Are Markets Amenable to Consequentialist Evaluation?

Abstract

There is an ongoing debate over the moral limits of the market. Many participants endorse the plausible idea that a market’s moral status depends, at least in part, on its consequences. For example, Satz holds that markets whose operation undermines citizens’ ability to interact as equals are bad. And Brennan and Jaworski maintain that markets trading in any good or service permissibly possessed may be arranged to operate without bad consequences. This plausible normative claim about markets depends on a descriptive one. Namely, that individual markets have descriptive properties which would provide a suitable basis for their consequentialist evaluation. This descriptive claim, I argue, is false. Markets’ consequences are a joint production. There is no principled means by which the consequences of one may be distinguished from those of another. Thus, the plausible idea is false. A market’s moral status cannot depend on its consequences.

Some hold that different markets differ morally. Most are fine. No one cares that paperclips are for sale. But others—think of those trading in women’s sexual labor, human kidneys, and surrogacy services—strike many as troubling. They are bad markets. What explains their badness? And what, if anything, should be done about it? People disagree. Thus ensues a debate over the moral limits of the market.¹

Here’s a plausible idea. Whether a market is good or bad depends, at least in part, on its consequences. This is a natural suggestion. We often take consequences to figure in moral evaluation. It’s also modest. It doesn’t say anything about which consequences matter, and it’s consistent with holding that other things matter too.

Many who think about the moral limits of the market accept the plausible idea. Kanbur (2004) does. His theory evaluates markets along three parameters: the extremity of their consequences, the agency of those effected, and the inequality of their relations. Satz (2010) accepts the plausible idea, too. On her theory, markets are objectionable when their operation undermines the conditions necessary for people to interact as equals. To identify those that offend, Satz, informed by Kanbur, proposes her own parameters for evaluating markets. Brennan and Jaworski (2016) also accept the plausible idea. They defend the thesis that there are no inherent moral limits to the market. For any good permissibly possessed, they maintain, a market trading in that good may be arranged to operate permissibly. Others who think about the moral limits of the market accept the plausible idea too, at least implicitly (Semrau 2017; Moriarty 2017; von Platz 2017; Panitch 2020; Silver 2023).

If what’s plausible is true—if a market’s moral status depends, at least in part, on its consequences—it must also be true that:

Markets are Amenable to Consequentialist Evaluation (MACE): Markets have descriptive properties which would provide a suitable basis for their consequentialist evaluation.

MACE is descriptive, not normative. To accept it is not to accept consequentialism. One could deny that theory, even deny that consequences matter at all, yet affirm that individual markets have descriptive properties that would provide a suitable basis for their consequentialist evaluation.

The plausible idea, which assumes MACE, should be distinguished from a different idea, which doesn’t. This is the idea that, whether a market should be introduced, regulated, or prohibited depends, at least in part, on the consequences of doing so. This idea rests on the assumption that:

Market Interventions are amenable to Consequentialist Evaluation (MICE): Acts of market intervention have descriptive properties which would provide a suitable basis for their consequentialist evaluation.

MICE says that certain acts have certain descriptive properties. Since many arguments about markets are practical, MICE-assuming claims are common. For example, Zwolinski (2007) opposes the prohibition of sweatshop labor, anticipating adverse effects on workers’ autonomy and welfare. The anticipated consequences, notice, are not consequences of the market in sweatshop labor. Indeed, that market wouldn’t exist when they eventuate. Rather, they are the consequence of the act that prohibits sweatshop labor. That act has the descriptive property <adversely effects workers’ autonomy and welfare>. For another example, Radcliffe Richards et al. (1998) favor the introduction of a kidney market on the grounds that many lives would be improved and extended. This is a claim, not about the consequences of the kidney market, but the consequences of adding that market to the rest. It asserts that a certain act of market intervention has the descriptive property <improves and extends lives>; it does not assert that any market has that property. Zwolinski’s and Radcliffe Richards’ claims, and every other claim about the consequences of intervening in markets, assume MICE. MICE is true.

Is MACE true? I argue not. There are countless markets, each exerting some influence on the world. MACE-assuming theories presume that, for any one of them, we can identify its effects. But we can’t. Markets’ consequences are a joint production. We cannot, as MACE requires, decompose them into the consequences of the individual markets that together brought them about. Consider, for example, the market in women’s sexual labor. Some deem that to be a bad market, claiming its operation diminishes women’s status. Yet, women’s status is what it is on account of many markets. People pay to look at pornographic pictures and to watch women dance on polls, sometimes topless. They pay for smartphones, then hook up with strangers met on Tinder. They pay for alcohol to make hooking up easier and condoms to make it safer. And when they are unsuccessful, they pay for ebooks authored by so-called ‘pick-up artists.’ Plausibly, these markets also influence women’s status. What are the effects of each? How much does the market in pornography contribute? Is it more or less than the market in smartphones? What about the market in alcohol? Ebooks? If MACE were true, these questions would have answers. But they don’t. Events diminish women’s status, but events are not caused by, or attributable to, any individual market. MACE is false.

