By Emma Elfversson & Kristine Höglund

October 27, 2021

As the recent takeover by the Taliban in Afghanistan unfolded, Afghan cities featured prominently in the reporting. City after city, and finally the capital Kabul, fell into the hands of the Taliban. In many cases, these takeovers happened without major violence or resistance from the security forces of the former government. But Afghan cities have also been arenas for large-scale fighting and attacks during the war between the internationally-supported former government and the Afghan Taliban. This violence has included terror attacks against civilian targets, a form of violent tactic that often targets cities. Recent wars in Ukraine and Syria have similarly seen major battles in cities. Examples like these, as well as broader debates in the literature on urban violence, suggest that cities are increasingly becoming the main battlegrounds in contemporary armed conflict. However, is there evidence to support this analysis?

In a recent study, we set out to answer this question. We investigated global patterns of conflict using geo-referenced data. We focused on armed conflicts between states and between states and rebel groups, and looked at the proportion of conflict-related violence that has taken place in cities over a period of almost 30 years (1989–2017). We included cases where actors involved in armed conflicts carried out violence against civilians, including terror attacks such as those by the Islamic State in cities like Mosul during the war in Iraq, and by al-Shabaab in Mogadishu, Somalia.

In our analysis, we do not find any evidence that armed conflicts are increasingly affecting cities. If anything, the share of violence in cities incurred by armed conflicts declined. These findings should lead us to reflect more about whether and how our “urban age” affects the role of cities in war. Importantly, the findings speak against an all-embracing securitization of cities as means to address urban violence. Instead, it calls for more nuanced understandings of how urbanization promotes or works against different forms of conflict and its varying effects on cities at war. 

The urban dimension of armed conflicts

In our analysis, we used data from the Uppsala Conflict Data Program on incidents of fatal violence in armed conflict. We looked at which of these incidents took place in cities with at least 100,000 inhabitants. When charting the total levels of conflict violence in the world – the number of fatalities in armed conflict, and in cities specifically – we found that it varies considerably. We find that levels of conflict fatalities in cities were very high both in the beginning and in the end of the period that we selected. In our main analysis, we shift the focus to the share of total fatalities that took place in cities. For each country in conflict, we look at the proportion of fatalities that took place in and outside of cities, over time. We found that since 1989, the average share of lethal violence in cities during conflicts across the world has decreased.

These findings may sound surprising given the global trend of increasing urbanization, and widely held perceptions of more urban warfare as a result. However, there are many reasons to expect that the role of cities in wars has been relatively constant over time. First, cities (and particularly capitals) have always been central targets of armed conflict. Even in armed conflicts that are largely rural, and where armed opposition mobilizes in peripheral areas, a key objective is often to secure (or maintain) control of major cities. The war in Afghanistan highlights that cities may be particularly important during certain stages of the conflict. Second, research underlines that cities are often safe havens or spaces for resistance to violence. So clearly, the effect of urbanization on the nature of warfare is more complex than mere causing an increase in urban warfare.

We also note that the pattern we see resembles a U-curve: the share of conflict-related violence in cities has been on the rise over the last decade or so, but the levels were even higher in the early 1990s. Put differently, in the time frame of our study, armed conflicts first became less and less urban in nature, but in the past 10 to 15 years they became more urban again. Specific moments in history, thus, appear to be associated with more urban violence. In particular, the early 1990s witnessed several highly urban conflicts, such as the war in Angola. In this armed conflict, siege warfare was a major feature in the battle between the government and the rebel group UNITA and centered on key cities like Huambo and Cuito.

The broader perspective

Our study focuses on a particular form of urban violence: armed conflict that involves fighting between states or between government forces and organized rebel groups. For this reason, the findings do not necessarily contradict broader claims that pertain to how urban contestation is increasingly common. Such claims often refer to broader categories of violence, including criminal violence, violence carried out by political militias, and police violence against civilian protests. We do not assess the claim by military strategists, and urban and development scholars, that the nature of political conflict has shifted from conventional warfare to forms of violence that are inherently more urban in character. What we do argue is for the importance of distinguishing and systematically assessing the urban dynamics of clearly defined forms of violence. Different dynamics drive different forms of urban violence.

We also observe additional patterns of armed conflict violence in cities that warrant more attention in future studies. For one, we observe regional patterns: armed conflicts in Latin America have been comparatively less urban than other parts of the world, and in Africa, we do find an indication that the share of conflict violence in cities has increased, but the increase is not statistically significant. Such nuances are in line with recent research on Africa that has emphasized the rising importance of cities on the continent. We also see a slightly different pattern for armed conflicts that are more internationalized. For instance, in the Middle East, civil wars with international involvement have experienced more fatalities in cities. One potential factor driving the urban nature of conflict in the Middle East could be that the highly internationalized war on terror has to a large extent been fought in this region.

So, to move forward in the study of armed conflict and cities, we need to pay closer attention to the circumstances under which armed conflict becomes more urban. This endeavor entails considering facets such as the city-level pace of urbanization, target proneness, level of internalization and conflict dynamics over time. Such analysis can lay the foundation for conflict-prevention strategies rooted in more nuanced appreciation of the factors driving particular armed conflicts with a prominent urban dimension. 

Emma Elfversson is Assistant Professor at the Department of Peace and Conflict Research, Uppsala University. Her research interests concern ethnic politics and communal conflict, the role of state and non-state actors in addressing communal conflicts, and rural/urban dimensions of organized violence. You can follow her on Twitter @emmaelfversson

Kristine Höglund is Professor of Peace and Conflict Research at Uppsala University. Her research interests cover urban conflict, violence and conflict resolution, issues related to the dynamics of peace processes, peacebuilding and transitional justice, and the causes and consequences of electoral violence. You can follow her on Twitter @krishogl

---

Latest posts

• Military Interventions in Policing in Brazil and Mexico
• In the barrios of Medellín, criminal groups set the rules for the tourism sector
• A Bargain with Narcos? Challenges for Colombia
• ‘The Heavy Footed State, Which Made a Mess’
• Criminal Armed Groups Compete with the State and Each Other

Disclaimer: The views and opinions expressed in the Word on the Street blog are those of the authors and do not necessarily reflect those of the Urban Violence Research Network.