The Accelerationism Events Dataset: Geographic and Time Trends
Samantha Fagone
--
This essay is the fourth contribution in a multi-part series of the analytical exploration of the newly launched Accelerationism Events Dataset (AED), which captures events of far-right militant accelerationism and proto-accelerationism from January 1990 through November 2023. Focusing on trends emerging from the 265 cases that comprise the AED, this essay explores features of place and time particular to accelerationist felonies.
To learn about the overall structure of the project, including the history of militant accelerationism and inclusion criteria for the AED, please read the introductory piece for this series; to learn about the tactics, techniques, and procedures seen throughout cases, please see the second contribution to the series; and to learn about the demographic breakdown of defendants and how they compare to the broader tPP dataset, please see the the third contribution to the series.
Methodology and Vernacular
To examine geographic trends we partitioned the United States into four regions per Census Bureau divisions: Midwest, Northeast, South, and West. We also performed a state-by-state breakdown to determine the densest areas of activity based on population. It is important to note that we recorded location variables (i.e., country, state, and city) based on the place the felony occurred, not where the defendant originated or resided. Thus, our analysis does not consider any travel between place of residence and location of crime.
For time-based trends, we examined the data to determine the years within the analyzed timeframe with the highest frequency of crime, as well as changes in regional trends and group affiliation over time, starting with the first incidence of a felony-level proto-accelerationist crime in 1995. We then more closely analyzed the five-year period with the highest level of activity (2018-2022) and determined trends in intended target (i.e., whether people or property were targeted), physical target (i.e., what, specifically, was targeted), and ideological target (i.e., why this target was chosen), along with trends in group affiliation. In 95% of cases, the date indicates the issuance of the indictment or complaint, with only 5% of cases indicating the date of crime, arrest, or sentencing when charging documents are not available. While this approach does create a lag between the time of offense and recorded date, it more accurately captures the political moment of the listed felony, indicating when the case was ready to be taken to trial. For example, while all punishable offenses from the January 6 attack occurred on January 6, 2024, the indictments span four years, reflective of such factors as case priority, investigative hurdles, and availability of defendant information. Thus, to encapsulate those unique factors and keep time recording as consistent as possible across cases, we use the date of indictment when such information is available.
Where relevant, we compare accelerationist felony cases to cases in the broader Prosecution Project (tPP), which tracks and analyzes all political violence in the United States. For this comparison, only tPP cases that were fully coded, audited, and with finalized sentencing were drawn from the database (n=3,976).
As a brief note, 30% of recorded accelerationist cases were from the January 6, 2021, attack on the U.S. Capitol. The vast majority of these cases targeted non-military, non-judicial sites (i.e., the U.S. Capitol building), were property crimes, and were affiliated with the Proud Boys. Additionally, those cases all occurred in Washington, D.C.—regionally in the South—with indictments still being filed in 2024. Thus, when these cases have the potential to skew findings, we present the data, first, as including Capitol cases and, second, while excluding them to demonstrate their impact on the overall dataset.
Overall Geographic Trends
All Capitol arrest cases have been removed from the dataset for this visualization in order to mitigate their outlier effect, resulting in a total of 186 data points for Figures 1 and 2. As shown in Figure 1, the states with the highest frequencies of accelerationist felonies are Michigan (n=31), California (n=24), Texas (n=15), Georgia (n=15), and Virginia (n=11). There are sixteen states in which accelerationist felonies did not occur during the timeframe of the analysis (1990-2023).
AED Heat Map
Figure 1: AED crimes by state, not adjusted for population
Of the states in which accelerationist felonies occurred and were prosecuted, Michigan and California experienced the largest proportion of activity. Heightened activity in Michigan may be explained by the group prosecutions in the state, including the 2020 case against eight members of the Wolverine Watchmen and the 2021 case against seven members of the Three Percenters. California’s high case frequency in California, however, may be partially explained by that state’s greater relative population density as compared to most other states represented in the AED.
