacted stood at a distance from the state’s own undertakings and from areas the government chose to notify, and its extension to public sector undertakings came only in 1991. Its subject was the private concentration of a licensed economy, and the state that licensed the economy was not originally an object of its suspicion. The newer statute begins from the opposite posture and then qualifies it: an enterprise is defined to include government departments and undertakings when they engage in economic activity, and the qualifications are placed in the open, in the exclusion of sovereign functions, in the power of exemption in section 54 and in the saving in section 3(5) for the reasonable conditions a holder of an intellectual property right imposes to protect that right, a carve out with no equivalent in the older statute and one that has generated its own line of argument about where the protection of a right ends and the restraint of a market begins. The pattern is consistent with everything else in this comparison. The older statute’s boundaries were boundaries of administrative convenience in a system the state ran. The newer statute’s boundaries are stated as propositions, about sovereignty, about security and public interest, about the rights of innovators, and are therefore arguable in a forum, which is where the newer regime prefers its questions to be settled.

Competition law among its neighbours

Neither statute operated alone, and part of what changed in the transition is the company the competition law kept. The MRTP Act belonged to a family of control instruments that included the licensing system of the Industries (Development and Regulation) Act 1951 and the price and distribution controls available under the Essential Commodities Act 1955, and its concentration provisions made sense as the member of that family charged with watching the size of the firms the other members were licensing and supplying. Within that family, the state’s posture was uniform: the significant decisions about capacity, entry and price were taken administratively, in advance, by the government, and the competition statute’s role was to police the accumulations that the family’s own decisions produced. The arrangement had a logic and a fatal circularity, since the licensing decisions were the main source of the market power the MRTP Act was then asked to restrain, and the restraint consisted of further administrative decisions by the same apparatus. The Competition Act belongs to a different family, that of the sector regulators created from the 1990s onward for electricity, telecommunications, securities, insurance and ports, bodies which assume that entry is free, that prices are formed in markets, and that the state’s role is to set the terms of conduct and to intervene on evidence. Its jurisdictional borders with those regulators have had to be litigated, most prominently where a sector regulator and the Commission both claimed a question, and the Supreme Court’s resolution in the Bharti Airtel case, requiring the sector regulator’s determination first on the matters within its technical remit, is the kind of boundary law that the older family never needed, because within that family the boundaries were internal to government. The change of family is not a detail of administrative history. It is the institutional form of the change of theory, and it explains why the newer statute’s difficulties are the difficulties of coordination between independent bodies, while the older statute’s difficulties were the difficulties of a single apparatus marking its own work. A reader deciding between the regimes is in the end deciding between those two kinds of difficulty, and the comparative record of the two families, in India and elsewhere, is that coordination problems are corrigible by doctrine and by practice whereas self review problems are corrigible only by changing the institution, which is what, in 2002 and again in the staged commencements that followed, the Indian state eventually did.

The chronology is itself an argument

Did the Competition Act change anything on the day it was passed?

Very little, and the chronology shows why. Assent in January 2003 created a statute and, in October of that year, a Commission, but the prohibitions on agreements and abuse commenced only in May 2009 and merger control in June 2011, so for six years the new regime was an institution in preparation rather than a law in action.

Set the two statutes out as a single line of dates and the line argues before any commentary is added. The older Act commenced on 1 June 1970. Unfair trade practices were added to it in 1984, the year its consumer chapter began. The asset threshold was raised in 1985, an admission inside the statute’s own lifetime that the line had been drawn where ordinary growth crossed it. In 1991 the concentration provisions were deleted and the Act was extended to public sector undertakings, so that in a single amendment Parliament removed its core, widened its coverage and left its remedies untouched, a combination that tells a reader more about the politics of that year than about any theory of competition. In 2000 the Raghavan Committee reported, on 22 May, and recommended that the Act be repealed and a competition law enacted. The Competition Act followed, assented to in January 2003 as Act 12 of 2003. In 2005 the Supreme Court’s observations in Brahm Dutt v Union of India, on the separation of the Commission’s adjudicatory and regulatory functions, set in train the amendment of 2007, Act 39 of 2007, which created the Competition Appellate Tribunal and recast the selection and structure of the Commission, before a single prohibition had been enforced. The conduct provisions commenced on 20 May 2009. The older Act was repealed on 1 September 2009, thirty nine years after it commenced and eighteen after it had lost its core. Merger control commenced on 1 June 2011. The appellate tribunal was merged into the National Company Law Appellate Tribunal in 2017. The amendment of 2023, Act 9 of 2023, added the deal value threshold, leniency plus, and settlement and commitment, commenced in stages between May 2023 and September 2024.

Read as a narrative, that line has an odd shape. The decisive events are not the enactments but the deletions and the commencements, and the longest intervals are the ones in which the law on the book and the law in action diverged: the eighteen years of the older Act without its core, the six years of the newer Act without its prohibitions, and the twenty years across both in which no merger control existed. A comparison that confined itself to the enacted texts would place the transition in 2002 or 2003 and would be wrong by nearly a decade in both directions. This is why the brief for the present article insists on the three phase history and on the interval, and why the verdict distinguishes so sharply between design and sequencing. The design changed once, on paper, in a single statute. The law in action changed at least five times, in 1984, in 1991, in 2009, in 2011 and in the staged commencements of 2023 and 2024, and each of those changes redistributed rights and exposures among firms, complainants and the state. Statutory comparison that ignores commencement is a comparison of intentions. The intentions are worth comparing, and this article has compared them, but the reader who wants to know what the law was at any moment in the last half century needs the second list, and needs to be told plainly that for long stretches of that half century the honest answer is that the function in question was performed by nobody.

