Law and Economics

Prior to critiquing Carson's argument, it is helpful to explain how different law and economic approaches can be applied to the same subject matter. There are many different ways to categorize theorists within law and economics,¹¹ however for the purpose of this paper we shall concentrate on different approaches to interactions between laws and markets. There are three different core beliefs on the interaction between the two. Firstly, Coaseans^{Reference Schwab12} hold that markets will produce efficient outcomes regardless of the law or informal norms.^{Reference Coase and Schwab13} This approach is based on a series of assumptions, the key of which are zero transaction costs and symmetric information. Coasean conclusions do not always hold when these assumptions are not correct.¹⁴ Secondly, Posnerians believe that markets tend to, but do not always, produce efficient outcomes. Where they do not, positive law can provide a state-backed assistance towards efficiency. Posner has stated “many legal doctrines rest on inarticulate gropings towards efficiency” — this is especially thought to be the case for common law rules.^{Reference Posner15} Thirdly, Ellicksonians hold that markets sometimes do not provide efficient outcomes, and where they do not then positive, state-enforced and centralised law is not useful: instead the primary way to solve such inefficiencies is by use of informal norms, or social norms, in which society finds its own way to exert pressure on markets to correct themselves to efficient outcomes.^{Reference Ellickson and Posner16}

Carson's argument is fundamentally Ellicksonian in tone — condom use will not be attained on its own, and social norms in the form of the condom code are required to enforce it. He also begins the article by debunking a Posnerian approach — formal, centrally enforced laws do not result in condom usage.¹⁷ However, Carson does not address the possibility of a Coa-sean response to condom use other than to assert that a problem of free-riders means that the market cannot operate efficiently in epidemiology. However, what if neither formal nor informal norms are responsible for condom use — could condom use simply be a private rational course of action which require no laws, formal or informal?

What are Collective Action Problems?

To commence this paper's critique of Carson's analysis it shall examine collective action problems to establish whether, economically, it is inevitable that HIV prevention through condom use by MSM falls within this category. The two most common examples of collective action problems from an economic perspective are the prisoners' dilemma and the tragedy of the commons.

The prisoners' dilemma is a classic and well known economic tool.^{Reference Peterson18} It has multiple iterations,^{Reference Poundstone19} but fundamentally proceeds on the following terms: two individuals (the prisoners) are arrested. Their options are to stay silent or to incriminate the other. There are therefore four possibilities: prisoner A stays quiet as does prisoner B, prisoner A incriminates prisoner B but prisoner B stays quiet, prisoner A stays quiet but prisoner B incriminates prisoner A, or each prisoner incriminates the other. If each of them stays silent then each of them will serve one year in prison, if each betrays the other then they will each get two years in prison and if one betrays the other then the betrayer is freed but the betrayed spends 3 years in prison. A tabular portrayal is therefore:

Efficiency can mean various different things economically,^{Reference Hardman20} but under any definition from the prisoners' perspective it is efficient for them to minimize their aggregate time in prison. The optimal outcome is, therefore, that both stay silent, as then an aggregate of two years is spent in prison. The least efficient outcome is both betraying the other, as then an aggregate of four years is spent in prison. However, from the perspective of either individual prisoner if they stay silent then they will either spent 1 or 3 years in prison (an average of 2 years), whereas if they betray then they will either spend 0 or 2 years in prison (an average of 1 year). It is therefore rational for a prisoner to betray the other. The issue arises that, economically, each prisoner behaving rationally will result in an inefficient outcome. If the figures were different and aligned the rational actions of the parties with an efficient outcome, then there would be no dilemma and thus no collective action problem. It is the mis-alignment between individual choices and efficient outcomes, and only this, which produces collective action problems.^{Reference McAdams21}

The same can be seen in the tragedy of the commons.^{Reference Hardin22} In such circumstances there is a common resource, such as a field or a lake, which various farmers exploit, by grazing their cows or fishing respectively. In such circumstances, once more the interests of participants are not aligned with overall efficiency of the market: each farmer is incentivised to maximize their own profit from the common resource, which they do by maximizing their use of the common resource past the sustainable levels of use. As a result, the common resource becomes depleted. Once more, it becomes rational for each actor to behave in a manner which is, overall, inefficient.

Carson's first economic pillar is that condom use for HIV prevention amongst MSM is a collective action problem. From the foregoing we can see that for this to be correct, there must be a misalignment between rational actions of participants and efficiency as a whole. If, instead, rational actions of market participants produce an efficient result, then there is no collective action problem and so Carson's social norms are not required to ensure condom usage. Is there, therefore, a way in which we can consider the use of condoms in hook-ups to be a rational actions?

