Competition policy and the consumer welfare standard

Abstract

This paper, which was given as the 2024 Bellamy Lecture, reviews the law and economics of the consumer welfare standard in competition policy, particularly in relation to mergers and abuse of dominance. With qualifications relating to input markets, the consumer welfare approach is defended against arguments for its relaxation, and against contrary criticisms that it is unduly permissive.

1. INTRODUCTION

I first met Christopher Bellamy in the context of Big Tech. In 1981, I was an economics research assistant for Derek Morris on the case brought by the European Commission against IBM. Their leading counsel was Mr Jeremy Lever QC (as he then was), who had quite a team, including David Edward, John Swift, Richard Fowler, Stephen Richards, and Nicholas Forwood, as well as Christopher. Little did I anticipate being appealed to him 20 years later, repeatedly, still less being here this evening. Meeting Jeremy Lever in connection with the IBM case changed my life, and I pay tribute to him.

The theme of this lecture is the consumer welfare standard, which has been and remains a guiding principle—arguably the guiding principle—of competition law and policy. At the Office of Fair Trading, next door in Salisbury Square, our strapline was ‘making markets work well for consumers’. That was intended not only to embrace consumer policy but also to express the essence of much competition policy as we saw it.

The consumer welfare standard is now under attack from various directions. For some, it is too interventionist, for example, in relation to mergers, and should be replaced or at least modified in the name of industrial competitiveness. Or it should be selectively disapplied, for example, in relation to agreements between competitors, to advance the net zero agenda. For others, the consumer welfare standard is too lax and wrongly ignores worker interests and concerns about inequality. It is partly to blame, some say, for ineffective competition policy during the ‘Rise of Market Power’ in the USA and other economies.

I will address these criticisms in this lecture. But first, we need more clarity about what ‘the consumer welfare standard’ is in economic terms, and the degree to which competition law reflects it, whether in statute or statutory interpretation. Needless to say, I am not qualified to provide legal analysis. My aim is rather to offer some economic perspectives on the law, and in the knowledge that the relationship between economic and legal concepts is always going to be imperfect.

2. CONSUMER WELFARE OR TOTAL WELFARE?

In his blast against prevailing US antitrust law—‘a policy at war with itself’—Robert Bork declared that ‘the only legitimate goal of antitrust is the maximization of consumer welfare’.¹ But by what he beguilingly called ‘consumer welfare’ he appears to have meant the combined interests of consumers and producers. Economists call that ‘total welfare’ and reserve ‘consumer welfare’ for the interests of consumers leaving aside the profits of producers. I will therefore translate Bork as saying that ‘the only legitimate goal of antitrust is the maximization of total welfare’.

The textbook illustration of how the total welfare standard and consumer welfare standard can give different answers is a horizontal merger that (i) leads to significantly higher prices to final consumers through the lessening of competition, but (ii) big savings of fixed costs that would not otherwise be achieved. The consumer welfare standard, focused on (i), disapproves of such a merger. The total welfare standard, which weighs (i) plus (ii), likes it because the profit gain outweighs the consumer detriment, and shareholders are ultimately consumers too. (For simplicity, this leaves aside the issue of what the merger does to the profits of non-merging firms.)

A merger of this description would appear to fail the substantial lessening of competition (SLC) criterion in UK merger law because by assumption the price increase is significant. If, however, we amended the example so that the merger reduced marginal costs, there could be sufficient customer benefits for the merger to be allowed. In particular, the lessening of competition might fatten the markup of prices over marginal costs, but if marginal costs fell, the overall consumer effect could come out as positive. This is all consistent with the consumer welfare standard.

For total welfare devotees in the Bork tradition, this is too interventionist. Some mergers that would enhance overall economic efficiency are blocked. They ask: why leave the profit gains of the merging firms, which arise from the fixed cost savings, out of account?

One line of response to this (good) question appeals to considerations of income distribution. If profits tend to go to higher income groups than the consumers that lose from the SLC, then arguably profits should have lower weight in the welfare calculus. But should income distributional considerations play any part in competition policy? And even if they were to, why put zero weight on profits?