The word in the title

The two statutes announce their difference in their titles before a single section is read. The Act of 1969 is named for a thing to be restrained, monopolies, together with the practices, restrictive and later unfair, through which the restraint would work. The Act of 2002 is named for a condition to be protected, competition, and its long title speaks of the establishment of a commission for the prevention of practices having an adverse effect on competition, the promotion and sustenance of competition in markets, the protection of the interests of consumers and the freedom of trade of other participants. A statute named for a villain invites its administrators to hunt the villain, and size was the villain’s most visible feature. A statute named for a process invites its administrators to ask what has damaged the process, which is a question that cannot be answered by looking at the defendant’s balance sheet. Behind the older title stood the Constitution’s own language, in clause (c) of Article 39, directing the state to secure that the operation of the economic system does not result in the concentration of wealth and means of production to the common detriment, and the MRTP Act was understood in its era as an instrument of that direction. The newer statute does not repudiate Article 39, but it relocates its implementation: concentration is to be prevented from causing harm by policing the conduct and the transactions through which harm occurs, rather than by restraining concentration itself. Constitutional lawyers disagree about how much of the older reading survives in the newer scheme, and the disagreement matters in the argument over conglomerate power recorded above, since the gap case is at bottom a claim that Article 39 asks a question the Competition Act has declined to answer. The comparison need not settle that argument to learn from the titles. A draftsman choosing a name and a long title is choosing the first and most durable piece of advocacy the statute will ever have, and the difference between these two documents begins, as their readers begin, on the cover.

A register and gateways, elsewhere

India was not the only country to build a competition law out of a register and a set of gateways and then to abandon the architecture, and the parallel is close enough to be evidence rather than decoration. The United Kingdom’s Restrictive Trade Practices Act 1956 required restrictive agreements to be registered and brought them before a court, where they were presumed contrary to the public interest unless the parties brought them within one of a list of gateways and showed that the benefit outweighed the detriment, an architecture strikingly similar to sections 33 and 38 of the Indian Act, which was drafted in awareness of it. The United Kingdom replaced that system with the prohibition model of its Competition Act 1998, in force from 2000, under which agreements with an appreciable effect on competition are prohibited and may be exempted on stated criteria, and the register and the gateways disappeared together. Two jurisdictions, starting from the same mid century template, each concluded after long experience that classification plus justification was the wrong machinery and that prohibition plus effects assessment was the right one, and each discovered in the change the same costs recorded in this article: heavier evidential burdens, longer proceedings, and a jurisprudence that had to be built from the beginning. The convergence does not prove that the Indian change was right, since both jurisdictions may have followed fashion as well as experience, and the European model that influenced both was available to be copied. It does shift the burden of the argument. A defender of the register and gateway design must explain why the two systems that used it longest both retired it, and why the explanation should not be the obvious one, that agreements which restrict competition are too various to be listed and too consequential to be excused by a schedule of public interests drawn up in advance.

The strongest case for the older design, stated whole

The stance of this series requires that a contested question be given its strongest opposing version before a verdict, and the verdicts above will carry more weight if the case for the MRTP design is assembled in one place at full strength rather than conceded in fragments. That case runs as follows. India’s economy in 1969 was one of administered scarcity, licensed capacity and chronic shortage, in which the allocation of the right to produce was itself the principal economic decision and was made by the state. In such an economy, market power was not primarily created by conduct; it was created by the licence, and it was held by the houses that had obtained licences early and accumulated them. A competition law addressed to conduct in defined markets would have been a law addressed to the symptom. The asset based concentration provisions were addressed to the stock of advantage itself, and their crudity reflected the crudity of the underlying allocation mechanism they policed. Nor was the permission system the absurdity of later caricature: where capacity is licensed, some authority must decide who expands, and placing the concentration question inside that decision was administratively rational. The regime’s defenders add that its abolition was not preceded by the construction of anything in its place, that the markets it left behind were in several sectors highly concentrated, that the consumer jurisdiction it carried was the only one of its kind for a generation, and that the effects based regime, for all its analytical sophistication, has struggled with the problems the old regime was built for: the political weight of very large diversified groups, the dependence of small suppliers on dominant buyers, and the capacity of a group to be present at every point of a value chain without dominating any single defined market within it. Those are not nostalgic arguments. Each names a phenomenon that exists, and the dispute is about whether a competition statute is the right instrument for it.

The answer this article gives, and the reason the dimension count survives the counter case, is that the counter case proves too much about the old statute and too little. It establishes that the phenomena are real. It does not establish that an asset threshold and a permission queue were an effective response to them even in their own period, and the evidence of the period is against it: the licensing system the concentration provisions policed was itself the principal source of the concentration, the large houses grew under the regime and partly by mastering it, since an incumbent with a licence and a legal department was better placed to obtain the next permission than an entrant was, and the register of undertakings recorded an accumulation it did not arrest. A control that is captured by the structure it controls is worse than an absence, because it certifies what it fails to prevent. The newer regime’s blindness to cross market power is a genuine limitation, and it is the limitation of an instrument honestly confined to what it can measure, which is a better fault than the older instrument’s claim to measure what it could not.