Other Approaches to HIV Prevention

Philpson and Posner examine the issue from a different perspective.^{Reference Philipson and Posner23} Starting from the proposition that people prefer sex without condoms,^{Reference Mohr24} they are able to build a simple but compelling economic model: someone will participate in unprotected sex so long as the benefits to them doing so exceed the costs, where costs are calculated as a function of probability that the sexual counterparty has HIV (P1), the probability of transmission if they have (P2), the probability that the person themselves have HIV (P3), the seriousness of HIV (P4) and the treatability of HIV (P5).²⁵ As a result, when the probability that the counterparty has HIV increases by a general increase of HIV in the population, P1 will increase and so fewer people will opt to undertake unprotected sexual activity. The result is that people engage in risky sex if either both partners believe there is a high chance of both having HIV or both believe there is a low chance.²⁶ In turn, this will increase the private demand for HIV testing.²⁷ Philip-son and Posner's analysis produces a Coasean result: that any attempt of the state to interfere in HIV testing is likely to increase the likelihood of HIV rather than decrease it.²⁸ There are further complications to the model but, ultimately P1 is calculated by Philipson and Posner by reference primarily to the proportion of the population with HIV as a whole.²⁹

Schroeder and Rojas approach the issue from the perspective of game theory.^{Reference Schroeder and Rojas30} They model a game theory game that they refer to as the “risky sex game.” It has the following rules:

The preferences indicated are that actors of the same serotype will prefer RS over PS as will HIV + actors with HIV – actors, but that HIV – actors will prefer PS and NS over RS with an HIV + actor. Each actor has private information as to their own HIV status. Schroeder and Rojas use a Bayesian technique (i.e. that players update their assessment of the HIV status of the other actor) over the course of the game.³² This means that the course of interaction in preparation for a hook up can provide some insight into the HIV status of a potential counterpart: P1 is updated over the course of interaction between counterparts.

However, proposition of this paper is that it may be possible for an actor to be able to reach some insight as to the HIV status of their counterpart before commencing such interaction: for the purposes of Philipson and Posner it may be possible to identify the probability of P1 by reference to the initial market signalling received from a counterparty.

Further analysis has been advanced by Francis and Mialon.^{Reference Francis and Mialon33} Francis and Mialon integrate law into signalling, by examining the incentives that are created depending on what is criminalized. Their modelling demonstrates inefficiencies in laws — most US states criminalize transmission of HIV and exposure to HIV equally, which disincentivizes safe sex as it is equally as criminal. Secondly, most US states only criminalize knowing transmission/exposure, which disincentivizes testing. Francis and Mialon's analysis takes matters further by factoring in legal incentives to obtain HIV tests, and formulating the model laws to deal with this. Francis and Mialon state that “if the potentially infected individual knows his HIV status, his choice of whether to propose safe or risky sex may be a signal about his status.”³⁴ Whilst this is correct, it is limited to when actors know their own HIV status. We can use economic analysis to expand this to show that signalling has wider application.

Information Asymmetry

Philipson and Posner do not identify how P1 is ascertained, Schroeder and Rojas start their game with a known risk that their sexual counterpart may be HIV+, and update the probability that this is the case (i.e. P1) throughout the game by “in game” signalling, and Francis and Mialon highlight that there can be “pre-game” signalling if the actor knows their own HIV status. This paper's addition to the field is to use economic methods to demonstrate that there may be “pre-game” signalling which influences P1 which apply regardless of whether one actor is aware of whether the other knows their HIV status. Let us focus on the rise of mobile applications which rank other users by proximity, such as Grindr after which this article takes its name. Grindr has been said to result in increased number of “casual hook up”s, which are transactional in nature.^{Reference Licoppe, Riviere and Morel35} As a result, we can borrow from transactional literature to resolve whether HIV protection amongst the MSM community constitutes a collective action problem. Carson notes that location based applications assist with serosorting, in that they allow for inclusion of HIV status within the profiles and have instituted campaigns to encourage testing — stating “members have ample opportunity to inquire about a prospective partner's HIV status; whether or not they actually ask is not always clear.”³⁶ However, by viewing location-based applications as transactional we can add to this analysis to demonstrate how condom use in respect of such applications may not be a collective action problem: even without a market participant asking about the serostatus of their counterpart, initial signalling can affect P1.

