ACLED Conflict Severity Index
A New Measure of the Complexities of Conflict

19 January 2023

Click here to explore an interactive visualization of the Index.


Political violence was pervasive in 2022, and it will likely only continue to proliferate in 2023. Almost every country in the world experienced some level of political violence throughout the past year. Yet conflict is not equally distributed, and it can take many forms. To address the challenge of accurately accounting for and comparing levels of conflict severity, ACLED is launching a Conflict Severity Index. Drawing on the latest ACLED data, the Conflict Severity Index assesses four key indicators to identify the most severe forms of conflict, providing new insights into how and where severe conflicts occur. At the start of 2023, 46 countries and territories meet the criteria of at least one severity indicator. Of these, 19 currently meet the Index’s criteria for high or extreme levels of conflict severity, ranging from Afghanistan, Nigeria, and Ukraine to Mali, Myanmar, and Yemen.

Accounting for the Complexity of Conflict

Conflict is complex and widespread. Globally, more than 125,700 political violence events took place in 2022, leading to over 145,500 reported fatalities. In 2021, the latest year for which comparable population data are available, ACLED estimates that up to 20% of the world’s population – roughly 1.7 billion people – were exposed to political violence.1This statistic was calculated using ACLED data on locations with at least one political violence event in 2022 with WorldPop’s global population data collected at 1km² scale. These incidents ranged from mob attacks on political party supporters and assassinations of local authorities, to state killings, insurgent clashes, and cartel violence. Multiple types of violence often occur in the same country at the same time, exposing the public and state to several, simultaneous threats.

To mitigate these threats, all manner of stakeholders – including policymakers, analysts, civil society organizations, businesses, and the media – require unbiased, straightforward, and reliable assessments of conflict severity. Determining which countries meet a severity threshold has tangible implications for what kinds of violence are seen and responded to – and where. Such a measure must recognize multifarious local complexities and extract important patterns that can inform strategic and operational decision-making.

Many existing indices are based on event and fatality counts, which – when evaluated on their own, isolated from aggravating factors and local context – can often distort the size and impact of conflict for different communities and states. Others focus on characteristics and assessments of ‘fragility,’ and do not sufficiently account for political violence and its variations. In 2022, the most intense levels of violence measured by total event count alone were reported in Ukraine (over 34,400 events), Syria (over 10,400), Myanmar (over 9,300), Brazil (over 7,900), Mexico (over 7,100), and Yemen (over 6,400). The countries with the most conflict-related fatalities were Ukraine (over 28,000 fatalities), Myanmar (over 19,000), Nigeria (over 10,600), Mexico (over 7,700), and Yemen (over 6,700). Ukraine ranks at the top on both lists due to a devastating conflict between the Russian and Ukrainian state militaries. Many other countries experienced high event and fatality counts through more varied violent circumstances.

Myanmar’s conflicts, for example, involve hundreds of non-state armed groups fighting against a military junta and joining several ethnic-based rebel organizations that have been active for decades. In Nigeria, the Islamic State insurgency (formerly dominated by Boko Haram) in the North East region accounts for 13% of the country’s political violence, while the majority is committed by dozens of militias fighting for local authority and control over territory and populations, especially in the North West and Middle Belt regions. Mexico, meanwhile, is plagued by a political-criminal violence nexus that generated the largest number of civilian targeting events in the world in 2022, and which lacks a clear political or military/security solution.

Conflicts that have similar rates of intensity are rarely similar in other ways. In Haiti and Mozambique, for example, the conflict event rate is consistent (close to 500 political violence events per year), but the violence is dangerous to their respective communities and governments in very different ways. The violence in Mozambique is clustered in its northern, less populated areas and supported by a global jihadi network: the prospect of its future diffusion is what concerns most analysts. However, in Haiti, a much smaller country that is characterized by no current central government, gangs control much of the capital city of Port-au-Prince and several food and health crises have worsened due to persistent violence and unrest.

Conflicts can also differ with respect to the proportion of events and fatalities and the risks faced by different communities. For example, a person in Ukraine is currently twice as likely to be killed by political violence as a person in Mali or the Central African Republic.

Assessing Severity: The Index’s Four Indicators

The intention of ACLED’s Conflict Severity Index is to capture this complexity by focusing on the range and multifaceted nature of violent threats. The Index uses four indicators to measure conflict severity: deadliness, danger, diffusion, and fragmentation. Countries and territories can rank high on one or multiple indicators. Countries and territories that rank high on three to four indicators are experiencing the most severe and difficult-to-resolve conflicts. See table below for details.

