Externalities, Rights, and the Problem of Knowledge

On the Difficulty of Designing a System of Rights

Disclaimer: Today’s essay is more academic than usual!

Very short summary: This essay explains how the knowledge problem applies to the definition of jurisdictional rights. Jurisdictional rights define spheres of individual sovereignty. Rights are appropriately defined if they internalize all potential externalities. However, individuals may disagree about what counts as an externality. This disagreement stems from individuals’ preferences, which are typically dispersed and local. I discuss various solutions to this problem, including the use of polycentricity. Finally, I show the relevance of this issue for the debate about the proper formalization of rights in social choice theory.

---

I don’t generally advertise my academic work here, but I’ll make an exception today by mentioning the recent online publication in the political philosophy journal Res Publica of my article “Living in Disagreement: Public Reason and Jurisdictional Rights.” I mention it because it’s related to a few essays I published on this newsletter (here, here, and here) and because it touches upon an insight about the definition and justification of rights that I’m only beginning to grasp. Recent readings made me realize that this insight was somehow already absorbed by the public choice tradition – at least in the Ostroms’ Bloomington School.[1] However, it’s interesting that no explicit connection is made with a related debate about the conceptualization of rights in social choice theory (I addressed this debate several years ago here). What follows is an attempt to connect the dots.

The Knowledge Problem

Many readers will be familiar with the so-called “knowledge problem.” The knowledge problem was first identified in the economic calculation debate that opposed neoclassical economists like Oskar Lange to Austrian economists Ludwig von Mises and Friedrich Hayek. To summarize quickly, in the 1920s, von Mises asserted the economic impossibility of socialism on the ground that the lack of market prices would undermine entrepreneurial activity and the rational allocation of resources.[2] Socialist neoclassical economists responded that, based on general equilibrium theory, it is in principle possible for the socialist planner to simulate the market process and infer market prices analogically to the Walrassian auctioneer. Hayek decisively refuted this argument by pointing out that market prices can be found in the Walrassian model only because general equilibrium theory assumes that we can write a system of equations summarizing all the information about consumers’ preferences and producers’ production costs. But that’s precisely the information that the socialist planner would not have. This information is private, local, and dispersed. Market prices emerge from competition mechanisms that themselves create by incentivizing agents to make a rational allocation of resources based on information that reflects preferences and production costs. Market processes reveal information and knowledge that would otherwise remain concealed from agents.

The knowledge problem can be reformulated as follows. In any society with a sufficient degree of division of labor, individuals will pursue different plans based on private and local knowledge. Except in exceptional circumstances, these plans are partially incompatible and individuals pursuing them will enter into conflict. Because knowledge is private and dispersed, it is virtually impossible for a single agent to design the “optimal” configuration to solve these conflicts. In the economic domain, that means that there is no optimal allocation of resources to be discovered by the mind alone. This optimal allocation cannot be computed, and any attempt to enforce an allocation by design will illegitimately coerce people to act against their will.[3]

The knowledge problem doesn’t only apply in the economic domain – though it is crucial there because economic resources are instrumentally essential for individuals to pursue their various ends. All domains of society face the knowledge problem. People have tastes, values, skills, knowledge, and other relevant characteristics that determine what they do, what they want (or don’t), and what they expect others to do and want. Information about these characteristics is typically private, dispersed, and local. In many cases, it cannot be easily communicated because it corresponds to tacit knowledge, i.e., know-how based on practice and experience. This typically makes coordination difficult and likely to lead to conflicts. As market prices don’t exist in all domains of social life, society must find other coordination mechanisms.

Rights and Externalities

Among them are collective decision mechanisms such as majority rule. Note that these mechanisms are by definition coercive and, unless individuals happen by chance to all want the same things, a more or less large portion will impose its will on the minority. Other mechanisms promote coordination by turning private information into public information. Rituals, public events like the Superbowl, and platforms that rate users are all instances of such mechanisms that aim to create common knowledge.[4] Finally, we have rights, and especially jurisdictional rights that assign to individuals a private sphere of sovereignty. These include rights to use and to exclude others from using goods (property rights), rights to associate with anyone (and to disassociate from them) to pursue a common objective (association rights), and rights to keep information about oneself private (privacy rights).

The existence of property rights is directly related to market prices. To some extent, this is also true of association rights (e.g., people can be asked to pay a fee to become members of a club). With respect to the coordination issue, privacy rights are interesting because they seem to contribute to the very problem they’re meant to solve. By giving people the right to keep some information private, they seemingly sustain the dispersion of knowledge that causes coordination failures and conflicts. Of course, privacy rights respond in part to ethical considerations that have nothing to do with social coordination. But like other jurisdictional rights, privacy rights aim to achieve coordination by exclusion. The basic idea is quite straightforward: in many cases, individuals’ plans conflict because they involve aspects that concern other persons’ lives. By strictly delimiting spheres of sovereignty, we basically tell people that what they think or want about other persons’ lives is irrelevant . You may want me to believe in god and act on this basis, but my freedom of conscience and the related privacy right tell you that this is none of your business. What you think or want is irrelevant.