Moreover, there are cases where income distributional effects may well point the other way, with low-income producers supplying better-off consumers. Only last month the BBC reported that nail technicians planned to join forces to raise manicure prices.² A leader of this initiative helpfully explained: ‘It’s really beneficial that we are all raising our prices the same day and you know no-one is going to undercut each other’. One cannot fault the economic analysis here. They got a letter from the Competition and Markets Authority.

An entirely different argument for the consumer welfare standard versus the total welfare standard stems from the observation that competition policy does not determine which mergers happen, but which are allowed. Firms decide which mergers happen among those that are allowed. Suppose for simplicity that the most profitable permissible mergers are the ones that happen. Then even if the ultimate objective is total welfare, that objective is not best achieved by adopting a total welfare standard. Indeed, greater total welfare might be achieved if consumer welfare is the standard for deciding which are allowed.

To see the logic of this, suppose that there are two alternative mergers. Merger A creates profit of 2 and consumer benefit of 1. Merger B creates profit of 3 and consumer detriment of 1. Under the total welfare standard both are allowed, and the firms will pick B, for total welfare gain of (3–1) = 2. Under the consumer welfare standard, B is disallowed and they pick A, for total welfare gain of (2 + 1) = 3. The general point is that because firms are not pursuing total welfare, it is better for total welfare to set a standard suitably different from that ultimate objective. As Farrell and Katz put it, if you want to travel north-east ‘and firms always push eastwards, there is something to be said for someone adding a northerly force’.³

The wider point is that competition law is about which rules to adopt, not which commercial decisions to take. Though perhaps not a clinching point, this for me weighs against total welfare and in favour of consumer welfare as the better economic standard for competition policy, and not only in relation to mergers.

What though is the link, if any, between welfare standards and competition? One response is that they provide yardsticks by which to help judge what is, and what is not, anti-competitive. A related point is to keep in mind the question of competition to do what?

3. ARE BUSINESS CUSTOMERS ‘CONSUMERS’?

We next need to clarify which consumers matter in the consumer welfare standard. In many mergers and other competition cases, the customers are of course not final consumers but businesses. How should they count under the consumer welfare standard? Is it only the ultimate effects on final consumers that we really care about (as zero-weighting of corporate profits might imply) or do effects on intermediate businesses matter?

If as a practical matter benefits and detriments to intermediaries tend to be passed on to final consumers, this might be a distinction without a difference. Although one-for-one pass-on cannot generally be assumed, partial pass-through still ensures that effects on final consumers will go in the same direction as effects on intermediate customers. It may be noted, however, that practices such as volume discounts to downstream firms, which have been the subject of much antitrust attention in Europe, might be especially good for final consumers because of their effects on downstream marginal costs.

In any event, we need a linguistic stipulation on whether business customers are ‘consumers’. In line with the jurisprudence (see below), I will take it that they are.

4. IS THE CONSUMER WELFARE STANDARD REFLECTED IN STATUTE?

The consumer welfare standard appears in competition law, with qualifications, both in statute and via statutory interpretation. We have already seen how the central criterion in UK merger law is consistent with the consumer welfare standard at least in broad terms. What about the prohibitions of anti-competitive agreements and abuse of dominance?

Chapter I and Article 101 prohibit anti-competitive agreements (etc) unless they meet conditions for exemption, notably by contributing ‘to improving the production or distribution of goods or to promoting technical or economic progress, while allowing consumers a fair share of the resulting benefit’.

Thus, consumers feature explicitly in the exemption condition. It is clear, however, that the prohibition is not all about consumers because it expressly applies to agreements that ‘directly or indirectly fix purchase or selling prices or any other trading conditions’ (emphasis added). This formal symmetry between buying and selling makes clear that consumer welfare is not the whole story.

