“It is time to stop heaping abuse on competition policy by criticizing the abstractness of perfect competition.”
“[P]ure competition” is evidently a theoretical concept, and […] the practical minded economist is often ready enough to point out that “no one has ever advocated that it be established.” […] But ordinary (purely) competitive theory remains the chief source of our criteria as to what should be done if possible, and of the direction in which we should move so far as we can.”
“The main point I want to make is that the welfare ideal itself (as well as the description of reality) involves a blend of monopoly and competition […]. If this is true, it is no longer self-evident which way we should move […].”
Some authors complain that competition law is unfairly identified with the ambition of turning the actual economy into perfect competition. This criticism is at least partly justified. Indeed, perfect competition has been rejected as a normative benchmark by the case law and the European Commission guidelines (4.). However, there remains a considerable ambiguity about the meaning of this model and its relevance for competition law.
It is well admitted that perfect competition is not a description of the actual world, but this discrepancy is often wrongly considered as a flaw of this model which would limit its usability for competition policy (1.) Actually, the real reason why it should not be used as a benchmark in law is that perfect competition has no title to be considered as a welfare ideal. Although many authors seem to think that this model maximizes efficiency, it would rather produce the opposite effect (2.).
As a consequence, competition agencies should not even try to replicate it “as far as possible” in the real world (3.). The uncertainty around the legal status of this notion stems from the fact that no serious substitute has been offered to replace it as a theoretical basis for competition policy. This explains why many people are inclined to smuggle this model back into competition law––to bridge the gap in the theory created by its previous rejection (5.).
1. Perfect Competition is Not a Description of the Actual World (except for farms).
Conditions for perfect competition are not met in the actual world (1.1.), except for the (unfortunate) farm sector (1.2.). This point is well-known (1.3.), but its consequences for competition law are debatable (1.4.).
1.1. Conditions for Perfect Competition are not met in the Actual World.
It is obvious that perfect competition is almost never achieved in the real world. No market is perfectly transparent, almost all firms differentiate their products, and it is very rare to see an atomistic market. It follows that most firms can and do price over their marginal costs.
1.2. An unfortunate exception: the agricultural sector.
Perfectly competitive markets are exceptions rather than the rule. The agricultural sector is one of them. But it is a rather unfortunate exception, considering the low standard of life of many farmers. It is interesting to note that, both in America and in Europe, this sector is (totally or partially) exempted from competition law and that the law promotes more or less actively its cartelization.
1.3. It is well-known that Perfect Competition is not a description of the actual world.
It is rather well-known that the real world is not characterized by perfect competition. This sounds like an echo in antitrust literature. Let us quote some examples.
Professor Moritz Lorenz writes that “It should be noted upfront that the [perfect competition] model assumes a highly idealised situation that is rarely seen in practice […][I]n real life only very few, if any, industries are characterised by the features underlying the model of perfect competition.” Professors Alison Jones and Brenda Sufrin similarly hold that “in the real world perfectly competitive markets hardly ever exist.”
Professor Hovenkamp also explains that:
The world contains no perfectly competitive markets, and many markets do not even come close. Firms often differentiate their products from other firms; as a result, customers are no longer indifferent to the identity of the seller. Information about market conditions is always less than perfect; as a result many transactions take place at some price other than the market price […]. “Economies of scale”––the ability of larger firms to produce at a lower cost than smaller firms––may result in markets that have fewer than the number of sellers required for perfect competition to occur.
Likewise, Professors Bosco and Prieto report that “Starting with the works of the 30’s, Robinson’s and Chamberlin’s works led to stop reasoning exclusively on the perfect competition model. Attention moved to the reality of markets, most of which do not meet the conditions required by the theory,” and Professor Petit states that “Nowhere the prerequisites for perfect competition seem to be met.” And to nail it: “All the authors agree on this point.”
I naturally concur with these statements.
1.4. Has the Discrepancy Between the Real World and Perfect Competition any Implication for Competition Law?
Disagreement starts when one tries to draw the consequences of the “imperfection” of actual competition for competition law. It seems that many authors fall prey to the “is/ought” (or “is not/ought not” fallacy) (1.4.1.), although some qualifications should be brought up (1.4.2.).
1.4.1. The Is/Ought (or Is not/Ought not) Fallacy.
Philosophers and lawyers generally agree that what ought to be cannot be derived from what actually is. For instance, the real-world is not crimeless. Yet, nobody would ever claim that this statement implies that we should not try to bring about a crimeless world. The same reasoning applies to other bad things, such as cancers, wars, etc.
However, it appears that, in competition law, people think otherwise. From the absence of perfect competition in the real world, it seems to be (intermittently) derived that the real world ought not to be transformed into perfect competition. (The careful reader will have noticed that I am criticizing neither the premise, nor the conclusion, but only the logical link between them.)
For instance, Professors Alison Jones and Brenda Sufrin write that “The analysis set out of perfectly competitive markets above does, however, present a number of problems. The main problem is that in the real world perfectly competitive markets hardly ever exist.” Similarly, Professor Petit considers that “In any case, the models which are made-up are so far removed from reality that they can hardly be enforced.”
However, let us mention some reasons which could justify these statements.
1.4.2. Some Qualifications.
One might argue that, if perfect competition is not achievable (for example, because one of its prerequisites is necessarily lacking), the second-best solution would not consist in trying to approach it at all. In such a case, one would face the theoretical cul-de-sac pointed out below. However, although second-best problems are often discussed, they are not generally put forward to explain why the “imperfection” of the actual competition should bar competition agencies from trying to achieve perfect competition.
b) Costs of Enforcement.