Two theories of the state

Underneath the drafting differences sits a difference in what each era believed the state was good at. The state of 1969 believed itself competent to decide in advance which expansions of industry were in the public interest, and correspondingly distrustful of outcomes it had not authorised. Its competition statute is a branch of its planning apparatus, and reads as one. The state of 2002, legislating after a decade of liberalisation, believed itself competent to specify prohibited conduct and to punish it, and correspondingly willing to let outcomes it had not authorised stand unless they were produced by prohibited means. Its competition statute is a branch of its regulatory apparatus, and reads as one. Neither belief was naive, and each was partly vindicated and partly falsified by its own period: the planning state’s advance decisions were often wrong and always slow, and the regulatory state’s after the event enforcement has been slower than its design promised and has leaned heavily on appellate correction. What the comparison finally shows is that a competition statute is a wager on institutional competence, and that the two statutes wagered on different competences, the capacity to judge in advance and the capacity to prove afterwards. India’s experience supplies evidence against the first wager in its pure form, because the advance judgments were made by the same machinery whose errors created the problem. It supplies less conclusive evidence for the second, because the second has had less time and was given its full powers late. The verdict for design rests on that asymmetry of evidence rather than on any demonstration that effects based enforcement has yet achieved what it promises, and a careful reader should hold it at exactly that strength.

The conglomerate question, one more time

Because the conglomerate issue is the one place where the older statute’s reach exceeded the newer one’s, it deserves a final separate treatment rather than a paragraph in a list. The phenomenon is the accumulation of positions across many markets by a single group, no one of which amounts to dominance in its market, but which in aggregate confer advantages that a market by market analysis does not register: preferential access to capital at rates a single market rival cannot match, the ability to sustain losses in one market from profits in another while a contest is decided, a presence in the distribution chain at several levels, a stock of data about consumers gathered across services, and a standing with regulators and governments that is a function of total size rather than of any market position. Digital groups have made the argument vivid, because a platform’s position in search, in commerce, in payments, in content and in cloud services can each be short of dominance as a tribunal defines the market, while the combination is a fact of economic life that every participant in those markets experiences. The older statute could in principle reach the accumulation because its trigger was the group’s total assets. The newer statute cannot reach it as such, and must proceed market by market, which it does, with the result that the aggregate phenomenon is nobody’s statutory business. The responses available within the newer design are real but partial: the deal value threshold brings large acquisitions of small targets into merger review, the hub and spoke and leveraging provisions reach some cross market conduct, and the debate over an ex ante regime for large digital enterprises, carried in the draft Digital Competition Bill whose status the amendment account in this series records, is precisely the debate about whether a threshold based, obligation in advance regime should be rebuilt for a defined class of very large firms. That a Parliament which abolished the permission model should be considering an advance obligations model for a narrow class is not a refutation of the change of 2002. It is a recognition that the question the MRTP Act asked, what follows from being very large, did not die with the answer it gave, and that the newer regime’s confidence that the question need never be asked was the one point on which its design was less honest than its predecessor’s.

Agreements in practice: the register and the gateways against the presumption

The deepest practical difference between the regimes is visible in how each handled the ordinary case, an agreement between businesses that restricted competition in some respect. Under the MRTP Act a wide class of agreements had to be registered under section 33, and registration was not approval: a registered agreement could still be inquired into, and an unregistered registrable agreement was an offence, so that the register functioned as a census of arrangements the state wished to watch. When an inquiry came, the respondent’s defence ran through the gateways in section 38, a list of circumstances, ranging from the protection of the public from fraud or danger to the negotiation of fair terms with a powerful counterparty, any one of which, if established and not outweighed by the resulting detriment to the public, took the practice outside the condemnation. The leading gateway litigation, of which the Supreme Court’s decision in Mahindra and Mahindra in 1979 is the standard example, was therefore litigation about the character of the justification, argued in the language of public interest rather than in the language of market effect. A practitioner of that era needed to know the enumerated heads and the gateway jurisprudence, and needed to know very little economics.

The Competition Act inverted the practitioner’s task. There is no register, and no agreement needs to be lodged anywhere. Section 3 prohibits agreements that cause or are likely to cause an appreciable adverse effect, and the inquiry runs through the factors in section 19(3), which direct attention to barriers to entry, the driving out of existing competitors, foreclosure, the accrual of benefits to consumers, improvements in production or distribution and the promotion of technical and economic development. For the listed horizontal agreements the presumption in section 3(3) does the early work, and the parties must displace it with evidence. For vertical agreements, tying, exclusive supply and distribution, refusal to deal and resale price maintenance among them, there is no presumption and the effect must be shown. The 2023 amendment added the hub and spoke configuration, in which an enterprise that is not a party at the horizontal level participates in furtherance of the horizontal agreement, closing a gap that the enumeration of parties had left. The practitioner of this era needs the enumerated factors, the market definition case law and a working command of industrial economics, and needs the gateway jurisprudence not at all, because there are no gateways: the balancing that section 38 performed as a defence is performed inside the prohibition itself, as part of the question whether an adverse effect is appreciable.