Which consumers count in applying the test about fair share of the resulting benefits? This question was recently addressed by the UK Supreme Court in Sainsbury’s v MasterCard, a case about a two-sided market with different consumers—merchants and cardholders—on the two sides.⁴ In accordance with the Advocate General in an EU MasterCard case, the Supreme Court ruled that for the fair share test to be met, ‘the disadvantages suffered by consumers in the market where competition was restricted must be counterbalanced by advantages benefiting the same consumers’,⁵ that is the merchants in this case, without considering possible benefits to cardholders.

One reason given for this conclusion was that ‘the contrary view would result in competition authorities favouring one category of consumers at the expense of others, something which was no part of the function of competition law’. But the consequence of no favouring in these terms was that merchant interests alone counted for the fair share test, with zero weight on the welfare of cardholders (who are final consumers) on the other side of the market. Arguably, that is favouring in effect. But it gives the right answer on the sign, though not magnitude, of the overall effect on consumers if pass-through is positive but less than one-for-one.

Chapter II and Article 102 on abuse of dominance speak of ‘limiting production, markets or technical development to the prejudice of consumers’. They give the imposition of ‘unfair purchase or selling prices’, as types of abuse, again indicating formal symmetry between effects in input and output markets, which makes good economic sense.

In view of this symmetry, one could say that, conceptually, the consumer welfare standard should really be thought—and spoken?—of more broadly as a trading partner welfare standard, including trading partners in input markets. But, if the bulk of competition problems in fact concern output markets, it would seem reasonable to speak in shorthand terms of a consumer welfare standard so long as that is not taken as a claim that law and policy are only about consumers.

5. LABOUR MARKET ISSUES

Input markets nevertheless require comment, and labour markets tend to be the input markets of most policy concerns. (Though I have not forgotten groceries.) How does or should competition law apply to them? This question must be viewed in the context of the large body of law and regulation on employment rights, collective bargaining, and trades union law, which for obvious reasons lacks parallel on the consumer side. There is nonetheless scope for competition law and policy to apply (see CMA, 2024).⁶

A recent illustration is the 2021 US case of NCAA v Alston concerning the restrictions on education-related benefits to student athletes in the rules of the National Collegiate Athletic Association.⁷ The notoriously divided US Supreme Court was on this occasion unanimous that these rules violated antitrust law, and Justice Kavanaugh in a concurring opinion made some sharp wider observations. This is orthodox, not hipster, antitrust law being applied to a labour market question, as the Supreme Court unanimity confirms.

Turning to merger policy, horizontal mergers that lead to output reduction could well be bad for workers in the merging firms with sector-specific skills in relevant geographical markets because of heightened monopsony power. But the SLC test properly applied to output markets should catch many of them anyway. A question is whether there is a significant number of cases that are competitively benign in output markets but which substantially increase market power in relation to workers.

One can certainly invent examples, such as a merger between the only two employers on an imaginary island in the Outer Hebrides, one supplying tweed, the other malt whisky. And there is empirical evidence from hospital mergers in the USA of reduced wage growth where both mergers led to large increases in concentration and workers had industry-specific skills (Prager and Schmitt, 2021).⁸ So the issue is not just theoretical. But while there seems no reason to exclude labour market effects altogether from merger review—labour markets can presumably be ‘markets for … services’ within the meaning of the Act—I would be somewhat surprised if many mergers were problematic on that front without also being so in consumer markets.

Finally, in relation to labour markets, let me mention non-compete agreements, which are in the news following the aggressive line recently taken by the Federal Trade Commission. These do appear to be prevalent to a surprising extent, and deserving of antitrust scrutiny. But non-compete clauses can be pro-efficiency and pro-worker in a variety of circumstances, where training and confidentiality are concerned, so general hostility seems inappropriate on economic grounds. Moreover, non-compete clauses may be assessed under the ancient common law doctrine on unreasonable restraints on trade, as in the 2019 UK Supreme Court case of Tillman v Egon Zehnder.⁹

6. THE CONSUMER WELFARE STANDARD AND ABUSE OF DOMINANCE

Most of the discussion so far has been about mergers and anti-competitive agreements, moreover horizontal ones. What if any role does the consumer welfare standard play in law and policy towards abuse of dominance? As noted earlier, there is reference to ‘the prejudice of consumers’ in the statute but, as with US antitrust, we need to look at how the statutory text has been interpreted to gauge the significance of consumer welfare considerations in the application of Article 102.

