Inframarginalism Miscellany Monopolization Regulation

Damages as Personalized Pricing in Favor of Wrongdoers

All courts do all day in civil cases in which the remedy is money damages is to engage in personalized pricing in favor of consumers. The plaintiff is the producer, the defendant is the consumer. And the damages amount is the price charged to the defendant for whatever it is that the defendant has taken from the plaintiff in violation of law, whether dignity, reputation, an arm or a leg.

When private enterprise personalizes prices, it chooses the highest possible prices: price equal to the maximum that the consumer is willing to pay. That is, firms strive to engage in perfect price discrimination.

Courts do the opposite. They personalize the prices of legal wrongs to be the lowest possible prices consistent with compensating victims: price equal to the cost to the plaintiff of the violation of law, and not a penny more. That is, courts strive to engage in what I have called perfect cost discrimination.

That’s weird, when you think about it.

All lawbreaking amounts to a forced sale. The defendant who shoots off the plaintiff’s arm forces the plaintiff to sell his arm to the defendant, or at least to sell the defendant the service of having an arm shot off, and whatever attendant satisfaction that provides the defendant, whether in the form of a feeling of security, the pleasures of power and domination, revenge, or what have you.

The law, in prohibiting battery, recognizes in the plaintiff a right to payment for the service. And if the transaction were not forced, and the plaintiff were to have any amount of market power, which we would expect to exist in spades with respect to the subject of many prohibitions–very few people are willing voluntarily to part with their arms, for example–then the plaintiff would almost surely charge a price for the arm above the bare minimum necessary to compensate the plaintiff for the harm. That is, if the exchange were voluntary, the price would in many cases be much in excess of cost, and indeed much closer to the maximum that the defendant would be willing to pay. Indeed, it seems reasonable to suppose that the defendant forces the transaction precisely because the defendant hopes to avoid being charged a price equal to the maximum the defendant is willing to pay.

So you would expect the law to provide the plaintiff with something closer to the bargain that the plaintiff would have struck voluntarily with the defendant. That at least would ensure that the defendant enjoys no gain from breaking the law and forcing a transaction.

But the law doesn’t see it that way.

The “rightful position” principle in remedies teaches that courts should measure damages in order to put the plaintiff in the position that the plaintiff would have occupied if the defendant had not engaged in the bad act. That causes courts to set the lowest possible price for breaking the law, rather than a price that approximates the voluntary price. For the position that the plaintiff would have occupied without the bad act is assumed to be the one in which no transaction takes place at all and the harm of the transaction has therefore not been inflicted. So damages under this measure just equal the amount necessary to compensate for harm. That is, the cost of the transaction to the plaintiff.

Law and economics scholars have made much of this cost-based baseline, arguing that it leads to optimal deterrence. The idea is that it forces the bad actor to internalize the costs of his actions. And so he will only act to break the law if the gains to him exceed the costs, which is to say, only if cost-benefit analysis shows that the action is efficient.

But that ignores something rather important about optimally-deterrent pricing: there isn’t just one optimal price. So long as the price the defendant pays for the forced sale is personalized, which it must be in a legal system in which judges award damages on a case-by-case basis, any price between cost and the maximum the defendant is willing to pay for the harm is optimally deterring.

Only a price above the maximum that the defendant is willing to pay–as opposed to cost–prevents the defendant from forcing the sale when the benefit exceeds the cost. So only such an extraordinarily high price is non-optimal. The maximum the defendant would be willing to pay is a measure of the benefit to the defendant. So only a price above that maximum drives the defendant away. There isn’t one optimally deterring price, but a range, that from cost all the way up to the maximum the defendant is willing to pay.

Where the courts set the price of illicit conduct within that range matters, because price determines the distribution of wealth between the plaintiff and the defendant, the victim and the injurer. By setting the price equal to cost, courts today achieve the perverse outcome of allowing the injurer to retain all of the gains associated with the forced transaction.