Indicator Measure Significance
Fatality Rate
Assessment of fatality counts. All fatalities from political violence are aggregated by country to determine a mean rate. Countries that fall above the mean are included in this indicator. The deadliness of a conflict is an important measure of the direct and immediate human cost of violence. Fatality levels are subject to significant bias, under-reporting and inflation. To mitigate these challenges, ACLED uses a standard, global methodology around the world that defaults to the most conservative estimates.2For more information, see FAQs: ACLED Fatality Methodology. This does not fully erase general fatality count inaccuracies, but can mitigate the extremes that result from excluding conflicts that do not meet certain fatality thresholds or including highly biased fatality rates suggested by warring parties.
Violence Targeting Civilians
Assessment of the direct danger to civilians. The number of violent events targeting civilians is aggregated by country and standardized by population (events per 100k people). Countries with a rate above the median are included in this indicator. The proportional rate of violence against civilians in conflicts can vary by country, group, or location. Some armed groups target civilians more than others. These groups are typically weaker but can be highly active and take on anti-civilian actions for other groups (like governments or other larger non-state actors).3For more information, see: Raleigh, C. (2012). “Violence against civilians: A disaggregated analysis.International Interactions, 38(4), pp.462-481.
However, almost all armed, organized groups generate some level of direct harm against civilians. The rate at which they attack civilians can change over the course of a conflict, largely based on territorial control, how armed competition with other groups is proceeding, or how many other groups are active.
Subnational spread of conflict
Assessment of the geographic distribution of conflict. For each country, the percentage of first level administrative units with a high average weekly event count is calculated. Countries with no administrative divisions meeting this threshold are excluded from this indicator. A ‘high’ average weekly event count is defined as 1.5 or more events per week. This threshold is consistent with ACLED’s Volatility & Risk Predictability Index. The geographic scope and spread of conflicts affecting communities and governments can be estimated by how widespread political violence is within a country, and how many administrative divisions experience violence.
Problems can be clustered or diffuse, and more or less dangerous due to the exposure of larger civilian populations to insecurity. Geographic variations present unique operational challenges for states, armed groups, and communities under threat.
Number of organized violent non-state groups
A count of all armed, organized active rebel and political militias per country in 2022 (excluding unidentified armed groups and including pro-government militias). In addition, a maximum of one communal militia is counted per first level administrative unit per country. Due to the heavily skewed distribution of active groups within countries, countries that fall within the top 80% of actor counts qualify for this indicator. The fragmentation of a conflict environment indicates the number of distinct threats and agendas that are accumulating in a given context and posing harm to communities and state institutions. It also indicates the number of distinct political motives and opportunities to form an armed group. A singular consolidated armed group can be a serious challenge to governments, but can take part in effective negotiations and engagements. A highly fragmented environment, in contrast, makes it more difficult to engage the necessary actors in effective negotiations and may indicate multiple overlapping conflicts that are more challenging to simultaneously resolve.

All countries and territories are assessed according to each indicator, and those that meet or exceed the designated threshold are flagged for that indicator category. The visual below shows the number of countries flagged in each indicator category and each combination of indicators. Countries are then ranked by the number of indicator thresholds they meet: one indicator represents limited severity; a combination of two represents moderate severity; a combination of three represents high severity; and a combination of all four represents extreme severity.

Forty-six countries and territories currently meet the criteria for at least one of the four indicators. Based on ACLED data for 2022, the following countries and territories rank from extreme to limited severity:

Extreme Severity
(Present in all 4 indicator lists)
High Severity
(Present in 3 out of 4 indicator lists)
Moderate Severity
(Present in 2 out of 4 indicator lists)
Limited Severity
(Present in 1 out of 4 indicator lists)
Colombia Afghanistan Bangladesh Armenia
Haiti Brazil Cameroon Azerbaijan
Mali Burkina Faso Ethiopia Belize
Mexico Democratic Republic of Congo Honduras Burundi
Myanmar India Jamaica Central African Republic
Syria Iraq Libya Egypt
Yemen Nigeria Mozambique El Salvador
Pakistan Niger eSwatini
Philippines Palestine Guatemala
Somalia Sudan Indonesia
South Sudan Venezuela Israel
Ukraine Kenya
Puerto Rico
Trinidad and Tobago

The 46 countries and territories represented here account for 95% of all political violence events recorded by ACLED in 2022. The remaining 5% were distributed across 119 other countries and territories. Approximately 30% of all political violence took place in the countries with extreme conflict severity levels (4/4 indicators met); 53% in countries with high conflict severity levels (3/4 indicators met); 8% in countries or territories with moderate conflict severity levels (2/4 indicators met); and 4% in countries or territories with limited conflict severity levels (1/4 indicators met).

Shifts and Trajectories: Are Conflict Trends Improving or Worsening?