On this, the Court of Justice in SEN (2022) recently made the forthright statement that:

the well-being of both intermediary and final consumers must be regarded as the ultimate objective warranting the intervention of competition law in order to penalise abuse of a dominant position.¹⁰

I appreciate that ‘well-being’ and ‘welfare’ might not have identical meanings, but this surely indicates that the consumer welfare standard is centrally important for understanding the law on abuse of dominance.

It is not hard to see how consumer welfare features in cases about exploitative abuse, but most cases, rightly, are about exclusionary abuse. Broadly speaking, there seem to be two ways in which consumer welfare considerations could matter in such cases.

One is consumer welfare as a counterbalancing factor. Thus, the Court of Justice in SEN went on to say that a dominant firm:

may show that an exclusionary practice escapes the prohibition laid down in Article 102 TFEU by, inter alia, demonstrating that the effects that could result from the practice at issue are counterbalanced or even outweighed by advantages in terms of efficiency which also benefit the consumer in terms of, specifically, price, choice, quality or innovation.

Note here that the counterbalancing is in terms of efficiency advantages that also benefit the consumer, implying a distinction between overall efficiency and consumer benefit. It is also made plain that consumer benefits go way beyond price.

If consumer welfare matters as a possible counterbalancing element, we need a prior determination of what is exclusionary or anti-competitive in the first place. Is consumer welfare relevant to that question too? In any case, it is a fundamental question that has to be answered. In a lecture 20 years ago, I canvassed three candidate answers.¹¹

The first candidate was the Sacrifice Principle. In general terms, this asks whether the conduct at issue would be profitable but for its tendency to eliminate or lessen competition. Below-cost pricing by a dominant firm, for example, fails this test because (subject to possible qualifications) it is loss-making. This form of the Sacrifice Principle aligns with the As-Efficient Competitor Principle discussed below.

A strict form of this principle might be regarded as abusive conduct by a dominant firm where there existed alternative conduct that was more profitable both for the firm and for its rivals. This would condemn much conduct that was profitable, though not maximally so. But this seems like a recipe for chilling competition, to the detriment of consumers. And in an uncertain and dynamic world, assessment of the profitability of different courses of conduct would seem hopelessly difficult, perhaps unless (wrongly) focused on the short run. There would be a real risk of attributing anti-competitive explanations to all sorts of business practices—a danger noted by Ronald Coase, who got tired of antitrust in the 1960s ‘because when the prices went up the judges said it was monopoly, when the prices went down, they said it was predatory pricing, and when they stayed the same, they said it was tacit collusion’.¹²

Another challenge for the Sacrifice Principle is to avoid circularity. To say that conduct is anti-competitive if it is unprofitable but for its anti-competitive effect does not advance understanding of what it is to be anti-competitive. Being bad for rivals cannot be the criterion because competition itself is bad for them.

The second candidate was the As-Efficient Competitor Principle. This answers the question ‘Whose exclusion by dominant firms should the law prohibit?’ by saying ‘Firms that are no less efficient than the dominant firm’. Normal competition may well lead to the exclusion of the inefficient, so this principle draws the line between normal competition and abuse in terms of as-efficiency. The principle means that the most efficient firm can win the relevant custom without making a loss. Of course, how to implement the principle in practice is another matter and much disputed. It does have the merit of being capable of self-assessment by the dominant firm because it does not require information about how efficient actual rivals are.

Predatory pricing illustrates the relationship between the as-efficient competitor principle and consumer welfare. Low prices are good for consumers, but below-cost pricing by a dominant firm is, subject to conditions, seen as detrimental because of the implication(?) of subsequently higher prices. It blocks the opportunity for as-efficient rivals to serve consumers without making losses.