Compare how social decision mechanisms like majority rule on the one hand, and jurisdictional rights on the other, solve coordination problems. The former imposes conformity by coercing the minority to follow the will of the majority. The latter, quite the contrary, maintains diversity. Ideally, if the allocation of rights is accepted by everybody, they permit “unanimity without conformity,” to quote Milton Friedman.[5]

However, note the proviso: if the allocation of rights is accepted by everybody. I said above that jurisdictional rights achieve coordination by exclusion. This is true only if rights effectively create exclusion. This will not always be the case, as we know from property rights. Externalities precisely exist when a system of (property) rights fails to internalize all the effects that the use of rights generates. Market prices then fail to reflect all the information and improperly incentivize agents, leading to coordination failures and disputes about the compensation of external costs. What is true of property rights is true for all jurisdictional rights: if there are external effects, then coordination will fail, and conflicts will ensue. We may then have no choice but to use some collective decision mechanisms.

“Women Reading a Letter,” Johannes Vermeer (1663)

Back to the Knowledge Problem

All this is fairly uncontroversial and well-known. What is less often noticed is that this leads us back to the knowledge problem. Put yourself in the shoes of a hypothetical social planner who wants to design a system of jurisdictional rights that eliminates external effects. Several difficulties emerge. First, as there are transaction costs (even more so for association and privacy rights), we know that the initial allocation of rights matters – the main insight of “Coase’s theorem.”[6] Second, the social planner may naively think that externalities can be objectively identified. But that’s a mistake: external costs (as all opportunity costs) are always relative to individuals’ preferences. Third, the social planner doesn’t know those preferences and cannot therefore identify and measure externalities before they are revealed by individuals’ behavior.

The classical examples of externalities illustrate these three points. Insofar as pollution resulting from economic activities harms individuals, the harm is relative to subjective perceptions and individuals’ willingness to “trade” this harm against something else (e.g., money). Not everybody will be impacted in the same way by the same quantity of pollution. Some people have health issues that make them more sensitive than others. Individuals don’t give the same importance to the future, and some are likely to discount future harm caused by pollution more than others. The difficulty here (and this is of course one of the main objections to Pigouvian taxes) is that the social planner hardly has the information about individuals’ “marginal rates of substitution,” i.e., they don’t know how different persons are subjectively harmed by pollution. That’s an advantage of Coasean-type solutions: that they in principle don’t need this information – in principle only, because in practice the fact that transaction costs are positive means that how we initially allocate rights matters.

The bottom line is that individuals will, most of the time, disagree about the value or even the very existence of externalities. Because one of the purposes of (jurisdictional) rights is to internalize externalities, disagreement about externalities is likely to trigger disagreement about how rights should be allocated. There are then two possibilities. Either rights are alienable and can be traded against each other, in which case Coasean bargaining solutions are available, still with the proviso that transaction costs can be quite high. Or rights are not alienable (e.g., the right of free speech), which therefore excludes any kind of bargaining. How do you allocate rights in these circumstances, where knowledge about subjective valuations is dispersed and local?

Disputes about rights are mostly due to the fact that we disagree about what counts as externalities and how to value them. This disagreement means that, even if we all agree that individuals should all have their private sphere of sovereignty, where these spheres stop is a matter of contention. Hence, a contradiction emerges: on the one hand, a universalist moral stance indicates that we should all have the same jurisdictional rights (we are all moral equals); on the other hand, the fact that we disagree about externalities suggests that rights should be allocated differentially. The knowledge problem makes the latter solution hardly implementable.

What About Polycentricity?

There may be a third way. Let people freely associate based on the system of rights that they consider best in light of their subjective preferences. The advantage of this polycentric solution is twofold. First, it relies on spontaneous sorting mechanisms that help to reveal individuals’ subjective evaluations of externalities. For instance, people who voluntarily regroup in a jurisdiction where the right of free speech is unlimited signal that they don’t perceive speech as causing any harm, contrary to individuals who choose to live in a jurisdiction where speech is severely constrained. Therefore, polycentricity can partially overcome the knowledge problem. Second, it limits disputes over rights as, by assumption, individuals can “exit” jurisdictions where the prevailing system of rights is causing externalities to them.

Still, polycentricity is not a panacea. Besides the usual concern with mobility costs that may impede individuals’ ability to exit and move where the right system is more appealing to them, externalities are unlikely to stop at the (territorial or not) borders of jurisdictions. This is particularly the case if individuals from different jurisdictions have to interact. More generally, polycentric solutions for externalities, including public goods (or “bads”), work well when what Vincent Ostrom, Charles Tiebout, and Robert Warren call “packageability” is possible, i.e., when external effects can be kept within the borders of the jurisdictions.[7] The possibility of packageability depends, however, on people’s perceptions and preferences, again information that nobody has ex ante. In all likelihood, there are external effects that cannot be entirely packaged in such a way, meaning that the design and implementation of jurisdictional rights is itself a public, rather than a “club,” good.