The starting point is George Akerlof's 1970 article “The Market For Lemons: Quality Uncertainty and the Market Mechanism.”^{Reference Akerlof37} Akerlof uses the example of the automobile market:

A purchaser of a new car will not know whether they are buying a good car or a lemon, but after they have owned it for a while they will know the quality of their car. This causes issues with the used car market. Used cars are worth less than new cars. A buyer will be unaware whether the car is a good car or a lemon, whereas the seller will know. Good cars are worth more than lemons. As the buyer will not know whether a car is a good car or a lemon, good cars and lemons will sell at the same price. This will result in the price for used cars being the average price between the value of a good car and the value of a lemon: owners of good cars can only sell cars for less than they are worth, whereas owners of lemons can sell cars for more than they are worth. This means that owners of good cars will not sell their cars at the price of average quality, whereas owners of lemons will be keen to sell their cars, causing a lowering of quality of car in the market, which in turn causes a lower average price. As a result, only lemons are sold in the used car market. Akerlof also applies this to health insurance for the over 65s in America — no matter how many checks are undertaken, the insured will know more about their own health status than the insurer. This will result in the price of insurance being the average amongst the same demographic, i.e. healthy and unhealthy. For those that are healthy, therefore, the price will be more than they are likely to have to expend in the absence of insurance, making them less likely to take out insurance. For those that are unhealthy, the price is less than their likely expend in the absence of insurance, making them more likely to take out insurance. The result is that insurance prices increase, which further exacerbates the problem.³⁹ To generalize, information asymmetry between the buyer and the seller as to the quality of the product can result in the exit of high quality products from the market place. In a market where the seller knows the quality of their good, the objective value of such good and the price at which they are able to sell, this process can be instantaneous. The analysis behind the market for lemons applies when exchanges apply, goods can be of different quality and market participants have a way to know this.

What application, however, does this analysis have to the subject matter at hand? Information asymme-tries apply to serostatus when meeting putative sexual partners for casual meetings. Even despite a lack of an overt exchange of a product for cash, market principles still apply for casual “hook-ups.” Each participant is willing to trade something, their giving of pleasure (in whatever form that may take), for something, their receiving of pleasure (in whatever form that may take). There is therefore an exchange — rather than a definite “buyer” and a definite “seller,” however, each participant is, at the same time, a buyer and a seller: buying pleasure for them in exchange with granting pleasure. The market for lemons can be directly applied to the market for unprotected anal intercourse (UAI) in hook-ups. Each participant knows either their serostatus or their historic risk profile in respect of UAI. However, they do not know their counterparts' serostatus, nor their historic risk profile. We therefore have an information asymmetry.

Where the Grindr market is more complicated, however, than the automobile market is the knowledge that market participants have of the potential for an information asymmetry. In the automobile market, this arises naturally due to price: prices are set on average value, sellers know the value of the goods they are selling and can directly compare the price that they would receive for entering the market to the value of their good. This is not the case for the market for UAI. However, a directly analogous position arises. Public knowledge about the risks of UAI^{Reference Kann, Anderson, Holtzman, Ross, Truman, Collins and Kolbe40} alert market participants of the potential of HIV risks from using UAI. This public knowledge of the risk results in, at the margin, those without HIV exiting the UAI market, whereas those with HIV have no need to leave the UAI market. When coupled with information asymmetry as to the serostatus of your counterpart, this will result in a rational HIV negative actor adopting a cautious attitude to UAI, as those without HIV will be aware that, at the margins, other HIV negative actors will have left the market for UAI.

Let us make the following assumptions about the market for MSM hook-ups. Firstly, each participant either has HIV or does not have HIV. Secondly, an information asymmetry as to serostatus (or historic risk profile) occurs between the parties. Thirdly, that UAI is preferable to protected anal intercourse (PAI), but that the costs of contracting HIV outweigh the benefits of UAI over PAI. Fourthly, that the only risk of UAI is HIV.^{Reference Muñoz-Perez, Rodrigues-Pichardo and Camacho Martinez41} Fifthly, that there is only one strain of HIV (the combination of the fourth and fifth assumptions meaning that once you have HIV there is no additional risk of UAI).^{Reference Dyer42} This means that HIV negative participants will want to undertake UAI if they can be certain that their counterpart is HIV negative, whereas HIV positive participants will always want to undertake UAI. In the presence of a risk that their counterpart has HIV, the rational HIV negative actor will chose PAI over UAI (in line with the assumptions advanced by Shroeder and Rojas). The result of the information asymmetry, together with knowledge of the risks of UAI, will result in those without HIV from exiting the market for UAI. This results in an analogous situation to the market for lemons — as those without HIV exit the UAI market, the proportion of those remaining in the UAI market with (or who are more likely to have) HIV increases. In the same way that the information that good cars are leaving the market will filter into the price for used cars, so the knowledge that those without HIV have left the UAI market will. This therefore provides strong signalling to the market that those in the UAI market have a higher chance of having HIV than those in other markets, including the PAI market. In turn, this disincentivizes those without HIV from entering the UAI market. The result of this feeds in to the previous analysis: in addition to being able to update your assumptions about the probability that your counterpart has HIV based on the population percentage (as per Philipson and Posner) and interactions that you have with them (as per Shroeder and Rojas), the fact that they are in the market for UAI at all indicates a higher probability of being HIV positive, regardless of whether they have private knowledge as to their own serostatus (thus extending from Francis and Mialon's analysis). This means that we can ascertain that P1 is greater for those in the market for UAI than elsewhere, in turn further driving those without HIV out of the UAI market. This disagrees with the approach taken by Francis and Mialon, who argue that criminal sanctions may mean that an offer for PAI may signal HIV status.