Another explanation could be that it would be too costly to try to attain perfect competition. In such a case, one should undertake to find the optimal trade-off between the perfection of competition and the cost of its enforcement. But it does not appear to be an apt explanation. When cost of enforcement problems are discussed in competition law, the goal is generally to justify an easing of competition agencies probationary burden. In contrast, costs are rarely brought up in order to displace perfect competition as the goal of antitrust.
A variation on the latter explanation would be that, because of the gap between the actual economy and perfect competition, it would be too exacting for firms to attempt to enforce this goal (the human cost would be too high). Private law does not usually force people to do as best as they can. It rather sets negative rules prohibiting very harmful actions. Competition law is a remarkable exception. Nevertheless, even considering the exceptional nature of this legal discipline, it could seem excessive to inflict punishments to individual firms which fail to behave as if they were in perfect competition, even though one believes that doing so would improve the total welfare.
Once again, I do not think that this idea can explain why so many antitrust lawyers detect a problem with the perfect competition model.
One could also argue that competition agencies have no sufficient means to transform the actual economy into perfect competition. The industrial base would be too rigid to be remodeled along those lines. Thus, perfect competition would remain the ideal, but the brief of competition agencies would only be to get “as close as possible” to it. This idea is discussed below.
All of the aforementioned explanations are in my view at least partly defensible, although they are not generally brought up (except for the last one) by those who think that there is a problem with perfect competition. In reality, it seems that the most important problem lies in the very desirability of perfect competition as a goal. To say it otherwise, the real issue is not that the conditions for perfect competition are rarely met in the real world, but that the conditions under which perfect competition would not be a state of radical inefficiency almost never exist.
2. Perfect Competition Is not an ideal to be replicated in the real world.
Perfect competition would maximize allocative efficiency but, although this point is not well understood (2.2.), it would destroy overall efficiency (2.1.).
2.1. Perfect Competition would maximize allocative efficiency But Would Destroy Overall Efficiency.
Perfect competition is often said to maximize efficiency. However, it only maximizes allocative efficiency, which is a mere component of overall efficiency. Thus, perfect competition would maximize overall efficiency only ceteris paribus. The crucial point is that it is not possible to keep all other things equal while satisfying perfect competition prerequisites. I will briefly discuss the two main components of efficiency which could not remain constant.
Beforehand, I would like to make it clear that I will only comment below on perfect competition. My point will not be to deny that there might be some optimal mix between perfect competition and monopoly. I will refer only to the pure model of perfect competition. Because it is confined to one extreme side of the trade-off (perfect competition), this enquiry is rather modest and can be treated succinctly.
That being said, let us first discuss productive efficiency. In order to transform the actual economy into perfect competition, one would need to achieve atomicity and thus to fragment all existing firms in small parts. It would annihilate scale economies. That is why Schumpeter wrote that:
The theory of simple and discriminating monopoly teaches that, excepting a limiting case, monopoly price is higher and monopoly output smaller than competitive price and competitive output. This is true provided that the method and organization of production—and everything else—are exactly the same in both cases. Actually however there are superior methods available to the monopolist which […] are not available at all to a crowd of competitors [….] Whenever this is so, then that proposition is no longer true. In other words, this element of the case for competition may fail completely because monopoly prices are not necessarily higher or monopoly outputs smaller than competitive prices and outputs would be at the levels of productive and organizational efficiency that are within the reach of the type of firm compatible with the competitive hypothesis.
Obviously, it would be harmful to force firms to reduce their price-cost margins if doing so generated far higher costs. There is therefore an unavoidable––and well-known––trade-off between allocative efficiency (the maximization of which requires that perfect competition be brought about) and productive efficiency (which requires in particular that firms achieve as many scale economies as possible). In perfect competition, which is one extreme side of the trade-off, overall efficiency is clearly minimized. Moreover, many complex goods (airliners…) could not be produced at all in perfect competition. It is not “just” an issue of relative efficiency in producing given goods.
It is also straightforward to show that the pure model of competition would destroy dynamic efficiency. All the dynamic gains generated by competition (better and cheaper products) result from the fact that firms try to surpass their rivals and to get larger market shares. However, firms will not make efforts to succeed if they fear that, once they do, they will be stabbed in the back by competition agencies. Now, perfect competition requires precisely that no firm be better than any other and that none of them gets any significant market share. Therefore, the model of perfect competition would thwart the benefits expected from ordinary competition.
Furthermore, over-efficient firms, whatever the cause of their superiority, naturally tend to acquire larger market shares and thus to extend the field of their efficiency. Hayek phrased it: “Although […] production is not likely to be more efficient because it is conducted by a monopoly, it will often be conducted most effectively by one particular enterprise that for some special reason is more efficient than other existing ones.” Thus, regardless of the incentive issue, the perfect competition model would harm the consumers by stunting the growth of over-efficient firms.
Now, the radical inefficiency of perfect competition has been emphasized by famous antitrust specialists. For instance, Robert Bork wrote that “A determined attempt to remake the American economy into a replica of the textbook model of competition would have roughly the same effect on national wealth as several dozen strategically placed nuclear explosions.” According to him, the problem with perfect competition is not just that it fails to optimize overall efficiency. It is that it would annihilate it and starve us to death.
Put differently, the real issue is not that conditions for perfect competition are rarely met in the real world, but that the conditions under which perfect competition would be compatible with a state of reasonable efficiency almost never exist. That being said (which, I repeat, only deals with the pure model of competition), how do contemporary authors view this pure model?
