Smith rightly concludes that breaking up big firms is not a perfect solution to the monopoly problem. (He thinks, incorrectly, in my view, that breakup is too hard; the real reason not to break up big firms is that they are often more efficient than small ones.) And he rightly gives a list of alternatives to breaking big firms up, including unions, minimum wage laws, putting workers on corporate boards, and imposing tougher labor standards on large firms than small. But he doesn’t seem to see where all of these alternatives point.
Where do minimum wage laws and applying tougher labor standards to large firms point?
To rate regulation, of course. To that approach to governing the market that once — in the decades following World War Two — stretched from securities brokerage to railroads to telephones to airlines.
In a regulated industry, a government administrative agency dictates prices and performance standards to the privately-owned firms that compete in the market. Applying tougher labor standards to firms with monopoly power, a proposal that Smith attributes to Nick Hanauer, is a shade of the old rate regulation, which was often imposed on monopolized industries, such as telephone service, to restrain the power of large firms.
Minimum wage laws are themselves a form of blunt price regulation, blunt because they are imposed on a one-off basis by legislatures instead of by expert administrative agencies with authority to revise the prices dynamically in response to changing circumstances. And both unionization and putting workers on corporate boards are even blunter forms of rate regulation, in that they hope that by increasing the bargaining power of workers, workers will succeed at negotiating the higher wages and better working conditions that a regulator would be empowered to impose by fiat.
True, most of Smith’s proposals are aimed at softening the consequences of labor market monopsony, whereas rate regulation was generally aimed at softening the consequences of consumer market monopoly. But there’s no reason why the Department of Labor couldn’t apply the tenets of rate regulation to labor markets.
Rate regulation is the most developed form of intervention in markets, one that encompasses all the other forms, but also goes beyond them, so it’s the natural choice for achieving just market-level distributions of wealth where unregulated markets fail to do so. A rate regulator can unionize an industry if the regulator wishes, just as the ICC effectively cartelized long-haul railroads to stabilize their prices: the regulator simply insists on approving only a wage tariff that is uniform for all workers, effectively forcing workers to bargain collectively with their employers. But a rate regulator can do more than that, regulating market entry to strike a balance between job security and competitiveness, insisting that workers offer certain bundles of skills, and even imposing workplace safety and benefits standards.
Once we start to believe that markets are failing, and that just breaking up big firms won’t achieve distributively fair market outcomes, as economists seem to be concluding, the door is open to market intervention, and at that point it makes sense to use the best tool for the job. The one-off ad hockery of minimum wages won’t do. Nor will strengthening unions — if you make them strong enough to really succeed, you make them strong enough to oppress investors and consumers. What you need is a politically accountable agency empowered to make markets work for all market participants.
That’s what rate regulation was, and could be again. Let’s stride to it, not slouch.
In a market economy in which dominance often rests on intellectual property, rather than on an installed base of industrial equipment, breaking up a large firm is as easy as ordering compulsory licensing, and letting markets do the hard work of pulling the rest of the firm apart. Breaking up large firms isn’t hard — it’s easy.
Noah Smith repeats the old fallacy that breaking up big firms, or reversing consummated mergers, is difficult, putting the divider in the position of creating two new companies from scratch. He writes:
It would be great if big companies could simply be divided into the competing rivals that existed before a merger took place. But once two competitors join, they tend to merge their sales departments, their engineering departments, their management structure and almost every other facet of their business. Antitrust regulators can’t easily order the merged company to split itself back into its constituent parts, because those parts no longer really exist.
Economists should know better than to make this mistake, because it involves ignoring markets. To break a firm up, all you have to do is to seize and divide up the asset that is the source of the firm’s advantage over competitors — force the licensing of key intellectual property to a new entity, for example — and the market will take care of the rest of the breakup.
The owners of the pieces of the divided asset will access markets on their own to assemble their own sales departments, engineering departments, management structures, supply chains, and so on — often, but not necessarily, by hiring away staff from the original firm that is the target of the breakup. Antitrust enforcers don’t have to worry about getting their hands dirty figuring out whether Bonnie in sales should go to the new firm, or Mark in accounting should stay with the old one. So long as antitrust enforcers divide the valuable asset properly, to ensure that the new companies are both financially viable, markets will take care of the rest. Bonnie may get a job offer from the new firm, and Mark may choose to stay put.
This messy breakup fallacy got a lot of air time twenty years ago, when a district court ordered the breakup of Microsoft. But Microsoft actually presents an excellent example of why breaking up should be easy to do in a market economy such as our own. The heart of Microsoft’s business wasn’t (and isn’t) its sales department, or even Microsoft Windows, but rather Microsoft Office, a program that had, and continues to have, a lock on virtually the entire word processing market thanks to a combination of consumer familiarity and the difficulty of exporting documents into competing systems. To break Microsoft up, all the court had to do back in 1999 — or, for that matter, would need to do today — was issue an order forcing Microsoft to release the full Microsoft Office source code and all future iterations. The court could then have just sat back and watched the company be devoured by a million startups, each offering a new flavor improving on the code.
Indeed, the easiest way to break up a big firm is to force licensing of its most valuable intellectual property assets. Because intellectual property doesn’t have a geographic location — ideas live in the ether — the problems of continued regional concentration that Smith also worries about don’t arise from licensing-driven breakups. And the beauty of it all is that in the Data Economy, intellectual property is the key to the dominance of most large firms. The age of behemoths deriving their power from vast installed bases of industrial equipment — the Standard Oils and the AT&T local loops — is gone. And so too any messiness associated with industrial deconcentration.
It’s time to recognize antitrust’s messy breakup fallacy for what it is.
Once upon a time, Chicago to New York was a cheap rail fare, whereas the much shorter trip from Chicago to Peoria was expensive, because in a dense rail network there were lots of ways to get from Chicago to New York, and therefore lots of competition on that route, whereas there were only a few ways to get from Chicago to Peoria, and therefore much less competition. The big city people who rode Chicago to New York had more money than the small city people who rode Peoria to Chicago, but the big city people paid lower fares, because the big city people benefited from competition.
Competition gave the big city folk alternatives, and that strengthened their bargaining power vis a vis the railroads. The lack of competition denied the small city folk alternatives, and that reduced their bargaining power vis a vis the railroads. But the changing of alternatives is taxation.
Suppose instead that rail competition were somehow equal in both markets, but the government were to tax small city riders by a certain amount and redistribute that amount of money to the big city riders. The tax would effectively drive up the fare paid by the small city riders and drive down the fare paid by the big city riders, achieving the same result as did the unequal levels of competition on the big and small city routes. Inequality in competition is tantamount to inequality in pricing, which is tantamount to tax and transfer.
Put another way, the presence of competition on the big city route and absence on the small city route caused the railroads to collect a large share of their revenues from small city riders. Small city riders subsidized big city riders, in a sense. The railroad acted as a kind of taxing authority, redistributing from small city riders to big city riders, but the railroad’s tax policy was determined by the competitive environment, not the railroad itself, by the presence of competition on the big city route and the absence of competition on the small city route.
Flip the competitive configuration — make the small city route less competitive than the big city route (perhaps by allowing cartelization of service providers on the big city route) — and now the big city riders will contribute a larger share of the railroad’s revenues. Now the big city riders might be said to pay the subsidy and the railroad to be taxing the big city riders for the benefit of the small city riders.
Depending on how antitrust divides up markets and goes about promoting competition in them — as I discuss briefly in another post — the competitive terrain created by antitrust represents a tax system and a specific set of distributive results.
From this perspective, if your goal is to redistribute wealth, you want more competition in some markets — the ones inhabited by the poor — and less competition in other markets — the ones inhabited by the rich. That’s just the kind of policy that progressives advocate: antitrust for big firms, cartelization for workers and small businesses.
Note that this is not an argument about the ability of competition to force a particular distribution of surplus within a market, between buyers and sellers. This is about the ability of antitrust to alter relative outcomes in different markets by altering relative levels of competition. This is about how antitrust makes price discrimination possible, and how that price discrimination redistributes, even when the different prices charged are charged by different firms in different markets.
The emails show that in 2013 Facebook cut off Twitter’s access to its users’ Facebook friend lists to cripple the growth of Twitter’s once-popular short-form video sharing service, Vine, which Twitter shuttered in 2016. The emails also show that Facebook used acquisition of the startup Onavo to spy on users, identifying WhatsApp as a serious threat in the process, and later acquiring that company, presumably to eliminate it as a competitor.
Both of these actions harmed competition, by eliminating what antitrust lawyers call “nascent competitors,” firms that could have matured into serious competitive threats to Facebook. Vine might have helped Twitter develop out of its microblogging niche into a full-fledged social media platform in direct competition with Facebook. And the same might have been true for WhatsApp, which could have leveraged its huge user base and privacy commitment to expand beyond chat into Facebook’s social media heartland.