To fully appreciate this perversion, imagine that you decide voluntarily to sell your house. You could sell it at cost, including a reasonable return on investment. But that would be disappointing. What you’d like to do is sell it at the highest price anyone is willing to pay for it. If you do, then you extract all of the value created by the transfer. The buyer obviously places a higher value on the house than you do, otherwise he wouldn’t buy and you wouldn’t sell, and because you charge the highest price the buyer is willing to pay, you cause the buyer to pay out all of that excess value over to you.

By contrast, if you sell at a price equal to cost, including a reasonable return on investment, you don’t extract any of the excess value buyers place on the house. What you paid plus a reasonable return is the value you place on the house, the reasonableness of the return being enough to make you sell at that price. So when you sell at that price, the buyer pays you your valuation, and not a penny more.

Selling at a price equal to cost, including a reasonable return on investment, doesn’t therefore enrich you at all. It just lets you break even in a sense: you give up your house in exchange for a price equal to the value you place on the house.

But now suppose that you decide not to sell the house. You don’t like the price the buyer is offering. You believe the buyer is willing to pay more and you want to hold out until he does. And the buyer responds by bursting in your door one morning, holding a gun to your head, and telling you to clear out permanently, which of course you do, before filing a lawsuit. Now the buyer has forced a sale, and the law of trespass allows the court to dictate to the buyer the price that he must pay for your house.

Under current rules on the measurement of damages, the court would award you cost plus a reasonable return on investment, and not a penny more! The buyer could walk away with all of the gains from trade.

(Let’s put aside the fact that almost any court would issue an injunction here allowing you to repossess your house. Perhaps you’re emotionally scarred and don’t want to live there anymore, so all you demand is money damages. And let’s suppose also that your lawyer commits malpractice and fails to request punitives or damages for emotional distress.)

Which means that current damages rules turn over the entirety of the surplus generated by a violation of law to the wrongdoer! They embody the policy that the wealth generated by illegal transactions should be allocated to the scofflaw.

Which, again, is weird.

Now, you might object that courts award damages equal only to costs because the maximum that the wrongdoer would be willing to pay for the privilege of breaking the law is a thing difficult to calculate.

But so too are costs.

For costs are themselves maxima that someone would be willing to pay. The cost of an injury is the maximum that the victim would be willing to pay to avoid the injury. The cost of your house is what you paid for it plus a reasonable return on investment only because that is the maximum that you would be willing to pay to avoid having it destroyed or taken from you. More than that and you could buy a better house. And there is a subjectively element in that cost calculation: the reasonableness of the return is subjective. Current rules in theory should force courts to take that subjective element into account in awarding you compensation for harm equal to cost. And if courts can do that, they should be able to answer the question what the maximum that the wrongdoer would be willing to pay might be, including any subjective element thereof. (Indeed, courts should already do this in restitution cases, of which more below.)

You might also object that the maximum that the wrongdoer would be willing to pay is always less than the cost to the victim, because otherwise the wrongdoer would just be able to enter into a voluntary transaction with the victim to inflict the harm.

But I don’t think that’s right, at least if we want to maintain the fiction of rational decisionmaking that is all of the fun of law and economics and which itself underpins the whole theory of optimal deterrence I wish to complicate here.

The wrongdoer knows that undertaking the bad act will result in liability, and so when the wrongdoer acts, he does so knowing that he will pay a price. If the price is too high, which it will be if he inflicts a harm for which he would not be willing to pay, then he will not act. The courts therefore never can extract damages from wrongdoers in amounts above those which wrongdoers are willing to pay. If they do, wrongdoers simply will not act.

The economic problem for the courts is precisely to find the price that is high enough to ensure that the wrongdoer will not act unless he values the harm more than the victim, but not so high as to prevent the wrongdoer from acting when he does value the harm more than the victim. The trouble is that under current damages rules the courts always choose the lowest possible price.