The third candidate was the Consumer Harm Principle. This answers the ‘Whose exclusion …?’ question by saying ‘Firms whose presence is good for consumers’. As a necessary condition for condemning conduct, this is consistent with the As-Efficient Competitor Principle. But as a sufficient condition, it goes further because the presence of inefficient competitors can be good for consumers. It would seem formidably difficult to make this principle operational without lapsing into competitor protection of a detrimental kind. Relative to the as-efficient competitor principle, it could soften competition even to the benefit of the dominant firm itself.

7. AS-EFFICIENT COMPETITORS IN THE LAW

Twenty years ago, EU law and policy on predatory pricing and margin squeeze in abuse of dominance cases could be said to be in accordance with the as-efficient competitor principle, but things were less clear beyond that. In the paper just mentioned, which argued for more economic and less formal approaches, I mused that ‘it will be interesting … to see whether the as-efficient competitor principle gains more extensive recognition as the case law evolves’.

It turns out that it has. Important impetus came from the European Commission’s 2009 guidance on enforcement priorities when applying the prohibition on abuse of dominance to exclusionary conduct, which has recently been amended. Then came a series of judgments from the Court of Justice, of which I will mention three.

In Intel (2017), the Court ‘further clarified’ the Hoffman La-Roche case law on loyalty rebates by saying that where an undertaking has been submitted on the basis of evidence that its impugned conduct was not capable of restricting competition, then the Commission is among other things ‘required to assess the possible existence of a strategy aiming to exclude competitors that are at least as efficient as the dominant undertaking from the market’.¹³ In economic terms, one can regard this as applying the predatory pricing test to contestable demand. The Court said further that the ‘balancing of the favourable and unfavourable effects of the practice in question on competition can be carried out in the Commission’s decision only after an analysis of the intrinsic capacity of that practice to foreclose competitors which are at least as efficient as the dominant undertaking’.

In Unilever Italia (2023), the Court stated that the Intel clarification must be understood as applying to exclusivity clauses as well as to rebate schemes.¹⁴ The use of an as-efficient competitor test (as distinct from the principle) is optional, and may be inappropriate for certain non-pricing practices, but even there the relevance of such a test cannot be ruled out.

Non-price conduct was the subject of the SEN judgment cited earlier. The case concerned the discriminatory supply of customer data by SEN to another part of the ENEL group, but not to rivals, when energy markets were liberalized in Italy. The Court said that, while it is in no way the purpose of Article 102 to ensure that less-efficient competitors remain on the market, the dominant firm must use means that are competition on the merits. A practice that ‘holds no economic interest for a dominant undertaking, except that of eliminating competitors so as to enable it subsequently to raise its prices’ is outside the merits. So is conduct that an as-efficient competitor cannot imitate or replicate because it relies on resources or means inherent to the holding of the dominant position (whatever that may mean). Replicable conduct, in contrast, is on the merits. The relevance of (ir)replicability by a hypothetical as-efficient competitor is said to be ‘clear from the case-law on practices both related and unrelated to prices’.

Needless to say, these judgments leave many open questions. They do however amount to a quite comprehensive endorsement by the Court of the relevance of the as-efficient competitor principle, and much more warmth towards economic approaches than was evident 15 years ago.

The European Commission on the other hand shows some signs of cooling its enthusiasm for the approach set out in the 2009 guidance on enforcement priorities in relation to exclusionary abuse. At least that is an impression that may be gained from the Commission’s 2023 amendments to that guidance.¹⁵ It will be fascinating to see the Commission’s draft Guidelines on exclusionary abuse, which are due soon. Of course, it is ultimately for the Courts, not the competition authorities, still less the economists, to say what the law is.

8. CONSUMER WELFARE AND COMPETITIVE STRUCTURE

A question referred to the Court of Justice in SEN was whether Article 102 is intended to protect consumers or to protect the competitive structure of the market. The interesting discussion by the Advocate General concludes that these objectives are not mutually exclusive but inseparably linked—to the ultimate purpose of consumer well-being, ‘the leitmotiv of any intervention of Article 102 and of competition law more generally’.

