In the latest edition of National Review, Charles Fain Lehman cites me, alongside noted authors Fordham’s John Pfaff and Harvard’s Thomas Abt, as subscribing to what he calls the “legitimacy hypothesis,” and uses this as a springboard to attack the idea that police behavior and perceived legitimacy may affect public safety. The question of police or governmental legitimacy's criminogenic effect is an interesting one, but Lehman's piece ultimately fails, in part, because he’s attacking something none of the three of us, as far as I know, have claimed: that eroded police legitimacy has sparked a national “crime wave.” Lehman’s argument doesn't confront the actual literature discussing legitimacy, instead focusing on the politics of it.

Lehman also prefaced his piece by saying the legitimacy hypothesis is an explanation for people “on the left.” I can’t speak directly to the political self-identification of either of the other men as I have not contacted them about this blog post—and, I should note, their credentials and accomplishments far surpass mine—but having spent the bulk of my career at market-friendly think tanks, including me among “the left” would make the categorization so inclusive as to render it almost meaningless. In the spirit of charity, however, I will grant that my worldview probably puts me to the left of the median NR subscriber these days. (I have since lost my once-treasured copy of the 1996 special issue titled “The War on Drugs is Lost.”) But our political affiliations are hardly the biggest flaw in Lehman’s argument.

Pfaff out of context

Lehman doesn’t directly quote Abt or me, choosing instead to summarize positions we’ve written elsewhere.  But he does quote Pfaff, saying that he

Has argued that violent crime “is the product of anger at police forces that kill far too many Black men, as well as at the remarkably violent, riotous way the police responded to last summer’s protests.”

This quote struck me as odd. For those not familiar with Prof. Pfaff, he is the author of Locked In: The True Causes of Mass Incarceration—and How to Achieve Real Reform. Among other things, the book blew up the widely held and disseminated belief, particularly among libertarians and progressives, that U.S. prisons are filled with nonviolent drug offenders via the War on Drugs.

Although a law professor, Pfaff is an econometric researcher who collects data and analyzes them in context. The data in Locked In unquestionably show that drug offenders, while making up a significant portion of federal inmates, comprise a small portion of the state prison population, where the vast majority of prisoners are held in the United States. The inescapable finding is that seriously reducing the number of incarcerated people would require policymakers to release a large number of people convicted of violent crimes and, going forward, incarcerate “violent” offenders for shorter periods of time.

Any responsible empiricist hesitates to draw broad conclusions without clear and convincing data. Far more common—as data are often messy and complicated—are qualified explanations or lists of possibilities that may explain trends in any given data or sets of data. Pfaff’s quote, as presented, thus appears to be far more sweeping than I am used to seeing from him.

Very curious, I investigated the quote’s origins. Thankfully, it was quite easy to find. Pfaff’s quote, in appropriate context, was in a lengthy article this past June in The New Republic:

The causal story linking violence to the protests is also plausible, but more complicated. The most common story is the so-called Ferguson Effect, which argues that police respond to protests by working less hard, which in turn leads to more crime. Rigorous studies, however, have generally failed to find any real connection between this sort of “depolicing” and homicide rates. There is also evidence that people’s views about the legitimacy—or illegitimacy—of law enforcement may influence their willingness to carry and thus use guns even more than concerns about being stopped by the police or victimized themselves. This would suggest that at least some of the lethal violence is the product of anger at police forces that kill far too many Black men, as well as at the remarkably violent, riotous way the police responded to last summer’s protests. To use this violence to justify doubling down on conventional approaches gets the lesson exactly wrong. (Emphasis added.)

While technically this excerpt puts forth an “argument,” it is identifying one contributing factor among many regarding increased violence and protests. This clip and the rest of Pfaff’s essay are far more careful and limited considerations of the available data—including the complicated relationship between Black Americans and police forces, and the different times and factors that led to local spikes in homicides—than Lehman’s selective quotation would lead a dispassionate reader to believe. Indeed, Pfaff notes the relevant data are “unambiguously ambiguous” insofar that “it will be years before we have a clear answer” as to what precisely is causing the increase. Pfaff spends hundreds of words in this essay explaining why simple answers are insufficient and yet he—and we—find ourselves accused of promoting a universal theory of causation for the nationwide homicide increase.

Blanks out of context

This is what Lehman wrote about my contribution to the legitimacy hypothesis:

Criminal-justice-policy analyst Jonathan Blanks advanced a similar notion last summer, writing that the protests show that the police have lost public trust — a loss linked to rising crime.

That em dash and the subsequent passive voice is doing a lot of work here, as I have no recollection or evidence of assigning cause of the 2020 crime rise (besides some police-triggered unrest itself) to legitimacy or any other factor. Admittedly, when many thousands of people take to the streets all over the country promulgating phrases like “ACAB”—all cops are b[ad word]s—I didn’t think noting an absence of trust among some Americans was a particularly controversial observation.

