I Richard Posner on same-sex marriage

While he did comment on same-sex marriage briefly in his text Economic Analysis of Law (Reference Posner1998), in his topical treatise Sex and Reason (Reference Posner1992) and more recently in the final chapter of Overcoming Law (Reference Posner1995) and on his blog (2005a, 2005b, 2008) shared with economist Gary Becker,Footnote ² Posner’s only published work dedicated solely to same-sex marriage is his review of William Eskridge’s 1996 book The Case for Same-Sex Marriage, which appeared in the Michigan Law Review in Reference Posner1997. Therein, Posner makes interesting arguments concerning the rhetorical structure and strategy of Eskridge’s book, which are not my concern here. Rather, it is the economic nature of Posner‘s responses to Eskridge’s arguments in favour of same-sex marriage, along with Posner’s other writings on the subject, that show the shortcomings of the economic approach to law. This is not to say, however, that Posner disagrees with Eskridge’s position; in fact, he writes ‘I find Eskridge’s argument for recognizing homosexual marriage quite persuasive’ (1997, p. 1584). But the issue is why: Posner’s reasoning is illustrative of the law-and-economics mindset and its failures in addressing what is essentially a matter of principle.

To Posner and other law-and-economics scholars, legal decision-making is a matter of efficiency, understood as maximising (or at least increasing) wellbeing. To the economist or economics-oriented legal scholar, wellbeing is usually taken to be a function of preference satisfaction; this is the basis of mainstream welfare economics, defended most forcefully in the context of the law by Kaplow and Shavell (Reference Kaplow and Shavell2002). For instance, in the blog post, Posner writes: ‘Obviously there are benefits to homosexual couples from marriage – otherwise there would be no pressure to extend marriage rights to them. … Some of these benefits appear to impose no significant costs on others and thus are clear social gains: an example is that a married person does not have to have a will in order to bequeath his property at death to his spouse’ (2005a). So, if same-sex marriage has clear benefits to some, while imposing few costs on others, it is welfare-improving or efficient in the Kaldor-Hicks sense, in the same way that it is efficient to build a new bridge if the benefits of the bridge to some exceed its costs to others.

Of course, the benefits and costs of a bridge are mostly financial in nature: providing access to homes and businesses on the one hand, while displacing property owners and affecting property values on the other. But the most significant benefits and costs of same-sex marriage are not usually considered to be financial. Certainly, there are legal advantages to marriage – such as the aforementioned matter of wills – which have financial aspects, but these considerations usually do not drive the debate, and could hardly explain the passions raised on both sides.Footnote ³ Instead, most law-and-economics scholars frame efficiency in terms of preferences, recasting benefits and costs as increases and decreases in preference satisfaction (respectively). And if people have preferences for opposite-sex marriage exclusively, legalising same-sex marriages would impose costs of them; as Posner writes, ‘the principal opposition to homosexual marriage comes from people who believe that for the state to recognize such marriage would be to place a stamp of official approval on homosexuality’, which they oppose (by assumption) (1997, p. 1583). Hence, elaborating on the earlier statement from his blog, Posner concludes: ‘Unless “outrage” costs are high, such benefits would, in an economic analysis, warrant recognizing gay marriage’ (2005a). Therefore, according to the economic approach, same-sex marriage should be legalised provided the people in opposition to it do not experience costs in excess of the benefits to homosexuals who wish to marry (or wish to have the right to marry). Therein lie several problems.

1. Preferences as wellbeing

Economists consider preference satisfaction to be equivalent to wellbeing or welfare. This position reflects a subjective notion of wellbeing which does not judge the content or nature of preferences, but simply takes them as given and assumes that persons will be better off if their preferences are satisfied as much as possible.Footnote ⁴ While the spirit of this reasoning is commendable, it is also naive; not only do some preferences, when satisfied, have no effect, or a negative effect, on wellbeing, but true wellbeing consists of much more than preference satisfaction alone. The first point is obvious: some preferences are self-destructive, especially preferences that we ourselves disavow, such as an alcoholic’s preference for a drink or a dieter’s preference for doughnuts. We may also judge other people’s preferences to be destructive to them – such as a mother’s disapproval of her son’s ‘reckless’ desire for a motorcycle – though this reflects her sense of his wellbeing, not his own. Some scholars recommend modifying preferences, rendering them ‘rational’ or ‘true’, as if made by fully informed minds not in a state of temptation or cognitive dissonance; while such a method would bring a person’s preferences closer to her wellbeing, it takes them further from what she actually prefers – and how she actually makes choices.Footnote ⁵