But what is protection of competitive structure? The Advocate General is clear that Article 102 ‘is not aimed at, or rather is no longer aimed at protecting competitors’. It is hard to see how structure for these purposes can reliably, or even usefully, be measured in terms of numbers of competitors or relative market shares. Rather the core idea seems more to be about opportunities to offer customers good deals.

Competition to do that is after all the primary form of competition that law and policy aim to protect. In these terms, competition is not only good for consumers, but competition-to-offer-consumers-good-deals is the competition that matters.

9. ‘COMPETITIVENESS’ VERSUS COMPETITION

As the Advocate General’s leitmotiv remark succinctly illustrates, consumer welfare has become quite embedded in competition law and policy. Yet the consumer welfare standard itself—not just the way it has been applied—has recently been the target of attack from various directions, with many urging a more interventionist standard, and some the opposite.

One line of the latter criticism, hardly new, is that pro-competitive policies have blocked efficient consolidation and the realization of scale economies. On this view, competition policy should be dialled down to promote international competitiveness. Or more broadly it should be disapplied when necessary to promote other objectives such as financial stability.

Before summarizing why I regard such arguments as unpersuasive, I should note that there are related points that are at least coherent. One, discussed earlier, is the view that consumer-oriented merger policy gives insufficient weight to fixed cost savings, to the detriment of overall efficiency. Another, for which I do not see evidence and which is not a criticism of the consumer welfare standard itself, is that competition authorities tend to define geographic markets too narrowly when assessing mergers, and would be more permissive if international competition was properly taken into account.

Overall, however, the ‘competitiveness’ argument for relaxed competition policy is weak and prone to be captured by vested interests. It is weak because competition—and again I mean competition-to-offer-consumers-good-deals—is generally positive, not negative, for productivity, and it is hard to see what ‘competitiveness’ usefully means if not productivity.

Furthermore, it is striking how often proponents of policy intervention in the name of ‘competitiveness’ mean the competitiveness of their own sector, not of the economy as a whole. For example, easing capital requirements on financial institutions may be good for their ‘competitiveness’ because it increases the implicit subsidy to them from taxpayers, but the competitiveness of the economy requires precisely the opposite policy, to reduce crisis risk. As a general matter, it is worth remembering that my subsidy is your tax.

The banking sector provides a perfect example of the unwisdom of relaxing competition policy in the name of other objectives. In the autumn of 2008, Government and Parliament exempted Lloyds acquisition of HBOS from merger control. The merger obviously raised competition problems, so this was bad competition policy. But it was worse financial stability policy. However much policymakers might have hoped that the takeover would spirit away the evident problems of HBOS, the fact was that it worsened the position of Lloyds TSB, which was of much greater systemic importance, with highly damaging consequences all round.

10. GREEN ANTITRUST

What about relaxing policy on mergers and agreements to promote environmental objectives? Again a sceptical attitude is warranted. As Schinkel and Treuren have argued¹⁶:

well-intended, green antitrust risks damaging both competition and the environment. It will suppress … market forces for companies to produce more sustainably, overburden competition authorities, invite abusive cartel greenwashing, and give the part of government that should promote sustainability further excuse to shun their responsibility for designing proper regulation.

In short, we can be confident about the anti-competitive effects of competition policy relaxation, but not at all about pro-environmental effects. Tirole discusses wider dangers of giving agencies multi-factor missions, including loss of accountability, institutional conflicts, and lack of policy coordination.¹⁷

11. INCOME DISTRIBUTION REVISITED

From other directions come calls for more interventionist competition policy than implied by the consumer welfare standard, including in the cause of reducing income inequality. But just as competition policy is not a good tool for the pursuit of environmental aims, neither is it for income redistribution. Indeed, there are strong arguments that on principle it should have nothing to do with that. In any event, as we have seen, the consumer welfare standard already has distributional implications by virtue of its zero-weighting of profits. For Bork and his followers, in contrast, the proper and neutral view would be the total welfare standard instead.

