Terrorism reaches far beyond the place where an attack happens, reshaping everyday life, public policy and people’s sense of security. Yet its impact is highly uneven: a small number of countries, organisations and exceptionally deadly incidents account for much of the recorded harm. These charts follow the changing global picture, compare national experiences and examine who is targeted, which groups are implicated and how European authorities respond. Together, they show why attack counts, deaths and arrests tell different parts of the story. Behind every number are people and communities whose losses cannot be captured by a single measure of violence.
Historical charts use GTD records through 2020. The impact map and country table use GTI 2026 scores for 2025; EU arrests run through 2025. These sources cover different definitions and periods and are not joined into one continuous series.
Statistics
- The rise and fall of recorded terrorism deaths
- How the geography of terrorism changed
- Where terrorism has the greatest impact
- The countries moving in opposite directions
- More attacks do not always mean more deaths
- When attacks and deaths take different paths
- Who and what are attacked
- The organisations associated with recorded deaths
- A small share of incidents causes much of the harm
- How the European response is changing
- Country comparison
1. The rise and fall of recorded terrorism deaths
Worldwide · 1970–2020
Recorded deaths reached their highest point in this series in 2014, when the database records 44,648 fatalities. By 2020 the total had fallen to 22,847, but that decline still left a substantial human toll. The mid-2010s surge coincided with severe violence in Iraq, Syria, Afghanistan and Nigeria, showing how a few major conflicts can reshape the global picture. The subsequent fall should therefore be read alongside regional and national trends: improvement in one large hotspot can outweigh deterioration elsewhere. These totals include perpetrators where the source records them, so they measure the lethality of incidents rather than civilian losses alone.
GTD May 2022 release. 1993 is missing, not zero. Collection changes in 1998, 2008 and 2012 affect long-run comparability.
2. How the geography of terrorism changed
Recorded deaths by region · 2000–2020
The global total hides several different regional stories. The Middle East and North Africa drove much of the mid-2010s surge, while South Asia remained a major centre of recorded fatalities as that surge receded. Sub-Saharan Africa also accounts for a substantial part of the later burden. This matters because a falling worldwide total does not necessarily mean that violence is retreating everywhere. Armed organisations can lose territory in one theatre while other movements intensify elsewhere. Regional comparisons are therefore most useful as a guide to where the human burden is shifting, rather than as evidence of a single worldwide trajectory.
GTD regions; smaller regions are combined as Other regions. Coverage and classification differ from the newer GTI map.
3. Where terrorism has the greatest impact
Global Terrorism Index 2026 · scores for 2025
Pakistan has the highest impact score in this edition, followed by Burkina Faso, Niger and Nigeria. The prominence of several African countries shows how far the centre of gravity extends beyond the conflicts that dominated international headlines a decade earlier. Persistent insecurity around borders and contested rural areas can give armed groups room to operate, although national scores cannot establish the causes in each place. A high score also reflects the continuing weight of recent years, rather than only the latest year’s deaths. The map therefore captures accumulated impact; it should not be read as a traveller’s personal probability of harm.
Score 0–10; higher means greater impact. GTI weights incidents, deaths, injuries and hostages over five years. Grey means not covered; pale blue means zero. Boundaries imply no position on sovereignty.
Source: Institute for Economics & Peace, Global Terrorism Index 2026.

Remembering people behind the numbers
A memorial offers a different perspective from a statistical series. At the Oklahoma City National Memorial, a place associated with sudden violence has become a setting for remembrance and reflection. The photograph provides a pause in the comparisons on this page: each total represents individual lives and the families, colleagues and communities affected by them. Numbers can show where losses are concentrated and how patterns change, but they cannot measure grief or the long process of recovery. Reading the charts alongside places of remembrance helps keep that distinction in view, without reducing very different national experiences to a competition over suffering.
4. The countries moving in opposite directions
Largest absolute changes in recorded deaths · 2019–2020
Afghanistan had the largest increase among these countries, rising from 8,257 recorded deaths in 2019 to 10,081 in 2020. Ethiopia and the Democratic Republic of the Congo also show large increases, while Yemen and Iraq recorded substantial declines. These movements help explain why a global trend can feel disconnected from events in a particular country. A change may reflect a prolonged campaign, an exceptionally deadly episode or altered reporting conditions. Sri Lanka’s decline after the 2019 Easter attacks is a reminder that a sharp improvement can follow an extraordinary spike; it does not by itself demonstrate that every underlying risk has disappeared.