2.2. It is not well understood that perfect competition is not an ideal Goal.
Most authors affirm in their introductory chapters that perfect competition maximizes welfare (2.2.1.). They often admit, at a later stage, that there is a trade-off between the various forms of efficiency, but they rarely incorporate it in their assessment of perfect competition (2.2.2.).
2.2.1. Modern authors routinely affirm that perfect competition maximizes welfare.
Bork’s Antitrust Paradox is a must-read, not a marginal book. Therefore, any competition law specialist must have read the aforementioned deprecating comment on perfect competition, which anyway should be common sense. So, one may wonder, how do contemporary authors position themselves regarding this highly negative judgment? Do they approve it completely? Do they think that it is sound but exaggerated? Do they consider it to be totally wrong and thus criticize it harshly?
None of that. They just ignore it and go on undisturbedly claiming that perfect competition law maximizes welfare, at least in their introductory chapters. Let us quote some of these authors.
Professor Moritz Lorenz writes that “The economic model of perfect competition illustrates a world in which the benefits from competition are fully maximised.” According to Professors Jones and Sufrin, “The theory of perfect competition presents a model of a market on which efficiency is maximized and cannot, therefore, be improved by the application of competition rules.”
Similarly, Professor Hovenkamp holds that “Perfectly competitive markets are generally thought to be “efficient” because they do the best job of providing consumers with goods at the cost of producing them.” As a result, competition maximizes the total value of goods produced in society.” Professors Prieto and Bosco also consider that “In perfect competition, consumer welfare is maximized.”
How to explain that so many authors mistake a state of minimal efficiency for one of maximal welfare?
2.2.2. Most modern authors admit that there is a trade-off but do not incorporate it in their assessment of perfect competition.
It appears that many authors admit that there is a trade-off between allocative efficiency and productive (and dynamic) efficiency––they reproduce Williamson’s graph––but do not really incorporate it in their assessment of perfect competition. Textbooks generally begin their presentation with a bold and erroneous claim (perfect competition maximizes welfare) and, although they later mention (more or less ambiguous) qualifications, they never formally modify the initial statement.
Professors Jones and Sufrin’s handbook is an interesting example. It is more skeptical than the average and avoids to a large extent the flaw that I am discussing, but it remains somehow contradictory.
On the page 7, these authors write, as mentioned above, that “The theory of perfect competition presents a model of a market on which efficiency is maximized and cannot, therefore, be improved by the application of competition rules.” However, just after pointing out the factual discrepancy between the actual economy and perfect competition, they comment that:
Important caveats must thus [sic] be attached to the theory that the perfectly competitive market is superior. In particular, in most markets economies of scale make the attainment of perfect competition impossible. In some markets monopolies or different structures are ‘natural’ because of the MES. […] The reality is that most markets lie somewhere between perfect competition and monopoly […].
It seems that, as discussed earlier, these authors mix what is and what should be. They show that the actual economy differs from perfect competition and deduce from this statement (“thus”) that perfect competition is not necessarily “superior.”
More importantly, after conceding that “Important caveats must thus be attached to the theory that the perfectly competitive market is superior,” they open a new section by writing that “A perfectly competitive market maximizes welfare because it leads to efficiency.” It looks as if they did not give any weight to the “important caveats” that they had just mentioned.
Their book, however, does not stop at this point. In the section which begins with a celebration of perfect competition, the authors discuss again the welfare trade-off and quote several economists arguing that perfect competition does not necessarily maximise productive efficiency. Yet, they do not update their synthetic assessment of perfect competition and they fail to emphasize that this pure model of competition actually annihilates welfare instead of maximizing it. Above all, as will be shown below, they still consider it to be a relevant “benchmark.”
3. Perfect competition is not even a goal that should be Pursued “as far as possible.”
Some authors claim that perfect competition should be pursued “as far as possible” (3.1.), only a few of them clearly deny that one should do so (3.2.), and most are just equivocal about the relevance of perfect competition for competition law (3.3.).
3.1. Some antitrust practitioners think that perfect competition should be Pursued “as far as possible.”
Professors Prieto and Bosco complain that “competition policy is often caricatured and stigmatized by being reduced to the perfect competition model.” This is at least partly true. But it should be noted that this “caricature” is sometimes propounded by competition law officials.
For instance, a member of the French competition agency holds that “the quest for conditions ‘as close as possible” to perfect competition is the main concern (enjeu immédiat) of competition law rules and, beyond, of competition policies.” Similarly, a previous agent of the French agency declared that “Promoting competition means furthering the intervention of the State, through competition agencies, and these competition agencies are going to try to re-establish the conditions for perfect competition on a market which is naturally imperfect […].
It seems that the relevance of perfect competition for competition law is far from clear.
3.2. A few authors clearly state that one should not even try to Achieve perfect competition “as far as possible.”
I can name only a few authors who unambiguously state that one should not try to realize perfect competition, even “as far as possible.” Bork is one of them. Schumpeter is another. Let us quote him:
[I]t is not sufficient to argue that because perfect competition is impossible under modern industrial conditions—or because it always has been impossible—the large-scale establishment or unit of control must be accepted as a necessary evil […]. What we have got to accept is that it has come to be the most powerful engine of that progress and in particular of the long-run expansion of total output not only in spite of, but to a considerable extent through, this strategy which looks so restrictive when viewed in the individual case and from the individual point of time. In this respect, perfect competition is not only impossible but inferior, and has no title to being set up as a model of ideal efficiency. It is hence a mistake to base the theory of government regulation of industry on the principle that big business should be made to work as the respective industry would work in perfect competition.