Why is MACE false but not MICE? After all, acts’ consequences are also a joint production. That’s true, but the two are different. MACE-assuming theories assert that markets have certain evaluative properties—good or bad. This requires that markets have certain descriptive properties such as <diminishes women’s status> which, as I shall explain, they lack. By contrast, MICE-assuming claims assert that market interventions have certain deontic properties—right or wrong. This requires that acts have certain descriptive properties such as <improves and extends lives>. And they do. The fact that acts’ consequences are a joint production poses no problem. For, the relevant descriptive properties depend, not on the outcome actualized, but on a comparison of that outcome with those that would be actualized by its alternatives. Suppose a kidney market is introduced. If lives are then better and longer than they would be otherwise, that act has the descriptive property <improves and extends lives>. All simultaneous acts actualize the same outcome—their consequences are a joint production—but, since each act has its own alternatives, each has the descriptive properties MICE requires.

The article’s contention, that MACE is false, is a theoretical claim, not a practical one. The coming argument, if sound, implies nothing about the goods and services that ought to be for sale. But it’s still interesting. For, if MACE is false, then the plausible idea is false. So too are all MACE-assuming theories. If individual markets have evaluative properties, they can’t depend on their consequences. And that, if true, would have significant implications for the debate over the moral limits of the market.

The article is in three parts. The first presents Satz’s MACE-assuming theory. The second, using that as an example, argues that MACE is false. The third evaluates three attempts to save MACE. None are successful.

A MACE-ASSUMING THEORY

Satz’s theory of the market’s limits sets out to complete both theoretical and practical tasks (2010, 91). It assigns evaluative properties to markets and provides insight into how to respond to bad markets.

To complete its theoretical task, the theory requires an account of what’s morally relevant for market evaluation. Satz endorses a relational conception of equal status in a democracy (2010, 100–104). Markets are objectionable when they undermine this ideal. To discover which are offending, Satz proposes four parameters. Two pertain to markets’ consequences. First, some markets have extremely harmful outcomes for individuals. Satz’s examples include markets that “lead to the depletion of the natural resource base of a country or to the fueling of a genocidal civil war” and “a stock market transaction that wipes out a person’s resources” (2010, 94). Second, some markets have extremely harmful outcomes for society. The social conditions necessary for people to interact as equals may be eroded by markets that “undermine the capacities that a person needs to claim her rights or to participate in society” (2010, 95). Bonded labor is one example. Of special concern are markets that lead to status inequalities in the domain of democracy, such as a market in votes. Two more parameters pertain to a market’s source. Some who engage in exchange do so with weak agency. They may be mistaken about the facts or otherwise ill equipped to pursue their interests. This concern is acute when transactions involve future performance, such as in commercial surrogacy. Also concerning are markets in which participants act on behalf of others, as in child labor. The final parameter is vulnerability. When participants negotiate from very different positions, the weaker party, because desperate, may be induced to transact on any terms. This may be true, for example, of those who would sell a kidney. Markets that score highly along these parameters have descriptive properties such as <has extremely harmful outcomes for individuals>, <has extremely harmful outcomes for society>, <involves weak agency>, and <involves vulnerability>. Such markets are apt to undermine citizens’ equal status. Thus, completing its theoretical task, they are assigned the evaluative property <bad>. Since Satz’s theory makes a market’s evaluation depend, at least in part, on its consequences, it assumes MACE.

To complete its practical task, Satz’s theory provides guidance about how to respond to bad markets. Importantly, the fact that a market is bad does not imply it should be closed. That may make things worse from the perspective of the same values that motivate intervention (2010, 110). Rather than aiming to close bad markets, sometimes we should aim to make them less bad (2010, 104). To that end, we attend to the features of the market—corresponding to the four parameters—that rendered it bad in the first place and pursue intervention that would improve them. Notice, here Satz’s theory shifts to the evaluation of acts rather than markets. The appropriate response to a market depends on a comparison of the conditions with intervention and without it. Accordingly, the judgments it now calls for are of the MICE-assuming variety.

Having sketched how Satz’s theory completes its theoretical and practical tasks, consider its application to the market in women’s sexual labor:

Selling Sex. Many who sell sex, living in poverty, are economically desperate. Most are under the control of a pimp, and so exercise little choice over who becomes a customer. Many have their agency and welfare interests set back by their participation in the market. Women are subject to negative stereotyping, marginalization, and stigmatization. The status of women is very low.