But most antitrust policymakers today are unlikely to see either Facebook’s calculated crippling of Vine, or the company’s snooping on nascent competitor WhatsApp, as problematic. For antitrust policymakers today, refusing to share and espionage are examples of the kind of no-holds-barred striving to win that ensures that competition yields results for consumers. As the greatest living antitrust scholar today, Herbert Hovenkamp, put it in a recent treatise, making firms share with competitors — which is what Facebook refused to do when it cut Vine’s access to friend lists —
is manifestly hostile toward the general goal of the antitrust laws. It serves to undermine rather than encourage rivals to develop alternative[s] . . . of their own.
Fortunately, there is actually a strong case to be made that Facebook’s treatment of Vine, at least, violated existing antitrust laws. But before getting to that case, let’s look more closely at exactly what Facebook did to Vine and what’s wrong with antitrust’s prevailing approach to that kind of conduct.
It’s clear that access to Facebook friend lists was key to Vine’s growth, because that allowed users in effect to port part of their existing social network from Facebook over to Vine, and then to use it to do something — post short-form videos — that Facebook at the time did not yet allow users to do.
By in effect preventing users from porting their network to Vine, Facebook denied Vine an essential input — the infrastructure to port the Facebook network into Vine — that was key to allowing Vine to break into the social media market.
Two Minds About Sharing
Refusals to deal have long vexed antitrust enforcers because they appear to be at once good and bad for competition.
They are bad for competition because if the input is truly essential, then the refusal to supply it to a competitor is fatal to the competitor. Indeed, if “input” is defined broadly enough, all anticompetitive behavior amounts to a denial of access to an essential input of one kind or another. You cannot harm competition any other way.
At the same time that refusals to deal appear bad for competition, however, they also appear to be good for competition, albeit competition of the bare-knuckle sort.
The toughest races are those in which you can expect no help from the other participants. The refusal of a firm to deal with competitors just creates an incentive for those competitors to go beyond the withheld input in question to find a new way to survive, to innovate, to create, to surpass.
This view of the virtues of no-holds-barred competition serves as the basis for the current ascendancy of the “Colgate Doctrine,” the antitrust rule that a firm has no general duty to deal with competitors. The doctrine takes its name from a 1919 case in which the U.S. Supreme Court permitted Colgate, charmingly described by Justice McReynolds as “a corporation engaged in manufacturing soap and toilet articles and selling them throughout the Union,” to refuse to sell its products to discounters.
[i]n the absence of any purpose to create or maintain a monopoly, the [antitrust laws do] not restrict the long recognized right of trader or manufacturer engaged in an entirely private business, freely to exercise his own independent discretion as to parties with whom he will deal.
For most of the century during which this language has been on the books, antitrust enforcers quite reasonably read the paean to business freedom in the second clause in conjunction with the first — “[i]n the absence of any purpose to create or maintain a monopoly” — to mean that the right to refuse to deal, whatever its extent, has no purchase whatsoever on the antitrust laws, which are dedicated to preventing the creation and maintenance of monopoly.
But in recent decades, the courts have preferred to drop the qualification contained in the first clause altogether, and to recognize a general right to refuse to deal even when the creation and maintenance of monopoly are rather baldly at stake. As Justice Scalia put it in an infamous 2007 opinion,
Firms may acquire monopoly power by establishing an infrastructure that renders them uniquely suited to serve their customers. Compelling such firms to share the source of their advantage is in some tension with the underlying purpose of antitrust law, since it may lessen the incentive for the monopolist, the rival, or both to invest in those economically beneficial facilities.
This notion that triumph of any kind in the free market is a necessary incentive for progress filters our understanding of competition through the fearsome metaphor of natural selection, survival of the fittest, the war of all against all, the Origin of Species.
According to this view, the lion did not need antitrust restrictions on refusals to deal to evolve out of the primordial soup, and if the lion goes extinct because humans fail to share habitat, that represents a triumph of competition, because the lion will then be replaced with a creature that obviously represents an evolutionary advance: us. Moreover, the argument goes, if the lion had been forced to share with the ape back when the lion was the king of beasts, the ape likely would never have needed to learn to walk upright, to heave javelins at passing herds, and eventually to invent the computer.
The natural selection metaphor is a big mistake, because the apparent virtues of natural selection are subject to severe survivorship bias. We’re here, and living and thinking, so our evolution must have been a success. But all those creatures who never came to exist — imagine whatever god or fairy you wish, so long as the creature is better than us according to whatever metric you prefer — aren’t here to observe the failure of their natural selection, because they never came to be.
The only thing we can say for sure about natural selection is that it selects; we cannot say that it selects well, for the criteria according to which it selects are unregulated. Natural selection is undirected, and therefore unreliable, selection. The rumpled paper airplane that is natural selection spirals off in whatever random direction the environment happens to impose upon it, with no guarantee that the direction is good, let alone the best, according to any metric we as human beings might hope to use as measure. Climate change, and the very real prospect of the imminent termination of life on earth, is a convenient reminder that the direction of evolution — evolution that has led to us — may be very bad indeed.
It follows that to consign our markets to the same law of the jungle that has produced us is a big mistake. Indeed, it is the sort of mistake that would scandalize our forebears, who, living closer to that state of nature themselves, understood our human advantage to be our capacity to choose the criteria according to which selection proceeds, rather than to submit to the random criteria of the jungle. Our talent for directing our own selection, not to mention the selection of other creatures (think of your dog) is our great advantage. (Indeed, our forebears understood this perhaps too well, leading to an excessive affection for absolute monarchy and planned economies. Ancient Egypt springs to mind, with its conscious glorying in divine kingship as antidote to the chaos of the natural world.)
To continue to escape nature, we must continue to choose the criteria according to which we select ourselves, and that is as true when we structure our markets as when we design our education system. Markets are themselves just machines for the selection of the things we want the economy to produce, with profitability determining winners and bankruptcy determining losers. These machines are useful to us only to the extent that they select for the characteristics that are most helpful to us. A market that selects for sloth, or for behavior designed to take wealth from others without providing a quality product in exchange, is not a useful market. The way to make markets select for desirable characteristics is to ensure that the undesirable characteristics provide no advantage.
The question that refusals to deal really pose is whether permitting firms to horde essential inputs selects for characteristics that are good for the economy. And here the answer must be no. If the input denied to competitors is truly essential, then there is no obvious way to invent around it, and so the characteristic that legalizing such refusals selects is talent for identifying and appropriating essential inputs that deliver the firm from having to compete hard on all the other characteristics that we really value, such as good management, incessant innovation, quality, distribution, and low costs. Allowing refusals to deal unlevels the field.
Selecting for skill at destroying competition may of course incidentally sweep in some characteristics that we care about — ambition, of course, and innovativeness aimed at finding or creating the essential inputs — but the presence of this anticompetitive selector pulls the market out of focus, sapping competitive energies away from the things we care about — low prices and high quality — and toward monopoly.
Sometimes the question is muddied by the need to ensure that innovative firms are able to cover the costs of research and development before competitors appropriate their innovations, pile into the market, and erode profit margins. In these cases, it is the refusal to deal that keeps the playing field level, instead of skewing it, by ensuring that innovators get the proper rewards. But true refusal to deal cases are different. True refusal to deal cases involve a refusal to supply an essential input when doing so facilitates supracompetitive profit taking, a dominance of markets that is not necessary to help firms cover their costs. Antitrust policymakers today would treat every refusal to deal as if it were necessary for firms to cover research and development costs, a conceit that is necessary only because the reality of almost never condemning a refusal to deal is so unjustifiable.
The Surprisingly Apt Sports Metaphor
Ensuring that undesirable characteristics provide no advantage is just what we do when we level a playing field in sports. Take a soccer game played on a hillside, for example. The inclined field gives one side — the side with the higher goal — an advantage based on luck, or the ability to strong arm the other team when sides are chosen before play, instead of based on characteristics that we want to promote, such as training, endurance, and the ability to bend a football into a net from twenty yards out.
To avoid this sort of adulteration of play, we insist on level playing fields in sports. It’s the reason we recoiled from steroids in baseball, for example, because all those home runs created an advantage that made for boring, uni-dimensional, play. Indeed, we feel the same way about all doping, because it leads to selection based on chemistry, rather than on the endurance and coordination that we value in sports. Only the level playing field produces the fittest players, just as it produces the fittest firms.
Just as we expect opposing players to help each other up off the ground when they have fallen — because losing a player makes for less satisfying play — we should expect firms to help each other to enter markets, when that would make for tougher, and therefore more productive, competition.
