If you prefer to listen rather than read, this blog is available as a podcast here. Or if you want to listen to just this post:

Or download the MP3


This might be a weird post. I have an idea for what direction I want to go in, but it’s also something of an exercise in thinking out loud as well. This whole exercise got started during a conversation with a friend of mine. I forget how we got on the topic, but he mentioned that from a domestic standpoint right wing extremism was a far worse danger than left wing extremism. In support of this statement he offered up the figure that 90 police officers had been killed by right wing domestic terrorists. My immediate reaction was to call “BS” on that figure, mostly because I could hardly imagine that the deaths of 90 police officers would have “flown under my radar” when the best known instance of right-wing violence, the murder of Heather Heyes by James Alex Fields Jr. when he drove his car into a crowd after the Unite the Right rally in August of 2017, had gotten so much attention.  So I challenged him to produce his source.

Eventually he pointed me at a report from the Anti-Defamation League (ADL) titled Murder and Extremism 2019 which includes a chart (page 21) showing that 59 police officers had been killed by right wing domestic extremists between 1965 and 2019, as compared to 44 who were killed by left wing extremists. (Fun exercise: See if you can spot the math errors on the chart…) As you might imagine 90 is more than 59, still as these things go it’s close. Actually the biggest discrepancy between the impression I received when having the conversation and the actual numbers was the time scale. At no point in our conversation did my friend mention that to arrive at the figure one had to go all the way back to 1965. I assumed he was talking about in the last few years, or the last decade or at a stretch maybe since 2000. And to be fair to him it’s difficult to be completely precise in off-the-cuff conversations like that, so I don’t blame him for misremembering the number or not ever mentioning the time frame. Also if we were to restrict ourselves to the period since 2000 things actually look worse for the right. With 36 deaths on that side vs. 10 on the left. In any event I’m going to admit that I was partially wrong, the claim wasn’t complete BS, but it was substantially different than what I understood his claim to be.

Of course the accuracy of that number is not what anyone is really interested in, it’s just a way of trying to get at the answer for our true question, which is: should we worry more about right wing extremism or left wing extremism? Knowing that 59 police officers were killed by right wing extremists over the last 55 years, as opposed to 90 over a shorter period of time has some utility in arriving at that answer, but that utility is surprisingly small. Primarily because the information we lack is still vastly greater than the information we have. To have any kind of confidence in an answer to our primary question of which side should worry us more, we would ideally have answers for all of these secondary questions:

  • How does violence against police compare to violence against everyone else?
  • If those are the numbers since 1965 what do the numbers look like more recently? Which way is the trend headed?
  • When the numbers are tallied how are people and incidents bucketed into left and right?
  • Out of all the harm caused by ideological extremism, what percentage of it is due to violence by extremists of that ideology, and what percentage of it is due to other factors?
  • Let’s take these questions in turn and see what we can glean from that initial paper, and maybe a few other sources besides.

How does violence against police compare to violence against everyone else?

From our initial paper we read that in 2019 there were 42 deaths from extremist violence (of which 22 came from the El Paso Walmart shooting). And that out of those 42 deaths 81% could be attributed to white supremacy. And, finally that only one death was a police officer (not part of the 81% by the way, for the first time the ADL put a police killing in the category of “other extremists”.) 

Expanding the time horizon, for the period 2010-2019, the ADL counts 330 deaths due to extremism of which they say that 78% were attributable to white supremacy. Out of that 330 deaths 21 police officers were killed (or close to that, the police numbers start in 2011 not 2010) and 11 were attributed to right-wing extremists, or 52%. So the answer, according to the paper, is that as a percentage, right wing violence against everyone else is worse than their violence against police. Though a lot then depends on how the ADL decides to classify something as right wing or left, a point we’ll get to. 

If those are the numbers since 1965 what do the numbers look like more recently?

Whenever you’re talking about numbers in this fashion, there is the temptation to shift them in your favor by choosing an advantageous starting (or ending) point for your count. As far as the police shootings, they probably start things in 1965 because the numbers aren’t available any earlier than that, but given that the left wing numbers are much higher in those earlier years it’s worth asking if the percentages would tilt even more towards the left if we went back to 1955 and the beginning of the Civil Rights Movement, or to 1920 for a full century of data, or to the turn of the century when the Propaganda of the Deed was at its height and numerous governmental officials were being killed by anarchists, including a President. And note that I’m still talking about killings of police, nor am I attempting to pass any kind of judgement on what oppression may or may not have been happening, it would just be interesting to see the ADL apply the same methodology to a much bigger data set.