To assess the trajectory of the conflicts identified by the Index, ACLED compared the change in severity scores for these countries and territories from 2018 to 2022. Country trajectories were ranked according to one of four classifications:

  • Improving – the country or territory moved from a higher severity level to a lower severity level (e.g. from Extreme Severity to High Severity)
  • Worsening – the country or territory moved from a lower severity level to a higher severity level (e.g. from High Severity to Extreme Severity)
  • Consistent: limited-to-moderate – the country or territory did not change severity level, and remains in either the Limited or Moderate Severity categories
  • Consistent: high-to-extreme – the country or territory did not change severity level, and remains in either the High or Extreme Severity categories

In total, 53 countries and territories met the threshold for this comparison in 2018 and/or 2022: 16 are consistent: limited-to-moderate; 10 are consistent: high-to-extreme; 13 are improving; and 14 are worsening. These findings indicate that most countries identified by the Index are experiencing sustained or escalating levels of severe violence.4Six countries – Lebanon, Nicaragua, Russia, Saudi Arabia, Thailand, and Zimbabwe – met the thresholds for inclusion for various indicators in 2018, but not in 2022. Conversely, six other countries – Burkina Faso, eSwatini, Indonesia, Kenya, Mozambique, and Niger – qualified for inclusion for various indicators in 2022 but not in 2018.

ACLED Index: 2022 Severity and Change in Weighted Severity Scores since 2018

Some countries in the worsening and consistent: high-to-extreme categories are home to the headline conflicts of 2022: Ukraine, Myanmar, parts of the Sahel, and Haiti. Others have received less attention, including the Democratic Republic of Congo, Yemen, Syria, and Iraq. Yet, in each case, deadliness, danger, diffusion, and/or fragmentation persisted or increased in the five years between 2018 and the end of 2022.

During this period, conflict levels remained high and severe in Bangladesh, Cameroon, Ethiopia, Honduras, Palestine, and Sudan, but they also stabilized. There are important lessons to be learned from these cases about the nature of conflict within countries that have at least partially functioning governments: they largely can and do continue to function with high levels of violence, despite the costs and harms faced by the general population. While these types of states are often regarded as ‘fragile’ or ‘failing,’ they typically persist despite instability, and their relative weakness is rarely the root cause of severe violence levels. In these countries, conflict infrequently occurs solely because of an absence or a failure of state power, but instead more commonly emerges as a result of competition for such power, shaped by local politics.

While several countries have seen notable improvements between 2018 and 2022, many of these improvements are positive only in relative terms, and few show signs of being permanent or secured. Improving countries identified by the Index include those that experienced temporary surges in violence around elections and other contests or crises during this period, such as Burundi, Lebanon, Thailand, and Zimbabwe. In other cases, specific aspects of the conflict environment, such as civilian targeting rates, have trended in a positive direction, even as overall conflict levels have remained severe. For example, Somalia saw a marginal decline in violence in 2022 compared to 2018 – falling from extreme conflict severity to high conflict severity – although it is still one of the most violent places in the world.

What is Happening in Countries with Severe Violence Levels?

Many countries in the highest severity categories share a number of characteristics:

  • They are home to multiple conflicts, often with limited overlap.
  • Their conflicts involve many active groups, and these groups are fighting each other as frequently as they clash with state security forces.
  • Militias play a leading role in the violence.
  • Peace agreements and post-conflict arrangements have changed the nature of the violence, but have failed to fully resolve or reduce it.
  • They are middle-income, have relatively high rates of development, and have adopted some democratic features. The largest growth in conflict is occurring in middle-income, democratizing countries.5This includes countries with hybrid systems, which are mixed types of political systems often created as a result of an incomplete transition from an authoritarian regime to a democratic one (or vice versa). More information on these common regime forms can be found here.

These characteristics are analyzed in further detail below.

Conflict Occurs in Many Forms

Of the 46 countries and territories that meet the threshold for inclusion on the Index, almost all host multiple concurrent conflicts that often do not fully overlap. Several countries with the highest severity levels are experiencing at least one rebellion or insurgency, where a non-state armed group is violently challenging the national government. Examples include Somalia’s al-Shabaab, Nigeria’s ISWAP, the Democratic Republic of Congo’s M23, and Islamic State and al-Qaeda affiliates in Mali. With the exception of Yemen, all of the civil war-affected countries with the highest severity levels are in Africa.6Countries are considered here to be in ‘civil war’ when the main threat to the national government (in terms of fighting proportion) is posed by a single armed, organized group with an agenda to replace the government (or a coordinated association of groups with a leading entity). In many cases, countries may face a range of threats, and the rebel entity seeking to replace the government does not constitute the main threat, nor engage in the most violence with national security forces. Nearly all of the countries experiencing insurgencies or civil wars are home to multiple other conflicts, and their civil wars often involve insurgents fighting with many other groups in addition to the government.