Now, I don’t mean to suggest that the law is entirely deaf to the problem of gains from trade. One can almost always bring an unjust enrichment action and obtain the remedy of restitution, which does provide the plaintiff with the gains from trade.

But here’s the thing: restitution is an alternative remedy. Either you get restitution, or you get damages, but you don’t get both.

So a plaintiff can receive compensation for the costs to the plaintiff of illegal activity, or the gains enjoyed by the defendant, but not both. Whether the plaintiff opts for one or the other, therefore, the plaintiff will never receive a price for what he gives up equal to the maximum that the defendant is willing to pay, because the maximum that the defendant is willing to pay must equal both the cost to the plaintiff–the value the plaintiff placed on the harm–and the gains to the defendant of inflicting the harm, the excess over plaintiff’s valuation that makes the rational defendant willing to break the law in the first place.

Do punitive damages pick up the slack? It’s true that the pleasure a wrongdoer derives from inflicting harm is in itself probably sufficient to convert an intentional tort into one of malice, and that in turn can lead to punitive damages. But the doctrine of punitive damages suffers from terrible incoherence; we know that it is meant to punish, but does that mean to take some of the ill-gotten gains, or all of them, or to take more than those gains? Unless we are very lucky, punitive damages will either leave some gains with the wrongdoer or charge the wrongdoer a price in excess of willingness to pay, preventing the wrongdoer from engaging in efficient conduct.

Only a reconceptualization of the “rightful position” principle to require that courts measure damages by the maximum the defendant is willing to pay, rather than the cost to the plaintiff, would ensure that defendants do not enjoy gains from the illicit trade that is every offense under the law.

In closing, a word on the relevance of personalized pricing. Why does it matter here that, in engaging in case by case adjudication, judges effectively personalize the price of offenses?

It matters because personalized pricing is efficient whether the price charged is equal to cost or to the maximum the buyer is willing to pay. When prices can’t be personalized, and price is therefore one-size-fits-all for an entire market of buyers and sellers, then there is likely only one price that does not price some buyers or sellers willing to engage in mutually beneficial trades out of the market. That’s the price equal to marginal cost, the competitive price. And that price distributes the gains from trade between all buyers and sellers in the market in a single unique way. Try to change that distribution, by raising or lowering the price, and efficiency suffers: some buyers or sellers will be priced out of the market.

With personalized pricing, however, the court can vary the price charged to one buyer-seller pair–the defendant and plaintiff before the court–without changing the price charged to other pairs, so regardless the price the court chooses in one case, buyers and sellers won’t be priced out of the market in other cases. So the case-by-case character of adjudication opens up a world of distributive options with respect to the market for illegal activity that would not exist if the courts were to engage in one-size-fits-all damages calculations.

It’s a world that the law has failed so far fully to recognize and exploit.

Antitrust Inframarginalism Monopolization Regulation

Wherein Henderson and Kaplan Confuse Value and Cost

Or Why We Need More Inframarginalism

Todd Henderson and Steven Kaplan commit one of the more basic economic mistakes I have encountered, one all the more embarrassing because they are Chicago lawyers and economists.

They write that the private equity industry should not be judged based on its low returns net of fees because “[w]hile this is the appropriate metric for the decision about whether an individual should invest, what matters for society is how much wealth they create above the next-best alternative.” If you don’t net out the fees, they argue, then private equity shows large returns, and those returns reflect the creation of social value.

What Henderson and Kaplan have done here, in case you missed it just now, is to argue that an industry is productive by redefining a cost—and not just any cost, but that sacredest of sacreds, the fund fee—as social value.

But if they really mean to do that, which I doubt, then they’re actually making the case that private equity earns excess—read unnecessary—profits. Profits that represent a redistribution of wealth from consumers to private equity firms.

Unfortunately, Costs Are Costs

Let’s say that you decide to build a fence, but you’re terrible at it. You nail in all the slats askew and some of them fall off on the way to market. The cost to you was $50 in materials and $30 in labor, judged by the wage in your next best alternative line of employment.