Neither architecture is costless, and the comparison should say what each costs. The gateway architecture placed the justification in the open, as a discrete question with the burden on the respondent, and a reader of the order could see exactly which public interest had been accepted and why. The effects architecture folds the justification into an overall assessment, which is more faithful to how competitive harm actually works and less transparent about the values being traded, because a conclusion that an effect is not appreciable can carry within it a judgment about consumer benefit that never has to be named. The older architecture was also more forgiving of weak evidence, since classification and justification could be established on documents and argument, whereas an effects case can fail, and has failed, for want of a properly evidenced market. What the newer architecture buys with those costs is the ability to reach conduct that nobody enumerated, and in markets that change faster than a schedule can be amended, that ability is worth more than the transparency it trades away.

The proof problem: who had to show what

Under the older statute the Commission’s task in a restrictive practice inquiry was to bring the practice within a head and then meet the gateway case if one was raised, and its investigative support came from the Director General of Investigation and Registration, whose office also maintained the registration machinery. The standard of persuasion was that of a civil inquiry, but the decisive questions were questions of characterisation, and characterisation can be settled on the documents. Under the newer statute the Director General investigates on the Commission’s direction under section 26 and reports, the parties are heard on the report, and the decisive questions are frequently questions of measurement: the boundaries of the market, the shares within it, the height of its barriers, the direction of its prices. The Commission is therefore dependent on an investigative and economic capability that the older body did not require, and the quality of its output varies with that capability in a way the older output did not. The Supreme Court’s insistence, in the Bharti Airtel litigation decided in 2018, that the jurisdictional facts be established and that the sector regulator’s findings be awaited where the dispute turned on them, and its confinement of the penalty base to relevant turnover in Excel Crop Care, are both instances of the appellate courts constructing, case by case, the evidential disciplines that a statute of this kind needs and does not itself supply. The older statute generated a thinner and more stable jurisprudence, because it asked thinner questions. Judged as law, the newer body of doctrine is incomparably richer and still visibly unfinished. Judged as a service to a business that wants to know where it stands, the older jurisprudence had the merit that it could be stated completely in a short book, and it is not a small merit.

What changed in the courtroom

The brief for this comparison identifies the change in the character of the case law as the practical measure of the statutory change, and the measure holds up. Gateway adjudication produced judgments about language and about public interest. Effects adjudication produces judgments about evidence: about whether the market was correctly defined, whether the data supported the inference, whether the presumption was displaced, whether the penalty base was correctly confined. The lawyer’s craft changed with it, from advocacy about the meaning of an enumerated phrase to advocacy conducted through expert economic evidence and cross examined data. Two consequences follow that a verdict should weigh. The first is that the newer regime’s outcomes depend far more heavily on the resources of the parties and of the authority, since an effects case is expensive to build and expensive to answer, and a small complainant’s information may open an inquiry but cannot carry it. The second is that the newer regime is more corrigible: an error about a market can be identified as an error on appeal and corrected by better evidence in the next case, whereas an error about the meaning of a gateway hardened into doctrine. The older case law is closed and complete, a finished body of learning about a repealed statute. The newer case law is open, growing and revisable, which is what a living instrument’s jurisprudence is supposed to look like, and the comparison counts that openness on the newer regime’s side while noting its price.

The transfer scheme of 2009 as evidence about the two regimes

Statutes rarely confess their own character, but a repeal and transfer provision comes close, because it has to sort the old regime’s unfinished business into the new regime’s categories and the sorting exposes what does not fit. Section 66 of the Competition Act, in force from 1 September 2009, repealed the older statute, dissolved its Commission, and divided the pending caseload: the monopolistic and restrictive practice matters, with the compensation applications attached to them, went to the appellate tribunal of the new regime, while the unfair trade practice matters went to the consumer commissions under the consumer statute, with the Director General’s pending investigations divided on the same principle. Article 243 owns the mechanics of that division, and the comparison needs only its meaning. Of the three species of business the old Commission carried, exactly one species, the practice complaints between traders, had a home in the new competition institution, and it arrived there as an inheritance to be wound up by the tribunal rather than as a jurisdiction the new Commission would itself exercise. The consumer species was exported. The concentration species did not exist to be transferred, having been abolished eighteen years earlier. A reader who had nothing but section 66 could reconstruct the argument of this article from it: the new regime took the conduct core, declined the consumer jurisdiction, and had no use for a permission apparatus, and the transfer scheme is documentary evidence that the change was one of object and not merely of vigour.

Verdicts by constituency

A dimension count establishes capacity. It does not establish who was served, and the three constituencies a competition law exists for were not served in the same proportions.