There are also preferences that are misinformed. For example, a person who is sincerely concerned with her health, but misunderstands a recent report she heard regarding the deleterious effects of cholesterol, may form a preference for fatty foods, thinking they will improve her health. Though this is her preference, and she will make choices based on it, it is clearly not in her own interest as she understands it; the preference was based on misinformation. Posner addressed this issue when he dismisses preferences against same-sex marriage based on the belief that official sanction will ‘turn’ children gay. Given the large body of evidence pointing to a largely genetic basis for homosexuality, this concern can be judged to be misinformed – children are unlikely to be ‘turned’, though children who are gay may feel more comfortable being honest about their orientation – but nonetheless it is a preference, though not linked to wellbeing due to its falsity.

Finally, there are preferences which do not affect one‘s wellbeing in any significant way. For instance, Jane may have a preference to see Topeka, Kansas, host the Summer Olympics someday (even after her death), but whether it does or not, it does not affect her wellbeing (especially after her death, assuming no afterlife). Likewise, John may have a preference to see same-sex marriage legalised, but as a straight man with no close gay friends or family members, it will have no direct effect on him, and will not impact his wellbeing in any significant way. The people who oppose same-sex marriage for fear that it will legitimise homosexuality have similar ‘detached’ preferences; the preference or belief may be very strong, but whether it is satisfied will likely have no significant impact on their wellbeing. Measuring a person’s wellbeing according to either self-destructive preferences or detached preferences will not accurately represent any reasonable account of wellbeing, even if they do help explain choice.

2. Treating all preferences equally

The existence of detached preferences not only helps sever the link between preference and wellbeing, but it also affects the way preferences are used in the analysis of economic welfare or efficiency. Posner confronts this issue in his blog post in reference to preferences against same-sex marriage:

He is correct that Mill’s is not a good economic argument, but that is a strike against economics, not Mill. Posner himself acknowledges as much in Sex and Reason, giving a qualified defence of Mill’s point (1992, pp. 201–204), and in the conclusion writing that ‘economics does not draw a satisfactory line between self-regarding and other-regarding conduct, that is, between conduct that is personal, and not a fit subject for public regulation, and conduct that is public – that affects strangers palpably – and that therefore is a fit subject for regulation’ (1992, p. 438).Footnote ⁶ In Overcoming Law, Posner reiterates his agreement with Mill on this, but declares that ‘to defend it … would carry [him] outside the boundaries of economics as they are generally, even generously, understood today’ (1995, pp. 571–72). The downside is that Posner discards Mill’s position in his economic analysis, and maintains that all costs and benefits deserve equal consideration in welfarist calculation; in a later book review (of Evan Gerstmann’s 2004 Same-Sex Marriage and the Constitution), he wrote: ‘the question is what weight the strong public disapproval of homosexual marriage should be given in balance with the interests of homosexuals’ (2003, p. 35).Footnote ⁷ (Further investigation of this issue in the next subsection will lead us to question the practice of welfarist calculation itself.)

No one will deny that some psychological or emotional harms are more serious than physical ones, but it is questionable whether the knowledge that homosexuals will be able to marry each other counts as psychological or emotional harm in persons not directly affected by it. A significant number of people disagree with it, to be sure; some vehemently so. There may a few who are disturbed by the possibility to such a degree that they experience some real degree of psychic harm, but may we assume that this number is small enough to be negligible? And even if the number is not negligible, might there not be a comparable number of people who would experience intense elation at the prospect of same-sex marriage, whether or not they are directly affected by it? If we are going to count emotional responses, these would need to counted as well; after all, there is no reason ‘elation benefits’ should count less than outrage costs. That said, preferences regarding same-sex marriage on the part of people not directly affected by such a change may be of interest and concern, but should not be considered instances of significant benefit or harm.Footnote ⁸

But why shouldn’t these effects matter, mild though they be? After all, as Posner writes, ‘there is nothing in principle to disqualify disgust … from counting as an external cost to which a polity dedicated to economic efficiency should pay heed’ (1992, p. 202). At bottom, it is not the relative intensity of the feelings that rule them out – it is their detached nature. Amartya Sen made this point in his 1970 paper ‘The Impossibility of a Paretian Liberal’ with the example of one person who wanted to read Lady Chatterley’s Lover and another, a ‘prude’, who didn’t want anyone to read it. Sen characterised the prude’s preference regarding others reading the book as a ‘nosy preference’, and showed that counting such preferences in Pareto judgments would conflict with the liberal ideal that each person is entitled to freedom of action within a protected sphere of privacy. Ronald Dworkin made a similar point, drawing the distinction between ‘external’ and ‘internal’ preferences, and arguing that if external preferences regarding other people’s actions are considered, double counting will result; those persons will have influence over not only events that affect them, but also events that affect others (1977, pp. 234–38). Both scholars confirm Mill’s basic position that there is a sphere of privacy within which others’ opinions do not matter – especially when acting through the state.