My best guess for Lehman’s reference is this piece I wrote for Democracy: A Journal of Ideas:

As American cities start to pick up the pieces from days of protests and unrest following the police killings of George Floyd and Breonna Taylor, policymakers and particularly police departments need to reflect on how their actions and policies have led to this moment. While it is clear that some individuals came to these protests intending to foment chaos, too many police departments shamed themselves and their profession with their actions and policy decisions. Moreover, officers’ widespread and reckless use of violence and noxious chemical agents has given the most anti-police voices the best evidence they could have hoped for that American police are irredeemable. For all the world to see, American police departments damaged the most valuable resource they possessed: their own legitimacy.

In the piece, I explain the basic differences between what I simplified as the reform versus abolitionist approaches to policing and crime policy. I confess my strong preference for reform while retaining respect for the abolitionist position despite deeply held disagreements—namely, my belief in the necessity of police as an institution. But at the core, I argued for simple policy changes police departments can make to mitigate the past and ongoing damage to their credibility.

The limits of available data

To be fair to Lehman, I have tweeted (in 2017) an observed correlation between spikes in homicides in a limited number of cities that have had both: a) recent high-profile police misconduct or violence; and b) long, well-documented histories of police corruption and/or abuse of Black people. Yet, I recognized at the time and have often lamented in the years since that we lack extant metrics to verify or disprove causation because we don’t have the granular data necessary to do so. As such, I’ve never made a causal claim about legitimacy and homicide, though a reasonable observer could read that implication into a handful of my tweets over the past few years. I have, however, explained how police tactics and their perceived illegitimacy might have an effect on public safety.

Nevertheless, Lehman’s NR piece uses data that, at best, show that a widespread lack of legitimacy cannot be observed in certain Black communities’ calls for service (i.e., 911 calls). That is fair enough, but as Pfaff noted in the lengthy piece from which Lehman pulled the misleading quote:

Defenders of the status quo are likely to point to two recent polls to argue that support for conventional policing remains high, even among Black communities. A 2020 Gallup poll, taken in June and July—so after George Floyd’s murder and the resulting protests—found that only 19 percent of Black Americans wanted a reduced police presence in their neighborhoods; 20 percent preferred more. A 2021 Data for Progress poll reported similar results with an awkwardly phrased question, answers to which suggested that two-thirds of Black respondents would feel safer with more police patrols.

The results, however, are far more nebulous than they first appear. Take the Gallup poll. When Gallup broke out its results by exposure to police, Black respondents who had regular contact with officers were about four times likelier to say they wanted less contact with them than those who had little contact (34 percent versus 8 percent). Even a survey that separates responses by race is aggregating Black respondents who have had very different experiences with law enforcement. In the Gallup poll, at least, it appears that much of the support for policing, even from Black respondents, comes from those with the least contact with police officers. The demands for significant reform thus seem to remain strong among those most affected by policing. These are people who live disproportionately in neighborhoods with higher rates of violent crime as well—which means they experience most immediately the trade-offs posed by reformers. It is essential to amplify their voices in the debate over police.

Furthermore, at the same time that a majority of Black respondents, even those with regular contact with police, said they wanted the same or more policing, they expressed little to no confidence in the institution itself. In another Gallup survey taken at almost the exact same time, only 19 percent of Black respondents said they had confidence in the police, and only 11 percent in the criminal legal system overall. Support for the police appears to be less a positive desire for more enforcement and more like grudging acceptance borne of limited options. (Emphasis added.)

In simple terms, Black folks’ views of police are dynamic and complicated, not monolithic and universally hostile. While some anti-police activists may overstate the disconnect between a police department and the Black communities in its jurisdiction, it is likewise a mistake to read too much into community members picking up a phone in a moment of crisis and using that to measure the trust and cooperation that police need to solve crimes.

An overreliance on police violence as legitimacy variable

Part of Lehman's argument rests on the idea that unnecessary police violence is overrepresented in high-profile cases and thus drives perceptions of police illegitimacy out of proportion to normal police behavior. I believe that's probably true, to an extent.

Lehman writes:

Proponents of the legitimacy hypothesis argue that if police did not do all the horrible things they are accused of, then their legitimacy would not fall and crime would not rise. The standard examples are high-profile police killings. These are taken to be proof of widespread systemic abuse.

But such anecdotes are unrepresentative. In 2018, just one in four Americans over 16 reported any contact with the police; of those, just 2 percent experienced any force. Ninety-nine percent of arrests involved no force; among those that did, 98 percent resulted in mild or no injury.