The small competitor

The small competitor’s characteristic injuries are a cartel among its suppliers or rivals, a dominant buyer’s discriminatory terms, a dominant rival’s exclusionary conduct, and a rival’s false claims in the market. Against the first three, the newer regime is incomparably the better instrument, and the reasons are mechanical rather than rhetorical: the cartel prohibition carries a presumption and a penalty that makes the practice dangerous to engage in, the leniency provision gives the small firm’s own evidence a route into an investigation it could never have funded, and the abuse provisions name the exclusionary techniques, denial of market access and leveraging among them, that the older statute reached only by forcing them into the language of restrictive or monopolistic practice. Against the fourth injury, the false claim, the older regime was the better instrument, because an unfair trade practice complaint lay to a commission that could direct the practice to stop, whereas the newer regime offers the small trader no competition remedy at all and leaves it to the consumer fora, where a trader’s standing is awkward, or to an action in passing off. The verdict for this constituency is therefore for the newer regime, clearly, but with the loss named rather than absorbed: the injury the new law abandoned is a real and common one, and it is abandoned not because it was judged trivial but because it does not fit the definition of the new law’s object. A small competitor menaced by a cartel has never been better served in India. A small competitor menaced by a liar was better served before 2009.

The consumer

The consumer’s interest in a competition statute is mostly indirect, in prices, choice and quality over time, and is direct only where the statute gives the consumer a forum. On the indirect interest the verdict follows the dimension count: a regime that can penalise cartels, induce their disclosure and review mergers is a regime capable of protecting the conditions under which prices are formed, and the older regime, without a penalty and without merger control after 1991, was not capable of it in the same degree. On the direct interest the verdict reverses. From 1984 the consumer had a named place in the older statute, and the unfair trade practice provisions were used by individuals in a way the Competition Act’s information procedure is not, because the Commission’s question is never whether this consumer was deceived but whether competition in a market was appreciably harmed. The consumer’s direct protection was rebuilt elsewhere, in the consumer commissions and after 2019 in a central authority with powers over misleading advertisements and product recall, so the consumer as plaintiff is not obviously worse served across the whole legal system. But that is a verdict about the system, not about this change of statute, and within the four corners of the comparison the consumer lost a forum and gained a guardian of process. Guardians of process deliver their benefits diffusely and over years, and they cannot be cross examined by the person they protect. The verdict for this constituency is the most qualified in the article: better protected as a member of a market, worse served as a complainant, and the balance between those two depends on how much weight is given to diffuse long run gains against a concrete short run remedy, which is a political judgment the comparison can frame but not make.

The honest large firm

The honest large firm is the constituency the older regime treated worst and discussed least, because a regime organised against accumulation cannot distinguish in its machinery between the firm that grew by efficiency and the firm that grew by capture: both crossed the same asset line, both needed permission to expand, both faced delay whose length had no necessary relation to any question about their conduct. For such a firm the newer regime’s advantages are the advantages of a rule of law over a discretion. Its obligations are stated as prohibitions on conduct. Its size is not an offence. Its mergers are reviewed by a specialist body against published factors within a statutory period at the end of which silence is consent, and the deal value threshold added by the amendment of 2023, set at two thousand crore rupees and in force from 10 September 2024, extends review to acquisitions whose asset and turnover figures are small but whose value is not, which is a widening of the net that a firm planning a digital acquisition must price in and can at least price in, because the threshold is a number. The costs the newer regime imposes on this constituency are the costs of the effects test itself: a dominant firm’s ordinary commercial decisions, its discounts, its exclusivity arrangements, its refusals to supply, are all potentially examinable, and the boundary between vigorous competition and abuse is a matter of assessment rather than of a line. A firm that wants certainty above all would prefer a bright line, and would be wrong to prefer the old one, because the old bright line was drawn around the wrong thing and was crossed by the mere fact of success. The verdict for this constituency is the clearest of the three: the newer regime, and not by a narrow margin, because it is the only one of the two under which a firm can know that its compliance problem is about what it does.

The honest limit: capacity, not outcomes

Everything asserted so far is a claim about the law’s capacity, and the distinction matters enough to carry its own section. It is not possible to establish from the record whether market outcomes in India improved because of the change of statute. The change coincided with the liberalisation of industrial licensing and trade, with the entry of foreign capital and technology, with the growth of services and of digital markets, and with two decades in which, for much of the time, neither regime was fully in force at all. A figure for concentration, entry, margins or prices across that span is a figure about all of those causes together, and attributing it to the statute alone, in either direction, is an error that partisans of both regimes commit with equal confidence. Defenders of the change point to the growth and diversification of Indian industry after 1991; the growth followed the removal of licensing and the opening of trade at least as directly as it followed any competition provision, and for the first eighteen years of it the operative competition provisions of the new Act were not in force. Defenders of the older regime point to the consolidation of the interval and to the scale of the large groups; the consolidation occurred under no merger control, which is an argument about sequencing and not about the virtues of asset thresholds. The strongest defensible claim, and the one this article makes, is that the newer statute is better designed for the task both statutes professed, that it possesses capabilities the older one lacked on seven of nine dimensions, and that the transition was executed in a sequence that wasted twenty years. A claim about outcomes would require a counterfactual India that does not exist, and a comparison article earns its verdict by declining to invent one.

What the MRTP experiment teaches about drafting competition law

Four lessons survive the repeal, and they are lessons about drafting rather than about the politics of either era.

The first is that a statute should measure the thing it fears. The asset threshold was chosen because it was measurable, and it governed for two decades because it was administrable, and both virtues were purchased by regulating a proxy that diverged further from market power with every change in the structure of industry. Draftsmen reach for proxies under pressure of administrability, and the MRTP experience is the standing Indian demonstration that a proxy in a statute does not stay a proxy: it becomes the law’s actual object, is planned around, is litigated as though it were the object, and outlives the reasoning that produced it.