Success and Excellence
The individual firm must therefore be governed by an ethic of excellence, rather than an ethic of success. For only the pursuit of excellence causes firms to affirmatively seek to bring competition upon themselves, whereas an ethic of success causes firms to seek only to win, rather than to win by being the best. We want the great athlete, who wants to run the hardest race against the toughest competitors, not the slouch or the crook, who celebrates when the going gets easiest.
This distinction, between the pursuit of success and the pursuit of excellence, may be loosely, and probably unfairly, associated with the divergent outlooks of the two great civilizations of European antiquity, the Romans and the Greeks. Ancient Greek culture focused on the struggle with the self, the desire to go beyond mortal limits through exposure to competition of the highest order, a desire reflected in the tradition of the Olympic Games.
The pursuit of success over excellence is a recipe for long-term failure of industry, and a threat to American national security in a world in which America is no longer clearly the most technologically advanced nationor the strongest economy, a world in which the failure to demand that our firms strive to be the best, even when they could succeed with less, could well mean the difference between victory and defeat in the next war. (True, Rome built a more enduring empire than did the Greeks, but that is only because internally, in their training and organization, the Romans were Greek.)
Which takes us back to what Facebook did to Vine. By killing Vine off via refusal to deal, Facebook prevented Vine, and Twitter, from morphing into genuine challenges to Facebook’s dominance as all-purpose social media platform.
That means that today Facebook doesn’t face the kind of competition it needs to continually improve, the competition on everything from likes to privacy that can come only from doing battle with other firms on an equal playing field, the competition that affects characteristics that matter. Instead, Facebook competed on one characteristic alone — the ability to build the largest network first — and used that high ground to defeat a more tech-savvy competitor.
That’s a recipe for the long-term decline of American social media, and of American tech savvy more generally.
The Antitrust Case against Facebook’s Treatment of Vine
Facebook’s killing of Vine should be the easiest of antitrust violations to prove, but instead the case can be made only through the luckiest of coincidences. Luck is needed because of the current ascendancy of the Colgate Doctrine: the right of any business to refuse to deal, even if that would create a monopoly.
Under the influence of economists and lawyers associated with the Chicago School, the courts have all but eliminated any liability for refusal to deal, allowing it only when the refusal represents the termination of a prior profitable course of dealing. The idea behind narrowing liability to this unusual set of facts is that only when the refusal to deal amounts to a choice to forego a current profitable relationship can enforcers be absolutely certain that the motivation for the refusal is to earn even greater profits from the destruction of competition. As Justice Scalia put it in that 2007 case,
The unilateral termination of a voluntary (and thus presumably profitable) course of dealing suggest[s] a willingness to forsake short-term profits to achieve an anticompetitive end.
Motivation should have no place in the resolution of antitrust cases, because the antitrust laws are not about policing morality, but about guaranteeing the vigor of the economy. What matters in antitrust are outcomes, not whether businesspeople act with virtuous or heinous intent. But this perversion of the law is of no consequence in the case of Facebook’s treatment of Vine. Miraculously, the question of Facebook’s motivation is subject to no doubt here because the British have provided us with emails pregnant with anticompetitive intent:
Justin Osofsky — Twitter launched Vine today which lets you shoot multiple short video segments to make one single, 6-second video. As part of their NUX, you can find friends via FB. Unless anyone raises objections, we will shut down their friends API access today. We’ve prepared reactive PR, and I will let Jana know our decision.
MZ – “Yup, go for it.”
But even if there were no such evidence of intent, Facebook’s actions meet the prior profitable course of dealing standard imposed today by the courts.
Facebook’s sharing of friend lists with Vine allowed Facebook to collect valuable data about which Facebook users were using Vine. Facebook’s termination of that sharing therefore represented the termination of a prior profitable — in data-denominated terms — course of dealing, the unmistakable sign the court demands that the motivation was to earn even greater profits — here in the form of monopoly-level access to users’ social networking data — that come from squelching competition in the market.
So as luck would have it the case against Facebook fits squarely within the sliver of an exception to the Colgate Doctrine currently tolerated by the courts. But the fact that we need to fit the case into that sliver tells much about the extent to which antitrust has been failing in recent decades in its duty to ensure level competitive playing fields.
Integrating into Espionage
The situation is even worse when it comes to Facebook’s snooping on, and eventual gobbling up of, WhatsApp.
The story of global merger enforcers’ disastrous failure to block the WhatsApp acquisition due to a failure to appreciate that consumers pay for both Facebook and WhatsApp in data, making the two companies rivals, and the merger the brazen elimination of a nascent competitor, has already been told. But Onavo’s role in helping Facebook identify WhatsApp for acquisition points to another failure in contemporary antitrust: the death of vertical merger enforcement.
Onavo’s app is properly understood as a component of the social media product offered by Facebook , one that includes not just liking and photo sharing, but also privacy services. As such, Onavo and Facebook stood in what antitrust lawyers call a “vertical” or supply-chain relationship, producing components of a common end product — the social media experience — that is sold to consumers. And Facebook’s acquisition of Onavo was therefore a vertical merger.
But in the 1980s antitrust enforcers abandoned vertical merger enforcement entirely, on the assumption that innovation and efficiency always result when businesses in a vertical relationship work together to serve consumers. The district court’s stinging and misguided rebuke of the Justice Department’s recent attempt to revive vertical merger enforcement by challenging AT&T’s acquisition of TimeWarner shows how alien the old learning regarding the threat of vertical mergers has become to the courts in recent decades.
There might well have been some synergies between Onavo’s analytics services and Facebook’s social media platform, but the role the acquisition played in enabling anticompetitive snooping makes clear that the dogma that vertical mergers are always good for the economy must go.
In everyday life economics presents itself to us first and foremost as a problem of distribution, not of output, not of what economists would call “allocative efficiency.” This pizza place is ripping you off. That employer is lowballing you. In our intuitive economics, higher prices mean someone is taking money out of our pockets and putting it in theirs. And low prices mean we are taking money from someone else’s pockets and putting it in ours.
Only once we’ve considered these distributive consequences do we go on to consider the effect of pricing on output, and even then we usually do so only in the context of bargaining over distributive outcomes. Thus we might say: “If he doesn’t lower the price of a slice, I’m going elsewhere.” Now, that’s an effect on output: the high prices cause you to buy less. But even when we make threats like that, we think of them as a bargaining position: we threaten to go elsewhere, so that the price will go down, and we can therefore be richer.
Yes, maybe we go elsewhere because we can’t afford it, but then we rue the fact that we are not richer — that more wealth has not been distributed to us — so that we would be able to buy. We don’t think to ourselves: “this price tells me that I don’t care as much about this product as the person who produced it, and therefore it is right and proper that I not buy this product and buy something less expensive instead.” That would be a view focused on allocative efficiency. We think instead that we should be given more money, or should find a way to get it ourselves.
Our way of thinking about prices and markets is through and through about the distribution of wealth. But introductory economics takes almost no account of this. Instead, economists introduce their subject primary by reference to competition, not monopoly. (The wildly popular Varian Intermediate Microeconomics textbook, for example, teaches competition before monopoly.)
But theories of competition are not about distribution at all.
In competitive markets, there is no give whatsoever in price. Price is uniquely determined by supply and demand to be just high enough to cover the cost of production and just low enough so that everyone who is willing to pay the cost of production, but not a penny less, is able to buy. As a result, there is no distributive question in competitive markets. Sellers never make any profit, because they sell at cost. And buyers never get away with good deals, because they always pay the maximum that they would be willing to pay for the good.
Indeed, in a world of competitive markets, you cannot, must not, ever think of prices as being too high or too low. If the pizzeria raises prices, it must be because the costs of production have gone up. If your employer reduces your wages, it must be because the value of your labor to the employer has gone down. The pizzeria is not trying to redistribute wealth to itself, but only responding to changes in the value that society places on the ingredients of pizza — cheese, tomatoes, and labor. Society values some of these more than before, which is why their costs have gone up, and that in turn is why the price of your pizza must go up. If you can’t afford the new prices, then you must not care as much about pizza as society cares for the cheese, labor, and tomatoes, and so it is right and proper that you not buy. To complain about the higher prices is not just foolhardy — nothing you say or do can bring those prices down, unless you are able somehow to change the value that society places on those ingredients — complaining is also evil, because it amounts to demanding that you be given something that you don’t deserve. If the pizzeria were to listen to your blandishments and give you a discount, then someone who values the ingredients of pizza more than you, someone who is willing to pay more for them than you, will not be able to buy them, and society as a whole will be worse off, because it will have taken things from someone who values them more and given them to someone who values them less. To hold out for a better deal is to try to pervert a mechanism that, if left to its own devices, ensures the greatest good for the greatest number.