Rights in Social Choice Theory

To conclude, let’s relate this to the debate over the formalization of rights in social choice theory. This debate has been triggered by Amartya Sen’s “Paretian-liberal paradox.”[8] Sen demonstrated that there are circumstances where the requirement of Paretian-optimality – here understood as a unanimity criterion – is inconsistent with “minimal liberty.” By minimal liberty, Sen means that each individual should be decisive over at least one pair of social states in the sense that the social choice between these two social states should align with the preference of the sovereign person. As an example, Bob may be sovereign over whether to have long or short hair, while Ann can be sovereign over whether to wear trousers or skirts. In this simple example, there are four social states:

Here, Bob’s domain of sovereignty is over the pairs {(W, Y), (X, Z)} while Ann’s is over {(W, X), (Y, Z)}. If Bob prefers having long hair and Ann prefers wearing trousers and both don’t care about the other’s hair or clothes, then the social choice is Y. In this scenario, there is no difficulty because respecting each individual’s jurisdictional rights is compatible with unanimous preferences. Suppose, however, that Bob considers that women should wear dresses and that he is willing to cut his hair to have this preference satisfied. Hence, Bob’s preference ordering is Z > X > Y > W. In the meantime, Ann would prefer that Bob have short hair, so that her preferences are W > X > Y > Z. Therefore, Ann and Bob agree that X should be socially chosen over Y. But, given the system of rights and Ann’s and Bob’s preferences, social preferences must be such that Y is preferred to X.[9] We have a contradiction.

The roots of the problem lie in the fact that Ann’s and Bob’s preferences are conditional, i.e., they range over what the other is doing within their sphere of sovereignty. As Allan Gibbard has demonstrated, the paradox disappears if preferences are unconditional, that is, if each individual is indifferent about other individuals’ choices within their sphere.[10] How can we interpret conditional preferences? John Stuart Mill’s On Liberty can be viewed as defending the thesis that such preferences are normatively irrelevant and should be ignored. That seems plausible in cases like those described in this simple example. However, conditional preferences can also reflect the perception of externalities that are not properly internalized by the right system. In my example, substitute “doing research to develop nuclear weapons” or “sacrificing animals based on religious beliefs” for Ann and Bob’s choices. Is it still clear that the respect of rights should prevail over unanimity? Would the definition of rights be appropriate in these circumstances?

Whatever one’s stance on this, it’s a virtue of Sen’s social-choice theoretic formalization of rights to help us understand this kind of issue. Sen’s insight is that there is a deep connection between rights and preferences: the justification of a system of jurisdictional rights is conditional on individuals’ preferences and the status of these preferences. This insight is completely lost with the alternative, so-called “game form,” formalization of rights.[11] the game-form approach assumes that rights can be identified independently of preferences (a “game-form” is a game minus the matrix of payoffs representing preferences). While this may be relevant to capture “natural” rights, this is not the right approach to understand the problems I’ve discussed above. As soon as we accept that rights are socially designed (or evolved) and must be justified to everyone, we must account for the likely disagreement over externalities. Rights must be responsive to this disagreement, and any approach that ignores the connection between rights and preferences cannot explain why.

---

[1] In particular Paul Dragos Aligica et al., Public Governance and the Classical-Liberal Perspective: Political Economy Foundations (Oxford University Press, 2019).

[2] Ludwig von Mises, Socialism: An Economic and Sociological Analysis (Liberty Fund, 1922 [1981]).

[3] See particular the chapter on planning and democracy in F. A. Hayek, The Road to Serfdom: Text And Documents--The Definitive Edition, New edition (University of Chicago Press, 1944 [2007]).

[4] See Michael Suk-Young Chwe, Rational Ritual: Culture, Coordination, and Common Knowledge (Princeton University Press, 2003).

[5] Milton Friedman, Capitalism and Freedom: Fortieth Anniversary Edition, First Edition (University of Chicago Press, 1960 [2002]).

[6] R. H. Coase, “The Problem of Social Cost,” Journal of Law and Economics 3 (October 1960): 1–44.

[7] Vincent Ostrom et al., “The Organization of Government in Metropolitan Areas: A Theoretical Inquiry,” American Political Science Review 55, no. 4 (1961): 831–42.

[8] Amartya Sen, “The Impossibility of a Paretian Liberal,” Journal of Political Economy 78, no. 1 (1970): 1.

[9] Indeed, if you look at Bob’s rights and preferences, we have the following social preferences: Y >* W and Z >* X. Looking at Ann’s rights and preferences, we must have W >* X and Y >* Z. Hence, Y is socially better than W and Z, which itself is socially better than X. By transitivity, Y is socially better than X.

[10] Allan Gibbard, “A Pareto-Consistent Libertarian Claim,” Journal of Economic Theory 7, no. 4 (1974): 388–410.

[11] Wulf Gaertner et al., “Individual Rights Revisited,” Economica 59, no. 234 (1992): 161–77.