Because your fence is a disaster, however, you are only able to sell the thing for $70, resulting in a loss of $10. Economics teaches that your fence business is a waste of economic resources. You expended $80 in combined value of resources to generate a product that created only $70 of value for consumers.

But Henderson and Kaplan say no. You have created $20 in value, the difference between the price of $70 paid by consumers and your materials costs of $50, because, well, if we ignore your $30 in labor costs, then you did!

What they don’t seem to realize is that the only way you can actually make that $30 in labor costs evaporate is if you don’t actually have an opportunity cost there for your labor; no one would have paid you a dime at any alternative employment. But if that’s true, and your costs really are just $50, then you didn’t need to charge $70 for the fence in order to have an incentive to build it. You just needed to charge $50, and so your $20 in profits are pure and unnecessary appropriation of surplus.

Which means that Henderson and Kaplan are inadvertently arguing that private equity is overpaid.

The Distinction between Value and Cost

But I really don’t think that’s what Henderson and Kaplan mean to argue. I think they are just confused about the relationship between value and cost, a confusion that is, alas, all too common in debates regarding law and economics, as I outline in a recent law review article.

The distinction between value and cost turns in fact on another distinction, that between utility and value.

The fence, even a badly constructed fence, has some utility for consumers, and that utility is measured by the maximum price that consumers are willing to pay for the fence: $70. In trying to avoid netting out costs and focusing instead on gross magnitudes, Henderson and Kaplan seem to be trying to say that utility and social value are one and the same.

But that $70 doesn’t represent value for society, because it does not account for the costs—the disutility—associated with generating it. If society must give up $80 in order to make a $70 fence, then society loses. Utility and social value just aren’t the same thing, as any careful undergraduate economics student should know.

To figure out how much value a business creates, you have to compare the utility the firm generates for those who use its products with the disutility—the costs!—the firm must create in order to produce those products. That is, value is a net quantity, it’s the difference between the maximum that consumers are willing to pay for the product and the cost of producing it. So the social value of private equity isn’t measured just by the gross returns that it generates, but by the returns it brings in net of costs.

All costs.

Fund Fees Are Costs

Including fund fees.

Costs in the economic sense are all harms that must be suffered in order for production to take place. The lost fees associated with not engaging in their next best alternative mode of employment outside of the private equity industry represent a cost, a harm, incurred by private equity funds in pursuing their work of privately acquiring and running firms. The fees that private equity firms charge must therefore be high enough fully to compensate them for this harm, otherwise they would not do private equity.

Henderson and Kaplan simply cannot ignore those fees in calculating the social value of private equity. They measure the harm of opportunities foregone to engage in private equity, the very harm of not sending physicists and engineers into physics and engineering, but instead allocating them to private equity funds, that critics of private equity decry.

If private equity can’t generate a decent return after netting out those costs, then private equity is social waste.

Unless They Represent Redistribution

The only way private equity fees don’t count as costs is if they not only fully compensate private equity firms for not engaging in some other line of business, but go beyond that to provide additional compensation. In which case some portion of the private equity fee can only represent one thing: an appropriation by private equity of the social value that private equity generates.

That is, private equity fees can only be ignored in the calculation of social value, as Henderson and Kaplan argue that they should be, if they represent an appropriation, by the private equity industry, of social value, defined as the value generated by their activities in excess of costs. And because Henderson and Kaplan appear to argue that we can count all private equity fees as social value, they are arguing that all private equity fees represent pure redistribution of social value from consumers to firms.

But precisely because social value is value in excess of cost, defined as the minimum necessary to compensate for all harms, it is value that does not need to be paid to firms in order to induce them to create social value. (Okay, it is necessary to pay private equity a penny more than cost, so that doing private equity makes firms strictly better off than they would be in their next-best alternative employments. Or just a ha’penny. Or a mill. But you get my point.) So what Henderson and Kaplan are arguing, in effect, is that private equity is taking more out of markets than is necessary to induce them to do private equity.