The market described in Selling Sex scores highly along all four parameters (2010, 144–50). Sellers are vulnerable, and their agency is weak. The market also has extremely harmful consequences for both individual and society. In particular, Satz faults the market for diminishing women’s status. Completing its theoretical task, the market is assigned the evaluative property <bad>. Yet, the proper response is not to close it. That may make things worse (2010, 150–53). Rather, it should be improved. Its participants, under the control of a pimp, are vulnerable and act with weak agency. Regulation might provide protection and legal recourse, thus eliminating the need for pimps. This, in turn, might reduce sellers’ vulnerability and strengthen their agency. Thus, the theory completes its practical task.

There’s an instructive objection. Consider a contrarian who holds that markets are morally neutral. She concedes that bad things happen in Selling Sex but denies that this is the fault of the market. She points to many other possible causes such as social norms, human nature, and government policy.² Any of these things, or something else, she insists, might be responsible for the lamentable state of affairs. Why, she asks, should we believe that the market in Selling Sex is a bad market, and not merely a conduit for other bad things? The trouble with Satz’s theory, she explains, is that, overlooking this distinction, it mistakenly attributes to the market the faults of its environment. To make her point, the contrarian observes that in some places where knives are sold people die in stabbings. The problem here, she contends, isn’t the market in knives, it’s the homicidal maniacs wielding them. They are bad; the market is not. In order to draw conclusions about markets, she continues, we must isolate their contribution. And when we do—when the other bad things in the world are corrected—we’ll find none left over. Markets, she concludes, are morally neutral. Call this the Other Bads Objection.

Replying to the Other Bads Objection clarifies Satz’s theory. We need not assume, with the contrarian, that a market’s moral status is a function of its contribution under ideal conditions. We might prefer a theory with practical application. So, we might prefer to evaluate markets on the basis of their contribution under actual conditions. This is Satz’s view. On her theory, markets have the moral status they do precisely because, given the social norms, human nature, and government policy where they operate, they have the consequences they do. A market that is ‘a conduit for other bad things’ is, for that reason, a bad market. This seems like the right result for a theory that evaluates markets by their consequences. If the contrarian remains skeptical, insisting that people who stab people are the problem, she, herself, may be overlooking something. Satz’s theory, like all theories, operates within a domain. It evaluates markets. This does not preclude the evaluation of other entities in other domains. We are not choosing, as the contrarian seems to think, between judging that the market in knives is bad, or judging that people who stab people are bad. The same state of affairs can sustain both judgments. A theory of the markets’ limits delivers the former, a theory of bad people delivers the latter. Similarly, we can hold that the market in Selling Sex is a bad one, while allowing that other things such as social norms, human nature, and government policy may also be bad. Acknowledging the existence of these other bads is compatible with holding, at the same time, that the market in Selling Sex, qua market, is bad.

Satz’s theory assumes MACE. It assigns evaluative properties to markets on the basis of certain of their consequences. With this theory in hand, I now argue that MACE is false.

THE ARGUMENT AGAINST MACE

We have focused on a simple case, Selling Sex, describing a single market. In reality, every market operates alongside the rest. So consider a fuller description of the case:

How Things Are. Almost anything that can be sold, is. Women are subject to negative stereotyping, marginalization, and stigmatization. The status of women is very low.

The Other Bads Objection asks why we should think a specific market is bad when there are other bads in the neighborhood. The challenge here is to explain why, of all the markets in the neighborhood, we should think one in particular is bad. Suppose the contrarian, having thought things over, concedes that a market would be bad if it diminished women’s status. And she agrees that, in How Things Are, women’s status is very low. But she isn’t convinced that this is a consequence of the market in women’s sexual labor. She notes that many other markets may contribute to the lamentable state of affairs. What about strip clubs, phone sex, and affair facilitation services? What about hookup apps on so many phones and sexist tropes always on television? Perhaps these markets also diminish women’s status. Internet forums host misogynists chatting about men’s rights. YouTube hosts their videos. Does that matter? Cosmetics, lingerie, and plastic surgery are all for sale. Might markets in these goods also play a role? The market for sex robots is small but growing. Is it part of the problem (Danaher 2017)? Pornography is big business. Perhaps, as Satz suggests, it contributes to women’s very low status (2010, 146). Does it? How much? The contrarian’s questions are not unserious. If MACE is true, they should have answers. But they don’t. What transpires in How Things Are cannot be reduced to the consequences of individual markets operating there. As I shall explain, attempts to do the impossible fail in one of two ways. Sometimes, being arbitrary, the consequences of many markets are attributed to just one. Other times, double counting, they are attributed to more than one. These mistakes are what we should expect if markets’ consequences are a joint production yet, believing MACE, we attempt to attribute them to individual markets.