The second is that remedies determine what a prohibition is worth. The older statute’s conduct provisions were not badly drawn, and several of them anticipate the structure of the newer prohibitions, but they were attached to a remedial ceiling of discontinuance and compensation on application, and a prohibition that cannot cost the infringer more than compliance would have cost is an instruction rather than a deterrent. The single most consequential drafting decision in the Competition Act was not the definition of its prohibitions but the penalty in section 27, and the single most consequential omission in the older Act was the absence of one.

The third is that repeal and commencement are part of the design and not an administrative afterthought. The interval from 1991 to 2011 was not the product of a decision that India needed no merger control; it was the product of deleting a control in one statute in 1991 and commencing its replacement in another statute in 2011, with no provision bridging the two events. A draftsman changing a regime owes the interval the same attention as the destination, because the interval is where the regulated parties actually live, and twenty years is longer than most of the statutes in this series operated in their intended form.

The fourth is that a statute should say what it is not doing. The older statute accumulated functions, concentration control, trade practice adjudication, consumer redress, a register, because each was added to the institution that happened to exist, and the accumulation left no function with an institution designed for it. The newer statute’s refusal of the consumer jurisdiction is the clearest instance in Indian economic legislation of a function being declined on principle and assigned elsewhere, and the resulting clarity about what the Commission is for has been worth more, in the quality of its jurisprudence, than the convenience of the combined forum ever was.

The verdict of the whole comparison can be stated in the terms the brief for it prescribes. The new regime is better at everything except being early. It defines its concepts where the older statute enumerated them, measures market power where the older statute measured assets, reaches conduct abroad where the older statute stopped at the border, penalises where the older statute directed, induces disclosure where the older statute offered nothing, and reviews mergers on their effects where the older statute approved them on policy. On the dimension of consumer complaint it relocated a function rather than improving it, and on the dimension of conglomerate power it abandoned a question rather than answering it, and both of those belong in the record against it. And the transition left India for two decades with no merger control of any kind and for four decades with a competition authority that could not impose a fine, so that the honest verdict is that the design change was clearly right and the sequencing was clearly costly, and that a reader who keeps both halves of that sentence has understood the comparison better than a reader who keeps either half alone.

Frequently Asked Questions

Q: What is the fundamental difference between the monopolies and competition regimes?

The MRTP Act 1969 was organised around the concentration of economic power, measured by the asset value of an undertaking, and it treated accumulation itself as the harm to be prevented, through prior approval of expansion, new undertakings, mergers and takeovers by undertakings above a threshold. The Competition Act 2002 is organised around conduct and transactions that cause or are likely to cause an appreciable adverse effect on competition in a defined relevant market. Size is not an offence under it, and dominance is not prohibited, only its abuse. The first regime asked how large a firm was. The second asks what a firm has done, and the difference governs every other contrast between them.

Q: How did the two regimes treat size and market share?

The MRTP Act treated size as presumptively suspect. An undertaking whose assets with its interconnected undertakings crossed the prescribed line, twenty crore rupees at first and one hundred crore rupees after 1985, attracted the concentration provisions and needed prior approval to grow, while a dominant undertaking was identified arithmetically, by a share of one third, later one fourth, of the supply of goods or services. The Competition Act is formally neutral about size. Market share is one factor among those listed in section 19(4) for inferring dominance in a relevant market, dominance itself is lawful, and liability begins only with the abuses listed in section 4. Duties follow market power rather than wealth.

Q: How did merger control differ between them?

Under the MRTP Act in its full form, an amalgamation, merger or takeover involving an undertaking within Part A of Chapter III required the prior approval of the Central Government under section 23, a decision of industrial policy taken inside government. The amendment of 1991 deleted that requirement along with the rest of Part A. The Competition Act restored merger control as mandatory pre notification to a specialist regulator, the Competition Commission, with a standstill obligation until clearance or the expiry of the statutory period, assessed on the factors in section 20(4). The decisive fact is the interval: the combination provisions commenced only on 1 June 2011, so from 1991 to 2011 India had no merger control under either regime.

Q: What enforcement powers did the MRTP Commission and the CCI have?

The MRTP Commission could inquire into monopolistic, restrictive and unfair trade practices, direct that a practice be discontinued and not repeated, grant an interim injunction under section 12A, and award compensation under section 12B to an applicant who proved loss or damage. It had no power to impose a penalty at any point in its existence, and appeals from it lay directly to the Supreme Court. The Competition Commission of India can impose a penalty of up to ten per cent of average turnover under section 27, and for cartels the greater of three times the profit or ten per cent of turnover for each year of continuance, order the division of a dominant enterprise under section 28, and grant interim relief under section 33, with appeals to an appellate tribunal.

Q: What happened to unfair trade practices after the change?