In all the conflicts represented by the countries on the Index, 23% of aggregated violent events involve a rebel or insurgent group, and 12% of these events (over half of all rebel/insurgent violence) are interactions with opposing state security forces. The remaining 77% of all conflict events involve other types of groups and violent interactions. Yet, much of the international community’s attention focuses on civil wars and insurgencies, elevating these cases based on the conflict agendas, rather than the proportion of violence they currently account for.7Civil wars became the dominant form of internal conflict in the post-cold war period, as international and proxy conflicts abated. Again, these changes manifested most strongly in states that democratized and developed: as political competition shifted, so too did political violence. A poor measure of change in violent group activity results in overly deterministic data collection exercises and predetermined categories of ‘seen conflict.’ For more information, see: Miller, E., Kishi, R., Raleigh, C., and Dowd, C., (2022). “An agenda for addressing bias in conflict data.” Sci Data 9, 593. This leads analysts, policymakers, and media to overlook a great deal of conflict outside these contests, at times distorting wider understanding of the overall political violence landscape.

Many countries with the highest severity levels are home to conflicts that fail to fit a standard narrative of a single insurgent group primarily contesting a government: Yemen’s wars involve at least three groups that claim a government role; Haiti is riven by gangs that control much of the country’s territory and fight amongst themselves; the Assad regime has firmly retaken most of Syria, but continues to face pockets of rebel activity. Countries like Colombia, Myanmar, and Iraq also have multiple different conflicts, many of which are regionally bound and focused on local political competition. Across these cases, co-occuring forms of political violence are common.

Overall, in many places around the world, political violence is shifting and increasingly looking more like the diffuse and fragmented conflicts in Mexico and Myanmar, and less like traditional civil wars or insurgencies, as in Somalia, where a single insurgent group has extensive territorial control and authority.

Non-State Actors Are Proliferating 

ACLED collects conflict data on thousands of non-state armed, organized groups – a number which has nearly doubled since 2019, and now includes over 3,000 actors. Non-state armed actors come in multiple forms, including rebel groups, militias, gangs, and mobs. Non-state groups were involved in 64% of all armed, organized activities in 2022, and perpetrated 76% of all violence targeting civilians.8This number excludes the actions of unidentified armed groups from consideration due to the inability to confirm that they are acting separately from the state. The number and proliferation of active non-state armed groups indicates that political violence is now more decentralized in aims, scope, and practice. The spread of non-state actors within conflict environments increases fragmentation and makes negotiation and achieving effective control very difficult.

One of the most significant aspects of this trend – and a key contributor to rising insecurity levels worldwide – is the surge in active militias and gangs, specifically. These actors pose a particular danger because they are widespread, provide cheap violent labor, and maintain local leverage.9For more information, see: Raleigh, C. (2016). “Pragmatic and Promiscuous: Explaining the Rise of Competitive Political Militias across Africa.” Journal of Conflict Resolution, 60(2), 283–310; Raleigh, C., and Kishi, R. (2020). “Hired Guns: Using Pro-Government Militias for Political Competition.” Terrorism and Political Violence 32.3, 582-603; Choi, H.J. and Raleigh, C. (2021). “The geography of regime support and political violence.” Democratization, 28:6, 1095-1114. In many cases, militias have replaced or become substitutes for traditional insurgents and rebels. These groups engage in political violence that may have, in the past, developed into civil wars, but the political environments that give rise to direct threats to regimes, and armed national contests, have become less common. Current conflicts are more flexible in form and direction, and armed groups often shift their agendas, operations, and form to best survive in new political circumstances.10ACLED regularly analyzes shifts in agents and forms of political violence over time; see, for example, our annual Year in Review series; analysis of pro-government militias; special reports on evolving activity during the COVID-19 pandemic; and more.

Despite the number of non-state armed groups, however, these actors are often poorly consolidated and engage in infighting more so than contests with the government. In many cases, non-state armed groups combat each other as frequently as they fight state security forces.11For an examination of how these trends manifest in the Sahel, for example, see The Sahel crisis since 2012

The proliferation of non-state armed groups – as well as their infighting and variable political agendas – is connected to political and criminal elite competition over control of government offices, territory, and populations. Political and criminal elites increasingly use militia groups as part of this competition, as militias can be flexible with regards to who they work with, who they attack, and what type of violence they will use.12For more information, see: Raleigh, C. (2016). “Pragmatic and Promiscuous: Explaining the Rise of Competitive Political Militias across Africa.” Journal of Conflict Resolution, 60(2), 283–310; Raleigh, C., and Kishi, R. (2020). “Hired Guns: Using Pro-Government Militias for Political Competition.” Terrorism and Political Violence 32.3, 582-603; Choi, H.J. and Raleigh, C. (2021). “The geography of regime support and political violence.” Democratization, 28:6, 1095-1114. Militias organize, ally, and demobilize frequently. Their agenda is not to replace the government or central authority, but to use violence to compete for political gains. Indeed, many regimes and governments use militias rather than engage in local violence with official forces. Flexibility is key to their success, and political competition often rewards the use of threats and repression.