One problem for MACE is that the goods and services exchanged in one market may influence the consequences of another. The internet has brought about a dramatic increase in pornography consumption. Maybe so many people looking at pictures contributes to the very low status of women. To which market—the one in pornography or the one in internet access—are these consequences attributable? It may be tempting to fault the market in pornography as, in its absence, the market in internet access would not have such harmful consequences. Then again, the internet plays a crucial role in the increase in pornography consumption. In its absence, the market in pornography wouldn’t have such harmful consequences either.³ Both markets are required to account for the outcome they together produce.⁴ To fault just one would be arbitrary. Such problem cases, where what’s traded on one market influences the consequences of another, are more problematic when the markets interact non-additively to produce their effects. Perhaps the consequences of hard drugs alone, and the consequences of selling sex alone, are, in sum, less harmful than the consequences of selling sex while using hard drugs. Perhaps johns visiting drug-addicted sex workers are especially likely to regard women as having especially low status. Here, too, we can’t fault just one market. Consider yet another problem case. Consumers respond to changes in price (Hayek 1945). If tin becomes more expensive, frozen orange juice producers turn to plastic-lined cardboard for packaging. Now, suppose the internet lowers the price of pornography so some would-be johns stay home to look at pictures. And suppose with fewer visits to brothels, women’s status is somewhat improved. Is this welcome consequence attributable to the market in pornography, which kept the johns at home? Or is it the market in internet access, which made that pornography affordable? And if the market in pornography has this welcome consequence, should we revise our earlier judgment that it contributes to the diminished status of women? In these and other cases, the goods and services exchanged in one market influence the consequences of others. No individual market has the descriptive property <diminishes women’s status>.

Another problem for MACE is that the profits from one market may influence the consequences of another. Take, for example, Satz’s assessment of the market in conflict diamonds (2010, 98). The diamonds themselves are innocuous, but the proceeds from their sale are used to finance bloody civil wars. Thus, Satz concludes, the market in conflict diamonds has extremely harmful consequences. But notice, the money earned from the sale of these diamonds must be used in another market, let’s say the market in weapons, before any harmful consequences materialize. Apparently, the consequences of the market in weapons are traced back to the market in conflict diamonds that financed their purchase. This possibility causes trouble for a MACE-assuming theory. Consider some questions it raises. Under what conditions are the consequences of one market transmitted to another? Suppose the diamonds were purchased with money earned from the sale of real estate. Should we then trace back the harmful consequences one step further to that market? Is the housing market, via the market in conflict diamonds, now implicated in bloody civil war? If not, why not? Suppose after the war the weapons are sold to purchase life-saving medicine. Should we then revise our assessment of the market in conflict diamonds? Does it now have better consequences? If not, why not? And where does all of this leave the market in weapons? After all, it facilitates bloody civil wars, which have extremely harmful consequences. But those consequences were already attributed to the market in conflict diamonds. To cite them again would involve double counting. Yet, how could we leave them out? It’s doubtful these questions admit of non-arbitrary answers. Far more plausible is the view that the market in conflict diamonds has consequences for the market in weapons, which has consequences for people at war. Markets together have consequences that are not reducible to the consequences of any one market.

Yet another problem for MACE is that options made available by one market may influence the consequences of another. Sometimes an option can be unwelcome. If you have it, you might be pressured to take it or held responsible if you don’t (Dworkin 1982). Consider an objection that Satz levels against kidney markets (2010, 199–202): She notes that in some places those seeking loans who are unwilling to use their spare kidney as collateral may receive less favorable terms than they would were there no kidney market. This, in effect, is how they are held responsible for refusing the option made available by the kidney market. This is bad, Satz explains, because people shouldn’t be made to pay a price to keep their body parts. So, we arrive at a novel objection to the kidney market: it gives rise to an unfair pecuniary externality. This attribution, claiming the unfair cost is a consequence of the kidney market, is surprising. If we must choose just one, why isn’t the unfairness material for a novel objection to the credit market, which, after all, is the proximal cause of the cost? Indeed, compare Satz’s treatment of the bonded labor market (2010, 171–88). In that market creditors extend loans, often at usurious rates, to debtors who promise future earnings from their labor as collateral. Since such debtors may be made servants indefinitely, Satz claims that market has extremely harmful consequences. Notice, in this case, Satz faults the creditors who make loans that make debtors servants indefinitely, not the labor market that affords the collateral. Yet, when creditors make loans that make debtors pay a price to keep their body parts, she faults the kidney market that affords the collateral, not the creditors who make the loans. It is thus somewhat mysterious what counts as a consequence of the credit market and what doesn’t. The mystery deepens. Having faulted the kidney market for the unfair cost, where does this leave the credit market? When cataloguing its consequences, what should we think of the fact that creditors make some people pay a price to keep their body parts? We’ve already attributed that to the kidney market. To include it in our evaluation of the credit market would involve double counting. But, then again, how could we not include it? The offending consequence is suffered by those who don’t participate in the kidney market but do participate in the credit market. The trouble here, as before, is that the consequence at issue—the unfair cost—is not a consequence of an individual market. Predictably, attempts to fault just one, as MACE requires, seem arbitrary.