It is my sense that the main reason for which economics seems to turn off the students with the greatest interest in the field is that these students can find in theories of competitive markets no shred of the distributive intuition with which they are familiar. Moreover, the theory of competitive markets destroys the motivation to study economics of anyone who actually cares about economic outcomes. Because the theory of competition suggests that the economy is a machine, and not a social endeavor at all, which in turn suggests that non-interference is the proper way to interact with the economy. Do not complain about prices. Do not hold out. Do not bargain. Just accept the prices and wages that you are given, and make sure that everyone else does too, and you can be confident that the prices are just and the wages are just. But we do not become interested in, and study, things that we do not wish to interfere with. No. We become interested in, and study, things that we want to tweak, to improve, and so on.
So those who are most interested in economics come to economics (1) with a thirst to understand how economics drives distributive outcomes and (2) with a thirst to understand what can be done to improve those outcomes. Introductory economics courses tell these students: (1) economics poses no distributive questions at all and (2) to the extent that there are any problems with the economy, they can be solved only by doing nothing. No wonder that few economists seem really passionate about their subject. The way introductory economics is taught alienates all those students who are really passionate about the subject.
That is a shame, because the economic concept of monopoly speaks to all the frustrations of those who really care about economics. It’s a crime that the concept gets so little coverage in introductory economics courses, a crime all the more serious because not only do economists of all stripes agree — at least when pushed — that the monopoly concept is far more widely applicable to economic life than the competition concept (think product differentiation and monopolistic competition), but the monopoly concept is also the far older and more established of the two concepts in economics, stretching right back to Ricardo and beyond (Adam Smith talked about both competition and monopoly, but he didn’t have models for either). If there is one piece of good that might be done for the world in matters economic, it might well be to henceforth start all economics textbooks with monopoly, and leave competition for the last chapters, the way monopoly is left to the end today.
(This extends, as well, to the teaching of introductory game theory. The game theoretic counterpart of the monopoly concept is bargaining, also called cooperative game theory, because bargaining is all about the distribution of surplus, and not about immutable equilibrium results. Why, then, is it that Dixit only gets to bargaining in the final chapter of his seventeen-chapter introduction to the field, Games of Strategy? Do not tell me that it is because the mathematics of bargaining is harder. It’s only harder because distributive questions are underdetermined — any distribution is possible — which means mathematics isn’t all that useful for distributive questions to begin with. Distributive questions are political, not mathematical, questions. The responsibility of economics is to study the economy, and to emphasize the most important aspects of the economy, not to emphasize the aspects that are mathematically tractable. Economists ought to lose named chairs if they do otherwise.)
Why does the concept of monopoly speak to our basic distributive intuition about economics? Because the monopoly concepts admits that people can and do choose their prices, that the supernatural forces of the market do not choose prices for them, and therefore that the distribution of wealth matters. The pizzeria can choose higher prices, and in so doing can extract more wealth from us. And our employers can pay us lower wages, and deny more wealth to us. And when the pizzeria or the employer acts this way, the monopoly concept tells us that the prices the pizzeria charges and the wages the employer pays are not necessarily an accurate reflection of the cost of making pizza or the value of our labor. If the pizzeria raises prices, it might be because costs have gone up, but it might also be — indeed, it is most likely to be, if we accept that in a world of differentiated products every firm has some amount of power over price — because the pizzeria has decided to try to extract more wealth from us in exchange for providing a good that hasn’t changed in value at all. And now, knowing that prices can rise even when value does not, it makes sense that our instinct is immediately to bargain, to walk away, to hold out, in order to drive that price down. Indeed, knowing that prices can rise even when value does not, it makes sense now that we intuitively view all of our economic interactions first and foremost in distributive terms, because economic interactions are first and foremost about distribution. Any creature foolish enough to just intuitively accept prices as dictated was long ago flushed from the gene pool, having failed to reproduce, because the creature spent all its money on pizza, or failed to bargain for a higher wage.
The monopoly concept tells us that to complain about prices is not to threaten to upset a well-oiled machine, not to try to take from others what they value more, but to insist on a share of the productive pie. The pizzeria creates value — surplus — over the cost of production when the pizzeria makes pies. Why? Because the pies are worth more to us than cheese, tomatoes, and labor separately. That surplus is expressed in the maximum prices that we are willing to pay for pizza, maximum prices that exceed the cost of cheese, tomatoes, and labor. What to do with this surplus of pleasure (our pleasure) over cost, of this surplus of our pleasure over the pleasures foregone by those who would otherwise use the cheese and tomatoes, or spend their labor time on other pursuits? If the pizzeria charges a high price — a price equal to our maximum willingness to pay — then we give that surplus to the pizzeria. If the pizzeria chooses to pay a higher wage than the minimum that its workers are willing to accept, then the pizzeria passes some of that surplus along to its workers. Otherwise, the surplus goes to the owners. And if the pizzeria charges a low price, a price equal just to its cost of cheese, labor, and tomatoes, then we, pizza eaters, get all of the surplus. This is real economics, the economics of the everyday. It’s also a fundamental part of virtually all economic research today, but you wouldn’t know it from taking an introductory economics class. Persevere.
In the world of the monopoly concept, economics is inexorably political, there is always something to be done to improve the system, and doing nothing means catastrophe. Distribution is politics. Should the workers take more of the surplus? Should consumers? Are prices too high? Too low? Something must be done. The monopoly concept takes you straightaway to action.
There is much that can be done. You can lower prices by promoting competition in markets. Is it hard to find another pizza place charging a lower price, which is why your local pizzeria feels comfortable raising prices? Why, then, is it hard to find competitors? Are they all owned by the same firm, despite different branding? Do they use WhatsApp to fix prices? Are there just too few of them? To solve these problems, you need to advocate for greater antitrust enforcement.
But you can also lower prices through “rate regulation” — government setting of prices. This is a lot more common than you might think. States regulate the prices charged by power, gas, and water companies, among others. And have regulated prices in many more industries over the past century or so. Indeed, regulatory economics and auction theory are devoted to little else than finding low-cost ways for government to set prices (albeit with less attention to distributive concerns than there should be).
There are yet other ways to deal with prices. Thee is the political harangue. Studies show that when presidents complain about high drug prices, prices fall. There is also the reallocation of property rights. Indeed, the problem of pricing strikes deeper than antitrust, with its focus on anticompetitive conduct, deeper even than rate regulation, with its focus on dictating prices, and instead goes to the heart of our legal system, and in particular the part of it dedicated to property. The home you own monopolizes a bit of space in the world, and that is why, when that space is in great demand, homeowners can become enormously wealthy, because they exploit that monopoly to charge the highest possible prices to buyers. Maybe that’s a good thing if the homeowner started out with modest means. But the point is that property is monopoly (a thing that was said long before Eric Posner and Glen Weyl called attention to it in a recent book). One way to drive prices down, or more generally to redistribute wealth, is to divide up or redistribute property. And that bring us to tax policy, which strives to deprive you of the fruit of your property-based monopolies, albeit without tying to deny you title to them. Property tax makes you pay some of that higher price you can claim back to the state. Income tax makes you pay some of the higher wages you can charge due to your monopoly over your talents back to the state. Sales tax does almost the same thing as rate regulation, making the pizzeria pay some of the monopoly profits it earns from you back to the state (don’t be fooled by the way stores fight the tax politically by adding the tax onto your bill — the bill would be higher without the tax). Trust getting busted, price being set, Presidents complaining, property being expropriated, and everything being taxed. You don’t get any more political and interventionist than that.
Moreover, unlike in competition theory, in monopoly theory the alternative of doing nothing about prices is just not any option, unless you believe that owners should always enjoy all of the surplus generated by production. Why? Because all markets tend naturally toward monopoly. Firms acquire firms, and run others out of business. Over time the result is monopoly everywhere, and owners that charge consumers the highest possible prices, and redistribute nothing to their workers. Unless government does something, whether to promote competition through the antitrust laws or property reform, or to regulate prices directly as part of a rate regulatory regime, markets will tend naturally to allocate all wealth to business owners. We are very far now from the competitive world view in which any sort of intervention in the economy misallocates resources.Of course, the two models — competition and monopoly — do work together. If you drive prices too low, or wages too high, then the allocative effects that competition theory worries about start to kick in. The pizzeria does have costs, and if you drive price below costs, then you really will start taking cheese, tomatoes, and labor from those who value these things more highly than you do. But it’s a good rule of thumb to assume that the initial change in prices is a distributive change, not a competitive — or what economists would call an “efficiency”-driven change. Distribution comes first, which is why most of us intuitively think about it first when we consider economic issues, and why it ought to be taught first, too.