Peace Treaties and Post-Conflict Arrangements Are Impacting Violence Trends

In many post-conflict settings, violence has only evolved and adapted to the new political environment, rather than abated. In the extreme severity category, three of the seven countries are considered to be in a post-conflict or truce phase: Syria, Colombia, and Yemen (which saw temporary truce for much of 2022). Yet, despite these shifts, each country has several armed, organized groups that continue to operate and sustain multiple active conflicts. The end of a war – formal or otherwise – does not necessarily entail the end of severe political violence levels.

The prevalence and persistence of armed groups in post-war states is a key feature of modern conflict, and underscores how the end of wars does not equate to peace for communities and citizens. A clear example is Syria, where 140 distinct non-state armed groups were active in 2022 due to the fragmentation of previous allies into smaller units. Over 10,000 conflict events were reported in Syria during the year, of which more than 4,000 (or 39%) involved state forces. Of these, 1,100 events were engagements between regime forces and rebel groups (11% of events). A significant level of state violence continues to be directed towards civilians, with the aim of consolidating ‘post-war’ control. A large proportion of violent events involved infighting between rebel/opposition and militia groups (more than 2,200, or 21%). Violence targeting civilians by non-state armed groups accounted for a similar share (18% of all reported events). In post-conflict settings, non-state groups frequently morph and shift their structure to survive, and continue engaging in violence. Further, governments which endure through conflicts often continue repressive violence against their citizens. The peace agreements following major wars can indeed limit intervention possibilities to mitigate or lessen these new forms of violence, and increase the risk to communities.

In Colombia, 64 distinct non-state armed groups were active in 2022. Of the more than 2,600 total reported political violence events, over 200 involved engagements between state forces and rebel opposition actors, or approximately 9% of violence overall. A similar rate was reported between non-state armed groups. The vast majority of violence was directed towards civilians by non-state actors: more than 1,600 events, or 63% of all political violence in Colombia last year.

What do these cases suggest about the potential for improvement in post-conflict contexts? The existence of multiple armed groups contesting a post-conflict space and generating violence points to spoilers becoming the main obstacle, not anti-state opposition agendas. Engagements between non-state actors and government forces may decline, but the former continue to pursue other political agendas. Rebels placated or integrated through peace agreements often have incentives to change structure, adapt, and continue activity under a different guise. This continued activity, often in the form of multiple smaller conflicts, can limit the potential for additional negotiations and peacebuilding interventions, the provision of humanitarian assistance, and even internal consolidation.

The Sharpest Rise in Conflict is in Middle-Income, Democratizing Countries

Severe conflicts occur across different political and economic contexts. A focus on ‘fragile’ or ‘failed’ states foregrounded under-development and poor governance in conflict spaces, and reinforced perceptions that high conflict rates primarily occur in poor, badly run, and often ethnically diverse places. In turn, many mitigation efforts and counter-terrorism initiatives have included intensified poverty alleviation and democratization programs.13See, for example, 2022 Prologue to the United States Strategy to Prevent Conflict and Promote Stability

But current data indicate that conflict occurs far more frequently outside of ‘fragile’ settings, and in fact takes place across the national development and wealth spectrum. At present, conflicts are growing fastest in middle-income states, while simultaneously continuing in low-income states.14This analysis is based on a comparison of ACLED conflict data with the World Bank’s country classifications by income level: 2022-2023. Conflict patterns have shifted significantly over the past 20 years, and in 2022 and into 2023, violence growth in middle-income countries with forms of democratic governments is pronounced. This suggests that it is not possible to develop out of violence in the short to medium term. Instead, it is imperative to acknowledge that political violence is taking on an increasing diversity of forms, and to identify the likely trajectories countries can take that lessen the risk of deadly, dangerous, diffuse, and fragmented types of conflict.

Shifts in political power structures and authority incentivize new forms of violence and reduce the likelihood of others.15For more information, see: Choi, H.J. and Raleigh, C. (2015). “Dominant Forms of Conflict in Changing Political Systems.International Studies Quarterly, Volume 59, Issue 1, Pages 158–171; Raleigh, C., Choi, H., and Wigmore-Shepherd, D. (2022). “Inclusive conflict? Competitive clientelism and the rise of political violence.” Review of International Studies, 48(1), 44-66; Choi, H.J. and Kim, D. (2018). “Coup, riot, war: How political institutions and ethnic politics shape alternative forms of political violence.” Terrorism and Political Violence, 30:4, 718-739; Raleigh, C. (2014). “Political hierarchies and landscapes of conflict across Africa.” Political Geography. Volume 42, 92-103; Choi, H.J. and Raleigh, C. (2021). “The geography of regime support and political violence.” Democratization. 28:6, 1095-1114. Adopting democratic practices can therefore contribute to the emergence of new types of conflict within countries, rather than the elimination of conflict altogether.16This is most evident across African and Latin American countries where democratic practices and institution building have grown in recent decades; see, for example: Carothers, T. and Feldmann, A. (2021). “Divisive Politics and Democratic Dangers in Latin America.” Carnegie Endowment for International Peace; Schedler, A. (2014) “The Criminal Subversion of Mexican Democracy.” Journal of Democracy. Volume 25, Number 1; Zovatto, D. (2020). “The rapidly deteriorating quality of democracy in Latin America.” Brookings. For example, many countries that rank as ‘partly free’ according to classifications from Freedom House – states, with elections, leader turnover and removal, inclusive representation, and other features of democratic systems17This analysis is based on a comparison of ACLED conflict data with Freedom House’s Global Freedom Status rankings. – are represented on the ACLED Conflict Severity Index, with severe violence levels largely stemming from changes to political authority at the subnational level, such as who captures and wields local power.