Five largest increases and five largest decreases in absolute recorded deaths. Blue = 2019; orange = 2020. Unknown death counts are excluded from death totals.
5. More attacks do not always mean more deaths
Countries with recorded attacks · 2020
Countries with similar numbers of attacks can experience very different death tolls. That difference matters for understanding the burden of terrorism: repeated incidents with few fatalities and fewer, much deadlier attacks create different kinds of insecurity. Afghanistan stands apart in this year’s records, combining a high attack count with a particularly large death toll. The comparison also invites closer attention to targets, weapons and the wider conflict environment, all of which can influence how lethal an incident becomes. The relationship is descriptive, however. It cannot tell us that attack frequency itself causes greater lethality, or that countries with fewer recorded incidents are free from intimidation.
Each point is one country. Attack totals include incidents with unknown death counts; death totals exclude those unknown values. The focused view retains countries with fewer than 500 recorded attacks.
6. When attacks and deaths take different paths
Afghanistan, Iraq, Nigeria and Pakistan · 2010–2020 · 2010 = 100
Following attacks and deaths within the same country reveals changes that a single year’s comparison cannot show. When deaths rise faster than attacks, the recorded violence is becoming more lethal on average; when they fall faster, the balance moves the other way. The selected countries experienced different combinations of escalation and decline during this decade, reflecting the changing intensity of their conflicts and militant campaigns. Their starting levels also differ substantially, so a large percentage increase need not mean the largest absolute death toll. Read each country’s path as its own history, and use the preceding comparison to restore the sense of scale.
Each country and each measure is indexed separately to 2010 = 100. Country selection does not change the baseline year. Index values do not compare absolute national totals.
7. Who and what are attacked
Primary target categories by region · 2016–2020
The target mix shows that terrorism affects both everyday civilian life and institutions associated with state power. Attacks on private citizens and property can spread fear well beyond those directly harmed, while attacks on police, military personnel and government targets can challenge security and authority. Regional differences suggest that these purposes occur in different combinations. They also help explain why the word terrorism can evoke very different experiences in different places. A region dominated by attacks on security forces is not necessarily safer for civilians: displacement, disruption and indirect harms are not captured by counting the primary target of each recorded incident.
Attacks are counted once using their primary target category. Other / unknown includes every remaining target category. These are attack shares, not death shares.
8. The organisations associated with recorded deaths
Primary recorded perpetrator · 2016–2020
The Taliban and Islamic State of Iraq and the Levant account for large shares of the deaths attributed to named organisations in this period. Their prominence reflects the intensity of violence in the conflicts where they operated, rather than a uniform global presence. Boko Haram also figures prominently, reinforcing the importance of the Lake Chad region to the wider story. Unknown perpetrators remain a separate part of the total because uncertainty is itself significant: many incidents cannot confidently be linked to a named group. Attribution in the database is a research classification, and should not be mistaken for a judicial finding about every incident.
Each incident is assigned to its primary recorded perpetrator; additional co-attributions are not counted again. Other reported attributions includes smaller organisations and nonspecific labels, such as Fulani extremists. Unknown perpetrator remains separate. Separate named affiliates are not merged into a parent organisation.
9. A small share of incidents causes much of the harm
Incidents with known death counts · 2016–2020
The deadliest tenth of incidents accounts for about 65% of recorded deaths in this period. That concentration helps explain why a relatively small change in the number of exceptionally lethal attacks can shift an annual death toll so sharply. Many recorded incidents cause no deaths, but they may still injure people, damage livelihoods or intimidate communities. Looking only at fatalities would miss those effects; looking only at attack counts would miss the unequal distribution of lethal harm. The two measures belong together. Concentration also cautions against treating a quieter year as proof of a durable improvement when a few events can alter the picture.
Known death counts only, including zero-death incidents. The horizontal axis runs from the least deadly incidents to the most deadly. The diagonal shows equal concentration, not a forecast.