To be sure, I am not claiming that it is always easy to declare preferences to be nosy, external or detached, when a person may feel them very intensely. Many of us cringe when we see disturbing pictures of natural disaster or war, at which we may break into tears or vomit; disturbing political speech often draws similar responses. Are these simply nosy preferences regarding others’ behaviour, even though they affect us so strongly? While these effects are very real and sometimes very intense, we still regard them as detached, and our intuitions lead us to want to disregard them, or at least discount them, in any welfarist calculation. The issue would seem to be one not of whether the preferences are detached or personal, but rather whether one person’s preference should have an effect on another’s action. I may be repulsed at seeing a person spitting on the pavement or wearing a hateful politically themed T-shirt, but should that preference matter? One way to put this is that giving weight to these preferences seems to conflict with rights: rights to free speech, rights to read the books one chooses – and perhaps the right to marry the person of your choice, regardless of their sex.

3. Rights and the contingent nature of welfarist calculations

Insofar as the economic approach to law recommends maximising wellbeing or welfare, it is essentially utilitarian (or at least consequentialist) in nature. As Posner‘s statement above illustrates, law-and-economics would regard same-sex marriage to be efficient ‘unless “outrage” costs are high’ – if they are high, then no same-sex marriage. Setting aside the issues with outrage costs and nosy preferences detailed above, there is a conceptual problem with deciding issues like same-sex marriage on the contingent balance of benefits and costs: such issues are often considered to deal with rights, and rights by definition can conflict with utilitarian concerns of preferences or wellbeing.Footnote ⁹

Sadly, Posner‘s view is consistent with the broader law-and-economics approach to rights, subjugating them to considerations of efficiency: ‘when transaction costs are prohibitive, the recognition of absolute rights is inefficient’ (1983, p. 70). In Sex and Reason, he lists the costs and benefits of a right to same-sex marriage, acknowledging that there are benefits to such a right, ‘and the costs seem slight’ (1992, p. 311).Footnote ¹⁰ Later, Posner claimed that ‘we have to dig much deeper to answer the question whether there should be a constitutional right of homosexual marriage. We have to consider why homosexuals want the right, what the consequences of giving it to them might be, why the right is so strongly opposed by the public, and whether the Supreme Court is the proper institution to authorize homosexual marriage’ (2003, p. 35). But rights can hardly be considered rights if they can fall to marginal changes in relative welfare (Dworkin Reference Dworkin1977, p. 92), especially when based on sentiment. Whether marriage is held to be a right of all citizens (as same-sex marriage supporters maintain), or a right of only heterosexuals (as opponents believe), the argument for such a position is not properly made on the basis of utilitarian calculation, endlessly subject to the vagaries of measurement and estimation. Rights are not determined on the basis of efficiency, with the qualification ‘unless costs are high’; the strength of rights rests in their ability to trump considerations of costs. (For instance, certain types of speech can enrage us, but the right of free speech – in almost all cases – still holds.) Rights are ultimately grounded in moral principles that are largely (if not completely) insensitive to matters of utility, welfare or preference, and therefore present a serious methodological difficulty for mainstream law and economics.

II Darren Bush on Posner

In a 2001 article, Darren Bush criticises the assumptions that Posner makes in his economic analysis of same-sex marriage, but nonetheless endorses using the economic approach in general to argue for it. I argue that the shortcomings Bush identifies condemn the entire enterprise of applying an economic approach to same-sex marriage, because it emphasises that welfarist calculation of the costs and benefits of legal institutions is contingent on the measurement, estimation and conceptualisation of those costs and benefits, as well as assumptions made regarding them.