Indeed, I have come to believe that most of the differences I have with economists are attributable to the difference between the competitive and monopoly world views. Whether they admit it or not, all economists walk the earth with a default economic model in their heads, parsing and interpreting economic facts in the first instance through that model. And very often, precisely because they went through introductory economics courses that emphasized competition, economists’ default model is the competitive model, not the monopoly model. It’s fine to have a default model — all science has to start from priors — but for the reasons that I gave above, that economic reality is characterized in the first instance by pervasive monopoly, even if only of the differentiated product monopolistic competition variety, the default model that economists out to be carrying around with them is the monopoly model. But in fact economists don’t usually have the monopoly model in mind in the first instance, and sometimes they seem to have forgotten about it entirely when pressed on public policy matters, even though in their own advanced technical work they may use nothing but monopoly models themselves.
I struggled and struggled with his comments, trying to understand what I might be missing, until I realized that he was thinking in purely competitive terms. In his world, there simply was no surplus! Reduce the price of a pizza, and either someone gets something they don’t deserve, or — his point — the pizzeria will just produce less pizza, because any reduction in price must drive price below cost, making some production unprofitable. The notion that a price drop might have no effect on output was entirely alien to him. And yet at that very conference he had himself given a paper about monopoly in the finance sector! It wasn’t that he was somehow unaware of the monopoly concept, but that it wasn’t the default model that he useds when talking about policy issues.
Such are the wages of failing to emphasize the monopoly concept in economics education.
Baer kicked off his remarks by stating that antitrust’s consumer welfare standard — a target of OMI, and much discussed on panels earlier that day in Cambridge — should stay, because it’s the only administrable standard available to antitrust.
To see why the consumer welfare standard is hard to apply, consider the merger of AT&T and TimeWarner. Let us suppose that the merger would lead to reduced costs (because of the elimination of what economists call double marginalization), some improvements in program quality, because, for example, the combined firm can use viewing data to tailor content, and some increased market power, because TimeWarner can now raise prices to other content distributors safe in the knowledge that if negotiations fail and a blackout ensues, TimeWarner will still be able to continue to supply content to AT&T viewers.
The increase in market power suggests that consumers would be harmed by the deal, but whether that actually happens depends on whether either of two escape valves opens. First, cost reductions associated with the merger could make consumers better off, even after market power effects are taken into account, if some portion of those cost reductions are passed on to consumers in the form of lower, though still monopoly-power-inflated, prices. Second, even if any cost reductions are not passed on to consumers, the improvement in programming quality might itself ultimately make consumers better off, if the improvement is sufficiently large to offset any increase in prices. Given the existence of these two escape valves, determining whether consumers are harmed by the merger requires enforcers to predict the price effects of the merger, along with the dollar value of the improvements in programming quality brought about by the merger, and to compare the difference between the two, known as “consumer surplus”, with the original pre-merger difference between price and programming value to consumers.
That’s hard, because quantifying the value of programming to consumers requires enforcers to deduce the maximum prices that consumers would be willing to pay for the programming, rather than the real prices that consumers actually are paying.
Indeed, measuring consumer welfare is so hard that in practice enforcers don’t even try to measure it, the law be damned. Instead, they just test to see whether the merger will raise prices, pretending that price increases are a good proxy for consumer harm, which of course they are not. If the value of the product to consumers rises by more than prices, for example, then consumers benefit from the merger. By the same token, a merger could drive down quality — perhaps the union of AT&T and TimeWarner would unleash targeted advertising that actually reduces program quality, for example — to such an extent that consumers would end up worse off from the merger even if the merging firms were to share some of their cost savings with consumers by lowering prices.
Enforcers don’t try to measure consumer welfare because they can’t. And that tells us something important about whether the consumer welfare standard is as administrable as Baer says that it is: namely, that it isn’t administrable at all. Precisely because it is not clear in any case whether consumers are harmed, antitrust enforcers look to see whether prices would rise instead, since prices, thank goodness, are actually observable. Ostriches can relate.
In fact, enforcers don’t even proxy consumer welfare effects by looking exclusively at prices. Instead, they try to distinguish price effects unrelated to anticompetitive conduct, such as price hikes driven by higher energy prices, or other “exogenous” factors, and price effects that are attributable to the vigor of competition in the market. As I indicated in my earlier post on the CPI/CCIA conference, such an inquiry into what might be called “abnormal price effects” is really an inquiry into profit margins — increases in prices that are not driven by increases in costs.
And here is where the irony, and not just the falsity, of the claim that the consumer welfare standard is the only administrable antitrust standard shines forth. For the rule that antitrust should condemn anticompetitive conduct that increases profit margins is actually the old standard that the consumer welfare standard was fashioned to replace, the very standard in comparison to which the consumer welfare standard is supposed to be an improvement in administrability, practicality, clarity. The covert inquiry into profit margins that enforcers understand when they are supposed to be testing for consumer harm is nothing but the standard of the mid-20th-century golden age of antitrust. That standard prohibited all anticompetitive conduct, regardless whether the conduct harmed consumers or not, so long as that conduct could be expected to lead to, or protect, market power, defined as the power to earn abnormally high profit margins. The supreme inadministrability of the consumer welfare standard is actually expressed in the fact that enforcers don’t even follow that standard as a technical matter, but still follow the old standard that it was supposed to replace.
But if antitrust is still doing what it has always done in testing for abnormal profits, what explains the remarkable declines in antitrust enforcement since the Chicago School shifted antitrust to the consumer welfare standard in the late 1970s? The answer is that Chicago did not just change the standard on paper from harm to competition to harm to consumers, but also changed the burden of proof required to meet any standard. Thus while enforcers have continued covertly to apply the old standard — which looks at profit margins, not consumer welfare — they have done so with a level of skepticism about their own ability to identify increases in margins that did not exist before the triumph of the Chicago School in the 1970s.
To the extent that this skepticism is warranted, the consumer welfare standard is perhaps no more administrable than the margins alternative. But to the extent that the skepticism is not warranted, the consumer welfare standard is less administrable than the margins alternative. The fact that enforcers have sought in the measurement of profit margins a refuge from the challenge of measuring consumer welfare certainly suggests that margins are easier to measure, and that the consumer welfare standard is the less administrable standard. Either way, the consumer welfare standard is not more administrable than the profit margins alternative that came before it.
Another way to see this is to consider the role of the consumer welfare standard in basic antitrust doctrine. Before 1975, antitrust had two kinds of legal tests. The first, called the per se rule, condemned certain kinds of anticompetitive conduct full stop. The second, called the rule of reason, prohibited anticompetitive conduct by firms possessing, or acquiring through anticompetitive conduct, market power, understood to mean the ability to earn abnormal profits. The focus of the rule of reason on actually proving margins did not imply the unimportance of margins to the per se rule, only the willingness of enforcers to invest more time in proving margins in some cases (rule of reason cases) than in others (per se cases), for which latter it was hoped that proof of anticompetitive conduct alone would be sufficient to signal the existence of abnormal profit margins, at least on average.
Comes now the consumer welfare standard in the 1970s, which appears in the doctrine as an additional element required to meet the rule of reason test. Under that new rule of reason, three things were now required: (1) anticompetitive conduct, (2) market power, and now (3) consumer harm. Thus the consumer welfare standard created a compound test, one that requires both proof of abnormal margins and proof of harm to consumers.
But doing two things is not easier than doing just one of those things. The consumer welfare standard does not make it easier to do antitrust, but harder.
I put this point to the panel in Cambridge, but received only affirmations of faith in reply from several panelists, including former FTC chairs Jon Leibovitz and Bill Kovacic. Why does the consumer welfare standard seem to so many — and not just Baer — to be a practical standard? Why, because it’s an empyrean, an ideal, a beautiful but unobtainable thing. And we mistake the clarity of the vision for clarity of practice.
(Don’t the consumer welfare and market power (profit margins) elements in the new rule of reason test collapse into the same thing? No, for the same reason that consumer welfare can’t be proxied by price effects. Suppose that market power does allow AT&T and TimeWarner to raise prices after the merger, but also increases the value of programming to consumers by a greater amount. Consumer welfare increases, but margins also rise. Under the old rule of reason, which only looked at market power (profit margins), there is antitrust liability, but not under the new rule of reason, with its requirement of harm to consumers.)
The Challenges to Antitrust in a Changing Economy conference, put on by CPI and CCIA at Harvard Law School two weeks ago, was an opportunity for today’s antitrust establishment, on both the (center) left and right, to react to recent calls from activists and journalists loosely associated with the Open Markets Institute for a radical increase in antitrust enforcement. In particular, the conference provided a view of how establishment scholars have been processing OMI’s extraordinary influence on progressive thinking, not to mention the national press, over the past couple of years. (I don’t mean “establishment” pejoratively here, but only to signal that these are leading scholars in antitrust law and economics teaching at leading schools.)