MACE may be seen to be problematic from another perspective. To claim a market has certain consequences is to claim that its parts—the transactions it comprises—have certain consequences. So consider a final case, one depicting a single transaction in the market in women’s sexual labor:

John’s Visit. John, an avid pornography viewer, visits a brothel. Upon arrival, he uses hard drugs and pays for sex. He then regards women as having low status. Returning home, he is arrested for drunk driving. This motivates him to seek help for his drug use. Later, in therapy, John opens his eyes to his own misogyny.

John transacts in the market for women’s sexual labor, and then regards women as having low status. As before, the contrarian has questions. Are we sure that the latter is exclusively attributable to the former? What about John’s pornography consumption? Suppose he visits the brothel only to emulate those he’s seen in his magazines. Suppose while he’s there he’s thinking about certain pictures. Some pornographers, not the sex workers in front of him, are the intensional objects of his degradation. Might this have influenced John’s low regard for women? And what about the drugs? Suppose they lower his inhibitions and increase his desire. Suppose had he not used them, his later regard for women would be more egalitarian. How much did John’s transacting in that market contribute to his troubling attitudes? Then there’s the market for automobiles. John hates to walk. If he didn’t have a car, he would have stayed home to look at pictures. Doesn’t that matter? These are not idle questions. Lacking convincing answers, they make a point. We may grant that, as a consequence of John’s visit, women’s status is diminished, but we should deny that this is a consequence of any individual transaction.

I have argued that MACE is false. Markets’ consequences are a joint production. They cannot be decomposed into the consequences of the individual markets that together brought them about. The descriptive properties, such as <diminishes the status of women>, which Satz’s theory converts to evaluative properties, such as <bad>, are predicated of many markets together. If, insisting on MACE, we attempt to force the reduction, we should expect arbitrariness and double counting. The problem, to be clear, is not merely epistemic. I do not claim that it’s hard to know a market’s consequences. The problem is ontological. I claim that markets lack the descriptive properties MACE posits. Compare: My cat sleeps 16 hours a day. I do not claim that it’s hard to know how long her whiskers sleep. Her whiskers don’t sleep. The property <sleeps 16 hours a day> is predicated of her, not any of her parts. Though I have focused on Satz’s theory, the argument against MACE is independent of it. The problems I discuss will afflict any theory that says markets have descriptive properties which would provide a suitable basis for their consequentialist evaluation. Thus, they are problems for Kanbur’s account of obnoxious markets, which, in key respects, is the same as Satz’s. They are also problems for Brennan and Jaworski. If, as they claim, markets trading in whatever is permissibly possessed may be arranged to operate permissibly, and if, given the plausible idea, whether a market operates permissibly depends, at least in part, on its consequences, then the consequences of one market must be distinguishable from those of the rest. How else could they vindicate their thesis? If the foregoing argument is sound, these views, and every other MACE-assuming theory, must be revised or rejected.

THREE UNSUCCESSFUL ATTEMPTS TO SAVE MACE

Friends of MACE are unconvinced. To discover a market’s consequences, they say, one needs to look in the right place, where I have not looked. They make three suggestions. I now examine each. None supplies what MACE requires.

The first proposal, and perhaps the most obvious, simply treats a market’s transactions as acts, assesses those acts based on their consequences, and aggregates the results. This may be what friends of MACE have had in mind all along. On this suggestion, every transaction in a given market is identified with an act, such as consenting to the exchange, receiving the good or service, or making payment, which, by virtue of performing, one participates in that market. Since these acts correspond to the transactions that market comprises, their consequences are that market’s consequences. Thus, once each is evaluated, and the results are aggregated, we arrive at the consequences of that market. The resulting theory proceeds much like an act-consequentialist theory of right action. If it works, the thinking goes, then we can, as MACE requires, determine a market’s consequences. They are just the consequences of certain acts identified with it.