Αs seen in Figure 2, when adjusting the number of accelerationist felonies in each state by 2023 Census Bureau population estimates, the states with the highest per capita rates are the District of Columbia (n=3), Michigan (n=31), Delaware (n=2), Oregon (n=7), and New Hampshire (n=2). Once adjusted, the density of cases in some states with high frequencies (i.e., California, Texas, Georgia, and Virginia) more closely reflects the density of other states, though the relative number of accelerationist cases in Michigan remains noticeably high.
Population-adjusted AED Heat Map
Figure 2: AED crimes by state, adjusted for population
Michigan’s 30 year history of heightened militia and other group activity likely helps explain the state’s greater density of accelerationist felonies; this history may make MIchigan a particularly attractive place for far-right actors to organize, including defendants who reside elsewhere, as was the case with several charged in the Watchmen plot. Thus, although California and Michigan display similar frequencies of cases in Figure 1, Michigan’s history of highly coordinated, far-right activity may enable greater levels of accelerationist organization and lead to the more pronounced presence of codefendant clusters in the state.
Regional Trends
This section explores geographic trends in more detail, analyzing AED data by U.S. regional divisions. Extending the analysis of the prior section, Figures 3 and 4 exclude Capitol arrest cases to avoid overstating the magnitude of accelerationist activity in the South. Tables 1 through 3 indicate statistics for the South with and without Capitol cases, illustrating the effects of the insurrection on intended target, physical target, and ideological target.
As seen in Figure 3, 40% of accelerationist felonies in the dataset occurred in the South, followed in prevalence by the Midwest (25%), and the West (23%).
Regional Frequency
Figure 3: Number of cases by Census regions, excluding Capitol cases
Out of the cases in the South, Georgia, Texas, and Virginia account for 57% of all activity, pointing to the strong clustering of accelerationists by state in that region. Similarly, Michigan alone accounts for 67% of all activity in the Midwest, and California alone accounts for 56% of all activity in the West. From this, we see that although accelerationists are spread widely throughout the country, cases within regions tend to cluster together.
Although the South contains the greatest frequency of prosecuted activity (Figure 3), Figure 4 indicates that it also holds the fewest number of group prosecutions, i.e., prosecutions where there are at least two defendants for a given case. We find that 70% of cases in the Midwest, 58% of cases in the Northeast, 49% of cases in the South, and 61% of cases in the West have codefendants.
Regional Activity: Joint Prosecutions
Figure 4: Percentage of cases with codefendants by region, excluding Capitol cases
Revisiting the two states with the highest proportion of activity as reflected above in Figure 2, Michigan’s 31 cases include 29 (94%) with codefendants, while California’s 24 cases include 16 (67%) with codefendants. California’s codefendant cases are also markedly smaller, with the largest joint prosecution including only four defendants, relative to Michigan’s cases that include as many as eight defendants. When considering that most defendants in California (83%) and Michigan (97%) are involved in joint prosecutions, these differences could indicate a greater overall degree of organization behind accelerationist felonies in these states relative to others.
Regional Activity: Targets
For AED analyses, we assess the intended target (the who or what), the physical target (the where), and the ideological target (the why) of each felony in the dataset.
Table 1 indicates the most common intended target of accelerationist actors across regions, reflecting whether the offense was targeted toward human beings, physical property, both, or neither. An offense may be designated as a property crime even with harm to human beings if that harm is incidental. For example, a bombing of a clinic after hours would be considered a property crime even if a person was injured in the act but the defendant’s target was the building and not the person. Similarly, a case may be labeled as one affecting people so long as property damage remains collateral. If a defendant’s motives are not clear, or if the crime only serves to support political violence without harm to people or property (e.g., providing material support to terrorist organizations), the case would be determined as one with “no direct target.”
The Midwest, Northeast, and West are consistent in their intended targets across both the AED and tPP datasets (Figure 6). In the South, however, the most common intended target for accelerationist cases was people, whereas having no direct target was most common for the broader set of crimes represented in the tPP.