At the subnational level, democratic processes can create local, fragmented contests between elites for more public offices, authority, and control of spaces and populations. A rise in turf wars over political control have created incentives for competitive political violence to proliferate.

Further, in countries like Mexico, Colombia, and the Philippines, democratic and hybrid democratic/autocratic governments often share or cede control to criminal ambitions, which are at times intertwined with political motives to control office holders, territory, and populations. In these spaces, the competition for control is frequently won by non-state entities supported by political elites. Local communities are exposed to a consistent threat of violence from non-state groups that are backed or sanctioned by the state and other elites, or enjoy impunity for their actions.

The spread of these groups alongside the rise of political violence in middle-income and democratizing states underscores a central point: in many contexts, conflict is not a breakdown of governance but a contest of political authority and control. In turn, conflict does not always result from an absence of power, or take place outside areas of government authority – violent competitions can occur across all countries and at multiple scales.

The severe conflicts identified by the Index represent a range of evolving cases that require innovative approaches to thinking about, responding to, and engaging with the intended and unintended consequences of political, economic, and security changes.

Conclusion: Conflict Severity Impact

Using four key indicators – deadliness, danger, diffusion, and fragmentation – the ACLED Conflict Severity Index currently identifies 46 countries and territories experiencing severe levels of political violence. Severe conflicts are extremely difficult to resolve, respond to, and overcome, and these difficulties increase the higher they rank on the Index. An assessment of changes in Index scores from 2018 to 2022 finds that most of these countries and territories are experiencing sustained or escalating levels of severe violence.

Analysis of the latest ACLED data for countries and territories identified by the Index reveals important insights for efforts to understand ongoing evolutions in conflict and ultimately mitigate the risk of its most severe forms.

Conflict patterns are shifting away from being dominated by national-level civil wars or insurgencies. Increasingly, in many violent spaces, conflict is local and based on competitions oriented around local and regional politics. Violence is also increasingly likely to involve local militias working on behalf of political elites, governments, and other groups eager to seek control of offices, territories, and populations. These groups are not attempting to replace a government or create new states. Even in countries experiencing a civil war, other underlying conflicts can be more intense and have alternative aims.

The overall trajectory is that conflicts like those in Colombia, Mexico, and Myanmar are likely to become more common. These are conflicts of authority and control fought by and between non-state armed groups, with differing levels of direct state involvement. The war in Ukraine, while devastating in its scale of destruction and impact on world politics, remains an outlier within the broader global conflict landscape.

Conflict adapts to new political and economic circumstances, rather than dissipating. Countries with severe conflict can be found across the spectrum of economic and political development, and the sharpest rise in conflict events is occurring in middle-income countries that are democratizing or have to some extent democratized. This points to several significant conclusions: a country cannot develop out of conflict, conflict arises out of competitions for power, and there is no widely applicable solution to different forms of conflict. Positive shifts in political and economic development typically influence the form that conflict will take, but they do not necessarily resolve or prevent it. On the contrary, these shifts can result in more prolific political contests and related incentives for political violence, even if such conflicts and conflict actors fail to diffuse across the country or generate sustained danger to civilians. Shifts in political systems in recent decades have altered how political groups and elites compete. These changes have, in turn, led to adaptations in conflict and how groups engage in and with political violence.

The varied conflict landscape at the start of 2023 requires new thinking about how violence evolves and adapts to changes in political and economic environments; the diversity of forms that conflict takes; who fights, where, and when; the competition that underlies political violence; the ability of states and armed groups to consolidate and coordinate; and the spread and impact of the most severe types conflict. These are among the most critical factors to consider when responding to and analyzing modern political violence.

Appendix: Definitions and Methodology

In alphabetical order:

Cartels: Organized armed groups that do not seek to topple the state, but rather seek to control territory for the purpose of extracting exclusive economic benefits via illicit activities.18Lessing, B. (2015). “Logics of Violence in Criminal War.Journal of Conflict Resolution 59.8: 1486-1516. Examples include the Sinaloa cartel in Mexico.