This proposal can’t save MACE. On standard forms of act-consequentialism, an act’s consequences are objective, including the entire outcome it actualizes. So construed, an act’s consequences are identified with the possible world its performance would actualize.⁵ That this proposal proceeds on such a capacious understanding of an act’s consequences is, in a sense, desirable. If the consequences of a market are identified with the consequences of its transactions, which are identified with the consequences of single acts, then those acts’ consequences must accommodate a lot. For, many markets have enduring effects. Consider those trading in surrogacy services, education, and votes. If the consequences of, say, having a baby are to be reduced to the consequences of a single act, such a capacious understanding is in order. The trouble is this capacious understanding counts among a market’s consequences far too much. This proposal implies that all simultaneous transactions, whether involving jellybeans or nuclear weapons, have the same consequences. After all, they actualize the same outcome. It also implies that every transaction has among its consequences the consequences of every subsequent transaction in every market. In John’s Visit, for example, <John opens his eyes to his own misogyny> would be a consequence of every market that figures in the case, including the market in women’s sexual labor. The consequences then attributed to a market would be, in fact, the upshot of all future markets. These results are unacceptable for those theorizing about the moral limits of the market. That project is rationalized by the idea that different markets differ morally. Yet, the present proposal implies that the only things that differentiate one market from another are the number of transactions and the time of their performance. If we hope to save MACE, we cannot treat everything that follows a transaction as a consequence of it.

The second proposal aims to improve on the first. It retains the basic act-consequentialist form but counts less than everything that follows a transaction as a consequence of it. Intuitively, certain things described in John’s Visit are rightly attributable to his participation in the market in women’s sexual labor. That <John regards women as having low status> is one. Other things, like <John opens his eyes to his own misogyny> are not. What’s required, then, is a way of mapping consequences to transactions that accommodates these intuitive judgments. So, here’s the proposal: a suitable relation must obtain between a transaction and a consequence for the latter to be attributable to the former. This relation does not obtain between transactions and those consequences that, intuitively, don’t count. But it does obtain between transactions and those consequences that, intuitively, do count. For example, while John transacts in the market in women’s sexual labor, and later opens his eyes to his own misogyny, the latter, being incidental, does not qualify as a consequence of the former. His visit, though causally necessary for his change of heart, is not suitably related to it. That distinction goes to the market in therapy, which is. When this more demanding relation is adopted, the problem cases I have described cease to be problematic. The morally relevant facts those cases depict will be attributable to the right individual markets. That <John regards women as having low status> will be attributed to the market in women’s sexual labor. And that <John opens his eyes to his own misogyny> will be attributed to the market in therapy. This is how MACE is saved.

This proposal can’t save MACE either. The principal difficulty lies in specifying the necessary ‘suitable relation.’ That relation, whatever it is, must meet two conditions. First, it must have the right extensional fit. It must validate, at least to some extent, the judgments already made about markets’ consequences. To achieve this, it will not be enough to shift from objective to subjective consequentialism, for example, counting only expected consequences. Since the expected consequences of a transaction may include many other transactions, the same problems that afflicted the first proposal, only slightly diminished, will arise here too. (Anyway, what would we say when buyers’ and sellers’ expectations differ?) A more radical shift is required. So, suppose we adopt a relation of temporal and spatial proximity. Suppose we count only those consequences that follow a transaction immediately or occur in the vicinity. This won’t work either. It’s too restrictive. Routinely, consequences attributed to markets occur long after the exchange and somewhere else. Recall the market in women’s sexual labor. It may contribute to the unequal status of women, but that is not its immediate effect. Its consequences are mediated by countless other acts over time and space. The same is true of markets in surrogacy services, votes, education, and much else. In these and other cases, markets are claimed to have bad consequences even though they eventuate only later and far away. A relation of proximity won’t work. Something more radical still is required. So consider the notion of causation in law. Even if an act is causally necessary for an event, if coincidence or the voluntary acts of another intervene, the event does not count as a consequence of the act. Hart and Honoré (1959) describe a case in which a woman, fleeing her homicidal husband, seeks safety in a nearby house only to be injured there by a falling ceiling tile. Though her injury would not have occurred absent the husband’s act of aggression, his act, they insist, did not cause her injury. The falling ceiling tile, too coincidental, intervenes between the two, breaking the causal chain. This relation, broken by coincidence and the voluntary acts of others, won’t work either. It’s also too restrictive. Recall Satz’s treatment of the market in conflict diamonds. Profits from those sales are used to purchase weapons which are used in bloody civil war. Satz attributes the war to the market in diamonds. Yet, the voluntary acts of others, such as those buying and selling the weapons, intervene between the two. Or recall her discussion of the kidney market. She faults that market on the grounds that creditors impose an unfair cost on those who are unwilling to use their spare kidney as collateral. Yet, that unfair cost results from the voluntary acts of creditors who impose it. It would not redound to the kidney market. Neither of these candidate ‘suitable relations’ have the right extensional fit.