Bush correctly points out that the law-and-economic community is overwhelmingly supportive of freedom to contract in traditional marriage and family relations, but has been virtually silent on the issue of same-sex marriage. Finding Posner to be a rare scholar who does address this issue, Bush is disappointed to find that he makes many assumptions regarding the behaviour, lifestyle and even the overall life satisfaction of homosexuals, which introduce scepticism about the efficiency of same-sex marriage where no parallel concerns exist for traditional marriage.Footnote ¹¹ As Bush writes, ‘the most striking feature of Posner’s argument is its lack of economic foundation’ (2001, p. 125), including Posner’s statements that ‘permitting homosexual marriage would place the government in the dishonest position of propagating a false picture of the reality of homosexuals’ lives’ (1992, p. 312) and that ‘it would be misleading to suggest that homosexual marriages are likely to be as stable and rewarding as heterosexual marriages’ (pp. 312–13). Furthermore, Posner argues that same-sex marriage is unlikely to be regarded as ‘noble’ (p. 312) by the majority of the population, a standard which, as Bush notes, he fails to apply to other controversial contractual arrangements, such as surrogate motherhood.Footnote ¹² Bush also criticizes Posner for making the familiar but circular argument that homosexual married couples would be less monogamous than their heterosexual counterparts if allowed to marry because they are (allegedly) more promiscuous now (Bush Reference Bush2001, pp. 127–28, and references therein). If they were to marry, Posner supposes that some homosexuals will engage in serial marriage to exploit medical insurance for HIV-positive friends, an accusation that applies just as well to opposite-sex couples and other diseases and ailments (pp. 128–29, and references therein). Finally, Posner simply declares homosexuality to be less fulfilling than heterosexuality, and claims that gays and lesbians should want to change their orientation if it were possible (pp. 132–34, and references therein).

But even after all of this criticism of Posner‘s question-begging – judging same-sex marriage harshly as a result of holding homosexuals to a higher standard of conduct – Bush still defends the application of the law-and-economics methodology to the same-sex marriage debate (in particular, to argue in its favour). For instance, the first sections of Bush’s paper contain a very thorough analysis of the relative efficiency of traditional and same-sex marriage contracts and the potential with each for ‘market failures’ due to externalities.Footnote ¹³ But, as I argued above, this is all beside the point; it matters not whether marriage contracts among heterosexuals or homosexuals are efficient or not, if we are arguing about a right to marry the person of one’s choice.

Consider externalities, usually defined as harms or benefits accruing to persons not directly involved in a transaction. In cost-benefit analysis, these are harms equivalent to any other harm, and count equally against the benefits of a law, policy or institution. But properly understood, externalities are only a concern for public decision-making if the harms to third parties are wrongful – in other words, if they violate a right rather than expressing one. To use the example of disturbing speech again, the rantings of a racist on a street corner may enrage and harm passers-by, but the speech does not violate their rights; rather, it is an expression of the racist’s right to free speech. (And of course, the passers-by have the right to respond to the racist if they choose, but not to demand he be silenced.) If, instead, the racist, in the process of making his points, stands in the way of passers-by and refuses to let them pass, then he has violated their rights, and this is an externality the law is justified in addressing.Footnote ¹⁴

One category of externality that Posner proposes in opposition to same-sex marriage regards information effects (1992, p. 312). A positive information effect of same-sex marriage is represented by the more reliable signal of commitment that homosexuals would be able to make, were they allowed to marry legally. But he also lists costs: for example, a host of a dinner party will be less able to predict accurately whether a married male guest will bring a man or woman as his spouse. Bush correctly notes the triviality and ambiguity of this ‘problem’, but misses the larger point: hosts have no right, legal or moral, to precise information about the demographic make-up of their dinner parties that would justify granting legitimacy to such external ‘harms’ resulting from same-sex marriage. To provide another example, Posner also cites the ‘distorted’ information that people may infer regarding a homosexual married couple’s intent to procreate, while not recognising a parallel ambiguity with regard to heterosexual couples, or to unmarried couples (or singles) of any type. But once again, such a ‘harm’ would only be relevant to public policy if it violated some right, and no one could reasonably defend a right to infer the procreative (or adoptive) plans of a couple (or single). It is not the absurd triviality of these examples which is important, but rather their lack of relevance to legal decision-making. Not all externalities are created equal, for only some harms violate rights to be free of those harms, and the generation of other harms are actually protected by rights; and obviously, none of these rights can be justified or denied on the basis of benefit and harm.