The most serious challenge to the antitrust status quo as an intellectual matter has interestingly come not from OMI, but from finance economists, who have shown in recent years that firm margins, which are the difference between revenues and costs, have experienced an abnormal expansion over the past two decades or so, a period that corresponds uncannily to the period over which antitrust enforcement has been in decline. Margins are the profits of common parlance, and the implication of this work is that firms are generating greater profits than they ever could before, and have been doing it both in periods of recession — such as the Great Recession of 2007 — and in periods of economic expansion, such as that taking place right now.
These scholars — top flight economists all — have shown that none of the variables you might think would account for increased profitability, such as increased investment in new technologies, explain this trend. The explanation that leaps out, one that these scholars have not been able to explain away with their data, is that firms have been leveraging the greater market power permitted to them by declines in antitrust enforcement to extract more profits from markets.
Speakers at the CPI/CCIA conference two weeks ago pushed back against the evidence both of rising margins and of rising concentration. NYU’s Larry White kicked off the day with an attack on the margins evidence. He argued that the finance economists are missing something important that the industrial organization (IO) economists who traditionally have taken the lead in antitrust policy debates learned in the 1970s: namely, that margins can’t be measured.
The trouble, argued White, is that costs are difficult to define. Subtract away the costs of all physical inputs, compensation to workers, and the like, and you still might not end up with an accurate measure of margins, because some of the remaining amount may be necessary — necessary in the way that all costs are necessary to production — to serve as a cushion against an unexpected shock to revenues. Or to compensate innovators, or managers with special skills, and so on.
Invoking noted mid-20th-century IO economist Leonard Weiss, who was long an advocate of greater antitrust enforcement, White pointed out that it was Weiss who in the 1970s finally came to recognize that margin data were unreliable, and concluded that going forward antitrust policy would need to be based on observation of price effects, rather that margin effects. White’s point was that absent an accurate way to measure margins, antitrust policy must make do with looking to see whether prices, rather than margins, are rising abnormally in the economy. And prices, notably, have not been going up abnormally, creating no basis for increased antitrust enforcement. Finance economists, argued White, weren’t around for the bruising quarter-century-long quest to measure margins and relate them to concentration levels that took place in industrial organization economics during antitrust’s Postwar golden age, and therefore are making the same mistakes today that IO economists once made.
The trouble with White’s argument is that it proves too much, because antitrust is through and through dedicated to the measurement and prohibition of anticompetitively-generated margins, whether antitrust is willing to admit it or not. So giving up on the measurement of margins means giving up on antitrust. White himself seemed inadvertently to advertise this point at the end of his presentation. In the final portion of his remarks, White observed that the margins problem rears its head in antitrust today whenever the courts require proof of market power, because market power is the power profitably to raise price above competitive levels, and profits are margins. But precisely because the requirement of proof of market power is ubiquitous in antitrust law — a staple of the “rule of reason” standard applied to both collusion claims under Section 1 of the Sherman Act and monopolization claims under Section 2 — White’s skepticism about the possibility of measuring margins translates into skepticism about the entire antitrust project. Take White’s position seriously, and there should not only be no radical increase in antitrust enforcement, but no antitrust at all.
White likely didn’t see that implication because he believes, as much of the antitrust establishment seems to believe today, that it is possible somehow to use price effects as a substitute for margin effects in deciding whether firms have power over markets. The argument for using price effects goes like this. Instead of trying to identify markets in which price increases are profitable, and then to scrutinize the behavior of firms in those markets to make sure that they are not profiting by actively squelching competition, antitrust enforcers need only look to see whether suspect firms could increase prices over competitive levels in any of the goods they sell. If prices could go up, and statistical analysis shows that the increase would not be due to irrelevant factors such as an increase in input costs, then it is safe to assume that the increase in prices would be due to the anticompetitive conduct. The apparent beauty of this approach is that there is no need to measure margins.
Or is there? What antitrust economists all ought to know, but perhaps don’t want to admit to themselves, is that when they consider price effects they are always also implicitly measuring margins. How? When they control for changes in input prices, of course. Price effects can have many causes, and antitrust is not a price stability regime. Antitrust wants to condemn conduct — like horizontal mergers — that leads to higher prices only when those higher prices are a result of anticompetitive conduct, and not the result of increases in costs. But that just puts any student of price effects in the position of having to distinguish between price effects that are driven by higher margins — the channel through which all anticompetitive conduct affects prices — and price effects that are driven by costs or other extraneous factors. When an econometrician controls for input cost increases, the econometrician is really just measuring margins, implicitly using a metric that expresses margins as revenues less input costs. (The funny thing is that this simple approach to margins is precisely the one that White, and the Chicago School in the 1970s, so roundly criticized the earlier Postwar establishment for employing.)
In other words, margins in antitrust are everywhere, and unavoidable. Indeed, you cannot have antitrust without the measurement of margins, because anticompetitive conduct is uniquely identifiable through the abnormal margins that the conduct makes possible. Anticompetitive conduct that does not increase margins simply is not anticompetitive. Conduct must somehow fail to squelch competition, and therefore fail to enable the firm to extract more value from consumers, in order to be anticompetitive.
Of course, the reverse is not true, higher margins can be caused by factors other than anticompetitive conduct, but that does not permit antitrust to ignore margin effects; the subject of antitrust is precisely margins caused by certain types of conduct. To give up on the ability of economics to measure margins is to give up on antitrust. Despite declines in enforcement since the late 1970s, today’s antitrust establishment has been unwilling to give up on antitrust, and it has dealt with the immense cognitive dissonance associated with practicing a discipline that it believes impossible to practice by using the classic cognitive strategies of denial and avoidance. The establishment today acts as if the show can go on without the measurement of margins, which of course it cannot.
I put this problem to the panel, and the responses were highly instructive. Bruce Kobayashi, current head of the FTC’s Bureau of Economics, stated that “everything” the Bureau does involves the measurement of margins. Antitrust cannot function without it. And White, to his credit, threw up his hands, seemingly agreeing that if the measurement of margins really is impossible, then there can be no antitrust enterprise.
In a way, this debate cuts right to the heart of antitrust’s agony of the past forty years. Until the mid-1970s, antitrust enforcement in the U.S. was vigorous. The Chicago School attack that lowered enforcement was based primarily on radical skepticism about the ability of economic science to identify truly anticompetitive conduct, and that skepticism was in turn expressed in a skepticism about the ability of economics to measure margins. Perhaps finance economists will drive renewed faith in the power of economics to engage in such measurement, but even if they don’t, we need to come to terms with the fact that the actions of IO economists have already spoken louder than their words. In continuing to muddle along measuring margins while professing not to be able to measure them, IO economists have been telling us for the last thirty years that yes, you can measure margins, and run an entire policy sector based on them.
Recognizing that fact may be all we need to cure antitrust of its present timidity.
Some thoughts on one of Brian L. Frye’s Ipse Dixits, devoted to the art market. In the podcast, Tim Schneider explains that high-end galleries won’t even quote prices to wealthy buyers unless the buyers have standing in the social network that is the art market, because who buys now affects the prices that galleries and their artists can demand in the future. In fact, Tim observes, buyers who have the highest art market status often pay the lowest prices for art, because the implicit endorsement created by a purchase is so important to galleries.
The Peculiarity of the Art Market
Brian suggests that there is something odd about this situation, because in a well-functioning market you would typically want price fully to reflect the value that is being exchanged. But I wonder if this is just an illusion created by the way prices are charged in the art market. Price there is denominated in multiple currencies, one being dollars, and the other being prestige. Imagine that a high-profile buyer were to tell a gallery in advance that the buyer will announce at the time of purchase that the buyer believes the artwork to be horrifically bad, that the artist is trash, the gallery worse, and the buyer will liquidate the object itself posthaste upon taking title.
To the extent that this would diminish the prestige of the sale, you would expect the gallery to raise its price, and whatever that higher price would be, less the original prestige-weighted price, would be the dollar value of the prestige itself that the buyer would normally pay for the artwork along with the work’s dollar-denominated sticker price. If the gallery charges $1 million for the work, but would raise that to $1.5 million were the buyer to declare the work trash, then the value of the prestige earned by the sale of the work to the buyer absent the trash talk is $500,000.
Regardless whether the artwork is sold for a price denominated in two currencies ($1 million plus prestige) or one ($1.5 million), however, the artwork is still being sold for a price, and that price represents the total value of the work. The price of the work is always $1.5 million whether $500,000 of that price is paid in prestige or in dollars. The fact that galleries vary their dollar-denominated sticker prices based on buyer identity suggests that they know the prestige value that each buyer can convey and alter their sticker prices accordingly to maintain a constant overall (dollar plus prestige) price. That overall price is doing the job that prices are supposed to do, by representing the total value of the exchange.