That speculation aside, let’s look at some things we can do with the numbers we do have by declaring different start points. Fortunately for this endeavor I was able to find some other ADL reports on extremist violence, some from previous years, and one that was sort of summation, as you’ll see these additional reports ended up being both illuminating and confusing. I’m still going to stick with police killings just because it’s more manageable, and I’m guessing the data is cleaner as well. I’ve already talked about pushing the start date back, but what if we bring it really close? What if we look at the number of police killed in the period 2016-2019? 

Fortunately the 2016 ADL report on extremism has the same chart of police killings (it was missing from the 2015 version.) So how did things stand in 2016 as compared to 2019? As I mentioned 11 cops had been killed by right wing extremists over the last decade, as it turns out 10 of the 11 were killed before 2016. The left’s number for the decade was 8, and as it turns out, all of those deaths happened in 2016. What this means is that if we decide to just look at the most recent three years, eight times as many police officers have been killed by left wing extremists. 

To be clear, I’m not saying that this is the right way to look at things, but it is what happened during the “Trump Era”. And of course we could reverse things, if we wanted to do the same thing for the “Obama Era” and look at the period from 2011-2015 when right wing extremists had killed 10 cops and left wing extremists had killed zero. You can certainly imagine the ADL screaming about the dangers of the right wing in 2015, not knowing that in 2016 things would almost equalize. Picking start and end points matters a lot.

Moving back a little bit farther, as I’ve pointed out things are pretty close to equal for the most recent decade at 11 to 8, it’s the two decades before that where right wing extremism really looks scary. If we go all the way back to 1991 the ADL numbers climb to 52 deaths from extremists on the right vs. only 11 for those on the left. But here’s also where things get confusing. When you’re engaged in any project like this, you really want to see the underlying data, so you can independently check the numbers. I was particularly interested in the 2001-2010 period where the ADL is saying that 25 police officers were killed by right wing extremists. Especially since the ADL was only showing 16 police officer killings in the 90s which contained the most dramatic example of right wing terrorism, the bombing of the Oklahoma City Federal Building. What had happened in the 2000s that was even harder on law enforcement than the bombing of a federal building?

Fortunately I found another report on their website A Dark and Constant Rage: 25 years of Right-Wing Terrorism in the United States. This report covered the period 1993 to 2017 and included a section called: Right-Wing Terrorism Inventory, 1993-2017. This section detailed the separate incidents which went into their numbers. As you go through them you see a lot of cases where people were arrested based on their intent to commit some act of terrorism, but it never actually came to pass. A few quick examples:

  •  a plot to murder Muslims in upstate New York using a “death ray” device that would emit lethal radiation.
  • Schmidt possessed a large cache of weapons and a notebook with evidence that Schmidt was targeting Detroit-area Jewish and African American leaders. 
  • on federal firearm charges after receiving information that he was plotting to kill Governor Gary Locke. 

Given the inclusion of these numerous unsuccessful plots I assumed that if someone actually killed a police officer it would definitely make the inventory, and yet after combing through the 2001-2010 list, I could only come up with eight incidents where police officers had been killed:

For obvious reasons I would really like to know where they’re getting the other 17 killings from. I have no doubt that 17 police officers were killed between 2001 and 2010, and ended up being counted by the ADL, but why were their deaths not considered noteworthy enough to be included in the inventory? Is their right-wing connection more tenuous? To be honest I’m not entirely convinced that all of the incidents which did make the list should be considered examples of right-wing terrorism Poplawski was a domestic dispute that went horribly wrong, and the Turnidges were trying to rob a bank to pay off their debts. Regardless of whether you agree with me, I think you’d concede that it’d be easier to decide if those other 17 deaths also belonged on the list if you knew the details behind them.

