[At points in this post I felt like some statements needed elaboration, and so I’ve put footnotes in for those who are interested.]
The Place of Signaling in Economics
Recently I have noticed instances of technical terms from economics diffusing into mainstream political discourse. The term “rent seeking” is one example of this. In economics rent seeking describes a situation where a monopoly privilege is up for grabs and firms or individuals spend resources competing with one another trying to win it. Its colloquial meaning is more expansive: I have seen people label as “rent seeking” any earned income that doesn’t seem to be derived from any wealth creation .
A similar thing seems to have happened with the term “Signaling.” Before I explain what Signaling is, I’ll try to put the concept within the context of various economic models of information. The simplest possible model is perfect information—all relevant facts are known to everyone. An example of this would be a labour market where firms know the productivity of all workers. There are other cases which amount effectively to perfect information. If there is an unobservable trait of interest, and if both parties can communicate and had no incentive to lie to one another this would lead to effective perfect information.
Continuing with the example of the labour market, suppose firms are hiring workers and want to make a wage offer based on productivity. However, suppose productivity is completely unobservable (although the distribution of productivity is common knowledge to all parties). If this were the whole situation, firms would just make wage offers based on the known average of productivity of the population .
Now suppose there existed a voluntary, costless test that revealed your productivity level. What would happen? Naively, you’d think that firms would only know the productivity of those taking the test. But the logic of Signaling shows that by refusing to take the test you’ve effectively revealed your productivity. The result of this test is that only those of the very lowest productivity would decline to take the test. This can be easily seen to be an “equilibrium”—a situation in which there is no incentive for any parties to change behaviour . Suppose firms believed that if you refused to take the test you’re of the lowest productivity. Therefore anyone whose productivity was above the lowest level would be incentivised to take the test. This in turn validates the beliefs of firms. (This is also the only equilibrium, see footnote  for proof.)
This is the basic reasoning underlying “Signaling.” Firms don’t know productivity. But they do know the distribution of productivity and the incentives the workers face. They deduce that anyone whose productivity is above the lowest level stands to gain by taking the test. Therefore refusing to take the test is a signal of having the lowest possible productivity. Although you haven’t formally revealed your productivity, your actions betray your hidden knowledge.
A more general notion of Signaling is this. Signaling arises in situations where one party has more information than the other and some people do not have an incentive to be entirely truthful (and both parties know this). Typically the information in question is about the level of a trait which is, for all practical purposes, unobservable. Because of the incentive to lie, for information to be convincing it must be transmitted through (observable) actions. But not just any actions. Such actions must have the feature that the cost  of doing them is lower for people who are expected to possess more of the desirable trait in question. With education, for example, it is fair to assume that education is less unpleasant (and so less psychologically costly) for people who are smarter, more conformist and people who enjoy work more. All of these things are positively correlated with higher productivity, and so education is a signal of higher productivity—even if education itself has no effect on productivity . As a result, any worker above the lowest level of productivity may have an incentive to consume education to signal their productivity.
Something to notice in these examples is that the more everyone else is Signaling, the more you do too. This is because Signaling is always about communicating relative rather than absolute worth; it always involves trying to do things that distinguish you from people who possess less of the unobservable, desirable trait. In our earlier example, if you’re a lower-than-average productivity worker and nobody else is taking the productivity-revealing test, it pays you not to take the test. It’s only once a certain proportion of others are taking the test that it becomes in your interest to do the same.
Understanding Social Justice activism through Signaling Theory
I won’t bore the reader with an explanation of what “social justice” means. I assume that if you’re reading this article you’re already familiar. Instead I’ll go straight into how I believe you can understand the behaviour of social justice advocates through the idea of Signaling.
To construct a theory based on Signaling you need two key ingredients:
- an unobservable trait that, if observable, would be rewarded with greater amounts of the trait
- an observable action (or set of actions) which is expected to be less costly the more of the unobservable desirable trait you possess
That is, the trait to be signalled (1) and the signal itself (2).
A particularly lucid example of Signaling is the following. Suppose you’re a Catholic who wishes to signal your commitment to Catholicism for whatever reason (maybe stronger commitment gets you more respect). One thing you could do is refrain from murder, since this is an important Catholic doctrine. However this is an extremely weak signal. Why? Because the costs of refraining from murder are very low, whether you are a Catholic or not. If, however, you refrain from using a condom, that is a strong signal of Catholicism. Why? Because those with weak commitment or no commitment to Catholicism, and so do not believe so strongly in the immorality of condoms, regard the net costs from refraining from using a condom as significantly higher.