More surprisingly, Edward Chamberlin is a third one. Indeed, he held that:
The main point I want to make is that the welfare ideal itself (as well as the description of reality) involves a blend of monopoly and competition and is therefore correctly described as one of monopolistic competition […].
These commentators are exceptions. A great many authors are equivocal about the relevance of perfect competition for competition law.
3.3. Most authors are equivocal about the relevance of perfect competition for competition law.
Most authors on competition law are very ambiguous regarding the relevance of perfect competition for competition law. They are particularly equivocal as to whether it should be used as a normative benchmark to assess the legality of commercial practices.
The fundamental issue is whether “workable competition” is a welfare ideal or just an operational target, a milestone on the road leading to “perfect competition,” the latter being the ideal. In other words, is it desirable or harmful to go beyond workable competition and to approach perfect competition as far as possible? Is it safe to pursue perfect competition––even if it is an unattainable goal––or is this model as the sun, something dangerous to get too close?
Let us see how some authors deal (or refuse to deal) with this issue.
- Professors Jones and Sufrin.
Professors Jones and Sufrin’s handbook has already been extensively discussed. It has been shown that they state alternatively that perfect competition maximizes welfare and that it does not. One should not be surprised that they are as equivocal regarding the normative relevance of perfect competition for competition law.
After mentioning the “caveats” about the superiority of perfect competition, these authors conclude that:
The model of perfect competition is nonetheless still useful as a benchmark against which to measure the competitiveness of real markets. The difference between the perfectly competitive and monopolistic (or oligopolistic) market focuses attention on the crucial question: whether the firm or firms have sufficient market power to raise prices above the competitive level and keep them there.
This statement seems to imply that perfect competition could and should be used as a benchmark to assess the legality of commercial practices. But if, as they say, perfect competition is not necessarily “superior” to the actual market, there would be no point in using this model as a benchmark to assess legality.
However, even though the authors expressly use the word “benchmark,” they do not exactly say that it should be used for assessing legality––in clear, that this benchmark should be used for normative purposes. They merely write that it is useful “as a benchmark against which to measure the competitiveness of real market.” Although it is implicit that––contrary to Chamberlin’s opinion––legality should depend on how competitive a practice is (or on how far from perfect competition it is), it is not expressly said. But what would be the point of measuring competitiveness otherwise?
The authors further blur their pronouncement by adding that:
The value of any model lies not in the absolute fidelity of each element to real world phenomena, but in the model’s ability to make useful predictions and, more importantly, its ability to give us meaningful verbal accounts of our observations.
The problem is that competition law enforcers are not mere observers, business analysts or stock-exchange traders. They do not only make predictions. They are policemen who actively interfere with firms’ pricing policies and who wield state imperium. Thus, the real issue is whether perfect competition can be used as a normative benchmark. Unfortunately, the authors do not clearly take up this question.
- Professor Hovenkamp.
Similarly, Professor Herbert Hovenkamp (on which Professors Jones and Sufrin rely) states that:
In short, like all scientific models, the model of perfect competition applies only imperfectly in the real world; nevertheless it can be of great service to the antitrust policy maker in predicting the consequences of a certain action or legal rule.
Again, competition agencies do not just try to “predict” the course of economic events. They actively engage in altering it. In order to do so, they need to make predictions (in particular about the effects of their own behavior), but they also need to pass a normative judgment. Anyway, it is hard to understand how perfect competition could help making predictions about real events. It will be shown below that this model cannot even indicate the direction to follow.
- Professors Prieto and Bosco.
As mentioned earlier, Professors Prieto and Bosco complain that “competition policy is often caricatured and stigmatized by being reduced to the perfect competition model.” Yet, these authors are rather equivocal regarding the relevance of perfect competition for competition law. They seem to alternate between two opposite views.
On the one hand, they appear to think that perfect competition is a welfare ideal, a goal that competition agencies could safely pursue, despite the difficulty of achieving it. On the other hand, they seem to think, as Chamberlin did, that the welfare ideal, known as “workable” competition, is a mix between competition and monopoly. In this view, a market could be too competitive as well as too monopolistic and there would be no reason to pursue perfect competition.
They seem to follow the first view (perfect competition as an––unachievable––ideal) when they write that “In perfect competition, consumer welfare is maximized” or that “the perfect competition model remains a reference standard. Nevertheless, the goal is not to achieve it, but to get closer of it.”
They appear to follow the opposite view when they complain that “competition policy is often caricatured and stigmatized by being reduced to the perfect competition model” (why would it be stigmatizing if it were safe to pursue perfect competition?) or when they write that:
It is all about being balanced (affaire de mesure). Thus, competition policy should be mobilized only against firms having a rather significant market power. Otherwise, firms would lose any incentive to invest and innovate and the real interest of the consumers would be damaged. It is not market power per se which must be stigmatized, but a degree of market power which remains hard to determine. One is allowed to think that the debate has just been shifted. The true question remains untouched: to locate noxiousness, namely to know as from when the appropriation of consumer’s surplus becomes inadmissible.
This paragraph is close to what I consider to be the correct view.
- Professor Petit.
Professor Petit adopts a rather skeptical stance. He comments that:
If there is something that one must retain from neoclassical models, it is that competition and its antithesis, monopoly, have radically different results on social welfare. But it would be imprudent to conclude that these models should be used as a compass for competition agencies.