Government could, if Henderson and Kaplan are right, therefore dictate lower private equity fund fees without reducing social value one bit. Which sounds like a great idea to me.

Inframarginalists Don’t Make This Mistake

What really seems to have gotten Henderson and Kaplan into hot water is a lack of attention to the distribution of wealth between buyers and sellers in individual markets, what Michael Guttentag once described to me in conversation as “inframarginalism,” in contrast to the “marginalism” of a microeconomics that focuses on problems of efficiency.

What matters for efficiency-oriented lawyers and economists is that all units of output for which buyers are willing to pay marginal cost actually get produced. Which means that marginalists are interested in the cost-benefit analysis of the marginal unit of production.

Inframarginalists, by contrast, are interested in how the aggregate social value created over all of the other units produced by the firm—the inframarginal units—is distributed between buyers and sellers.

So social value is a bread and butter concept for inframarginalists. If they can’t define it properly—by netting costs out of willingness to pay—they can’t do their work.

And because inframarginalists know where social value begins and ends, they are unlikely to make the same mistake as Henderson and Kaplan.

Antitrust Inframarginalism Monopolization Regulation

Cost Discrimination

One hears constantly about the power of technology to enable the consumer-harmful practice of price discrimination, which is the charging, to each consumer of a given product, of a price equal to the maximum that the consumer is willing to pay for that product. But one hears very little about the power of technology to enable the consumer-beneficial practice of cost discrimination, which is the foisting upon each firm of a price equal to the minimum that firm is willing to accept in exchange for selling a given product.

That’s not because the technology isn’t there. In fact, because big business invested in supply chain automation long before the tech giants made possible the snooping needed to identify consumer willingness to pay, the technology needed for cost discrimination is more developed than the technology needed for price discrimination. The reason we don’t hear about cost discrimination is that the technology needed to implement it is in the hands of firms, rather than the consumers who would benefit from cost discrimination.

This state of affairs isn’t surprising, since firms are few relative to consumers, and therefore more likely to have the pooled resources and capacity for unified action needed to invest in and implement a discrimination scheme. Yes, consumers have review websites, and price aggregators, but that’s a far cry from the centralized acquisition and analysis of data, and the ability to bargain as a unit based upon it, that firms enjoy.

One way for consumers to implement cost discrimination would be by organizing themselves into data-savvy cooperatives for purposes of negotiating prices with firms. Another would be for startups to step in as middlemen, taking a cut from consumers in exchange for engaging in data-based bargaining on their behalf.

But another solution is for the government to create an administrative agency with the power to regulate consumer prices. It turns out that there is ample precedent for government price regulators to dictate cost-discriminatory prices.

Here, for example, is an account of the Federal Power Commission doing just that for wellhead natural gas rates in 1965:

Pricing designed to encourage supply could also create “economic rents” (profits above a normal return) for gas producers with old, inexpensive reserves. Neither the producers’ brief for fair field prices nor the staff’s preference for rates based on average historical costs seemed acceptable or sufficient. It was the young economist Alfred Kahn, serving as an expert witness, who suggested a two-tied pricing structure: separate prices for old gas and new gas. Here, from the commission’s perspective, was an ideal political, and perhaps economic, solution. “The two-price system,” wrote the commission, “thus holds out a reward to encourage producers to engage in further exploration and development while preventing excess and unnecessary revenues from the sale of gar developed at a period when there was no special exploratory activity directed to gas discovery.”

Richard H.K. Vietor, Contrived Competition: Regulation and Deregulation in America 113-14 (1994).

The old gas here corresponds to inframarginal units of production and the new gas corresponds to marginal units of production. Economists were once acutely aware of the problem that even under perfect competition the inframarginal units can enjoy a windfall at the competitive price, so long as the cost to their owners of producing those units happens to be below the cost of the marginal units, which determine the competitive price.