Any plausible ‘suitable relation’ must meet a second condition. Its adoption must be motivated. That relation will ask us to count only some of what follows a transaction as among its consequences. Why should we do that? Consider the shift from actual to expected consequences. Why, when evaluating markets, should we care at all about the consequences the participants expect? Isn’t it a virtue of markets that participants, as if guided by an invisible hand, bring about outcomes no one was aiming at (Smith [1776] 1976)? And isn’t it true that people, focusing too much on what’s seen, think too little about what’s not seen (Bastiat [1850] 1995)? This is all to say, it is desirable that markets’ consequences aren’t limited to what their participants expect. There are reasons to prefer subjective to objective consequentialism, but they appeal to considerations related to agency, such as friendship, moral luck, and moral knowledge (Driver 2012). None of those things are relevant for evaluating markets. The other two candidate relations—causation in law, and temporal and spatial proximity—are similarly problematic. Consider causation in law. That relation, in that context, makes sense. Its use reflects the aims of the legal system and the constraints of what’s feasible there. Intentions matter. Responsibility matters. There, it makes sense. Here, it doesn’t. A theory of the market’s limits has no feasibility constraints. It isn’t intended to influence people’s behavior or hold them accountable. It’s intended to explain why markets have the moral status they do. On Satz’s view, for example, ultimately what matters is whether markets undermine citizens’ ability to interact as equals. Yet, if that’s what matters, there’s no reason to discount the consequences that follow after the voluntary acts of another intervene, or to distinguish between the consequences of intentional acts and unintentional ones. The relation that counts only temporally and spatially proximate consequences fares no better. There is no reason to recognize any such distinction among consequences. The adoption of either of these relations would be unmotivated.

Although there are countless other possible ‘suitable relations,’ it is doubtful that any will prove acceptable. Any candidate must overcome a pair of challenges. There is, first, a worry about gerrymandering. Perhaps we can devise some very complex relation, having many exceptions and qualifications, that maps individual markets to the consequences we want them to have. It may be possible, by adding more epicycles, to specify a relation with the right extensional fit. Still, this would not yet save MACE. We need reason to believe that the relation actually obtains. Doubts about that grow with the suspicion that it’s contrived ad hoc. Candidate relations face a second challenge. Lacking rationale, all are vulnerable to counterexample. The ‘suitable relation,’ if it is to save MACE, will count less than everything that follows a transaction as a consequence of it. But suppose we discover that terrible things follow a market in curious widgets. Suppose everyone regards everyone else as having very low status. Suppose further that the ‘suitable relation’—whatever it is—does not obtain between that market and those terrible things. Since they are not attributable to it, they would not figure in its evaluation. That market would not be a bad one. But that’s incredible. The theory that says it’s true is false.

A third proposal abandons the basic act-consequentialist form adopted by the first two. It seeks to save MACE using something like the method of science. It isolates a market’s consequences by comparing How Things Are with it and without it. This is a sensible suggestion. If we want to know the effects of a drug, we do the same thing. If we administer a dose to half of our subjects, and they all develop slight headaches, and our undrugged subjects experience no symptoms, we infer that the drug causes slight headaches. The present proposal suggests we perform a similar test to identify the consequences of a market. We compare conditions with all markets in operation, to conditions with all markets except for the one to be evaluated. The difference between the two conditions is attributed to the market that made it. Call this the Difference Maker Test. By way of demonstration, consider the market in women’s sexual labor. To isolate its consequences, we first note conditions in How Things Are, where <the status of women is very low>. Next, we attend to the conditions that would result in the absence of that market. There, call it How Things Are Without Selling Sex, we suppose that <the status of women is somewhat low>. Comparing the two, we see improvement. Things are better in How Things Are Without Selling Sex. We then attribute this evaluative difference to the market in women’s sexual labor.