The main shortfall of classical and neoclassical theories is to abstain from formulating any recommendation to public authorities. In any case, the models which are made-up are so far from reality that they hardly lean themselves to be actually enforced. This lack of operational significance is perhaps fortunate given the radical consequences that would result from an attempt to legally enforce these classical and neoclassical doctrines.
I agree by and large with Professor Petit. I regret, however, that he does not enquire further about the implications for competition law of the unworkability of these models.
4. Perfect Competition is––at least officially––not considered as a goal or a Benchmark in actual law.
Both the European case law (4.1.) and the European Commission (4.2.) officially dismiss perfect competition as a goal or a benchmark for competition law.
4.1. The European Case Law Refers to Workable or Efficient Competition.
The European case law makes it clear that it does not require perfect competition and is satisfied with “efficient” or “workable” competition. What constitutes “efficient” or “workable” competition is a mystery. But it is almost certain that it does not amount to perfect competition.
4.2. The European Commission Guidelines Do not Equate the Competitive Level with Marginal Costs.
In neoclassic economics, perfect competition helps defining the concepts of “competitive level” and “market power” through the “marginal cost pricing rule.” But it appears that the European Commission, in its guidelines, severs the aforementioned concepts from the marginal cost rule.
The Commission writes for instance that:
Negative effects on competition within the relevant market are likely to occur when the parties individually or jointly have or obtain some degree of market power […]. Market power is the ability to maintain prices above competitive levels for a significant period of time […]. In markets with high fixed costs undertakings must price significantly above their marginal costs of production in order to ensure a competitive return on their investment. The fact that undertakings price above their marginal costs is therefore not in itself a sign that competition in the market is not functioning well and that undertakings have market power that allows them to price above the competitive level. It is when competitive constraints are insufficient to maintain prices and output at competitive levels [?] that undertakings have market power within the meaning of Article 81(1).
If the Commission is consistent, it should entail that it does not consider perfect competition as a norm and probably neither as a goal which should be pursued as far as possible. However, the Commission fails to set out any theoretical substitute to replace perfect competition.
5. The Dismissal of Perfect Competition Has Left a Gap in the Theory.
It should be of paramount importance, if competition law is to be a rational discipline, to define theoretically what “competition” and the “competitive level” are. The perfect competition model could have offered such a theoretical basis. However, as shown above, it is not retained by competition law as a normative benchmark.
This leads to the question: how to replace the discarded concept? The concepts of “imperfect” and “monopolistic” competitions cannot perform this function (no more than “efficient” or “workable” competition) (5.1.). In reality, it appears that this question has been left unanswered (5.2.).
5.1. The Concepts of “Imperfect” and “Monopolistic” Competitions Are Useless.
Many claim that the substitute could be “monopolistic” or “imperfect” competition. However, these concepts, in addition to being extremely vague (as vague as “workable” or “efficient” competition), cannot guide competition agencies.
Even Edward Chamberlin, who invented the notion of “monopolistic” competition, pointed out that it could not be used as a model for competition law. As he wrote:
[O]rdinary (purely) competitive theory remains the chief source of our criteria as to what should be done if possible, and of the direction in which we should move so far as we can. A striking instance is the subtitle of this part of the program of these meetings: “Can the American economy be made more competitive?” The implication is evident that if it can be it should. Now if pure competition is the ideal, the direction in which we should move is very clear. For it is easy enough to show that the actual economy is shot full of monopoly elements, and hence that any move to get rid of them or to diminish their importance is in the right direction.
The main point I want to make is that the welfare ideal itself (as well as the description of reality) involves a blend of monopoly and competition and is therefore correctly described as one of monopolistic competition. If this is true, it is no longer self-evident which way we should move, for it is no longer self-evident on which side of the ideal lies the actuality for which a policy is sought. It is possible that the economy should be made “more competitive”; but it is also quite possible that it should be made “more monopolistic” instead.
[S]ince we cannot measure monopoly and competition quantitatively, there is no way of comparing the actual with the ideal on any yardstick involving these concepts.
Milton Friedman also found that “The theory of monopolistic competition offers no tools for the analysis of an industry and so no stopping place between the firm at one extreme and general equilibrium at the other.”
Joan Robinson, a neo-Marxist author, initially adopted a different stance. She analyzed the actual economy as an “imperfect” state that it would be necessary to turn into perfect competition. In her mind, any discrepancy with perfect competition constituted “a nail in the intellectual coffin of free market capitalism.” She saw perfect competition as a normative model. However, she seems to have later changed her opinion. In addition, her theory did not offer any serious operational basis which could be used by competition law.
5.2. The Question of How to Replace Perfect Competition Is Generally Left Unanswered.
Generally, the question of how to replace perfect competition as a theoretical basis for antitrust is left unanswered.
For example, the European Commission, in its aforementioned guidelines, discards marginal cost (and thus perfect competition) as a way of defining the competitive level, but it does not offer any alternative criterion to explain what the competitive level is. It goes on referring to the concept of competitive level (and market power), but it does not elucidate its content.