The institutions behind the response
Europol’s headquarters in The Hague represents the cooperative side of Europe’s response to cross-border crime and terrorism. The arrests chart brings together reports from national authorities, whose investigations operate within different legal and institutional settings. That makes the figures useful for following recorded enforcement activity, while also making their limits important. An arrest does not establish guilt, and a change in arrests can reflect the timing or scope of investigations as well as changes in threats. The people represented by these figures are therefore distinct from attack totals and recorded fatalities, even when all three measures contribute to the public debate.
10. How the European response is changing
EU terrorism-related arrests by reported category · 2023–2025
EU Member States reported 486 terrorism-related arrests in 2025, compared with 449 in 2024 and 426 in 2023. Arrests classified as jihadist rose from 289 to 347 between the last two years and account for most of the increase. Other categories moved differently: right-wing and left-wing or anarchist arrests declined, while ethno-nationalist and separatist arrests increased. This is a picture of law-enforcement activity, influenced by investigations, legal definitions and reporting practices as well as threats. Arrests may concern support, recruitment or other offences, and do not establish guilt. They cannot be translated directly into a count of attacks prevented.
Europol TE-SAT 2026, annex, printed page 49. Reported arrests are not convictions or attack counts; categories reproduce Member State reporting.
Source: Europol, EU TE-SAT 2026.
11. Country comparison
Table 1 · Global Terrorism Index 2026 · 163 countries · 2025 scores
163 countries
| 1 | Pakistan | 8.574 | +0.201 |
|---|---|---|---|
| 2 | Burkina Faso | 8.324 | -0.257 |
| 3 | Niger | 7.816 | +0.029 |
| 4 | Nigeria | 7.792 | +0.159 |
| 5 | Mali | 7.586 | -0.310 |
| 6 | Syria | 7.545 | -0.461 |
| 7 | Somalia | 7.391 | -0.235 |
| 8 | Democratic Republic of the Congo | 7.171 | +0.285 |
| 9 | Colombia | 7.116 | +0.528 |
| 10 | Israel | 6.790 | -0.674 |
| 11 | Afghanistan | 6.678 | -0.610 |
| 12 | Cameroon | 6.593 | -0.390 |
| 13 | India | 6.428 | -0.136 |
| 14 | Myanmar | 6.245 | -0.696 |
| 15 | Mozambique | 6.022 | -0.229 |
| 16 | Iraq | 5.822 | -0.770 |
| 17 | Russia | 5.593 | -0.574 |
| 18 | Iran | 5.477 | -0.605 |
| 19 | Benin | 5.434 | +0.632 |
| 20 | Thailand | 5.275 | +0.446 |
| 21 | Kenya | 5.088 | -0.278 |
| 22 | Palestine | 4.800 | -0.223 |
| 23 | Philippines | 4.719 | -0.521 |
| 24 | Indonesia | 4.714 | +0.338 |
| 25 | Yemen | 4.653 | -0.427 |
| 26 | Chad | 4.625 | -0.777 |
| 27 | Chile | 4.553 | -0.615 |
| 28 | United States of America | 4.521 | +0.980 |
| 29 | Germany | 4.447 | -0.311 |
| 30 | Togo | 4.305 | -0.699 |
| 31 | Australia | 3.732 | +1.620 |
| 32 | Egypt | 3.465 | -0.951 |
| 33 | Burundi | 3.361 | -0.682 |
| 34 | Uganda | 3.250 | -0.452 |
| 35 | France | 3.224 | +0.477 |
| 36 | Turkiye | 3.212 | -0.756 |