I believe that a similar things is true for social justice advocates. The unobservable trait in question is fairly obvious: strength of devotion to social justice causes. Anyone who observes mainstream culture today understands that there is a social premium on being perceived to be supportive of social justice. To be more in favour of social justice is to be more moral and more respectable in today’s Western society, to obtain a higher position in an informal moral hierarchy. Even if many people dislike social justice advocates, if the advocates themselves regard their strength of belief as a virtue, they will try to signal it. Simply saying that you’re committed to leftist causes counts for nothing, because almost anyone can do that—so Signaling is required. That’s ingredient (1).
Ingredient (2) is less obvious. I believe the signal is simply the positions social justice advocates take. It seems plausible to me that many things that social justice advocates believe many people would find unpleasant to believe. Examples of this are the idea that the riots in Baltimore and Ferguson were a good thing and that whites should pay reparations to nonwhites. One source of unpleasantness is also believing things with less supportive/more conflicting evidence. This is not to say the average person has especially rational, evidence-based views, but intensely believing something that goes against common sense and isn’t really supported by experience is something most people find difficult.
How does a theory of social justice based on Signaling differ in predictions from one in which social justice activists have objectives and try to meet those objectives? Take the case of police shootings of black Americans, which has received a lot of media attention recently. If “social justice” were the goal of social justice activists, you’d expect them to focus their energies on cases where the evidence against the police is strongest. However the Signaling theory predicts the opposite. The stronger the evidence against the police, the more likely someone weakly committed to social justice (or not committed to social justice at all) will side against the police. Therefore the signal being sent is weakest in those cases. The signal is in fact strongest when the evidence against the police is weakest.
A case study in this is to compare the Michael Brown shooting with the death of Eric Garner. In the case of Michael Brown, all the evidence available supported officer Darren Wilson’s story. In contrast, in the Eric Garner case, many conservative figures such as Bill O’Reilly thought the police were at fault . The Michael Brown case caused riots and received much longer and more intense coverage than the Garner case. In fact, it seemed like the media attention on Garner evaporated the moment it was clear that conservatives were siding against police.
More generally, if “black lives matter” was really about protecting black lives, you’d think they’d be concerned that more black Americans are killed due to sneakers alone than are killed by police. You’d also think they’d be concerned about the so-called Ferguson effect. However, from a Signaling perspective, both of these blind spots make perfect sense. Everyone can see that deaths over sneakers and other black-on-black crime is bad, and it’s precisely for this reason that it’s ignored—it’s a weak signal.
Conversely, the Signaling theory explains why certain ideas, e.g. libertarianism, have failed to penetrate into mainstream social justice. Many libertarians  believe that libertarians should sell libertarian solutions to social justice problems. This sounds good superficially—you may be able to make a decent case that government intervention is bad for black communities, for example. However the Signaling theory predicts this will never succeed. The reason is that libertarian solutions can be (and are) advocated by people who have little to no commitment to social justice. From a Signaling perspective it is therefore a weak signal. It will always lose out and be displaced by stronger signals, such as calls for reparations to be paid to nonwhites .
This leads into why “social justice” is a black hole. Since social status is a purely relative thing, associating with social justice advocates—unless you step up your Signaling—pushes down your social status. This encourages more Signaling. Furthermore, to be an effective signaller means being as dogmatic in one’s thinking as possible (there is nothing worse than a devil’s advocate!). In this sense, entering into social justice is like crossing an event horizon. It is a self-reinforcing downward spiral of greater and greater Signaling. In principle a person deep within social justice could completely renounce social justice in all its forms and therefore escape, but in practice people generally change their views on a piecemeal basis. Since any marginal step away from social justice orthodoxy is harshly punished, many social justice activists are mentally trapped.
This theory I think is much better than alternative theories, such as that social justice advocates are stupid or irrational . It accurately accounts for a wide range of behaviour and can predict future behaviour. Recently there was a shooting in North Miami in which a black American was shot by police while lying down with his hands in the air. Because the evidence against the police is strong (or perceived to be, at least) I predict that there will be no riots, few protests and the media attention it has received will be fleeting. After reading this post, I hope the reader will begin to notice how well a lot of social justice behaviour, which may otherwise seem mysterious, fits the Signaling model.