One problem with this line of thought, however, is that Tim suggests that there are plenty of wealthy buyers out there who could afford to pay just about any price for art, but can’t even obtain a price quote, let alone a sale, from a high-end gallery unless these buyers belong to the art market network. If dollar-denominated sticker prices for art just represent the true price of the art less the dollar value of the prestige generated by the identity of the buyer, and non-members of the art market network simply have no prestige, then you would expect galleries to be willing to quote prices to them, albeit higher prices that reflect the absence of a prestige offset. That sort of thing happens all the time in consumer markets — a brand might charge one price to the average consumer but give a discount to a celebrity, because association of the product with the celebrity makes it easier to sell the product to others.
The fact that galleries won’t even quote prices to average customers can therefore mean only one thing: that the prestige offset is so high that no one can pay the true, full, prestige-value-inclusive, dollar-denominated price of high-end art. The losses to the galleries from selling to just anyone are so large that no buyer, however wealthy, would be able to compensate the galleries for what they would lose from democratizing the art market.
How can that possibly be so? (And at this point Brian’s intuition that there is something strange about the art market starts to make sense to me.)
Perhaps galleries don’t want just any buyer because modern art is junk, valuable only for the social significance of ownership, and not because it conveys any consumption value at all to its owners in the form of enlightenment, edification, catharsis, or what have you, and in this sense is the purest of Veblen goods. Modern art is junk because anything can be modern art (a point that modern artists admirably concede when they are not, Jeckyll-and-Hyde-like, running about pretending to see greatness in some spatters of paint but not others), and therefore the supply of modern art is infinite. Economists are fond of saying that scarcity is an iron law (there’s no free lunch), but in point of fact modern art, in virtue of the fact that everything counts, is the quintessential unscarce good, the only free lunch, the cornucopia, the bounty unique.
To put a finer point on it, there are about 10 to the 80 atoms in the visible universe. If anything can be modern art, then any of these atoms, plus any combination thereof, can be modern art, which is to say that the available supply of modern art in the visible universe is the power set of all the atoms in the visible universe, or 2 to the power of 10 to the power of 80. The reader might object that surely a hydrogen atom in some distant corner of the universe could never be modern art, or could never be bought or sold as modern art. And perhaps this is true, if we at least require that art be objects here and now on earth. But there are about 10 to the 34 atoms on the surface of the earth, and the power set thereof even larger, and the bounty no less mind-bogglingly great.
That modern art is junk and unlimited in supply explains why galleries cannot sell to just anyone, because things in unlimited supply have a market price of zero. It follows that in order for modern art to have a non-zero price, some method must be found to limit supply. When a firm like DeBeers wants to limit the supply of diamonds, it simply keeps them in the ground, or warehouses them. Supply can here be limited because supply is in a sense already limited. Diamonds are plentiful, but they can in fact run out, and so if DeBeers can own all the diamonds in their plenty, DeBeers can restrict access and in this way make diamonds genuinely scarce. Not so for modern art, however, because the supply of modern art is quite unlimited. To do to modern art what DeBeers does to diamonds, the galleries would need to buy up all the atoms in the universe (or at least on the surface of the earth) and horde them, which of course the galleries can never do.
If the standard method of limiting supply to support price — hording — cannot work for the galleries, then what? Another approach would be to impose standards of quality on art. But in the debased cultural environment in which we live today, there is no agreement about what constitutes good art and bad art. One man’s David is another’s toilet bowl, and vice versa. Were the galleries to impose standards, they would immediately become the old guard, the target of a million culture warriors, and their inventories not long after consigned to the fire sales of bankruptcy.
No. The only solution for the galleries was to tie art to the only thing in its universe that in the modern age remains in truly fixed supply, and that is prestige. Sell membership in an exclusive club. Make the ability to keep up with a constantly changing art market the ultimate secret handshake. Make the art market a luxury market — no different from the market for fancy handbags or sportscars — only more so, governed not just by wealth but by a willingness to adhere to a whole set of social rules. Make it the aristocracy to Prada’s bourgeoisie.
It should come as no surprise that one of the social rules that distinguish the art market from mere luxury markets, a rule Tim describes, is that art buyers must resell art only through the high-end galleries, rather than on secondary markets. Reselling on the secondary market gives any member of the club the right to admit new members, and that is a recipe for disaster, because it increases membership and allows in buyers not vetted by the galleries, buyers who may be equally willing to violate the rules of membership, accelerating damage to the club. Membership increases make the club less elite, which in turn reduces the value of membership to all other members, making each less willing to pay the high prices for art that are a necessary, though not sufficient, condition for membership.
In the rule against resale on the secondary market one sees particularly starkly that what is sold in the art market is prestige and not art. If what is sold were art, then expanding the market — allowing more people to bid on art — would be desirable for the galleries, because increases in demand increase profits, all else equal. But prestige behaves like a commons — let more people in and everyone suffers — which is why the number one club rule must be to leave it to some governing authority to police access. The galleries are that authority.
This brings me to the antitrust question that Brian poses in the podcast: Is there something wrong with the unwillingness of galleries to do business with buyers who violate club rules by reselling on the secondary market? If what the galleries are selling is art, rather than prestige, then there might be an antitrust case to be made against the galleries, but only if one of two crucial conditions holds.
Either any one art gallery refusing to do business with a reselling buyer must have at least a 75% share or so of the art market (or, more generally, power to profitably raise art prices above the levels that other galleries can charge in the art market). Or there must be some at least circumstantial evidence that galleries have explicitly agreed among themselves not to do business with resellers, and the group of explicitly colluding galleries must collectively have at least a 35% or so market share in the art market (or, more generally, some at least very weak power profitably to raise price). If the former holds — the gallery has a 75% market share or substantial power over price — then a claim for monopolization under Section 2 of the Sherman Act would lie. If the latter holds — a group of galleries with a 35% market share or some at least small amount of power over price have agreed not to do business with resellers — then a claim for concerted refusal to deal under Section 1 of the Sherman Act would lie.
The key to both claims would be recognizing that the buyer becomes a competitor of the galleries when the buyer starts to resell artwork. I have made much of this competitor status of resellers in the context of data-driven price discrimination, although it should be noted that the antitrust case against the art market could not successfully turn on a charge of price discrimination. Price discrimination is emphatically legal under current law in antitrust, unless it consist of volume discounts, something that appears hardly to be the rage in the art market. And even then, there has been almost no enforcement of cases of that kind since the early 1990s. The art market reseller is a competitor, but the antitrust case against the galleries is not that the galleries punish the reseller because the reseller is selling at a discount to buyers who might otherwise receive discriminatory prices from the galleries, since such discriminatory pricing is legal outside of the data-driven pricing context. Rather, the antitrust case is that the galleries punish the reseller because the reseller is competing the price of all art down by democratizing access to the art club.
Harm to consumers is also required by these claims, and here the antitrust case becomes potentially unwinnable. At first glance there does appear to be consumer harm, because all the rich social outcasts suffer from not being able to buy in to the club. But at second glance the harm falls away. For the value of the product is tied to membership, and therefore any attempts by the galleries to exclude competitors who want to undermine membership genuinely count as attempts to maintain the value of the product not just to the galleries but also to the other members, the in-group of buyers who play by the unwritten rules. If the art market democratizes, the value of the art to everyone falls to zero, the club is destroyed, the prestige is destroyed, and all art owners are left with is the junk that once was their scroll and key. So the exclusionary conduct protects the value enjoyed by consumers — indeed, gives the unlimited resource that is modern art its scarcity value — instead of harming consumers.
Antitrust has long distinguished between product-improving conduct that incidentally harms competitors and conduct that serves only to harm competitors, generally exempting product-improving conduct from censure and condemning only conduct that has no such redeeming feature.
A classic example of exempt conduct would be Apple’s introduction of the iPhone into the cell phone market in 2007. Any Apple refusal to license iPhone technology to Nokia may well have been the proximate cause of Nokia’s demise, but that refusal would be no antitrust violation, because the refusal would presumably be necessary to protect the value of the iPhone to consumers. If, the argument goes, Steve Jobs knew in 2005 that he would be forced to share iPhone technology with Nokia, destroying his competitive advantage and thereby his ability to recoup his costs through higher prices, then Steven Jobs would never have bothered to invent the iPhone and consumers would have ended up with nothing. (The intellectual property context is not unique here. Supermarkets would not trouble to exercise the foresight needed to build on land most convenient to consumers, the argument goes, if they could expect to be forced to share access to their real property with competing supermarkets.)
Similarly, the refusal of the galleries to do business with resellers is key to the ability of modern art to maintain its value as a signifier of elite club membership. The product is social prestige, and denying others the ability to sell access to that prestige is key to maintaining the value of the product. The galleries are harming competitors when they harm resellers, but they are not really harming consumers.