| 37 | Ecuador | 3.063 | +1.513 |
| 38 | United Kingdom | 2.936 | +0.154 |
| 39 | Ukraine | 2.927 | +0.924 |
| 40 | Greece | 2.788 | -0.196 |
| 41 | Tajikistan | 2.602 | +0.781 |
| 42 | Bangladesh | 2.286 | -0.677 |
| 43 | Oman | 2.282 | -0.645 |
| 44 | Jordan | 2.268 | -0.645 |
| 45 | Czechia | 2.261 | -0.645 |
| 46 | Sweden | 1.839 | -0.003 |
| 47 | Algeria | 1.766 | -0.649 |
| 48 | Poland | 1.682 | -0.353 |
| 49 | Peru | 1.572 | -0.588 |
| 50 | Tunisia | 1.522 | -0.662 |
| 51 | Austria | 1.498 | +0.916 |
| 52 | Netherlands | 1.475 | +0.073 |
| 53 | Canada | 1.333 | -0.617 |
| 54 | China | 1.311 | -0.552 |
| 55 | Belgium | 1.198 | -0.337 |
| 56 | Angola | 1.136 | -0.521 |
| 57 | Malaysia | 1.092 | -0.534 |
| 58 | Senegal | 1.070 | -0.508 |
| 59 | Libya | 1.007 | -0.605 |
| 60 | Italy | 0.999 | -0.280 |
| 61 | Djibouti | 0.925 | -0.536 |
| 62 | Brazil | 0.909 | -0.521 |
| 63 | Tanzania | 0.888 | -0.913 |
| 64 | Spain | 0.794 | -0.462 |
| 65 | Serbia | 0.782 | +0.200 |
| 66 | Bosnia and Herzegovina | 0.782 | -0.436 |
| 67 | United Arab Emirates | 0.749 | -0.429 |
| 68 | Switzerland | 0.749 | -0.516 |
| 69 | Norway | 0.725 | -0.473 |
| 70 | Denmark | 0.720 | +0.000 |
| 71 | Cote d’Ivoire | 0.702 | -0.752 |
| 72 | Lebanon | 0.648 | -0.589 |
| 73 | Slovakia | 0.616 | -0.407 |
| 74 | Finland | 0.582 | -0.367 |
| 75 | Japan | 0.571 | -0.378 |
| 76 | Central African Republic | 0.556 | -0.401 |
| 77 | Georgia | 0.506 | -0.335 |
| 78 | Kosovo | 0.465 | -0.317 |
| 79 | Argentina | 0.455 | -0.346 |
| 80 | Saudi Arabia | 0.443 | -0.402 |
| 81 | Portugal | 0.423 | +0.423 |
| 82 | Armenia | 0.423 | -0.297 |
| 83 | Ireland | 0.423 | -0.297 |
| 84 | Lithuania | 0.423 | -0.297 |
| 85 | Venezuela | 0.396 | -0.314 |
| 86 | South Korea | 0.333 | -0.249 |
| 87 | Mexico | 0.325 | -0.257 |
| 88 | Ethiopia | 0.288 | -0.499 |
| 89 | Nepal | 0.288 | -0.825 |
| 90 | Cambodia | 0.233 | -0.190 |
| 91 | Latvia | 0.233 | -0.190 |
| 92 | Cyprus | 0.176 | -0.171 |
| 93 | Azerbaijan | 0.123 | -0.110 |
| 94 | Belarus | 0.123 | -0.110 |
| 95 | New Zealand | 0.114 | -0.103 |
| 96 | Uzbekistan | 0.114 | -0.119 |
| 97 | Iceland | 0.059 | -0.064 |
| 98 | Eswatini | 0.044 | -0.043 |
| 99 | Bahrain | 0.030 | -0.029 |
| 100 | Albania | 0.000 | +0.000 |
| 100 | Bhutan | 0.000 | +0.000 |
| 100 | Bolivia | 0.000 | +0.000 |
| 100 | Botswana | 0.000 | +0.000 |
| 100 | Bulgaria | 0.000 | +0.000 |
| 100 | Costa Rica | 0.000 | +0.000 |
| 100 | Croatia | 0.000 | +0.000 |
| 100 | Cuba | 0.000 | +0.000 |
| 100 | Dominican Republic | 0.000 | +0.000 |
| 100 | El Salvador | 0.000 | +0.000 |
| 100 | Equatorial Guinea | 0.000 | +0.000 |
| 100 | Eritrea | 0.000 | +0.000 |
| 100 | Estonia | 0.000 | +0.000 |
| 100 | Gabon | 0.000 | +0.000 |
| 100 | Ghana | 0.000 | +0.000 |
| 100 | Guatemala | 0.000 | +0.000 |
| 100 | Guinea | 0.000 | +0.000 |
| 100 | Guinea-Bissau | 0.000 | +0.000 |
| 100 | Guyana | 0.000 | +0.000 |
| 100 | Haiti | 0.000 | +0.000 |
| 100 | Honduras | 0.000 | +0.000 |
| 100 | Hungary | 0.000 | +0.000 |
| 100 | Jamaica | 0.000 | +0.000 |
