A Theory of Cancel Culture

The standard U.S. real estate commission is 6%. That means a $1M house costs $60,000 to sell. How is this possible?! Two of my colleagues at UT’s Salem Center, John Hatfield and Richard Lowery, along with their co-author Scott Kominers, offer an explanation in their “Collusion in Brokered Markets.” Despite the vast number of realtors, HKL argue that collusion can be sustained because competing realtors typically need to cooperate to sell houses. It’s a classic network industry. Due to this interdependency, realtors can arrange the following “network exclusion” strategy to deter discounting:

(a) Refuse to cooperate with any realtor who discounts.

(b) Refuse to cooperate with any realtor who cooperates with any realtor who discounts.

In other words, punish the discounter and punish anyone who fails to punish the discounter.

Here’s how HKL explain (a):

In our setting, agents can maintain high prices by refusing to work with any agent who undercuts either of the “agreed upon” buyer and seller prices. This endogenously lowers the quality of a price deviator: a price deviator can no longer facilitate transactions between his buyers and another agent’s sellers (or between his sellers and another agent’s buyers). This implies that cutting prices by a small amount is not enough to attract buyers and sellers, as they understand that any price deviator will find it much more difficult to facilitate transactions. As a result, it is possible to maintain prices above marginal cost while ensuring that a price deviator who does attract buyers and sellers will not profit from his actions.

Here’s how HKL explain (b):

Of course, it must be incentive compatible for agents to refuse to work with a price deviator. Here, we rely on the repeated interactions among agents: A non-price deviating agent is willing to forgo working with a price deviator today if he is sufficiently rewarded for doing so in the future. To see why, we note first that the non-price deviating agent’s forgone profits are small, since the non-price deviating agent only has a small fraction of the buyers and sellers. Second, the non-price deviating agent is incentivized to follow the prescribed punishment strategy as, if he does not, future play reverts to a equilibrium in which he obtains no profits in every future period. By contrast, if every non-price deviating agent punishes the price deviator as prescribed, future play moves to a collusive punishment phase in which every agent other than the price deviator obtains positive rents. Thus, a non-price deviating agent will be willing to forgo working with a price deviator today for reasonably high discount factors—even as the number of agents becomes arbitrarily large.

If you prefer a diagram:

While I’ve yet to meet Kominers, I’ve had many arguments with both Hatfield and Lowery about the empirical relevance of their model. They’re brilliant scholars, but after years of discussion, their story continues to seem crazy to me. For two big reasons:

1. Organizing such omnipresent punishments would require a massive organizational apparatus, but there is only anecdotal evidence that even modest punishments exist.

2. Plenty of realtors publicly offer discounted commissions without being harshly punished.

HKL do offer a little empirical evidence, but to my mind, it’s paper-thin. You decide:

One might be incredulous that real estate agents would, in fact, engage in such behavior. However, such behavior is well-documented: In one particularly harrowing tale, Birger and Caplin (2004) reported that the proprietress of a new discount realty, You Win Realty, was not only boycotted by other realtors in the area but also harassed by threatening phone calls and the like; indeed, one competing realtor threatened to report her to the Federal Trade Commission(!). Taking a broader view, a report by the Department of Justice and Federal Trade Commission (2007) documented that such “steering” behavior (away from discount realtors) is commonly reported and an earlier investigation by the Federal Trade Commission (1983) found that 84% of alternative brokers reported frequent or occasional “refusals by other brokers to show homes listed by [their] business,” with 49% saying that this was a frequent issue. Relatedly, Barwick et al. (2017) documented that properties with lower commission rates are both less likely to sell and take longer to sell.

Remember the base rates for human misunderstandings: With over three million licensed realtors in the U.S., we should expect tens of thousands of intra-occupational accusations even if nothing unusual is afoot. If millions of realtors were part of massive multi-tier punishment strategy, we would see vastly more accusations of much more extreme behavior. Indeed, we should expect the news to be full of stories of the great realtor conspiracy spreading fear with a fearsome toolbox that includes a little arson and murder. At minimum, social media would be full of audio and video recordings of realtors threatening realtors. These wouldn’t just be threats against realtors who cut their rates. There’d also be lots of threats against realtors who fail to make threats against realtors who cut their rates!

Nevertheless, I maintain the HKL model provides deep insight into a totally different facet of the human experience: “cancel culture.” You know the drill: Someone says something “offensive.” A few activists declare themselves to be “outraged” and demand the offensive person’s professional ostracism. This, in turn, makes many onlookers feel like they have to echo the activists’ outrage and comply with the activists’ demand for ostracism. After all, if they continue to consort with the offensive person - or publicly defend him - they may share his fate. Every time one onlooker echoes and complies, this intensifies the pressure for other onlookers to echo and comply. If the stars align, you’ve got a mob going - and the offensive person proverbially “never works in this town again.”

Why is HKL plausible for cancel culture but not for real estate? For two big reasons:

1. While organizing such omnipresent punishments would require a massive organizational apparatus, the woke movement in culture, business, and government is precisely such an apparatus. They operate in plain sight en masse at the highest levels of society.

2. Few public figures outside of professional politics publicly say offensive things without being harshly punished. Even towering figures like J.K. Rowling have been harshly punished for absurdly mild heresies against the woke religion.

In sum, HKL is a classic case of an insightful theory applied to the wrong facts. Yes, social pressure can maintain dysfunctional social norms, even when the number of actors involved in large. Yes, to maintain these norms, you have to punish not only the initial norm violator, but everyone who fails to join in punishing the initial norm violator. There’s almost no sign this is happening in the market for realtors. The best story of 6% commissions, as I’ve explained, is the banal fact that lots of customers fail to shop around for a good price. With cancel culture, in contrast, the model fits like a glove.

Why the disparity? Simple: To maintain harsh punishments a la HKL - including harsh punishments against the merciful and apathetic - you need intense moral passion, which fuels moralistic aggression, which in turn fuels fear of self-righteous anger. Realtors, a famously affable lot, generate almost none of this. Fanatical religions - including secular religions like wokism - possess the whole package. Which is why reasonable people with tenure have a grave responsibility to show a little courage. I’m no Salman Rushdie, but mobs of angry cyber-strangers won’t keep me from speaking my mind.