Unfair trade practices were added to the MRTP Act in 1984, in section 36A, and the MRTP Commission adjudicated them until the repeal. The Competition Act contains no jurisdiction over misleading advertising or false representations as such, because its prohibitions are addressed to effects on competition rather than to deception of buyers. When section 66 repealed the MRTP Act on 1 September 2009, the pending unfair trade practice matters were transferred to the consumer commissions under the consumer protection statute, and the function has since been strengthened on that side, notably by the Central Consumer Protection Authority under the Consumer Protection Act 2019. The jurisdiction moved rather than improved, and a trader complaining of a rival’s false claim lost the forum it once had.

Q: Which regime handled cartels better?

The competition regime, and by a wide margin. The MRTP Act had no definition of a cartel and no offence of cartelisation. Collusive arrangements reached the Commission as restrictive or monopolistic trade practices, were tested through the gateways in section 38, and could attract no penalty, so a detected cartel risked a direction to stop and nothing more. The Competition Act lists price fixing, output limitation, market sharing and bid rigging in section 3(3) and presumes their adverse effect, attaches the heaviest penalty in the statute through the proviso to section 27, reaches a facilitating participant through the hub and spoke provision added in 2023, and induces disclosure through the lesser penalty in section 46 and the leniency plus mechanism, in force from 20 February 2024.

Q: Did the shift to competition law improve market outcomes?

That cannot be established from the record, and any confident answer in either direction should be treated with caution. The statutory change coincided with the liberalisation of licensing and trade, the entry of foreign capital and technology, the growth of services and digital markets, and a period of twenty years in which neither regime operated in full, so no figure for concentration, entry, margins or prices across the span can be attributed to the statute alone. What can be established is a claim about capacity: the newer statute possesses a penalty power, a disclosure incentive, market definitions and merger review that the older one lacked, on seven of the nine dimensions compared in this article. The strongest defensible verdict is about the law’s design, not about measured outcomes.

Q: What did India lose in the transition to competition law?

Three things of substance. First, a law addressed to concentration of economic power as such: the accumulation of assets across unrelated markets, which the MRTP Act could reach through its thresholds, is not a concern of the Competition Act unless it produces dominance in a defined relevant market, and commentators divide on whether that is a gap or a correct narrowing. Second, a forum in which an individual or a small trader could complain of a trade practice, including a rival’s false claims, and obtain a direction without proving an effect on competition in a market. Third, the register: a standing record of undertakings and registrable agreements, replaced by a regime that learns of arrangements only through complaints, filings and its own inquiries.

Q: Why did the MRTP Act measure economic power by assets instead of market share?

Because of the economy it was drafted for and the machinery it belonged to. In a licensed economy the allocation of capacity was an administrative act, the balance sheet was a document the state already received, and an asset figure could be applied by an official without defining a market or hearing economic evidence, which suited a provision meant to operate as a trigger for prior approval. Market share measurement would have required market definition, data on supply and a judgment about the boundaries of competition that the administrative system of the time was not built to make. The choice was administrable and durable, and its defect appeared only as industry changed: asset value and market power diverged, so the statute measured a proxy with increasing precision and decreasing relevance.

Q: What were the gateway provisions under the MRTP Act and what replaced them?

Section 38 of the MRTP Act listed circumstances, called gateways, through which a restrictive trade practice that fell within an enumerated head could escape condemnation if the Commission was satisfied that a gateway applied and that the resulting restriction was not unreasonable in the public interest. They included such matters as protecting the public from fraud or danger, negotiating fair terms against a powerful counterparty and preventing a substantial lessening of competition elsewhere. Nothing in the Competition Act corresponds to them by name. Their function, the weighing of a justification against a competitive harm, was moved inside the prohibition itself: an agreement is caught only if it causes or is likely to cause an appreciable adverse effect, assessed on the factors in section 19(3), which include benefits to consumers and improvements in production and distribution.

Q: Did either regime control conglomerate accumulation across unrelated markets?

The MRTP Act could, and the Competition Act cannot, at least not as such. Because the older statute’s trigger was the total assets of an undertaking together with its interconnected undertakings, a group accumulating positions in markets unconnected with each other crossed the line by the fact of accumulation and needed approval to expand, merge or take over. The Competition Act proceeds market by market: it asks whether an enterprise is dominant in a relevant market, whether a combination is likely to cause an appreciable adverse effect in a relevant market, and whether an agreement harms competition in a market. A group that is powerful in aggregate and dominant nowhere in particular presents it with no question to answer. Whether that is a defect is contested, and the debate has revived in discussion of large digital groups.

Q: What could a small competitor do under each regime against a dominant rival’s conduct?

Under the MRTP Act it could complain to the Commission of a restrictive or monopolistic trade practice without first establishing that the rival was dominant in a defined market, and it could obtain a direction that the practice stop, an interim injunction in a proper case, and compensation on application if it proved loss. Under the Competition Act it files information with the Commission, and the inquiry is the state’s rather than its own, which relieves it of the cost of proof but also of control over the case. It must translate its grievance into an abuse within section 4, which requires the rival’s dominance in a relevant market, or into an agreement within section 3. If the translation succeeds, the remedies are far stronger, including a penalty on the rival’s turnover. If it fails, there is no competition remedy at all.

Q: Which regime gave a large firm clearer notice of what conduct was prohibited?