Event: The fundamental unit of observation in ACLED is referred to as the “event.” Each coded event involves at least one designated actor – e.g. a named rebel group (or, in some cases, an unidentified group), the type of action carried out, a specific named location, a specific date, and other key variables. ACLED currently codes for six event types and 25 sub-event types, both violent and non-violent, that enable analysis of patterns of political violence and demonstration activity.

Executions/Assassinations: The killing of ruling elites, prominent members of society, or important political figures and dissidents. Examples include the assassination of political elites in Yemen and the assassinations of social leaders in Colombia during the COVID-19 pandemic.

Fatalities: In ACLED analysis, the use of the term fatalities always refers to reported fatalities recorded in the dataset, per ACLED fatality methodology.

Gangs: Organized criminal groups whose violent actions (e.g. battles and varied violence against civilians) can have clear political consequences despite a lack of an overt political agenda. They commonly cooperate with local elites and participate in income-seeking behavior, but vary in the degree to which they are organized and at what scale they operate. Examples include the Red Command in Brazil.

Gang Violence: Violence committed by criminal groups without an overt political agenda, also known as gangs (see above). Gang violence is only captured in the ACLED dataset in certain circumstances that have been determined to meet the parameters for inclusion based on the country-context, as outlined in the ACLED methodology brief, Gang Violence: Concepts, Benchmarks, and Coding Rules.

Insurgency: This term refers descriptively to violent activity carried out by an organized, armed, non-state group or groups internally against a governing authority, often to contest control over a territorial area. Use of the term is highly context-specific, and the aims, ideology, intensity, size, and geographic scope of insurgencies will vary.

Militias: The ACLED dataset categorizes two types of militia, political militias and identity militias, and acknowledges the actions of pro-government militias.

  • Political Militias: Armed, organized groups with political goals that use violence to advance those goals. Unlike rebel groups, political militias generally do not actively seek to topple or replace the national government using violence, though some are organized in opposition to government authority (e.g. anti-government militias in the United States that at times shift orientation based on the dominant national party or president). These groups often cooperate or ally with various domestic elites as well as external forces, albeit typically without a formal link. Examples include Loyalist militias in Northern Ireland, far-right militias and militant social movements in the United States (such as the Three Percenters or the Proud Boys), and the Bakassi Boys in Nigeria.
  • Identity Militias: Armed and violent groups organized around a collective, common feature including community, ethnicity, region, religion, or, in exceptional cases, livelihood. Therefore, identity militias captured in the ACLED dataset include those reported as tribal, clan, communal, ethnic, local, community, religious, and livelihood militias. Violent events involving identity militias are often referred to as communal violence, as these violent groups frequently act locally, in the pursuit of local goals, resources, power, security, and retribution. An armed group claiming to operate on behalf of a larger identity community may be associated with that community, but not represent it (e.g. Luo Ethnic Militia in Kenya or Fulani Ethnic Militia in Nigeria). Recruitment and participation are by association with the identity of the group. For more, see ACLED Codebook section on Actors.

Mob Violence/Vigilantism: Extrajudicial violence in response to crimes, real, perceived, or not yet committed.19Bateson, R. (2021). “The Politics of Vigilantism.” Comparative Political Studies 54.6: 923-955. This can also include the use of violence to punish social infractions or deviations from social norms. Examples include lynchings in Port au Prince, Haiti, or South Africa, as well as pandemic-related attacks against healthcare workers in India.

Organized Criminal Violence: Organized crime is the use of violence by highly organized groups in order to control (often illegal) markets and extract economic benefits. Resources that can motivate such violence can include timber, drugs, diamonds, minerals, and more. The designation of violence of this form is not based on the size of the organization or their scope. Organized violence can be localized but networked over larger spaces – such as the Mafia violence of Southern Italy;or specifically relating to a locality or entire region, such as the various agents and groups engaged in the Mexican drug war.

Political Violence: The use of force by a group with a political purpose or motivation. In analysis, this is a category used to refer collectively to ACLED’s violence against civilians, battles, and explosions/remote violence event types, as well as the mob violence sub-event type of the riot event type and the excessive force against protesters sub-event type of the protests event type. For more, see the ACLED Codebook.

Rebels: Armed, organized non-state actors that seek to challenge and topple or replace the government through the use of violence. These actors are also commonly referred to as insurgents or guerrilla fighters. They commonly coordinate with external forces and political militias/gangs. Examples include the Islamic State in the Greater Sahara, Allied Democratic Forces in the Democratic Republic of Congo and Uganda, and Southern Transitional Council armed forces in Yemen.

Rioters: Individuals who engage in violence during demonstrations and mob violence events. Violence can be directed against people, property, or both. Examples include the 2020 Delhi riots in India and riots against COVID-19 restrictions.

Vigilantes/Mobs: A group of individuals engaging in extrajudicial violence meant to punish a crime, perceived crime, or social infraction.20Bateson, R. (2021). “The Politics of Vigilantism.” Comparative Political Studies 54.6: 923-955. Examples include mobs engaging in the justicia comunitaria lynchings in Bolivia, especially within Indigenous communities, and lynching mobs in Bangladesh.