This proposal fails. The Difference Maker Test will not save MACE. It does not identify the consequences of a market. In the demonstration above, the test did not reveal the consequences of the market in women’s sexual labor in How Things Are. Instead, it revealed how conditions in How Things Are would change if the market in women’s sexual labor were eliminated, which isn’t what we were looking for. The fact that things would improve if the market were removed does not prove that the market is bad when it’s there. For, the way things are in How Things Are is the product of many markets and we cannot assume that when one is removed, those that remain have the same consequences as they did when it was present. Markets respond to changes in other markets. That’s the key to their efficiency. For this reason the drug testing analogy was inapt. In that case we could validly infer that the drug caused slight headaches because there were no other drugs present to have effects. But there are other markets present in our cases.⁶ And changes in these markets, and not only the market in women’s sexual labor, may contribute to the differences between them. Perhaps in How Things Are, johns—feeling guilty about dissipating kids’ college funds on sex workers—drink more and, when drunk, more readily judge women as having low status. Perhaps excited by their experiences with sex workers they view more pornography, further exacerbating their objectionable estimation of women. If so, the improvement seen in How Things Are Without Selling Sex may be partially attributable to changes in the markets in alcohol and pornography. We then cannot infer that the difference between the cases represents the effects of the market in women’s sexual labor in How Things Are. The Difference Maker Test does not reveal the consequences of a market.

There’s another way to see why the Difference Maker Test won’t do. Suppose in How Things Are <the status of women is precisely 37.9>. What does the test say about the consequences of the market in pornography? Performing it, suppose we find that in How Things Are Without Pornography <the status of women is precisely 37.9>. The conditions are the same. Perhaps former picture viewers visit brothels instead, with the result compensating exactly for the absent market. According to the Difference Maker Test, since there is no difference between the cases, the market in pornography in How Things Are has no consequences for the status of women. That’s false. Moreover, if the Difference Maker Test did reveal the consequences of a market, as the proposal claims, then we could, by its repeated application, discover the consequences of each market in How Things Are. Suppose, for simplicity, there are only two. We’ve already tested the market in pornography. Next, doublechecking our earlier work, suppose we retest the market in women’s sexual labor and discover that, in fact, in How Things Are Without Selling Sex <the status of women is precisely 37.9>. Perhaps former johns look at pictures instead, with the result compensating exactly for the absent market. According to the Difference Maker Test, like the market in pornography, the market in women’s sexual labor in How Things Are also has no consequences for the status of women. These results imply that the markets in How Things Are, in aggregate, have no consequences for the status of women. This is embarrassing for the test because it proceeds from the stipulation that, as a consequence of those very same markets <the status of women is precisely 37.9>. The Difference Maker Test does not reveal the consequence of a market. It does not provide what MACE requires.

Perhaps this proposal was doomed from the start. The Difference Maker Test, in effect, defines bad markets such that their closure necessarily has good consequences. If that were right, then the appropriate response to all bad markets would be to close them. Yet, many who think about the moral limits of the market deny that. Radin does. She calls this dynamic, wherein a market is bad, but closing it would make things worse, the ‘double bind’ (1987, 1915–17). Satz agrees. She claims that closing the market in Selling Sex may well be counterproductive (2010, 150). Given this, she must regard the results of the Difference Maker Test described above as erroneous. On her view, conditions in How Things Are Without Selling Sex are likely worse than those in How Things Are. Kanbur takes the same view of bad markets. He notes, for example, that sometimes banning child labor leaves children starving (2004, 39). Brennan and Jaworski apparently agree. They imagine a bad drug, one that shouldn’t be possessed, but which is less objectionable when available on a market than not (2016, 17). If any of these judgments are correct, then a market’s consequences are not equivalent to the difference it makes. The Difference Maker Test does not save MACE.

CONCLUSION

I have argued that MACE is false. That, if true, has implications for theories of the market’s limits that accept the plausible idea. If MACE is false, so too is that idea. But other theories of the market’s limits, non-consequentialist ones, may still be MACE-assuming. They may posit that markets have descriptive properties which would provide for their consequentialist evaluation, even if they do not call for that evaluation. What I have argued has implications for them too.

The article’s upshot, I should stress, is purely theoretical. It implies nothing about which markets should be introduced, regulated, or prohibited. Those practical claims, about the consequences of intervening in markets, are of the MICE-assuming variety. Many arguments about markets are MICE-assuming. For example, Wolff (2011) argues, appealing to efficiency and social solidarity, that it doesn’t matter so much which goods and services are available on the market, so long as there remains a significant non-market sphere. This argument depends in no way on the consequences of individual markets. Its conclusion amounts to the comparative claim that, with respect to efficiency and social solidarity, conditions with a significant non-market sphere are preferable to those without one. Or, for another example, Maguire and Brown (2019) argue that certain valuable signaling practices, pertaining to care, testimony, and esteem, may be undermined, or lost, if behaviors associated with them were available on a market. Their argument, like Wolff’s, presupposes nothing about the consequences of individual markets. They assert only that something of value, present now, would be absent if certain behaviors were for sale. Both of these views may be construed as claims about the consequences of certain acts of market intervention. They compare the consequences of all markets together under two conditions. Neither assumes MACE. The foregoing argument leaves these, and all other MICE-assuming claims about markets, untouched.