Turning to the other side, the big decade for the left was 1971-1980, when 25 police officers were killed by left wing extremists. How does the ADL do in accounting for those deaths? Well, I already mentioned the eight police officers that were killed in 2016 by left-wing extremists, in the report covering that year, here’s what the ADL had to say: 

None of the police officers shot by Long or Johnson [the two 2016 perpetrators] were themselves involved in any controversial shootings; they were blameless. The killings were acts of indirect retaliation aimed at local law enforcement officers because of earlier officer-involved shootings in Dallas and Baton Rouge. These killings represent the worst spate of black nationalist-related murders of police officers since the late 1960s and early 1970s, when more than two dozen police officers, and several more corrections officers, were killed by black nationalists, particularly from the Black Liberation Army and the Black Panther Party (no relation to the New Black Panther Party). 

This is immediately followed by the same chart of police officer killings we keep referencing, and as I mentioned it shows 25 killed by left-wing violence in the 70s, and 3 who were killed in the late 60s, so 28. The “more than two dozen police officers and several more corrections officers” would seem to get us to that figure. Meaning that while 17 right-wing murders were left unspecified by the ADL in the 2000’s, when supposedly things were at its peak, we have a basic description for all of the left-wing murders when they had their peak. Now I admit that this description still lacks specifics, but at least we have some idea of where to look, I have no idea where to go to find the missing 17 on the other side of things.

This section went longer than I expected, and illustrates what I mean by thinking out loud, so to sum things up: Deciding where to draw your line can make a big difference. There are points at which right-wing extremists are way ahead and points at which the left-wing extremists are. Again this is just if we look at killings of police officers, but one assumes that you’d find much the same thing with other measurements of extremist violence. A picture that looked very different depending on where you drew the line and what you expanded your search to encompass. And that the numbers and circumstances surrounding any additional incidents you decided to bring in would be even more ambiguous than the police killings we’ve been talking about. Which takes us to:

When the numbers are tallied how are people and incidents bucketed into left and right?

It’s interesting that not just the ADL, but everyone wants to classify violence as being either the fault of one side or the other. That the ADL spends so much effort talking about the evils of right wing extremism and not the evils of violent political extremism in general.

Alternatively, if you’re really trying to target ideologies that lead to violence, I think you’d want to get as specific as possible. As I mentioned, if you just look at the last three years not only are left-wing extremists responsible for more police officer killings, but all of those killings were carried out by black nationalists, furthermore, and as we saw from the quote, according to the ADL, the last big spike in police killings by the left, in the late 60s early 70s, were also carried out by black nationalists. Meaning that one fairly compelling interpretation of the numbers would be that the left in general is very good at eschewing violence, but we should spend a lot of resources specifically to prevent violence from black nationalists. Once again, to be clear, none of this is to deny the many grievances blacks currently have, or to say that oppression doesn’t exist, but if we’re looking for patterns in the numbers, which seems to be the whole point of these reports I’ve been referencing, this pattern of violence from black nationalists does seem like one we should be paying attention to, and in fact it’s the dominant pattern if we just look at the last few years and also a very significant one if we go back as far as we have numbers.

As long as we’re on the subject of black nationalists, the character of those incidents is different as well. The two incidents from 2016 (and many of the incidents from the 70s) were ambushes that were specifically designed to target and kill police officers. From that years ADL report:

Eight police officers were killed in two incidents this past year in which extremists deliberately targeted police officers for murder. In July 2016, Micah Xavier Johnson, who had ties to black nationalist groups such as the New Black Panther Party, killed five police officers (and injured nine others) in Dallas, Texas, in an ambush attack aimed at police maintaining public order at a Black Lives Matter protest. That same month, Gavin Eugene Long ambushed and shot six police officers, three of them fatally, in Baton Rouge. Long was an adherent of black nationalism as well as the anti-government sovereign citizen movement.

Contrast this with Richard Andrew Poplawski who I mentioned above. He also killed three police officers, but under very different circumstances, Wikipedia says that the shooting:

…stemm[ed] from a mother and her 22-year-old son’s argument over a dog urinating in the house. At approximately 7:11 a.m. EDT, 22-year-old Richard Poplawski opened fire on two Pittsburgh Police officers responding to a 9-1-1 call from Poplawski’s mother, who was attempting to get the police officers to remove her son from the home. Despite Poplawski’s mother telling the 9-1-1 operator that Poplawski had guns, the police officers were not told. Three police officers were ultimately confirmed dead, and another two were seriously injured.