Similarly, the OECD takes notice that the concept of market power as defined in microeconomics is not used in competition law, but it “omits” to specify the meaning of this notion in law. It contents itself with mentioning that “the notion of competitive prices in competition law is often less precise”:
[The] technical notion of market power defined as pricing above marginal cost may [sic] be distinct from the notion of market power as a competition law concept. Perfect competition is a theoretical concept that relies on several restrictive assumptions to be satisfied. Therefore, competition economists have suggested the concept of effective competition which is compatible with prices exceeding marginal cost. While in the economic discipline equilibrium prices under perfect competition are equated with marginal costs and may therefore also be considered competitive prices, the notion of competitive prices in competition law is often less precise [sic] and prices in the absence of competition law infringements may be called competitive prices even if they exceed marginal cost and the market cannot be characterized as perfectly competitive.
Let us note that, according to the OECD, prices may be considered as competitive if they exceed marginal cost but (only) “in the absence of competition law infringements.” One of the problems is that market power is generally considered as a criterion for a “competition law infringement.” Consequently, a key element of such an infringement is left indeterminate.
To sum up, the phasing out of perfect competition left a gap in the theory. The current competition law has dismissed marginal cost (and so perfect competition) as a yardstick to define the competitive level and has failed to offer any substitute. It goes on using the concepts of “market power” and “competitive level,” but it does not bother giving them any theoretical content.
The pure model of competition has officially passed away, but it seems to be still among us. As Frank Machovec wrote, “the ghost of the model of perfect competition continues to haunt antitrust policy.”
Everybody agrees that perfect competition is not a description of the actual world. However, the implications of this point for competition law are not clear. Many authors seem to consider that it renders it problematic to use perfect competition as a legal standard. However, such an inference is contestable, if not necessarily wrong, because the goal of law is precisely to modify the actual world and to turn it into something better.
The real problem with perfect competition lies in its very desirability. Although this model would maximize allocative efficiency, it would also destroy productive and dynamic efficiencies. Overall, it would result in a radical loss of welfare.
Unfortunately, this finding is contradicted by many authors, who claim on the contrary that perfect competition would maximize welfare. Paradoxically, the same authors generally admit that there is a trade-off between allocative and productive (and dynamic) efficiency, but they do not incorporate this consideration into their assessment of perfect competition, or only very ambiguously. It seems that the impossibility of achieving perfect competition has barred them from enquiring about its desirability.
Moreover, most commentators are equivocal as to the relevance of the pure model of competition for competition law. They generally seem to consider that it could somehow be used as a goal that one should try to approach “as far as possible.” A few authors have clearly criticized the desirability of perfect competition and affirmed that it is of no use for competition law.
In actual law, perfect competition has been (officially) dismissed as a goal. However, beyond some superficial catchwords as “workable” or “efficient” competition, no alternate theoretical concept has been substituted to the discarded one. Competition agencies go on employing notions such as “competitive level” or “market power” that are deprived of intellectual content since the model of perfect competition has been discarded.
 It is a highly debated issue to what extent the farm sector is exempted from the European competition law.
 Moritz Lorenz, An Introduction to EU Competition Law, Cambridge University Press, 2013, p. 5.
 Alison Jones and Brenda Sufrin, EC Competition Law, Oxford University Press, 3rd Edition, 2008, p. 11.
 Herbert Hovenkamp, Federal Antitrust Policy – The Law of Competition and its Practice, West Group, Hornbooks Series, 2nd Edition, 1999, p. 11.
 Catherine Prieto and David Bosco, Droit européen de la concurrence – Ententes et abus de position dominante, Bruylant, 2013, p. 291 (“A partir des travaux des années 1930, les travaux de Robinson et de Chamberlin ont conduit à ne plus raisonner exclusivement sur le modèle théorique de la concurrence pure et parfaite. L’intérêt s’est porté sur la réalité des marchés lesquels, pour la plupart, ne connaissaient pas les conditions requises par la théorie”).
 Nicolas Petit, Droit européen de la concurrence, Lextenso éditions, Montchrestien, 2013, pp. 127-128 (“Nulle part ne semblent réunies les propriétés de la concurrence pure et parfaite”).
 Ibid., p. 128, footnote 58.
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 11.
 Nicolas Petit, Droit européen de la concurrence, aforementioned, p. 129 (“En toute hypothèse, d’ailleurs, les modèles échafaudés se prêtent mal à une mise en œuvre pratique, tant ils sont éloignés de la réalité des marchés”). Likewise, according to Professor Moritz Lorenz: “The objective of EU competition law is the protection and promotion of ‘effective competition’. […] However, surprisingly neither law nor economics provides a clear definition of what ‘effective competition’ means. From an economic perspective, the discussion above has made it clear that effective competition cannot be equated with perfect competition in every market. Given that the assumptions underlying the economic model of perfect competition are often not fulfilled in real life, it would not be possible to achieve a perfectly competitive market outcome in every industry” (Moritz Lorenz, An Introduction to EU Competition Law, aforementioned, p. 21).
 John Clark is a noticeable exception (“[S]ome of the features listed as « imperfections » in our present theoretical scheme may turn out to have some positive use in actual situations. […] If there are, for example, five conditions, all of which are essential to perfect competition, and the first is lacking in a given case, then it no longer follows that we are necessarily better off for the presence of any one of the other four. In the absence of the first, it is a priori quite possible that the second and third may become positive detriments; and a workably satisfactory result may depend on achieving some degree of « imperfection » in these other two factors” – John M. Clark, Toward a Concept of Workable Competition, The American Economic Review, Vol. 30, No. 2, Part 1 (Jun., 1940), p. 242). This author is widely quoted, but he does not seem to be followed into this specific consideration.