Violence Targeting Civilians: A category that encompasses all events of political violence that target civilians. This includes a broader scope than the violence against civilians event type (sexual violence, attack, and abduction/forced disappearance sub-event types). It is inclusive of the aforementioned sub-event types, the excessive force against protesters sub-event type, as well as explosions/remote violence and riots event types which involve civilians or protesters, but excludes the peaceful protest and protest with intervention sub-event types. Violence targeting civilians is typically used in ACLED analysis because it reflects the widest scope of violence faced by civilians recorded in the dataset.

The aim of the index is to measure and compare conflict rates in countries accurately and systematically. We use four dimensions to compare conflicts over the course of 12 months: ACLED events from January 2022 to December 2022 are used to design and test the four indicators that compose the severity index. These indicator measures were then re-tested on close variants of these measures (June 2021-June 2022, October 2021-October 2022, etc.) for robustness.All ACLED events are collected using the same methodology, allowing a comparison of event numbers for armed, organised violence. The event types included in this index (and whose characteristics are the basis of the indicators) are political violence events by armed, organized groups.21This includes some events involving violent groups that are organized spontaneously which are recorded as mob violence (see ACLED Codebook).


Deadliness (How fatal are political violence events)? The amount of political violence-related fatalities can indicate how intense conflict is within a state. The deadliness indicator represents the number of reported fatalities per country from January-December 2022.

All fatalities from political violence events are aggregated by country to determine the mean number of fatalities. Countries that fall above the mean are included in this indicator.

Mean: 882

Standard Deviation: 3075

Number of countries with above average rates: 24

Danger (What proportion of political violence events are directed towards civilians?) Conflicts differ with respect to how much armed groups — including state actors – prey on civilians. Conflicts that have a higher rate of civilian violence are likely to continue and proliferate, in part because armed groups are not facing more active resistance from other armed entities.

To assess the danger to civilians, ACLED’s classifies events as‘violence targeting civilians’. These events are then aggregated by country and standardized by population (events per 100k people). Countries with a rate above the median are included in this indicator.

Mean: 3 fatalities per 100k people

Standard Deviation: 4

Countries with above average rates: 21

Diffusion (How many of a country’s subnational regions have a high level of violence?) Many conflicts can occur in a country simultaneously, adding to the geographic spread of conflict across states. This measure is an assessment of the geographic distribution of conflict. For each country, the percentage of first level administrative units with a high average weekly event count is calculated. A ‘high’ average weekly event count is defined as 1.5 or more events per week. This threshold is consistent with ACLED’s Volatility & Risk Predictability Index. Countries with no administrative divisions meeting this threshold are excluded from this indicator.

Mean: 34

Standard Deviation: 27

Countries with ‘high’ event count admins: 39

Fragmentation (How many non-state armed, organized groups are operating within the conflict?)

The fragmentation of a conflict environment indicates the number of distinct threats and agendas that are accumulating in a given context and posing harm to communities and state institutions. It also indicates the number of distinct political motives and opportunities to form an armed group. A singular consolidated armed group can be a serious challenge to governments, but can take part in effective negotiations and engagements. A highly fragmented environment, in contrast, makes it more difficult to engage the necessary actors in effective negotiations and may indicate multiple overlapping conflicts that are more challenging to simultaneously resolve.

This indicator is a count of all armed, organized, active rebel and political militias per country in 2022 (excluding unidentified armed groups and including pro-government militias). In addition, a maximum of one communal militia is counted per first level administrative unit per country. Due to the heavily skewed distribution of active groups within countries, countries that fall within the top 80% of actor counts qualify for this indicator.

Threshold: 25

Standard Deviation: 310

Countries with above average rates: 19

Ranking process:

Countries are ranked across these dimensions based on their value for each separate indicator. Countries can appear in none of the indicators (if below the threshold in all measures), or if they experience fewer than 10 political violence events in the course of the year. Countries can appear in one, two, three or all four of the indicators if they place above the respective thresholds. The number of indicator rankings that a country falls in determines the ‘severity’ of their conflict context.

Weighted scores:

Each country is also given a weighted score based on the number of indicators it is included in as well as where it ranks within that indicator. Each indicator list is first ranked in ascending order (with lower values ranked at the top of the list). Each country’s ranking is then aggregated across the indicators where it appears. Then, the aggregated ranking value is multiplied by the number of indicator lists the country appears in. For example, if Afghanistan has a ranked value of four, five, and six across three indicators, its weighted score is (4+5+6)*3=45.

This weighting system emphasizes countries that score both highly in the rankings as well as appear across multiple indicators. Still, it is possible for a country that is highly ranked in just two indicators to have a higher weighted score than a country that is ranked low across three indicators. The weighted scores are useful in differentiating the intensity of conflict within countries in the same severity category.