It later came out that Poplawski frequented right-wing web-sites and voiced racist views online but nothing about the actual killings was ideological in nature, and yet these three killings get counted and reported as being fundamentally identical in the chart we keep going back to, despite being very different. Long and Johnson ambushed the police officers they killed, Poplawski only killed them after they showed up at his house. Additionally it sounds like if the dispatcher had not neglected to tell the police that Poplawski had guns perhaps the killings wouldn’t have happened at all. This suggests that there might be a continuum when it comes to the circumstances of the killings as well, and more importantly another place where a line is being drawn. What characteristics does an incident have to possess in order to classify it as right-wing or left wing extremism?

I’ve provided a comparison related to our primary focus, police killings, what about if we widen it to other killings? Turning back to the 2019 ADL report, we see that it also has a section for incidents. These include people targeting synagogues or, the most horrible one from last year, when Patrick Crusius went into a Walmart intending to kill Hispanics and ended up murdering 23 people. These unquestionably are acts of violence in the service of extremist right wing ideology. But many of the incidents on the list seem less clear cut. This despite being listed under the heading, “The 2018 extremist-related murders preliminarily documented by ADL include:” (side note: 2018 is obviously a copy and paste error, all the incidents are from 2019) 

For example some incidents which also appear under that heading:

Anthony Voight, a member of the white supremacist Aryan Brotherhood of Texas, was arrested for the shooting murder of a man who accompanied Voight’s ex-girlfriend to retrieve her belongings from Voight’s home.

White supremacist Travers Proulx was arrested on first-degree murder charges after he allegedly stabbed his mother to death following an argument.

Keeton Waring, a reported member of the Southwest Honkeys, one of several Missouri-based white supremacist prison gangs, allegedly shot and killed another man during an argument over a missing cellphone. He has been charged with second-degree murder.

These three sound far more similar to the Poplawski murders than the Johnson and Long murders. 

Is it possible that the ADL has a bias? For many people the fact of their left-wing bias is so obvious that they wonder why I’ve wasted so much time quoting from them. For myself, I would expect that, if they were unbiased, that somewhere in that accounting of the incidents which occurred in 2019 that I would find reference to a marxist, or an anarchist or a black nationalist who stabbed their mother, or got into a violent argument over a cellphone. In the absence of that I’m inclined to believe that their statistics probably do have a left-wing bias, if that’s the case then using their biased numbers to answer our original question, “should we worry more about right wing extremism or left wing extremism?” leads to a biased answer.

Of course the idea that the ADL is biased shouldn’t be surprising. Everyone has biases. I’m sure you’d be quick to point out mine. But as a result of these biases any categorization of one murder as being a right-wing murder, while another is left wing murder (or bereft of ideology entirely) is bound to depend on where the person making that categorization drew a subjective line. But as I reflect on it, I think drawing the initial line separating the world into just two teams: right and left, may be the most damaging line of all. Rather than having everyone on the same team against all extremist violence, or indeed against all violence period, or, on the other hand, attempting to narrowly define the source of the problem so that our accusations against the innocent are minimized, we’ve got the worst of both worlds. We bind up half of everyone in our accusations, which must surely include some innocent people, while also just as surely overlooking some violence in the half we’ve declared to be “our team”.

To return to the question that started this section, “When the numbers are tallied how are people and incidents bucketed into left and right?” The answer is: subjectively.

Out of all the harm caused by ideological extremism, what percentage of it is due to violence by extremists of that ideology, and what percentage of it is due to other factors?

I’m already running long on this post, so I’ll try and keep this section much shorter than the previous two. I think the obvious answer to this question is that only a tiny fraction is due to violence by ideological extremists. Allow me to explain what I mean.

In our modern world the only people who identify as Nazis, or indeed as Communists, tend to be pretty radical (though more so in the former case than the latter), but there were times and places when such identification was not only mainstream but expected, e.g. the Third Reich and Soviet Russia respectively, and it was when these ideologies had triumphed, when they had gone from extreme to expected that the really horrific violence occurred. My takeaway from this is that what we’re really engaged in when we ask the question, “should we worry more about right wing extremism or left wing extremism?” Is a discussion on what the behavior of the extremists tells us about the direction we’re headed and the potential harm that could be inflicted once an ideology becomes more widespread. 