The competition regime, with one qualification. Under the MRTP Act the firm’s principal exposure was not to a prohibition at all but to a permission: it knew the asset line with precision, and knew that crossing it meant applying to government for approval of growth, merger or takeover, with the outcome resting on policy as much as on stated criteria. That is certainty about a procedure and uncertainty about a decision. Under the Competition Act the firm’s obligations are stated as prohibitions on defined conduct, its size is not an offence, and its combinations are reviewed against published factors within a statutory period after which silence is clearance. The qualification is that the boundary between vigorous competition and abuse of dominance is a matter of assessment, so notice is clearer about the kind of question than about its answer in a hard case.

Q: How did the burden of proof differ between the two regimes?

Under the MRTP Act the Commission had to bring a practice within an enumerated head of restrictive or monopolistic practice, after which the practical burden moved to the respondent to bring the practice within a gateway in section 38 and to show that the restriction was not unreasonable in the public interest. Under the Competition Act the Commission, assisted by the Director General’s investigation, must establish the relevant market and the appreciable adverse effect on the factors in section 19(3), except for the horizontal agreements listed in section 3(3), where the effect is presumed and the burden moves to the parties to displace the presumption with evidence. The change is from a burden about justification, argued in the language of public interest, to a burden about effect, argued in the language of evidence.

Q: What place did economic evidence have in adjudication under each statute?

Very little under the older statute, and a central place under the newer one. MRTP adjudication turned on classification and justification: whether conduct fell within an enumerated practice and whether a gateway applied. Those are questions for lawyers, answered from documents and argument, and the Commission’s orders reflect it. Competition Act adjudication turns on market definition, shares, entry barriers, countervailing power and the likely direction of prices and output, which are questions that documents alone cannot answer. The Commission maintains an economics capability, parties file expert economic evidence, and appellate decisions, including the Supreme Court’s confinement of penalties to relevant turnover, are decisions about the quality and reach of economic inference. The change in the evidence is the practical measure of the change in the law.

Q: What filled the gap during the two decades when neither regime operated in full?

Less than the question implies. From 1991, when the MRTP Act lost its concentration provisions, until 2011, when the combination provisions of the Competition Act commenced, no statute controlled mergers or the growth of large undertakings, and from 2003 to 2009 the Competition Act existed with a Commission established but its conduct prohibitions not in force. What operated in the interval was the remainder of the MRTP Act, adjudicating restrictive and unfair trade practices with a remedial ceiling of a direction to discontinue, the consumer commissions after 1986, the sector regulators created for telecommunications, electricity, securities and insurance within their own remits, and the general law of contract, tort and intellectual property. Industrial consolidation in those decades was supervised, as a matter of competition law, by nobody.

Q: What is a relevant market and why did the MRTP Act have no use for the concept?

A relevant market is the field within which competition is assessed, defined by the Competition Act in section 2(r) by reference to a relevant product market under section 2(t) and a relevant geographic market under section 2(s): the products regarded as interchangeable by consumers by reason of their characteristics, prices and intended use, in the area where conditions of competition are distinctly homogeneous. Every inquiry under sections 3, 4 and 6 is conducted within such a market, since an effect on competition, a dominant position and a combination’s impact are all meaningless without a field in which to measure them. The MRTP Act had no use for the concept because its trigger was the assets of an undertaking, a figure that requires no market, and its practices were condemned by classification rather than by effect.

Q: How did each regime allow the government to exempt an activity or a sector from its reach?

The MRTP Act began by standing back from the state itself. Section 3 kept undertakings owned or controlled by government, statutory corporations and certain other bodies outside the Act unless the Central Government directed otherwise by notification, so exemption was the default for the public sector and application the exception, a position reversed only in 1991 when public sector undertakings were brought within the conduct provisions. The Competition Act begins from the opposite default: an enterprise includes a government department or undertaking engaged in an economic activity, and activities that are sovereign in character fall outside the definition. Exemption is then a deliberate act under section 54, by which the Central Government may exempt an enterprise, a class of enterprises or a practice in the interest of the security of the state or of the public.

Q: Which regime served the consumer better, and by what route?

By different routes, and the answer depends on which consumer is meant. As a member of a market, the consumer is better served by the competition regime, whose penalty on cartels, leniency mechanism and merger review protect the conditions under which prices and choice are formed, and which the older regime, without a penalty and without merger control after 1991, could not protect in the same degree. As a complainant with a grievance of deception, the consumer was better served between 1984 and 2009 by the MRTP Act, which gave the complaint a named place in section 36A and a commission to hear it. That direct route runs through the consumer commissions and the Central Consumer Protection Authority under the Act of 2019, so the consumer’s protection was rebuilt in another statute rather than in the successor to the one that was repealed.

Q: What does the MRTP experiment teach about drafting competition legislation?

Four things. A statute should measure the thing it fears rather than an administrable proxy, because the proxy becomes the law’s real object and outlives its rationale, as the asset threshold did. Remedies decide what a prohibition is worth, and a prohibition whose ceiling is a direction to stop will be priced by firms as an instruction, which is why the penalty in section 27 mattered more than any definition in either Act. Repeal and commencement are part of the design, and deleting a control in 1991 while commencing its replacement in 2011 abolished the function for twenty years by inadvertence. And a statute should say what it is not doing, as the Competition Act did in declining the consumer jurisdiction, rather than accumulate functions in whichever institution happens to exist.