| 100 | Kazakhstan | 0.000 | +0.000 |
| 100 | Kuwait | 0.000 | +0.000 |
| 100 | Kyrgyz Republic | 0.000 | +0.000 |
| 100 | Laos | 0.000 | +0.000 |
| 100 | Lesotho | 0.000 | +0.000 |
| 100 | Liberia | 0.000 | +0.000 |
| 100 | Madagascar | 0.000 | +0.000 |
| 100 | Malawi | 0.000 | +0.000 |
| 100 | Mauritania | 0.000 | +0.000 |
| 100 | Mauritius | 0.000 | +0.000 |
| 100 | Moldova | 0.000 | +0.000 |
| 100 | Mongolia | 0.000 | +0.000 |
| 100 | Montenegro | 0.000 | +0.000 |
| 100 | Morocco | 0.000 | +0.000 |
| 100 | Namibia | 0.000 | +0.000 |
| 100 | Nicaragua | 0.000 | +0.000 |
| 100 | North Korea | 0.000 | +0.000 |
| 100 | North Macedonia | 0.000 | +0.000 |
| 100 | Panama | 0.000 | +0.000 |
| 100 | Papua New Guinea | 0.000 | +0.000 |
| 100 | Paraguay | 0.000 | -0.073 |
| 100 | Qatar | 0.000 | +0.000 |
| 100 | Republic of the Congo | 0.000 | +0.000 |
| 100 | Romania | 0.000 | +0.000 |
| 100 | Rwanda | 0.000 | +0.000 |
| 100 | Sierra Leone | 0.000 | +0.000 |
| 100 | Singapore | 0.000 | +0.000 |
| 100 | Slovenia | 0.000 | +0.000 |
| 100 | South Africa | 0.000 | +0.000 |
| 100 | South Sudan | 0.000 | +0.000 |
| 100 | Sri Lanka | 0.000 | +0.000 |
| 100 | Sudan | 0.000 | +0.000 |
| 100 | Taiwan | 0.000 | +0.000 |
| 100 | The Gambia | 0.000 | +0.000 |
| 100 | Timor-Leste | 0.000 | +0.000 |
| 100 | Trinidad and Tobago | 0.000 | +0.000 |
| 100 | Turkmenistan | 0.000 | +0.000 |
| 100 | Uruguay | 0.000 | -0.059 |
| 100 | Vietnam | 0.000 | +0.000 |
| 100 | Zambia | 0.000 | +0.000 |
| 100 | Zimbabwe | 0.000 | +0.000 |
The country comparison puts the map’s pattern into a form that can be searched and checked. A positive score change means that measured impact increased; a negative change means it fell. Rank describes a country’s position relative to others, whereas the score change describes movement in its own measured impact. Those are different questions. Countries with a zero score share rank 100 in this edition, which does not mean that they have never experienced terrorism. The five-year weighting also means that an earlier attack can continue to influence a score after the immediate annual toll has fallen.
Source: Global Terrorism Index 2026, Appendix A, printed pages 83–84. Score changes are not rank changes. Annual attack and death figures are not inferred from index scores.
Other pages
Espionage · Russia–Ukraine war · European Union
Sources and image credits
GTD death totals sum known nkill values, including perpetrators where recorded; unknown death counts are excluded rather than treated as zero. All source-coded incidents are retained, including cases with source uncertainty. The database does not measure all conflict deaths or all forms of political violence. GTI and Europol use their own definitions. The pack includes aggregate data and a methodology note.
Banner: National Park Service, Statue of Liberty and World Trade Center fire, 11 September 2001, public domain. Original image: 350 × 265 pixels. Body images are credited beside each photograph. Map geometry: Natural Earth, public domain.
