AOC is right – we shouldn't accept the heckler's veto
On Saturday, a group of congressional Republicans wrote to President-elect Biden, asking him to dissuade Speaker Pelosi from pursuing a second impeachment of President Trump — “in the spirit of healing”.
In an interview on ABC’s This Week, Representative Alexandria Ocasio-Cortez responded to the letter:
Rep. Ocasio-Cortez pushes back against some GOP lawmakers suggesting a second impeachment of Pres. Trump would threaten unity: "The process of healing is separate and in fact requires accountability... because without it, it will happen again." https://t.co/wbex0bS4Sb pic.twitter.com/6QcDczitua— ABC News (@ABC) January 10, 2021
You can click above to watch the full clip, but here’s an excerpt:
“the process of healing is separate and in fact requires accountability… if we allow insurrection against the United States with impunity, with no accountability, we are inviting it to happen again”
She’s absolutely right. The subtext of the Republicans’ letter is essentially: we can’t impeach President Trump for inciting an insurrection… that’ll make him do it again!
It’s a version of the heckler’s veto: a concept from First Amendment law that describes the government restricting free speech due to the anticipated or actual reactions of a third party.
The heckler’s veto is unconstitutional in the United States. But even without the protection of the First Amendment, there’s an obvious reason why governments should not accept this logic: it incentivizes threats of violence!
In fact, it might incentivize actual violence, since threats alone may not be deemed credible.
If the Democrats decide not to impeach Trump a second time because they fear that he or his supporters will react violently, they are creating incentives for those individuals to continue issuing violent threats. As AOC said, it would be inviting it to happen again.
And yet, while we might find it easy to dismiss an argument made by congressional Republicans - hardly known for good faith debate - this same type of argument crops up in many social science papers.
Typically, it goes like this:
- Researchers demonstrate that a policy has an unintended negative effect
- Researchers take this unintended negative effect as fixed or unavoidable
- Researchers argue that policymakers should consider whether the policy is still attractive once the unintended negative effect is taken into account
Steps #1 and #3 are actually pretty innocuous. It’s step #2, which is often left implicit, where the trouble comes in.
Let me show you what I mean with a couple of examples from recent social science research.
Example 1: Policing the police
Consider this paper by Tanaya Devi and Roland Fryer, both economists at Harvard.
In the paper, Devi and Fryer look at how federal and state investigations into policing affect crime rates. Taking the data in aggregate, there is no relationship: investigations into policing do not affect subsequent crime rates.
However, a subset of those cases show a very different pattern. When the investigation was sparked by a “viral video” of a police shooting, there is a large jump up in homicides and total crime.
Here’s how Devi & Fryer interpret these results:
Put plainly, the causal effect of the investigations in these five cities – triggered mainly by the deaths of Freddie Gray, Laquan McDonald, Timothy Thomas, Tyisha Miller and Michael Brown at the hands of police – has resulted in 893 more homicides than would have been expected with no investigation and more than 33,472 additional felony crimes, relative to synthetic control cities.
To get a sense of how large this number is, the average number of fatal shootings of African American civilians by police officers in Baltimore, Chicago, Cincinnati, Riverside and Saint Louis, per year, is 12. Thus, even if investigations cured these cities of all future civilian casualties at the hands of police, it would take approximately 75 years to ‘break even.’ Our estimates suggest that investigating police departments after viral incidents of police violence is responsible for approximately 450 excess homicides per year. This is 2x the loss of life in the line of duty for the US Military in a year, 12.6x the annual loss of life due to school shootings, and 3x the loss of life due to lynchings between 1882 and 1901 – the most gruesome years.
Devi & Fryer are quite open about the mechanism here: investigations following a viral incident lead to a decrease in the “quantity of police activity”. As they put it: “[i]f the price of policing increases, officers are rational to retreat”.
The glaring omission here is the authors’ refusal to see police as strategic actors, despite a long history of police unions engaging in work stoppages as a form of protest.
To cite just two examples:
In 2014, the NYPD engaged in an illegal work stoppage to protest Mayor Bill de Blasio’s response to the murder of two cops. Arrests plunged by 66%.
In 2020, Minneapolis police slowed down their response to 911 calls in the district of Steve Fletcher, a city councilman who had sought to prioritize violence prevention over hiring new officers. Fletcher said the police union was operating “a little bit like a protection racket”.
With this context, we might be inclined to think that what police do when facing extra scrutiny is not retreat, but retaliate.
This distinction matters when it comes to policy recommendations. If you view police work stoppages as a “rational retreat”, you might advise policymakers to internalize those costs, and to avoid investigating police departments after a viral incident.
But if you see this behavior as retaliation, then the policy recommendation is quite different. The fact that police engage in illegal work stoppages after government investigations is not a reason to avoid investigations! It is further evidence that many police departments are rogue institutions that operate outside of democratic accountability.
It demands more action, not less.
Example 2: Sexual harassment training
Here’s another example, this time from the field of sociology.
Frank Dobbin and Alexandra Kalev have written several papers arguing that mandatory diversity training and sexual harassment training programs are ineffective and may lead to reduced numbers of women and minorities in management positions.
The suggested mechanism for this effect is the idea that mandatory training programs - especially those that might portray employees as potential wrongdoers - can cause a backlash.
Here’s how Dobbin & Kalev put it in an article for the Harvard Business Review:
Trainers tell us that people often respond to compulsory courses with anger and resistance — and many participants actually report more animosity toward other groups afterward.
Now there are all sorts of important issues here. One is that Dobbin & Kalev never really grapple with the fact that “diversity training”, especially considered over a 50-year period, is not a well-defined intervention. There’s huge variation in goals, content, and implementation. More importantly, the quality of these training programs likely correlates with some unobserved characteristics of the company, making a causal interpretation very tricky.
Leaving those issues aside, though, what we have here is a classic case of the heckler’s veto. Dobbin & Kalev argue that mandatory diversity training for employees may cause some of them to become angry - so angry, in fact, that they might create an even more hostile environment that will lead to even fewer women in management positions years later.
In fact, I saw how this type of argument can affect real-world decision-making when my department was embroiled in a sexual harassment scandal a couple of years ago.
While many (including me) felt that this decades-long cover-up warranted an external review by an independent investigator, others argued that such a review might provoke a backlash among senior faculty and administrators, who might feel victimized and refuse to provide information. (Indeed, Frank Dobbin himself was consulted on this point.)
In the end, Harvard President Larry Bacow chose a compromise between these two positions. He initiated an external review, but with an important carveout:
we are not asking the committee to review the behavior or decisions of individual members of the Harvard community in regard to the Dominguez matter
Now I know that the word “Kafka-esque” is overused… but this was, and remains to this day, one of the most ridiculous things ever written. Honestly, they should pay university administrators more, just to compensate for the abject loss of dignity that comes from having to write a sentence like that.
I mention this as a cautionary tale. This is what happens when we accept the logic of the heckler’s veto.
Don’t scrutinize powerful people, in case they stonewall the investigation. Don’t make men attend diversity training, in case they get angry. Don’t investigate police brutality, in case they hold an illegal strike. Don’t impeach Trump, in case his supporters riot.
Rejecting the heckler’s veto doesn’t mean we should be indifferent to the consequences of policy choices. It means that we should think of these consequences the same way we think about everything else: a product of circumstances!
If a policy intervention produces an undesired result, then perhaps it should be abandoned. On the other hand, perhaps it needs to be strengthened, or supplemented with a separate intervention.
As researchers, we choose to consider certain social phenomena as “fixed” or “variable” in different contexts, often for convenience. But if these assumptions are not made explicit, they can end up distorting our policy recommendations.