 See for example Murray Rothbard, Man, Economy, and State – A Treatise on Economic Principles, Ludwig von Mises Institute, 2nd Edition, Scholar’s Edition, 2009, p. 729 (“to apply the concept of pure competition to existing firms would mean, for example, assuming a very large number of similar firms producing the identical product. If this were done, say, with General Motors, it would mean that either GM must conceptually be divided up into numerous fragments, or else that it be multiplied. If divided, then unit costs would undoubtedly be higher, and then the “competitive firm” would suffer higher costs and have to subsist on higher prices.”).
 Joseph Schumpeter, Capitalism, Socialism, and Democracy, Routledge, Taylor & Francis e-Library, 2003, p. 101.
 This point is obscured by the fact that perfect information, which is required by perfect competition, would imply that firms adopt the best productive methods. This element makes some people say that perfect competition would also maximize productive efficiency. This statement is confusing for a twofold reason. First, information cannot be increased––it is rather reduced––by competition agencies. Agencies have some leverage on the size of firms or their ability to coordinate, but almost none on the degree of information. Second, a sound formulation of this thesis would be, as enunciated by Schumpeter, that perfect competition (or more exactly perfect information) maximizes productive efficiency given the techniques of production and methods of organisation compatible with the condition of atomicity. Once this qualification is taken into account, it appears obvious that the level of efficiency attainable under perfect competition must be considerably inferior to the one resulting from a private monopoly.
 Friedrich Hayek, Law, Legislation, and Liberty – A New Statement of the Liberal Principles of Justice and Political Economy, Routledge, 2013, p. 411. Of course, this argument does not apply to public monopolies.
 Robert Bork, The Antitrust Paradox – A Policy at War with Itself, The Free Press, 1993, p. 92.
 Moritz Lorenz, An Introduction to EU Competition Law, aforementioned, p. 5.
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 7.
 As if there were only one cost of producing a good…
 Herbert Hovenkamp, Federal Antitrust Policy, aforementioned, p. 11.
 Catherine Prieto and David Bosco, Droit européen de la concurrence, aforementioned, p. 290 (“En situation de concurrence pure et parfaite, le bien-être du consommateur est au maximum”).
 As Hovenkamp writes, “The welfare trade-off model that antitrust uses most often was developed in the 1960s by economist Oliver E. Williamson. The model is […] probably the most frequently reproduced graph in all of the antitrust literature” (Herbert Hovenkamp, The Antitrust Enterprise – Principle and Execution, Harvard University Press, 2008, p. 26).
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 7.
 Minimum efficient size. The MES relates to efficiency.
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 12.
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 13.
 As the authors write, “Although theory accords prime position to allocative efficiency, improving overall efficiency may require a trade-off between different types of efficiency […]” (Ibid., p. 13). See also, page 8, “The relation of dynamic efficiency to the concept of perfectly competitive markets is complex for, as we see below, it can be argued that innovation may be better delivered by monopolistic rather than competitive markets and that the ability to achieve market power is an important spur to innovation.” In the previous paragraph, where the authors dealt with productive efficiency, they did not make such a caveat. On the pages 15-16, they reproduce a large quote by D. Hildebrand saying that the welfare loss associated with market power is “concerned solely with the allocation of resources in the context of fixed technology and a given cost situation” and that it is “crucially predicated on the assumptions that the costs of the firm concerned do not fall due to production rationalization.”
 Catherine Prieto and David Bosco, Droit européen de la concurrence, aforementioned, pp. 288-289 (“La politique de la concurrence est souvent caricaturée et stigmatisée en étant réduite au modèle de la concurrence pure et parfaite, alors que celui-ci n’en est que le point de départ théorique à la fin du XIXème siècle jusque dans les années 1930”). Professor Prieto also wrote that “Robinson and Chamberlin demonstrated, as from the 30’s, that the model were never met. So, it is time to stop heaping abuse on competition policy by criticizing the abstractness of perfect competition prerequisites” (Catherine Prieto, “La culture de la concurrence,” in Martine Behar-Touchais, Nicolas Charbit, and Rafael Amaro (eds.), A quoi sert la concurrence ?, Institut de droit de la concurrence, octobre 2014, p. 723 (“[R]obinson et Chamberlin ont établi, dès les années 1930, que les présupposés du modèle n’étaient jamais réunis. Il est donc temps de ne plus invectiver la politique de concurrence en dénonçant le caractère abstrait d’une concurrence pure et parfaite”).
 Sandra Lagumina et Olivier Béatrix, “Droit de la concurrence et politique de concurrence : Un couple en quête de légitimité,” in Martine Behar-Touchais, Nicolas Charbit, and Rafael Amaro (eds.), A quoi sert la concurrence ?, précité, p. 430 (“[la] recherche des conditions « aussi proches que possible » de la concurrence parfaite est donc l’enjeu immédiat des règles du droit de la concurrence et, au-delà, des politiques de concurrence”).
 Irène Luc, “A quoi sert la concurrence ?” (interview), Vimeo, Concurrences TV, 0:35 – 1:28, link (“La concurrence… est aussi un objectif théorique. Un objectif, qui n’est pas un objectif ultime… C’est un objectif théorique parce qu’on constate concrètement sur le marché que la concurrence pure et parfaite n’existe pas, et on a affaire à des marchés caractérisés par la concurrence imparfaite. … Promouvoir la concurrence, ça veut dire favoriser l’intervention de l’Etat, par l’intermédiaire d’autorités de concurrence, et ces autorités de concurrence vont essayer de rétablir les conditions de la concurrence pure et parfaite sur un marché par nature imparfait, ce qui veut dire préserver la loyauté… dans les relations commerciales”)
 Joseph Schumpeter, Capitalism, Socialism, and Democracy, Routledge, Taylor & Francis e-Library, 2003, p. 106.