David Ricardo famously explained all of English aristocratic wealth in these terms. The aristocrats take the best land by force, he observed, and cultivation of that land is relatively inexpensive, because it is the best land. The rest take land that is more expensive to cultivate. Because the competitive price for agricultural goods must be high enough to pay the higher cost of cultivating the poorer-quality, hence marginal, land, the price must then be above the cost to the aristocracy of cultivating the best land, leaving the aristocracy with great profits.

Just so, the FPC worried that the producers of the old gas, who had come upon the gas only as an accident as part of explorations for oil, and therefore had incurred a gas exploration cost of zero, would enjoy a windfall if prices were set to cover the costs of bringing new gas from the ground through dedicated and costly explorations. So the FPC approved prices that discriminated against the old gas producers based on their lower exploration costs.

Consumers don’t know enough about the costs incurred by the firms that sell to them to insist on low prices when buying from firms with low costs. Which is why I suspect that government price regulation will be the only way for consumers eventually to enjoy some of the pricing-based fruits of the information age.

Antitrust Inframarginalism

John Bates Clark’s Peculiar Case for the Distributive Justice of the Perfectly Competitive Market

The argument for the efficiency of the perfectly competitive market is familiar: large numbers of firms selling an undifferentiated product will compete price to marginal cost, ensuring that everyone who can afford to pay the cost of production gets access to the product.

But is the distribution of wealth between buyers and sellers that is created by marginal cost pricing fair?

Today, economists agree that the distribution of wealth created by marginal cost pricing is entirely arbitrary. For the distribution of wealth is determined by inframarginal units of production, not marginal units. The benefit to me of the tenth unit of production of a particular good might be $10, and that might be just equal to the marginal cost of producing the tenth unit, making the market price under perfect competition in turn $10. But the gain I get from buying all ten units is determined by the value to me of each of the first nine units I purchase, not the marginal tenth, and the value I get from the first nine may be very much higher than $10. If I get $20 of enjoyment from each of the first nine units, then my gain from my purchase of ten units, net of the price of $10 that I pay for each unit, is $90. If the marginal cost of production is a constant $10 over all units, the gain to the producer is $0. That’s hardly fair to the producer. All of the gains from trade, defined as the difference between the value conferred on consumers by production and the costs of that production, go to me. Even though price is set equal to marginal cost.

It is for this reason that a century ago economists rejected the promotion of competition as a means of guaranteeing a fair distribution of wealth. They embraced competition because it is efficient–I am able to purchase every unit for which I am willing to pay the cost of production, so the economy produces all of the gains from trade of which it is capable–but they recognized that some other means was needed to achieve a fair distribution of wealth. That other means was the tax system. Raise my taxes by $45 and reduce the producer’s by $45 and now the gains from trade are distributed equally between us.

Given our current understanding of the distributive importance of inframarginal units of production, it is startling to discover that a century ago John Bates Clark, a giant of conservative economics, made a vigorous case not just for the efficiency but also for the distributive justice of perfectly competitive markets. His distributive case was rejected almost as soon as it was made, and has since sunk into obscurity. But that leaves me wondering: How could Clark have been unaware of the fact that the distribution of wealth is determined by inframarginal units, not marginal units? Did he just not understand marginalist economics? The few contemporary scholarly discussions of Clark’s work that I have encountered fail to explain.

The answer, it turns out, is that Clark understood marginalism, and the argument that inframarginal units determine the distribution of wealth. But he thought he had a convincing rejoinder. Here he is in his magnum opus, The Distribution of Wealth:

The man that we are studying is a society in himself: he makes things and he alone uses them. [The value to him of the last unit that he produces] measures the effective utility of everything that he makes. Though [the value to him of the first unit that he produces] may measure the absolute benefit conferred by the loaf that satisfies hunger, the real importance of having that loaf is far less. If this necessary article were taken away, the man would devote a final hour to bread-making, and would go without the article otherwise secured by that final increment of work. Destroy his day’s supply of food, and what he goes without will be luxuries naturally secured by the terminal period of labor. [The value to him of the last unit that he produces] measures the utility of those luxuires, and it measures therefore the effective service rendered by the supply of necessaries that are produced in an equal period of work. Any [inframarginal unit] will have a true importance measured by [the value to him of the last unit that he produces]; since, if it were lost, there would be diverted to the replacing of it some work that would otherwise secure an article having an importance measured by [the value to him of the last unit that he produces]. As it is of no more real consequence to the man to keep one of these articles than it is to keep any other, [the value to him of the last unit that he produces] measures the subjective value of each of them. . . . Bread and the other necessaries of life are absolutely more important than jewelry and other luxuries; but in effective utility the complements are all on a par, since, if any one of them were destroyed, the result would be to make the community go without the last.

John Bates Clark, The Distribution of Wealth: A Theory of Wages, Interest, and Profits 385, 388 (1914).

Clark’s argument is that if any one of the inframarginal units (Clark calls this bread, or another necessity, to illustrate that it is valued more highly than the marginal unit) is destroyed, the only actual unit that disappears from production is the marginal unit (Clark calls the marginal unit a luxury good to illustrate the fact that the consumer values it less than inframarginal units), because one unit is subtracted from output. Because consumers lose the value of the marginal unit when the inframarginal unit is destroyed, it follows, according to Clark, that the true value of the inframarginal unit is actually the value to the consumers of the marginal unit.

That allows Clark to treat all inframarginal units as having no value to the consumer that is separate from the value of the marginal unit, which is to say that it allows him to ignore entirely the value of inframarginal units and to treat the value to the consumer of the marginal unit as the entire value of production. It then follows that because the competitive price equals both the value of the marginal unit to the consumer and the marginal cost of producing that additional unit, there is in fact no surplus generated by any transaction, and the consumer pays a price exactly equal to the value the consumer receives from production and the producer is paid a price exactly equal to the producer’s cost of production. Thus the problem of distributive justice is not so much resolved in a fair way by competitive markets as it is eliminated entirely, because under competition there are, according to Clark, no gains from trade at all, just a buyer and a seller who both subsist on a knife’s edge, buying and selling at a competitive price that leaves both just as well off as each would be had neither entered the market at all.

But is Clark right to argue that inframarginal units have no real value to consumers, because their disappearance would, individually, deprive the consumer only of the marginal unit?

Of course not. The enjoyment you get from eating your first scoop of ice cream is real, whether you eat a second scoop or not. And the enjoyment you get from eating your first scoop of ice cream really is greater than the enjoyment you get from eating your second scoop of ice cream, notwithstanding the fact that if your first scoop is somehow clawed back, you won’t be able to eat that second scoop and get the lesser enjoyment from it. Indeed, once that first scoop is clawed back, your second scoop becomes your first scoop, and your enjoyment of it goes up. That’s why monopolies restrict output. They know consumers place a higher value on the first few units they consume, allowing monopolies to charge them higher prices.

Clark’s argument implies that the more you eat, the less valuable your meal is to you, because the less valuable is your last bite. That’s highly counterintuitive. Few would prefer a cracker for dinner to a four course meal, even if the last bite of that four course meal is worth less to them than would the first bite of that cracker. So why did Clark come up with such a view?

It seems to me that the strangeness of Clark’s theory is a measure of the level of disappointment felt by those who believed in the justice of competitive markets at the implication of the marginalism that the distribution of wealth in competitive markets is arbitrary. True, marginalism validated the Adam Smithian faith in the efficiency of competitive markets. But what had always been at stake in economic debates was the morality of the market, and this marginalism could not prove.

Clark’s failure to prove the distributive justice of the competitive market is a warning to those today who would promote greater antitrust enforcement and competition more generally as a solution to economic inequality. Indeed, the progressives of Clark’s own day, who were profoundly concerned about economic inequality, tended to reject antitrust and competition as solutions.