Most Common Intended Target by Region
Table 1: Intended target by Census regions
When examining the physical targets of each felony, no dominant physical target emerged in the overall AED data (Table 2). The most frequent physical target in the Northeast was split evenly between Private site: business/corporate property, Individual person(s), No direct target, and Unspecified.
Type of Physical Target by Region
Table 2: Physical target by Census regions
When considering ideological targets of the AED dataset (Table 3), accelerationists demonstrate a target preference for those affiliated with the United States government, with the federal government taking a clear lead when including cases from the January 6 attack. Other accelerationist targets commonly fall in the unspecified category, indicating crimes with no clear ideological target.
Type of Ideological Target by Region
Table 3: Ideological target by Census regions
Accelerationist target types tend to be similar to the overall regional trends in broader felony political violence reflected in the tPP. However, the differences between the AED and tPP datasets in the Midwest are notable. Not only do accelerationist felonies in the Midwest have significantly higher rates of crime targeting individuals, but they are also markedly more government-oriented. This may reflect the greater level of group coordination found within the Midwest, relative to what is perhaps a more fragmented organization in other regions that is most commonly associated with accelerationist activity.
Temporal Trends
The trends in this section exclude January 6 Capitol indictments, as those cases skew post-2021 data and overrepresent the Proud Boys. Figures 5 and 6 range from 1995 to 2023 in order to illustrate the full progression of accelerationist activity in the AED. Figure 7 begins in 2018 for graphical clarity, and we restrict our analysis to the years 2018 to 2022 in Tables 4 through 6 in order to highlight years of peak activity.
As shown in Figure 5, accelerationist cases saw a spike in 2018 and continued to rise until peaking in 2020. The year 2023 saw not even a third of the activity that was present in 2020, but cases remained elevated compared to the pre-2018 era.
Case Frequency by Year
Figure 5: Number of accelerationist cases by year
Both the 2011 and 2020 spikes can be attributed to a higher incidence of group offenses, with 100% of cases (n=9) in 2011 and 63% of cases (n=59) in 2020 being associated with coordinated group activity. Most notably, 2011 saw the indictment of nine members of the Hutaree militia for a planned attack on law enforcement. Thus, we observe that periods with higher crime rates correspond to periods of greater group coordination.
Figure 6 illustrates the frequency of cases by region over time and indicates the consistent presence of proto and accelerationist activity within the South relative to other regions whose activity started later.
Regional Frequency by Year
Figure 6: Number of AED cases in each region by year
Not only do the first manifestations of this ideology appear in the South, but the highest share of cases during the peak frequency of accelerationist felonies is located in this region as well. The Midwest, which has the second highest incidence of cases in 2020 and 2021, does not see activity until 2011. From the regional analysis, we can see that, unlike the South, cases in the Midwest are more group oriented rather than individually initiated. Although the West has an isolated, group-led felony in 1999, activity in that region does not pick up again until 2018, and the peak of its cases in 2020 is outstripped by both the Midwest and the South. The Northeast, reflecting its relatively low overall case numbers, displays the latest beginnings of accelerationist violence. Activity in this region did, however, see an uptick in 2023, outpacing both the Midwest and the South for that year.
Figure 7 illustrates group affiliation over time and reflects the level of fragmentation specific to accelerationist ideology. Here, group affiliation indicates only if a defendant had known ties to an accelerationist organization and is not indicative of either group size or the existence of a codefendant. For readability, only the top ten groups have been graphed of the 44 recorded unique group affiliations.
Group Affiliation Over Time
Figure 7: Top Ten Most Common Group Affiliations from 2018-2023
Primarily formed and propagated in online spaces, accelerationist groups disband and reform regularly, resulting in the 40+ unique group affiliations represented in the AED dataset. The most consistent and common group attribution for an accelerationist case is nonetheless “no known affiliation,” reflective of how group belonging is less of a motivating factor for accelerationist violence than are ideology and symbology.
Table 4, indicating whether a defendant’s intended target was people or property, reflects increasing alignment between the criminal targets in AED and non-AED cases.