 Edward Chamberlin, “Product Heterogeneity and Public Policy,” The American Economic Review, Vol. 40, No. 2, Papers and Proceedings of the Sixty-second Annual Meeting of the American Economic Association, may 1950, p. 86.
 Is this level defined by perfect competition?
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 13.
 Alison Jones and Brenda Sufrin, EC Competition Law, aforementioned, p. 14.
 Herbert Hovenkamp, Federal Antitrust Policy, aforementioned, p. 11.
 Catherine Prieto and David Bosco, Droit européen de la concurrence, aforementioned, pp. 288-289.
 Ibid., p. 290 (“En situation de concurrence pure et parfait, le bien-être du consommateur est au maximum”).
 Ibid., p. 291 (“Le modèle de la concurrence pure et parfaite reste un « étalon » de référence. Néanmoins, l’objectif n’est pas de l’atteindre, mais de s’en rapprocher”). I often find it hard to ascertain whether these authors are exposing their viewpoint or just presenting other people’s. Here, I think they are stating their own opinion.
 Catherine Prieto and David Bosco, Droit européen de la concurrence, aforementioned, p. 293 (“Tout est donc affaire de mesure. Ainsi, la politique de concurrence ne devrait être mobilisée qu’à l’égard des entreprises ayant un assez grand pouvoir de marché, à défaut de quoi les entreprises perdraient toute incitation à investir et à innover au détriment de l’intérêt bien compris des consommateurs. Ce n’est pas le pouvoir de marché en soi qui doit être stigmatisé, mais un degré de pouvoir de marché qui reste difficile à déterminer. Il est permis de penser que l’on n’a fait que déplacer le débat. Le vrai questionnement reste entier : saisir la nocivité, c’est-à-dire savoir à partir de quand la captation du surplus du consommateur devient inadmissible.”).
 Unfortunately, the last word (inadmissible) tends to blur its meaning, because it implies that, up to this point, market power would merely be tolerated.
 Nicolas Petit, Droit européen de la concurrence, p. 127 (“S’il est une chose qu’il faut retenir des modèles néo-classiques, c’est d’avoir montré les résultats radicalement différents de la concurrence et de son antithèse, le monopole, sur le bien-être social. De là cependant à ériger ces modèles en boussole de l’intervention des autorités de concurrence, il n’y a qu’en pas qu’il serait imprudent de franchir”).
 Ibid., p. 129 (“La principale carence des théories classiques et néo-classiques, c’est d’omettre de formuler toute recommandation à l’adresse des pouvoirs publics. En toute hypothèse, d’ailleurs, les modèles échafaudés se prêtent mal à une mise en œuvre pratique, tant ils sont éloignés de la réalité des marchés. Ce manque de portée opérationnelle est peut-être heureux au regard des conséquences radicales qu’aurait emportées une transposition politique des doctrines classiques et néo-classiques”).
 See for instance Judgment of the Court of 13 July 1966, “Consten and Grundig” (“The situation as ascertained above results in the isolation of the French market and makes it possible to charge for the products in question prices which are sheltered from all effective competition”).
 See for instance Judgment of the Court of 25 October 1977, “Metro” (“20 The requirement contained in articles 3 and 85 of the EEC Treaty that competition shall not be distorted implies the existence on the market of workable competition, that is to say the degree of competition necessary to ensure the observance of the basic requirements and the attainment of the objectives of the Treaty, in particular the creation of a single market achieving conditions similar to those of a domestic market”).
 European Commission, Guidelines on the application of Article 81(3) of the Treaty, 2004, §25.
 Edward Chamberlin, “Product Heterogeneity and Public Policy,” aforementionned, pp. 85-86 (quoted in Don Bellante, “Edward Chamberlin: Monopolistic Competition And Pareto Optimality,” Journal Of Business & Economics Research Volume 2, Number 4).
 Quoted in Don Bellante, “Edward Chamberlin: Monopolistic Competition And Pareto Optimality,” Journal Of Business & Economics Research Volume 2, Number 4, p. 21. Similarly, Frank Machovec reports that Stigler “conceded that the ideas of Sraffa, Robinson, and Chamberlin made an indisputably valuable contribution to economics, but he argued that the perfectly competitive model should remain ‘the chief work of economic theorists’ because the model of monopolistic competition did not appreciably enhance the ability to make predictions (Stigler 1937:707; and 1968 : 320)” (Frank M. Machovec, Perfect Competition and the Transformation of Economics, Routledge, 2003, p. 274). According to this author, “the proposals of Sraffa, Robinson, and Chamberlin threatened the newly-won premier position of the perfectly competitive model. Yet when all the dust had settled, the model of monopolistic competition faltered because it was plagued with the curse of indeterminateness” (ibid.).
 Don Bellante, “Edward Chamberlin: Monopolistic Competition And Pareto Optimality,” aforementioned, p. 19.
 See Ignace De Leon, “Antitrust policy versus the rule of law,” p. 11.
 Ulrich Schwalbe and Frank Maier-Rigaud, Market definition, OECD, Policy Roundtables, DAF/COMP (2012) 19, 2012, pp. 24-25, quoted in Catherine Prieto et David Bosco, Droit européen de la concurrence, aforementioned, pp. 299-300.
 Frank M. Machovec, Perfect Competition and the Transformation of Economics, aforementioned, p. 221.