Recent scholarly work suggests a decline in violence rates generally but these conclusions are quickly being challenged. Recent predictions suggest that interstate, national and substate violence is likely to increase in the near future (see Figure 1a). The character of that conflict is unlikely to be the same as the past: fewer battles and typical ‘state vs opponent’ conflicts, and far more diffuse and diverse disruption. African states already experience multidimensional conflict that has change in character in the past ten years. Typical African states experience many forms of political violence and protest, with low levels of overlap between groups and scales (see Figure 1b). Battles continue to decrease, while rioting, protesting, bombing and violence against civilians continues. The number of events across Africa for 2016 was largely similar to that of 2015 (at 17,539 ACLED events in 2016 and 17,537 ACLED events in 2015). However, fatalities have decreased between these years, from over 36,000 to 30,000 (see Figure 1c).
What explains increased event numbers? On-going improvements in the reporting and coverage of violent conflict, as well as improvements to data collection techniques within the ACLED project, contribute to an increase in the number of recorded conflict events. Yet the rising trend in multiple forms of political violence stands (please see the “Agents of Violence Trends” for a discussion of dominant forms of political violence across Africa).
The magnitude of the increase in the past twenty years is unknown: due to changes in source availability that better record the activities of violent agents locally, recent years see an increase in reported conflict events. There is no way to verify how much conflict has increased or decreased over periods of decades or longer. What is clear is that the types of violence occurring in Africa during the recent past have changed, and that similar methods of collecting political violence and protest data over the past years demonstrates an increase and stable high rate of event occurrence.
Across Africa in 2016, approximately 41% of organized armed conflict events  involved battles between armed groups, approximately 13% involved remote violence, and approximately 45% involved violence against civilians (see Figure 1). These proportional rates have remained relatively consistent over time, especially in recent years.
The estimated number of reported fatalities stemming from organized armed conflict events decreased last year from 2015 and the proportion of reported civilian fatalities similarly decreased last year. Approximately 27% of all reported conflict deaths resulting from organized armed conflict events (over 8,030 fatalities) resulted from violence against civilians in 2016, compared to more than 37% in 2015 (over 13,400 fatalities) (see Figure 2).
Civilians are most at risk in Burundi, Somalia and Sudan; these countries experienced 707, 597, and 578 instances of violence against civilians in 2016, respectively (see Figure 3). In terms of civilian deaths, however, Nigeria continues to be the most dangerous by far, responsible for over 2,063 civilian deaths in 2016 (please see the “Violence Against Civilians Trends” and the “Violence by Country Trends” for a discussion of civilian targeting and political violence in Nigeria, respectively).
Remote violence tactics have continued to make up a larger proportion of conflict activity in Africa in 2016; over 13% of organized armed conflict last year were remote violence events; a similar level to 2015 and up from around 10% in 2013 (see Figure 1). These attacks have also become more deadly: the proportion of reported conflict-related fatalities attributed to remote violence increased slightly from 7.5% of reported fatalities in 2015 to 9.5% of reported fatalities in 2016 (see Figure 2).
The relative proportion of organized armed conflict events made up of battles between armed groups remained relatively constant from 2015 to 2016 (from 43.2% of events to 41.3% of events) (see Figure 1). The majority of battles continue to occur in Somalia (over 30% of all battles in 2016); these battles consist mainly of contests between Al Shaabab against military and AMISOM forces. Reported battle-related fatalities primarily occurred in Somalia (over 20% of battle-related deaths) – comprised mainly of contests involving Al Shaabab against military and AMISOM forces – and Sudan (15.8% of battle-related deaths) – comprised largely of contests between government and SPLM-N and SLM/A-Nur rebel forces.
In addition to organized armed conflict (i.e. battles between armed groups, remote violence, and violence against civilians), riots and protests contribute largely to political disturbance in Africa (see Figure 4). A riot is defined as a violent disturbance of the public peace by three or more persons assembled for a common purpose, and can be either spontaneous or organized. Protests are nonviolent, spontaneous organizations of civilians for a political purpose. While the above review examines organized armed conflict specifically, riots and protests comprise a large proportion of political activity in Africa (approximately 40% of political conflict in 2016).
The majority of rioting and protesting in 2016 occurs in South Africa (20.04% of riots and protests in 2016) and Tunisia (11.25% of riots and protests in 2016). In South Africa, riots and protests are largely driven by two dynamics: protests over poor service delivery and corruption that occurred around local government elections and student protests on university campuses under the auspices of the #FeesMustFall campaign. Tunisia, witnessed a large escalation in the rate of rioting and protest – particularly in January where protests reached levels unprecedented since January 2011 – as protest movements demanded development and employment.
Historically, 1999 witnessed the highest proportion of battles between armed actors, as well as the highest proportion of reported battle-related fatalities as a proportion of reported conflict deaths, largely due to violence in Angola, DR-Congo, and Sierra Leone. The years 2002-2005 witnessed the highest proportion of violence against civilians, driven by several contexts including the Zimbabwean elections, Ugandan LRA activity, the Darfur conflict in Sudan and the peace talks over the civil war in DR-Congo. 2011 witnessed the highest proportion of remote violence, as well as the highest proportion of reported remote violence-related fatalities as a proportion of reported conflict deaths, largely due to air strikes and bombings stemming from the Libyan civil war, as well as Al Shabaab activity in Somalia. 2016 witnessed the highest proportion of riots and protests, driven primarily by events in South Africa and Tunisia as outlined above, as well as protests that spread from the Oromia region of Ethiopia to the Amhara region.
A relative drop in violence is apparent from 2004-2006, when a number of high-intensity conflicts (e.g. civil wars in Liberia and Cote d’Ivoire) ended. Other areas (e.g. the Darfur region and the Congolese Kivus) also experienced a relative fall in violence during these years. Additionally, the operational and logistical environment in Uganda changed over this period, and the activity of the Lord’s Resistance Army – which had been a prominent actor in the region, and proportionally on the continent as a whole – fell with their diffusion into neighboring states, where they continued their activity but at a much-reduced rate and with significantly reduced capacity.
 Organized armed conflict events include battles between armed groups, remote violence, as well as instances of violence against civilians.
 Fatality figures report total fatalities from both sides combined for each violent event, and are not attributed to losses from certain sides. However, in instances of violence against civilians, as civilians (by ACLED’s definition) do not exhibit aggression, fatalities associated with these events can be attributed to civilian fatalities. As fatality figures are often difficult to obtain, verify, and cross-check, and are subject to higher levels of reporting bias than overall conflict events, ACLED codes the lower end of reported ranges, and in instances when only a total number of fatalities is given for multiple events taking place across more than one day or in more than one place, the number of fatalities is divided amongst the events evenly.
 By definition, protesters do not engage in violence. Hence, if violence occurs during a protest as a result of protesters’ actions, it is a riot and not a protest. If violence is done to protesters in the event of a protest, the event is deemed an instance of violence against civilians as, by definition, the protesters were targeted while not engaging in any violence.