Intended Target over Time*^
Table 4: Intended target by year, excluding Capitol cases
*Cases from 2023 are excluded from the Intended Target, Physical Target, and Ideological Target tables because only three 2023 cases in the tPP overall data were fully coded, audited, and including finalized sentencing, thus providing insufficient comparison for this year.
^In 2018 the most common target of incidents was recorded as “No Direct Target,” indicating the crime targeted neither people nor property and was likely a financial crime, such as theft.
From 2020 to 2022, the percent of felonies targeting people in AED cases is notably similar to the percent of crime targeting people in non-AED cases. Although 93% of AED cases targeted people in 2018 and 37% had no direct target in 2019, those differences were primarily caused by the Rise Above Movement in 2018 (comprising 8 of 14 cases) and Atomwaffen Division in 2019 (comprising 8 of 27 cases). Those differences in intended target point to the degree of influence specific accelerationist groups can have on the trajectory of violence. Overall, Table 4 indicates that accelerationists are not dramatically different from other criminals who commit political violence in terms of targeting people versus property; they are, however, differentiated by their physical and ideological targets.
Table 5 indicates the physical target most prominent in AED and non-AED cases from 2018 to 2022. Here, No direct target reflects a lack of physical target and encompasses such crimes as providing material support to a terrorist organization or possession of a firearm by a convicted felon.
Physical Target over Time
Table 5: Physical target by year, excluding Capitol cases
In non-AED cases, the physical target shifted from No direct target in 2018 and 2019 to Individual people in 2020 onward, a move toward felonies that tend to be more violent in their ends. AED cases, however, shifted toward No direct target in 2021 and 2022. Instead of indicating non-violence, however, the majority of those cases reflect instances in which defendants were arrested for building explosives or illegally obtaining firearms before being able to enact presumed plots, likely indicating cases that would have resulted in serious violence had plans not been successfully surveilled and stopped.
Table 6 shows the most common ideological target in AED and non-AED cases from 2018 to 2022. Unlike AED physical targets during this time frame, AED ideological targets do not exhibit strong temporal patterns.
Ideological Target over Time
Table 6: Ideological target by year, excluding Capitol cases
Although accelerationists are united in the belief that “there is no political solution” to perceived social ills, Table 6 implies the specific methods accelerationists use to criminally pursue that belief are markedly more fragmented.
Comparing each target type in the 2018-2022 window of peak activity to tPP data, we observe that accelerationist ideology has the greatest impact on physical targets and ideological targets. Although accelerationists’ intended targets generally follow broader trends in political violence, their physical and ideological targets reflect the varied motives and affiliations within accelerationist spaces.
Conclusions
The South experienced both the earliest instances and highest frequencies of AED felonies from 1990 to 2023.
Beginning with a spike in 2018, identified accelerationist felony activity rose until its peak in 2020, after which case numbers remained elevated relative to pre-2018 levels.
While activity in the South is diffused throughout multiple states, AED crime in the Midwest and West is most attributable to codefendant clustering in Michigan and California, respectively.
The regions with the highest density of per capita AED felonies also had the greatest number of group prosecutions, indicating a positive correlation between accelerationist felonies committed with codefendants and prosecuted accelerationist activity.
–
Samantha Fagone is a graduate student at the University of Kentucky studying Applied Environmental Science. For the past two years, Samantha has been a Coder and Researcher for the Prosecution Project, where she has previously studied crowd-sourced policing methods following the January 6 attack on the United States Capitol Building. Samantha’s research focuses on political violence, political economy, and policy modeling, and their intersections with extremism, markets, and the law.
This report is part of a multi-publication series led by Dr. Michael Loadenthal, and supported by research conducted by Mary Bennett Doty, Samantha Fagone, Grace Stewart, Olivia Thomas, and Bella Tuffias-Mora.
To check out the introduction to this series, read Dr. Loadenthal’s piece here.
To learn about the tactics, techniques, and procedures (TTPs) of the dataset, read Mary Bennett’s piece here.
To learn about the demographics of the dataset, read Grace Stewart and Mary Bennett’s piece here.
