In recent years, there has been widespread concern that democracy is threatened by the spread of misinformation and “fake news” on social media and other similar new technologies. Rick Hasen’s recent book Cheap Speech: How Disinformation Poisons Our Politics-and How to Cure It, is an important addition to the literature making this case. Hasen and others argue that the problem can be mitigated by public and private actions to restrict the spread of misinformation on social media (though Hasen advocates more modest regulatory measures than some other commentators).
In a recent symposium on Cheap Speech at the Balkinization blog, Harvard Law School Prof. Guy-Uriel Charles explains how this case is weakened by the reality that the the demand for misinformation may be a more significant menace than the supply:
I…. wonder how we ought to think about the problem of disinformation and misinformation if we assume that the market for political information is operating efficiently and that the problem is not one of market failure, which is how Rick frames the issue. Rick defines cheap speech as “speech that is both inexpensive to produce and often of markedly low social value,” (21) and frames it as a problem of political market failure caused by information asymmetry (30). He uses as his model a pathbreaking paper by George Akerlof, the Nobel Prize winning economist, entitled The Market for “Lemons”: Quality Uncertainty and the Market Mechanism. In that famous paper Akerlof explored how the information asymmetry between sellers and buyers with respect to the quality of certain goods might result in a market in which lower quality goods overwhelm high quality goods and in a reduction in the size of the market. For example, if you’re a buyer in the used car market, you can’t tell whether a seller is offering a reliable used car or a lemon, though the seller knows. To hedge the risk that you’re buying a lemon, you make a lower offer. Potential sellers of quality cars are less likely to enter the market because buyers are unlikely to pay their asking price. The absence of sellers of quality cars leaves sellers of lemons in the market…..
It is unclear to me that the Akerlof model, which assumes that consumers are in the market for quality cars, is the right frame for thinking about political misinformation and disinformation. There are certainly some voters who are interested in truthful political information. But there are certainly a, perhaps larger, group of voters who are not in the market for truthful political information. We know, for example, that there is a relationship between partisanship and misinformation (see, e.g., here, here, and here). There’s literature, and debate, on the role of motivated reasoning on assessing the accuracy of information (see, e.g., here vs. here). Moreover, as some researchers have demonstrated, the demand may be asymmetrical (see, e.g., here and here; conservative or Republican voters may be more likely to believe misinformation and there is evidence of partisan asymmetry with respect to cures to misinformation. If voters are filtering information based upon their partisanship or other identities that are salient to them or if they are seeking information that is consistent with their priors, then the Akerlof model is less apt….
To the extent that voters are seeking information that is consistent with their partisan identities or confirms their priors, then the market is working perfectly. There is no market failure, given that the market is supplying precisely what the people want. Republicans seek and get the information they like; Democrats seek and get the information they like. Everyone gets to live within their echo chamber, and no one must be confronted with ideas and information that makes them uncomfortable. Of course, this is no way to run a democracy.
Cheap Speech is extremely compelling on its own terms. At the same time, Rick’s exhaustive exposition raises the question whether we have the right model for understanding the problem. If the problem of misinformation presents a demand-side problem, or to the extent that there is both a demand-side and supply-side problem, supply-side only solutions are not likely to resolve the problem. Similarly, to the extent that we have a supply-side problem, then demand-side solutions are not going to suffice…
If it is the case that political disinformation is at least about voter preferences as it is about politicians and social media platforms, solutions to the problem are much more complex. Modern democracies are not very good about figuring out what to do when voters get exactly what they want and what voters want is actually bad for democracy. Tweaking the law and relying upon private ordering is less than optimal, if the goal is a resolution of the problem. Rather, the focus will need to be on structural political and economic reforms.
I agree with almost every point Charles makes, and have made similar arguments in my own writings, including here, here, and here. The popularity of political misinformation is indeed due primarily to demand, rather than supply, which is why the problem long predates the rise of modern social media, and might well have been as bad or even worse in earlier eras dominated by what we today call the “legacy” media of newspapers and radio. The lies and disinformation that promoted fascism, communism, and other enormously harmful ideologies spread without the aid of Twitter and Facebook.
Charles is also right to emphasize that Akerlof’s “market for lemons” is a bad analogy for the market for political information. Potential buyers of used cars generally want to know the truth about the condition of the vehicle in question. They have strong incentives to seek out the truth, because their decisions on whether or not to buy the car will make a big difference.
By contrast, the low odds that any one will vote will make a difference to the outcome of an election ensure that many consumers of political information are acting not as truth-seekers, but as “political fans” eager to endorse anything that supports their position or casts the opposing party and its supporters in a bad light. These biases affect not only ordinary voters, but also otherwise highly knowledgeable ones, and even policymakers and politicians.
This demand for misinformation is the real root of the problem. If it were lower, the supply would not be much of a danger, and at the very least would not affect many voters’ political decision-making.
In recent years, right-wingers’ susceptibility to disinformation that confirms their priors has been especially notable, as in the case of Donald Trump’s lies about how the 2020 election was supposedly “stolen” from him. But, unlike Charles, I’m not convinced that the left is generally less susceptible to this problem. Social science evidence indicates that bias in evaluation of political information is roughly equal across the political spectrum. Each side is relatively more susceptible to misinformation that confirms their priors. Examples that appeal disproportionately to the left include 9/11 “trutherism” (discussed in my book Democracy and Public Ignorance ), and claims that GMO foods should be banned or tightly restricted because they are supposedly more dangerous than “natural” ones.
Finally, Charles is absolutely right that dealing with this demand-side problem requires “structural political and economic reforms.” Restricting social media is unlikely to accomplish much, because the demand for misinformation can easily be met by other producers. Fox News, a traditional broadcast media operation, is likely a far more significant spreader of right-wing misinformation than anything on social media. Left-wing misinformation is also readily spread on more traditional media. Indeed, across the political spectrum, many more people get their political information from TV news or media websites than from social media.
In theory, government regulators could suppress misinformation across the board, regardless of whether the producers are traditional media or new ones. But, in addition to constitutional problems, such policies are objectionable because they would give political leaders vast power over the spread of information. It’s far more likely they would use it to promote narratives that support their parties and policies than “objectively” promote truth in a “neutral” way. Even if you trust the current Democratic administration to wield such power responsibly, you probably don’t trust the Republicans – and vice versa.
In previous writings, such as my book on political ignorance, and a more recent article in National Affairs, I have argued that the right structural reform is to shift more decisions to formats in which people can “vote with their feet,” and thus have stronger incentives to seek out information and evaluate it objectively than ballot-box voters. This can be accomplished by limiting and decentralizing the power of government, thereby enabling people to vote with their feet over more issues.
The Akerlof “market for lemons” problem is a good example of how foot voting in the market can mitigate information problems. Because used-car buyers have strong incentives to seek out the truth, over time market mechanisms have arisen to provide relatively accurate and unbiased evaluations of used cars. Despite Akerlof’s fears, good used cars have not been systematically driven out by lemons.
Thus, the last time I sold an old car of mine, the prospective buyer and I took it to an independent mechanic to have him evaluate its condition. The buyer had a strong incentive to take it to a neutral expert evaluator, and to assess the latter’s report in an unbiased way. Few voters are willing to make the time and effort to do the same with political information. A world with more “markets for lemons” and fewer political markets would be a world where misinformation is a less serious danger.