Despite criticising the ad hoc assumptions Posner uses to flesh out his economic analysis of same-sex marriage, Bush still defends the economic approach because it ‘forbids discussion of the murky and unproductive area of morality’ (2001, p. 137) by using cost-benefit analysis.Footnote ¹⁵ But cost-benefit analysis, like any variation of utilitarianism, is just as ‘murky’ and morally loaded as any other system of ethical judgement. The discussion above suggests several value judgements inherent in utilitarian calculation: (a) it counts all benefits and harms equally, regardless of their origin or normativity; (b) it does not recognise rights except insofar as they are supported by utility or preferences; and (c), interpreted more generally, it denies individuals the basic respect and dignity they are owed as persons, and implicitly grants the governing authority the right to manipulate resources and rights as it sees fit to maximise welfare. Considered more practically, the sensitivity of welfarist calculation to assumptions regarding the incidence and relative strength of positive and negative effects opens the door to the type of manipulation of background assumptions of which Bush accuses Posner. Since cost-benefit comparisons are by nature contingent on the measurement, estimation and conceptualisation of costs and benefits, these assumptions are essential and unavoidable; without pointing a finger at anyone in particular, it is clear that a researcher can always get the result he wants by manipulating the starting assumptions. As Eskridge writes, ‘Posner’s cost-benefit approach is not neutral and is severely inconsistent with his supposed libertarian presumption, because it privileges an antihomosexual status quo. Since Posner does the cost-benefit analysis using the entitlements and values that exist in the status quo ante, he ends up with policies that tend to mirror the values he started with, namely, those of compulsory heterosexuality’ (1992, p. 357).Footnote ¹⁶

For these reasons, I disagree with Bush’s endorsement of the standard law-and-economics methodology – conducted in a more sympathetic manner – in application to the same-sex debate, or any similarly principle-laden public issue. As explained above, the issue of same-sex marriage is an issue of rights, which should not be decided on the basis of preference satisfaction, wellbeing or utility.Footnote ¹⁷ There is no doubt that, just as Posner’s welfarist calculations lead him to be sceptical (although, in the end, supportive) of same-sex marriage, the same calculations can be conducted, using different starting assumptions and estimates of relative costs and benefits, to arrive at a more supportive conclusion. As Posner writes, ‘the benefits of [same-sex] marriage may outweigh the costs’ (1992, p. 313, emphasis mine) – but then again, they may not. Bush himself regards the inclusion of ‘discomfort costs’ as legitimate, but claims ‘there is no empirical evidence to support either that [they] are significant or that they outweigh the benefits the same-sex marriage will generate’ (2001, p. 135). But would he change his mind if such evidence was presented – and should he? Such a method of reasoning, contingent as it is on assumptions and estimates, cannot be used to ground an argument for rights; such arguments must be made on the basis of principle. Both sides in the debate can make their arguments on principle, which should not be dismissed as ‘murky’ discussion of ethics, but rather narrowing the focus of the debate on the essential issues, and not simply the issues that can be quantified, summed and maximised. After all, the different assumptions preferred by Posner and Bush can also be considered differences in principle that impact their different welfarist calculations. But the law-and-economics technique, dressed in its scientistic clothing, obscures the principled essence of the debate, and makes the entire argument seem to be one of mathematics, not rights.

Finally, since the argument over same-sex marriage is one of principles, not relative costs and benefits, the proper venue for state action on the issue is the courts, not the legislature (state or federal). Following the distinction drawn by Ronald Dworkin (Reference Dworkin1977, pp. 82–86), legislatures are best suited to handle matters of policy, based on oft-conflicting goals of different segments of the polity that require deliberation and debate before they can be prioritised, funded and implemented. But matters of principles or rights are different; since these concepts ground the legal/political system itself, it is not appropriate to subject them to democratic processes, whether direct or representative. If rights are in question, they should not be up for vote; a right either exists or it does not – it is not a matter of whether there ‘should’ be a right, as Posner says. And whether or not it does exist should be formally recognised by the part of the legal system responsible for deciding on issues of rights – the courts. To do otherwise is to invoke Mill’s ‘tyranny of the majority’ by endorsing a system in which the majority can vote to rescind essential rights claimed by the minority. Unfortunately, this is what Posner is endorsing, because he sees same-sex marriage as an issue of policy to be decided along consequentialist lines; if undertaken by judges, ‘they will have to go beyond the technical legal materials of decision and consider moral, political, empirical, prudential, and institutional issues, including the public acceptability of a decision recognizing the new right’ (1997, p. 1585).Footnote ¹⁸ But just as judges should not legislate or make policy, the standard legislative process cannot do justice to the principled matter of same-sex marriage, nor does it serve democratic principles to try.Footnote ¹⁹