As I mentioned in the opening paragraph of my Democracy piece above, the thousands of video-documented instances of wantonly violent and inexcusable behavior by police officers during the summer protests of 2020 supported the anti-police rhetoric to the detriment of police legitimacy. I have no idea whether those videos had any criminogenic effect, but at the same time, those instances make it hard to absolve police from their role in damaging the profession's credibility as Lehman attempts.

Nevertheless, much of my work that touches on the idea of police legitimacy—again, without asserting illegitimacy is criminogenic—concerns legal police actions involving nonviolent but nonconsensual interactions: investigatory traffic and pedestrian stops. Such stops are invasive, often conducted on a flimsy pretext, and usually concentrated in Black and Latino neighborhoods. Lehman suggests that a full 25 percent of Americans having contact with police over the year is a small number. I disagree, but even putting that aside, that number and description elides the frequency and severity of police contact for many Americans.

As I explained in my 2016 law review article,  geographic, racial, and socio-economic factors feature heavily on who gets stopped by police and why:

A 2006 study showed that 18-19 year old black men in New York City
had nearly an 80 percent chance of being stopped by New York City
Police in a given year; that figure dips to 50-70 percent when the age
group  expanded  to  18-to-24  year  olds  within  the  same  racial  demo-
graphic. For whites in these age groups, the percentages were 10 and
13 percent, respectively. (citations omitted)

I also explain in that article how the perception of disparate racial treatment can undermine police legitimacy. I cite the work of Charles R. Epp, Steven Maynard-Moody, and Donald P. Haider-Markel that shows run-of-the-mill traffic safety stops (e.g., for unambiguous speeding violations) do not have deleterious effects on perceptions of police. However, being stopped for minor infractions (e.g., license plate light burnouts or air freshener dangling on rear-view mirrors) that lead to questioning and consent searches do lead to resentment, particularly among Black drivers who are statistically far more likely to experience them.

Thus, people who experience police contact more frequently and/or are in the same social and familial networks as those who experience frequent contact do not need to rely on videos or talking heads to inform their opinions of police or their legitimacy. As Pfaff suggested, this frequent and unwanted contact may help explain the lower confidence that Black Americans have consistently expressed for police as institutions for decades.

So while Lehman may be correct that discrete, high-profile incidents of police violence have a disproportionate effect on their perceptions in the general public, it is hardly the only metric upon which to judge the actions of police and how they may affect community members' perception of legitimacy, particularly in highly policed areas most likely to suffer during homicide spikes.

Legitimacy as strawman?

Abt has speculated that illegitimacy has contributed to recent homicide increases, and is continuing to comment to that effect with similar “we don’t know exactly why” caveats in news reports. But it bears mentioning that none of the people Lehman named are legitimacy researchers or theorists. Abt and I have cited, inter alia, work by Yale social psychologist Tom Tyler and YLS professor Tracey Meares to support policy prescriptions that police should take steps to build trust with the communities who need effective policing the most. (I wrote a generally positive but somewhat critical review of Abt's book Bleeding Out for the Cato Journal.) Pfaff conducts quantitative research on incarceration and other criminal legal topics, sometimes using others’ research to contextualize findings. I’m sure we all stand by our own work, but it’s very strange that we are the targets as if we were forcefully pushing the legitimacy-as-homicide-driver in the media which, as far as I know, none of us nor the leading legitimacy scholars have done.

Perhaps we were just convenient examples put in during editing and I’m overreacting to our specific inclusions in the piece, but the decontextualized use of Pfaff’s quote is nonetheless troubling. It clearly represents a position far in excess of what Pfaff intended and, by our specific inclusion, lumps us in with a position that none of us have been demonstrated to hold. Indeed, unlike people who openly claim defunding the police has led to the homicide rise—which is not supported by available data—we have, to varying degrees, been cautious about assigning cause to such complicated phenomena.


The homicide rise is real and it is alarming. Some reformers have been guilty of downplaying its importance, while the traditional law-and-order camp has been using it as their long-awaited justification for their sky-is-falling fear-mongering against police and carceral reform. The simple fact is we don’t have enough data to determine which factors are fueling the rise with any certainty. As Pfaff noted in his commendable TNR piece, the timing of spikes and differing inputs vary across jurisdictions and thus singular explanations are wholly inadequate.

But responsible commentators aren’t claiming they have definitive answers to such an important and complicated problem. I very much welcome the rigorous testing of the "legitimacy hypothesis," but that's not what Lehman's article does. Indeed, Lehman cherry-picks quotes, asserts claims not made, and uses selected data to disprove a hypothesis so poorly defined and represented. As such, Lehman’s piece, to use his own words, “obfuscates more than it clarifies.”