Resources on Signaling
Michael Spence’s Nobel Prize Lecture
Less Wrong article on Signaling (see Robin Hanson’s posts in particular)
 Suey Park, who I’ve often seen labelled as a rent seeker, is not a rent seeker in the technical economic sense. Even though she creates no real wealth, her occupation is not shielded from competition so there is no “economic rent”.
 This could lead to high productivity workers leaving the market to find work elsewhere and low productivity workers entering the market. This would be a case of “adverse selection.”
 Let’s be more concrete about what is actually meant by equilibrium in a Signaling context. The recipients of the signal have beliefs about what information a signal conveys. That in turn determines the rewards given to different Signaling behaviour. That in turn determines how people with different levels of the unobservable trait will signal. Knowing the incentives faced by the signallers, the recipients of the signal form beliefs about what each signal means. This is a complete circle. Beliefs → Reward structure → Behaviour of signallers → Beliefs. There is said to be an equilibrium if the beliefs at the end are the same as the beliefs at the beginning. Interestingly, there can be (and often is) multiple equilibria—there exists more than one set of beliefs which is “self-validating.”
 Suppose some workers took the test and some didn’t. Those who did will receive wages according to their revealed productivities; those who didn’t will receive wages according to the average productivity of all those who didn’t take the test. Suppose for contradiction that amongst those who didn’t take the test, not all of them were the same productivity. Therefore some of them will find themselves with productivities above the average productivity of all those who didn’t take the test. They will therefore earn higher wages if they take the test, so they are incentivised to change their behaviour. Therefore under equilibrium all those who do not take the test must all be of the same productivity. Secondly, they must all be of the lowest productivity. If this were not the case, then there would be an incentive for those of the lowest productivity to take the test—which is a change of behaviour and so not an equilibrium. Therefore under equilibrium, the only people who refrain from taking the test are at the very lowest productivity, QED.
 “Cost” here does not necessarily mean monetary cost but could include any psychic costs associated with the action. We treat psychic costs as being equivalent to some quantity of money.
 Consider the simple case where there are two kinds of workers, low productivity and high productivity, and education has no effect on productivity. Suppose also there’s an “education threshold” E* where if you consume more education than this level, firms will assume you’re a high productivity worker, otherwise they assume you’re low productivity. What level of E* would lead to an equilibrium in which firms’ beliefs are continually validated? E* can be any level such that:
(1) for low productivity workers, the costs of consuming education beyond the level E* are more than the benefits in terms of wages.
(2) for high productivity workers, the costs of consuming education beyond the level E* are less than the benefits in terms of wages.
From this, it is never in the interest of a low productivity worker to consume more than E*, but it is in the interest of high quality workers. As a result, only high productivity workers consume above E* and only low productivity workers consume below E*. This validates the beliefs of firms, so it is an equilibrium. So by consuming more than E*, a worker convincingly signals higher productivity.
 In the real world, it is of course an empirical question to what degree the wage gains from education are merely due to Signaling. The economist Bryan Caplan is a particularly outspoken proponent of the idea that education is primarily Signaling, though many economists disagree. (If you’re asking for my personal view, I agree with Caplan.)
 As it turns out, I don’t actually believe the police were to blame for Garner’s death. But that’s beside the point here. What matters is that many people perceived the evidence against the police as being strong, and it’s perceptions that govern behaviour.
 e.g. the so-called “Bleeding Heart Libertarians”, people like Cathy Reisenwitz etc.
 the tendency of stronger forms of Signaling to displace weaker forms could almost be regarded as the social equivalent of the second law of thermodynamics: dS ≥ 0, where S stands for Signaling.
 They may be both of these things. But their irrationality is a kind of “rational irrationality” (see Bryan Caplan’s The Myth of the Rational Voter) driven by Signaling. There is method to the madness.
 I also like it because it shows how economic theory can be used to understand a lot of behaviour outside the market. It is also illustrative of Milton Friedman’s point (The Methodology of Positive Economics, 1953) that people don’t need to understand optimisation theory to obey economic laws (like how a billiards player doesn’t need to understand the relevant physics to be a good player). People do not need to have ever heard the concept of Signaling to be controlled by it.