Cities are often violent places – a social, ethnic and religious tinderbox of people piled up together with competing needs for space, housing or cash. Mostly the tension is contained, but not always – when and why does it spill over into bloody mayhem? That’s the question at the heart of a fascinating research project run by Caroline Moser, one of my development heroes, and Dennis Rodgers. The research team fed back on its findings in Geneva last week. They welcome any comments on the draft overview paper and welcome any comments by the end of June (as comments on this post, or if you want to get really stuck in, emailed to urbantippingpoint[at]Manchester[dot]ac[dot]uk). Here’s a summary of the discussion in Geneva.
The Urban Tipping Point project scanned the literature and identified four ‘conventional wisdoms’ (not always based on much evidence) on the causes of urban violence: poverty; ‘youth bulges’ (demographic, rather than waistlines); political exclusion and gender-based insecurity. It decided to test these with empirical research in four very dissimilar cities – Nairobi (Kenya), Dili (Timor-Leste), Santiago (Chile) and Patna (India).
The project title suggests its hypothesis – that violence does not arise in a continuous manner, but through sudden, discontinuous tipping points. The team also decided to borrow from commodity chains and explore whether ‘violence chains’ were a useful way to understand these outbreaks.
For me, this comes pretty close to being the perfect research project: a clear, important question; a challenge to conventional wisdom; mixed research methods combining participatory qualitative and quantitative; gender mainstreaming throughout; practical implications; the majority of the research carried out by a range of southern research institutions and lots of in-country consultation on the initial findings.
Each city study threw up something different and interesting: in Nairobi, the ‘violence chains’ discussions with focus groups revealed multiple forms of violence (landlord-tenant, sexual, criminal) beyond the political violence after the 2008 elections. These came together when they mapped them onto violence ‘hotspots’ such as a bridge in the Mukuru slum (see pic) infamous for rape attacks. Community leaders jumped on the information and started discussing improved street lighting and providing housing for police to make the area safe.
In Dili, the focus was on the outbreak of post-independence violence in 2006. A key cause was the ‘institutional multiplicity’ of overlapping authorities and security systems (police v burgeoning private security groups v martial arts associations) – the project found that violence routinely ascribed to mobs and unemployed youths is actually instigated by local politicians and elites conducting turf wars (the Nairobi study came to the same conclusion).
Patna was initially chosen as a success story – a formerly crime-ridden city that after 2005 had seen a dramatic fall in conflict. But the research found something altogether different – violence has not gone away, merely changed its nature (decline in organized crime, but overall crime rates have increased) and geography (now much more confined to the slums). The local authorities brought in a ban on gambling to (successfully) attack organized crime, but liberalized alcohol production and consumption to raise revenues, leading to an upsurge in violence against women. In the slums, the police seem to have ‘violence filters’, acting promptly to resolve disputes over water and sanitation, where caste tensions might spill over into the wider city, but largely ignoring more containable land disputes.
In Santiago, the main finding was that violence persists in poor, middle income and upper income areas, but in different forms: “In low-income El Castillo violence results from exclusion and a lack of opportunities; among the elites in La Dehesa it relates to accumulating and maintaining wealth and “a fear of the have-nots”; finally in the intermediate sector of Contraloría that violence occurs in a context where households struggle to improve their lot, suffering high stress levels and family breakdown.”
In terms of the four ‘conventional wisdoms’, the project found that politics and power, much more than poverty, are a driver of violence; that ‘youth bulges’ only become a problem when politicians start recruiting young men as footsoldiers; that political exclusion is less related to violence than elites stirring up violence in a negative variety of inclusion and that gender-based insecurity is universal, with clear links between it and more public forms of violence.
As for the ‘so whats’, I think there’s still a way to go in drawing out the excellent insights from the case studies, but some of the initial recommendations are:
1. Mainstream considerations of urban violence into wider processes like slum upgrading and urban planning
2. Engage far more with the police as development partners (the researchers were surprised by the level of interest from senior police figures in all the case study cities, and convinced that they were genuinely interested in finding solutions)
3. Use ‘participatory violence appraisal’ to identify hot spots and early warnings of brewing conflict
4. Land tenure systems are crucial, and are best dealt with by clarifying the ambiguities and overlaps in existing systems, rather than starting from scratch.
5. Similarly, overlapping authority systems (e.g. between traditional and modern; city and national) are a recipe for conflict, with more participation and active organization a healthy counterweight.
I think the project could be improved in two main ways. Firstly, it needs to get deeper into the idea of tipping points (where it seems to think that Malcolm Gladwell has the last word). It found these helpful in tying in localized outbreaks to city-wide social and political processes, but discontinuous change is generally symptomatic of complex systems, and the project could usefully draw on that literature to pay more attention to the nature of resilience and fragility (‘fragile cities’ – nice concept) and to think in terms of violence webs, not violence chains (life and conflict just aren’t as linear as the word ‘chain’ implies).
A complex systems approach would also suggest thinking less in terms of specific policy recommendations and more about ways of working – circuit breakers, faster feedback loops, multiple parallel experiments and learning to fail faster (see previous post).
It might also be interesting, as one participant suggested, to borrow from epidemiology, for example to identify the individuals who act as the main vectors of violence in given situations and focus on influencing them, or look at the 70% of young people who, even in the most violent communities, do not engage in violence – understanding their ‘resistance’ to violence would be a fascinating place to start.
Secondly, the research needs to talk more about how some new key actors need to be drawn into the process – informal authorities, faith-based and grassroots organizations, local NGOs and, as mentioned previously, the police. These can be part of the problem as well as the solution (in Nairobi, slum residents identified community based organizations as causes of violence, and NGOs as peace makers). Picking the right people to talk (and more importantly, listen) to is more important than having another blueprint in your back pocket.
This post also appears on the World Bank’s People, Spaces, Deliberation blog