It’s clear that people want to use evidence of extremist violence to act as a guide for where an ideology and society as a whole is headed. And more commonly, but less acknowledged as evidence to back up their impression of where they think it’s going. But the connection between ideological extremism (even if properly attributed) and the ideology itself and society more broadly is more tenuous than people think. Compare the heights of ideological violence from the reports, with the high points of those actual ideologies. The high point of leftist violence preceded a run of right-wing presidents Nixon, Ford and Reagen, with only Carter thrown into the mix to be really confusing. On the other side the peak of right-wing police shootings happened in the 2000s, and was followed by the election of Obama. The most extreme example of right wing violence, the Oklahoma City Bombing immediately preceded the re-election of Clinton. But people think there’s a connection between that and Trump being elected 20 years later?

Obviously I was being facetious just then. My claim is not that right-wing violence isn’t increasing, it might be, and also people might be drawing the line more expansively when it comes to classifying an incident as right wing vs. left wing, as I pointed out in the examples I quoted of cellphone fights and domestic disturbances. No, my claim is that the waters are very muddy, and I don’t think some of the connections being made by organizations like the ADL are as clear as they would lead you to believe, while on the other hand I think some very clear connections are being ignored entirely, for example the Ferguson Effect. Given that I’ve already talked about it at some length, I’m not going to rehash it, but it is worth looking at what happened in St. Louis last month

Before the Michael Brown shooting, monthly homicides in St. Louis averaged around 13. Afterwards that average went up to around 18. In July they spiked to 54. That’s a marginal increase of 36 deaths, over something that was already elevated, or in other words more deaths in one month, in one city, than the worst decade of police shootings from the right and left combined. If we assume that these deaths are due to left-wing ideology, which is at least a hypothesis that can’t be rejected out of hand, then, to refer back to the question that began the section, this would be an example of where the harm caused by violent extremists is a tiny, essentially negligible fraction of the violence caused by the ideology as a whole. And of course, as I see it, the whole point of this blog has been to map out the wider and less visible harms caused by technology, progress and yes, ideology.

Conclusion

I hope this post has been interesting for you, it was very interesting for me to deeply comb through a very small set of statistics, and it led to a major epiphany, though you’ll have to wait for my next post to discover what the epiphany was. It’s possible, perhaps even likely, that this post was just a vast exercise of my own biases, which is okay, that’s part of my epiphany. But to the extent we can look past my biases I think there are still a few objective principles we can take away:

  1. Even with a very narrow focus (i.e. just police shootings) you can still pull out numerous different interpretations of what we should “really be worried about”.
  2. One way to change the interpretation is to change the start or end date you’re using for the statistics on which that interpretation relies. 
  3. Interpretation and subjectivity operate at all levels of the discourse, from the classification of individual incidents all the way up into deciding that there’s really only two kinds of extremists. 
  4. This bilateral division might be the worst possible way to divide things, maximizing both the innocent people who are declared guilty and the guilty people who are declared innocent.
  5. A point I ran out of space for, but can be seen both from the Oklahoma City bombing and from the ambushing and assassination of police officers in 2016, violence is very much subject to tail events, or black swans as they are sometimes known. Where a large part of the harm comes from only one or two incidents.

Pulling all of this together and returning to that original question, “should we worry more about right wing extremism or left wing extremism?” I think there’s plenty of reason to worry about both, that in the process of declaring something as part of one side or the other lots of bias is brought to bear, and that all of our worries or alternatively all of our assurances could look silly in the face of some future extreme event. This is one of the points I’ve made again and again. That the future will be shaped by unforeseen, extreme events, that someday, probably sooner than we expect, some ideology will be responsible for the deaths of thousands if not millions and it will make our comparison of 59 right wing police killings to only 44 left wing police killings look both quaint and naive. But this is not all, running underneath these extreme events, are broad, implacable currents, which are ultimately just as impactful, but largely dismissed or ignored by people who want to talk about whether extremism was up or down in 2019. Together these two factors combine to determine the shape of the future, and we’re not paying enough attention to either of them.


Where you draw the line makes a big difference. I have decided to draw the line at never charging for my blog, but really hoping that other people will decide to draw the line at: worth supporting anyway!