Is it true that the right has committed more violence than the left
Executive summary
The preponderance of empirical research and multiple datasets indicates that, historically and in aggregate, right‑wing extremists have committed more frequent and deadlier acts of political violence than left‑wing extremists in the United States and globally [1] [2] [3]. Recent analyses note a measurable rise in some left‑wing incidents since 2016, but scholars and reviewers warn this uptick remains small in absolute terms and contested on methodological grounds [4] [5] [6].
1. What the hard data says: right more frequent, more lethal
Longitudinal and cross‑national studies find that left‑wing actors are less likely to use violence and that right‑wing attacks have produced more fatalities; UMD/PNAS research concluded left‑wing radicals were less likely to use violence than right‑wing and Islamist radicals in U.S. data and that left‑wing attacks are less deadly globally [1] [2] [7]. Contemporary reporting and expert summaries reach similar conclusions: outlets synthesizing terrorism databases report that right‑wing extremist violence has been both more frequent and more lethal than left‑wing violence in recent years [3] [8].
2. Recent shifts and the caveat: an uptick in left‑wing incidents, but from a low baseline
Analysts at CSIS and some news outlets documented an increase in left‑wing incidents since 2016 and noted that, in certain recent windows, left‑wing incidents have outnumbered far‑right attacks for the first time in decades — yet these events remained much lower in lethality and small in absolute count compared to historical right‑wing fatalities [4] [5]. The CSIS brief itself stresses that left‑wing violence rose from “very low levels” and that lethality remained low [4].
3. Why measurements disagree: counting, definitions, and selection bias
Disagreement among studies often stems from how incidents are defined, which events are included, and coding decisions — disputes that have practical consequences for headline claims [6]. Critics of the CSIS framing pointed to opaque inclusion choices and argued that excluding or classifying borderline cases can materially shift the apparent balance of incidents [6]. Academic work cautions that labels like “left” and “right” collapse diverse actors and that informal networks complicate neat categorization [9].
4. The role of lethality versus frequency: different threats, different impacts
Multiple sources emphasize that frequency and lethality are distinct metrics: far‑right violence has produced a disproportionate share of fatalities and high‑profile terrorist attacks, while far‑left incidents historically have focused more on property or fewer‑fatality actions [2] [8]. UMD data showed nearly no difference between Islamist and right‑wing likelihoods of committing violent acts, while left‑wing probability of violence was substantially lower in U.S. samples [7].
5. Politics, public perception, and the media ecosystem
Public perception of which side is more violent is sharply partisan: surveys show Democrats overwhelmingly view right‑wing extremism as the major problem, while Republicans are more likely to see left‑wing threats as comparable — perceptions that do not always map onto empirical incidence or lethality [10]. Media narratives and political statements can amplify isolated incidents and feed competing agendas, which makes careful attention to definitions and sources essential [9].
6. Bottom line and limits of current reporting
On balance, the best available comparative datasets and peer‑reviewed studies support the conclusion that right‑wing actors have committed more political violence and caused more deaths than left‑wing actors in the U.S. and worldwide, though some recent data show a rise in left‑wing incidents from a low baseline [1] [2] [4]. This assessment is subject to caveats about coding decisions, time windows, and evolving dynamics; critics warn that methodological choices can change headline claims, so continued transparent data collection and scrutiny are required [6] [7].