There is perhaps one exception, noted by Tirole, to the view that competition authorities should ignore social goals. In a world of limited resources, they must prioritize which cases to pursue, and it is arguably legitimate for wider considerations to influence enforcement priorities. Relevant to this issue is the fact that the public authorities do not monopolize enforcement. Entry barriers to private actions, which have long been a feature of US antitrust, are now much lower here than used to be the case.

12. THE ‘RISE OF MARKET POWER’?

Some of the pressure for radical reform of competition policy comes from an apparently widespread view, from President Biden down, that we have been living through an age of rising market power, at least in the USA. In short, the argument is that concentration and markups have increased and that this demonstrates rising market power and the weakening of competition. But is that true? The recent papers by Miller (2024) and Shapiro and Yurukoglu (2024) survey the state of play.¹⁸

The famous paper by De Loecker and others (2020) showed among other things that, for the US economy as a whole, the ratio of firms’ revenue to the accounting measure of ‘cost of goods sold’ has risen a lot since 1980.¹⁹ This is consistent with market power having risen but does not prove it, for at least three reasons, and of course what is true for the USA might not be so here.

First, the cost of goods sold might be a poor measure of variable cost. If other components of variable cost have become more important over recent decades, then the ratio of revenue to variable cost will have risen by less.

Secondly, whole-economy measures say little about developments in properly defined markets. As the mix of industries has evolved, including from manufacturing to services, the overall price/variable cost ratio might rise even though not generally doing so as much in individual markets. In other words, the microeconomic picture might differ, at least by degree, from the macroeconomic picture.

Thirdly and crucially, to the extent that the price/variable cost ratio has risen in particular markets, and/or concentration has increased, we need to ask why. It could be the anti-competitive exercise of greater market power or it could be the result of pro-competitive processes of more efficient firms winning business from less efficient firms.

Only microeconomic studies—of particular markets and industries—can discern which of these possibilities is grounded in evidence, and to what extent. Inevitably, the picture is mixed. For Miller (2024), a theme emerging from such studies is that technological advance matters a lot, whereas they ‘do not point to weak antitrust enforcement as contributing to greater market power’.

It therefore appears wise to reserve judgment both on whether there has been a comprehensive rise of market power in the USA, and on whether weak competition policy is responsible if so. Reserving judgment in that way is perfectly consistent with believing that law and policy, at least in the USA, need strengthening. However, it suggests that the better tool to apply is the screwdriver, not the hammer.

13. CONCLUSION

For competition law and policy, consumer welfare is not the whole story, but it should be and now is central to it. If, as I have tried to argue, attacks on the consumer welfare standard are not compelling, then it should not be dislodged. Major questions nonetheless remain, not just about the competition law treatment of particular types of conduct, but also about how the concepts of competition and consumer welfare relate to each other.

To illustrate, let me return to the IBM case. One of the alleged abuses was IBM supplying computers with main memory included. This ‘bundling’ was said to be anti-competitive towards rival suppliers of memory capacity, and an element of the 1984 settlement was IBM’s undertaking to supply memoryless computers on request.²⁰ It might seem bizarre to us now that this feature of computer design was challenged, and if you agree that it was not abuse of dominance, you can reach that conclusion in two ways.

One is to say that it was anti-competitive but outweighed by consumer benefits from computers working better with main memory included. Or you can say that it was not anti-competitive in the first place because it was a wholly reasonable way to compete to deliver what consumers wanted. The latter view, which I favour, involves consumer welfare in the notion of competition itself.

Acknowledgements

This article was the basis for the Bellamy Lecture given at the Competition Appeal Tribunal in London on 16 May 2024. It reflects the writings of, and conversations with, numerous economic and legal scholars, and the author is especially grateful to Mark Armstrong and Carl Shapiro. He is solely responsible for what is said here.

Footnotes