Rich social outcasts might think they would benefit as a group from an end to resale restrictions, but the benefits would be fleeting. The value of the artwork these losers buy, not to mention the social status it confers, would evaporate as more and more losers were to pour into the market, just as the marginal fishing boat destroys the fishery, or the marginal cow destroys the commons. Indeed, the destruction incident to a failure to police entry to the art market is worse — in fact, total — because unlike in fisheries or on town commons, for which the cost of a boat or a cow places a natural limit even on unregulated entry, the near-infinite abundance of modern art means that once the club doors are open the cost of entering can be driven all the way down to zero.
The only thing that could therefore prevent the galleries from winning the antitrust case is shame — the shame of raising as a defense the fact that modern art is junk and that they deal instead in prestige. I suspect that such an admission would have no material effect on galleries’ business. If buyers really are there for the prestige and not the art, as I think is the case, then acknowledging that fact deprives buyers of nothing, and so should have no effect on art prices. But such an admission could have an effect on pride. The galleries do claim to sell art; all really good sellers believe their pitches.
Samuel Cook defends advertising on the grounds that it: funds; drives culture; creates competition; and influences, rather than manipulates. I have argued at length against each of these apologies for advertising.
Funding. Yes, advertising does fund newspapers, online search, social media, and many other useful services. But there are better ways to fund services we love, such as charging full price for those services, or voting for government subsidization.
Why are those better funding methods? Because they are less economically wasteful. Funding through advertising involves two kinds of waste.
First, the money that advertisers are willing to pay for advertising represents waste, in the form of profits generated from inducing consumers to pay higher prices for products that those consumers do not actually prefer. In the information age, you don’t advertise to inform, but rather to manipulate (or as Cook would have it, to influence, of which more below), and you should never invest more in manipulation than you can hope to recover from being able to charge higher prices for lower quality products as a result of your promotional efforts. But those higher prices and reduce product quality represent waste — an advertising-induced misallocation of consumer resources. Consumers would be better off paying less for unadvertised products.
The second source of waste from funding through advertising is the expenditure of resources on advertising infrastructure itself. Given that advertising serves only to distort consumer choices, all the money required to actually create advertisements and target them at consumers itself represent a waste — a diversion of productive resources away from more socially valuable uses. Other funding models, such as subscriptions or tax-and-transfer, do not waste resources in these ways.
Suppose, for example, that a newspaper funds itself by selling $1 million worth of advertising. Advertisers would never be willing to spend $1 million on advertising unless they could earn a profit on that investment — through the higher prices for lower quality made possible by the advertising. If the advertisers hope to make a market rate of return on their investment of 8%, then we can infer that consumers lost $1.08 million to advertisers by buying higher-priced, lower-quality advertised products as a result of the advertising sold by the newspaper. To keep things simple, let’s assume that the newspaper offers production as part of its advertising service, so that creation of messages, images, and so on is all included in the $1 million the newspaper charge for advertising. If all these production services cost the newspaper $100,000 to provide, then the newspaper will net $900,000 from its advertising sales that it can spend on news gathering. The net result is that, to fund journalism to the tune of $900,000 using advertising, a total of $1.08 million must be wasted by consumers, and another $100,000 wasted by newspapers, which translates into a total loss to society of $1.18 million.
By contrast, if the newspaper were simply to charge full price to readers, and readers to pay it, then nothing would be wasted in the funding of journalism. Consumers would pay $900,000 for $900,000 worth of newsgathering (or perhaps a little less, to account for the costs of administering a subscription service). Granted, news is a public good for which consumers may not be willing to pay, even if they love it. In that case, the most efficient way to fund newspapers would be through government subsidy. Government would tax those $900,000 from consumers and pay them out to newspapers (less the cost of administering the tax system), again with none of the waste associated with advertising as a funding mechanism.
Drives culture. Cook’s argument here is that firms use advertising to convey messages of public importance: “Nike has, for years been on the frontline of effecting social change,” he writes. This is really a variant of the argument that advertising funds socially useful services — here the socially useful service is public messages that promote certain social values. The trouble with this apology for advertising is here again that advertising is a wasteful funding method. Ban commercial advertising, raise taxes by the $3.34 billion that Nike will spent last year on advertising, and then earmark that money for distribution to activist organizations for use in public service advertising campaigns. You would end up with a more powerful public service message, because Nike’s public service message is weakened by the fact that it’s tied to selling shoes, and consumers would end up better off because many would not be influenced into buying sneakers they don’t need as collateral financial damage associated with Nike’s public service advertising.
Looking beyond this economic argument, it is important to acknowledge here too the many rich critiques of advertising’s influence on culture. Some argue that advertising destroys culture by encouraging people to derive pleasure from consumption. Others argue that it simply crowds out messages that are not tied to commercial ends. Talented young artists simply can’t afford to put up billboards celebrating love, or run Super Bowl commercials celebrating social harmony, for example, because deep-pocketed advertisers can outbid them for that advertising space every time. Ban commercial advertising, and political advertising — something that I most certainly do not condemn — becomes far less expensive and more common. That’s good for culture and for democracy. The heart of my argument against commercial advertising is not, however, cultural, but economic.
Creates competition. Advertising certainly can create competition, by helping a startup, for example, challenge an entrenched brand. But advertising promotes competition in a way that deprives competition of the virtues we usually ascribe to it. The startup that advertises wins by influencing consumers, not by fielding a better product. When firms advertise to compete, they compete on advertising — which company has the more powerful influence campaign — and not exclusively on the variables that healthy competition is supposed to influence — price and quality.
In a world without advertising, startups have only one way to challenge the incumbent — by offering a better quality product at a lower price. That limitation ensures that competition will tend to minimize price and maximize quality. When firms compete on influence — via advertising — there is no guarantee that the firm that wins will offer the best product at the lowest price; the winner may win instead by advertising best. So while advertising may promote competition, it undermines healthy competition.
True, sometimes a brand is so entrenched that startups cannot enter the market without the aid of advertising, no matter how good their products and how low their prices. But the only solution to this problem that leads to health competition — competition that’s focused on price and quality — is to ban advertising and to use the antitrust laws to break up entrenched brands. Banning advertising itself undermines entrenched brands, because owners of those brands often use advertising to maintain their power. Action under the antitrust laws to compel licensing of iconic trademarks and other intellectual property would undermine brands that are entrenched for reasons other than advertising, such as advantages associated with being first to market. It has been done before.
Influences, notmanipulates. Cook takes issue with my characterization of advertising as “manipulative,” arguing that manipulation implies deceit. Good advertising, argues Cook, does not deceive, but rather “shar[es] passion” for a product in a truthful and honest way, “influencing,” rather than manipulating.
Certainly, much advertising influences without deceiving, and any case against advertising must explain what is wrong with such influencing. What is wrong is that truthful, non-misleading advertising can influence consumers into purchasing products that they do not really prefer. Current law regarding advertising, as well as Mr. Cook, assume that the only advertising that can cause consumers to buy products they do not really prefer is deceptive or misleading advertising, which is why the law prohibits misleading or deceptive messaging. But for several decades now, behavioral economists have been quite clear that truthful and non-misleading advertising can also induce consumers to buy products they do not really prefer, by operating on the habit-forming parts of the brain, to the exclusion of the deliberative parts of the brain. Just as countless tourists have died in London by looking — out of habit — in the wrong direction before crossing the street, making an entirely voluntary, though habit-directed, choice to die, even though every last one of them would certainly say that they did not prefer to die, millions of consumers may well voluntarily buy products that they do not really prefer, due to the influence of truthful, non-misleading advertisements that nevertheless appeal to the non-deliberative faculties of the brain.
The trouble with advertising is precisely that advertising influences, in Cook’s sense of the word. My argument is not that influencing itself is wrong. Influence is an essential part of political debate. We want great politicians to move us to be better than we might prefer to be. But commerce is different. The Invisible Hand only works when consumers are in the driver’s seat in markets, imposing their preferences on firms through the choices that they make. Every advertisement that influences a consumer into abandoning those preferences weakens the control of consumers over markets, putting markets into the hands of firms, and causing markets to serve consumers that much less well.
There is nothing wrong with having a passion for a product and sharing it with other consumers, but when you are paid to do that by a firm, and the firm necessarily uses its funds to give your passion an influence greater than it would ever have were you to express it unaided, as a private citizen, then your passion becomes a method for impoverishing consumers and undermining markets. It becomes a virtuous and well-meaning thing that has been put to wasteful ends.
Cash is just another input. Concentrations of wealth, concentrations of cash, throw firm surpluses to the wealthy. Deconcentrate wealth, deconcentrate cash, and corporate surpluses will fall more evenly across inputs to the firm.