In other words, whilst a different situation to the automobile market that created the market for lemons, the UAI market is analogous. Given our assumptions listed above, in the UAI market someone with HIV can be considered as providing a less valued product than someone without HIV,⁴³ as with a lemon in the automobile market. As with lemons in the automobile market, HIV positive entrants in the UAI market result in the exit of the other category from the market. In the case of HIV, this means that a signal from someone that they wish to undertake UAI (already a venture in respect of which an HIV negative person is aware of a risk) alone signals their increased chance of having HIV to a counterparty without HIV.

Why does this matter? This signalling is initial. It is, therefore, different than the game theory played out by Schroeder and Rojas which concentrated on signalling responses, and provides a nuance to the Philipson and Posner argument. It applies regardless of whether the signaller is aware of their own serostatus, and thus advances the Francis and Mialon proposition. It also makes the abstract risk of HIV more specific and concrete, and means that those without HIV are less likely to enter into the UAI market. In other words, it is rational for someone without HIV, who does not want to contract HIV, to avoid the UAI market, by utilizing condoms. In turn, this means that condom use in casual meetings on online platforms is rational. We have seen that collective action problems arise when it is rational for actors to behave in a way which is inefficient. If we view avoiding HIV, and therefore condom use, as efficient, then the above analysis shows us that rational actions of market participants lead to the efficient outcome. In other words, using condoms to minimize the risk of contracting HIV is not a collective action problem — it is the position that the market itself will provide once knowledge of the serostatus in the UAI marketplace percolates across market participants.

If using condoms is not a collective action problem, then informal norms are not required to solve it. This allows a different take on the use of condoms — rather than being an external pressure applied by the gay community which is internalized by market participants, instead using condoms becomes the rational choice that would be taken by market participants. Rather than social norms being the answer, instead perhaps simple market action is. In other words, we move from Ellickson law and economics to Coasean law and economics.

Instinctively, this makes sense: it is difficult to monitor breaches of the condom code. Whilst Carson cites issues with overt and public rejection of condoms,⁴⁴ decisions to use condoms are normally private to the participants in question. As failure to use a condom applies equally to that person who wished to not use a condom and the person who acquiesced to such request, there is little incentive for either party to publicize the lack of condom use for any particular sexual activity. Accordingly, it is difficult to monitor widespread compliance. It is not possible to establish lack of compliance from a positive HIV result, as HIV can be contracted from parents,^{Reference Betancourt, Abrams, McBain and Fawzi45} blood transfusions,^{Reference Moore, Herrera and Nyamongo46} sharing of intra-venous needles,^{Reference Valente and Vlahov47} and others. Conversely, not using a condom does not automatically result in HIV transmission. This combination means that in addition to not being able to privately track compliance with the condom code, it is not possible to establish it after the fact either. The analysis advanced by this paper therefore seems instinctively sensible — condom use arises from a private decision based on a cost/benefit analysis rather than by way of social pressure to follow norms to do so. Of course, it is not the case that such cost/benefit analysis and social norms exist in isolation, instead norms can be internalized and factored in to the cost/benefit analysis.^{Reference Francis and Mialon48} However, this internalization is just one factor, and may be less important than it seems to be under Carson's analysis.

Carson indicates that market mechanics (i.e. Coa-sean solutions for our purposes) do not work due to free rider problems — that condom use by others, or perceived condom use by others, will mean that people freely enter the market for UAI relying on the preventative measures taken by others. There is a risk, however, that this misstates the free riding problem. Free riding occurs when parties enjoy external benefits without the need to incur private costs. Carson indicates that the increased use of condoms by third parties lowers the prevalence of HIV in the market as a whole. This is correct, however it does not create an external benefit to a market entrant in the UAI market, it reduces the potential cost to the market entrant. Thus the presence of condoms, ART and PrEP lowers the risk of HIV transmission (P2 in the Philipson and Posner probabilities). However, it does not create an isolated benefit for UAI participants — it slightly lowers the risk. At the margin, this may be enough to convince some people that the risk is worth it, however does not fundamentally alter the analysis.

It can therefore be seen that the solution to condom use amongst MSM for hook-ups may be a purely rational response to the knowledge that HIV can be spread through UAI. As the market functions efficiently, there is no need for any informal norms to correct it. Market signalling mechanisms may provide an alternative way to view sexual behaviour which is of greater importance than informal norms.