Responding to Conflict in Asia: Why Good Data is Needed

December 16, 2015

By Thanit Herabat, Patrick Barron, and Bryony Lau

The new set of post-2015 Sustainable Development Goals includes for the first time a target that specifically sets out to promote peaceful and inclusive societies, marking an increase in awareness that peace and security is critical for sustainable development.

Violent conflict of multiple forms has negatively affected development trajectories in Asia for decades. The region has experienced national civil wars (Afghanistan and Nepal), secessionist subnational conflict (Philippines, Indonesia, Thailand, Myanmar, Sri Lanka, Bangladesh, and India), and widespread inter-communal riots and pogroms (India, Indonesia, Myanmar, and Malaysia).

Violent conflict of multiple forms has negatively affected development trajectories in Asia for decades, including in Thailand’s Deep South. Photo/Chandler Vandergrift

Beyond such escalated conflict, everyday localized violence over issues ranging from control of land and elections, to urban crime and gender-based violence, is pervasive and has immense cumulative impacts. These forms of violence are often inter-related, with localized contention feeding into broader identity narratives that underpin larger-scale unrest.

Developing effective policies to prevent and respond to such violence requires information on what types of violence are occurring where and how different forms are inter-related. Without better data, development assistance can, at times, despite its good intentions, make conflict worse.

Data is needed but also limited

Good data can show where and when violence is occurring, the impacts it is having, and what factors are driving it. It can also reveal patterns and trends over time, and allow us to better analyze the different forms of violence occurring. This in turn allows for more effective responses; for example, by allowing policy-makers to focus on hot spots. Looking at more peaceful communities can reveal what actions can prevent violence. Violence has a logic of its own, and data can help us understand it better. A number of large, cross-country databases have been set up with this goal in mind. However, most of these datasets have limitations.

The Uppsala Conflict Data Program (UCDP) has comparable information on violent conflict deaths across countries. Yet the UCDP datasets record only some forms of violence (largely related to battle deaths) meaning that everyday violence, for example cattle rustling in the southern Philippines, is missed. Further, UCDP and other global datasets such as the Armed Conflict Location and Events Data Project (ACLED) rely on sources far from the areas where violence occurs, meaning they are getting the information second-hand, and likely greatly underreporting incidents and impacts.

Other global datasets, such as that used in the UNODC Global Study on Homicide, track a wider range of violent deaths. However, the UNODC study only disaggregates deaths by three forms –interpersonal, socio-political and criminal – meaning that the underlying drivers of violence cannot be analyzed. And information is on overall patterns of violence rather than specific incidents.

In parallel, many early warning systems, which track incidents of violence in real time within countries, have also been established. But these typically lack consistent coding systems and quality control procedures. This results in underreporting of violence, faulty analysis, and, ultimately, poor policy responses.

Data innovations in Asia

The need for better data has led to the development of subnational violence incident monitoring systems in a number of countries in Asia. The Indonesian National Violence Monitoring System (NVMS) is the largest of these, including over 200,000 coded incidents stretching back to 1998. In the southern Philippines, International Alert and the World Bank have developed the Bangsamoro Conflict Monitoring System (BCMS). And the Deep South Watch (DSW) database plays a similar role in Thailand’s conflict-affected South.

These systems are radical improvements on what came before, with far more accurate information. In 2007, for example, DSW reported 12 times as many deaths as recorded in the UCDP battle deaths dataset (Figure 1); the UCDP data for 2002 records 10 times fewer deaths in Aceh as does the NVMS (Figure 2). The datasets also benefit from local ownership. DSW is run by the Prince of Songkhla University whose strong local networks channel data and analysis to local policy-makers. Since 2004, the system has developed consistently as violence has escalated and has tracked 15,123 violent incidents and over 6,480 casualties so far.

Figure 1: Deaths from violence in Southern Thailand

Figure 2: Deaths from violence in Aceh

These monitoring systems are also different because they are incident-based. Data for each incident is collected systematically with robust quality control measures in place. Information coded includes when and where the incident took place, its impacts, the actors involved, and the causes. This allows for data to be easily disaggregated. One can easily focus, for example, on violence of just one form (e.g., riots), over one issue (e.g., land), involving one set of actors (a former insurgent group), or in one area. It is also possible to track the ways in which different types of violence are related (or not). Such finely-grained and scalable systems make it possible to see violence trends across time and space, filling huge gaps in policy research and applications.

Where these systems are in place, they can be of great use to policy-makers. For example, following the 2005 peace agreement in Aceh, data was used to determine the amount of support given to each victim and to assess emerging post-conflict hotspots. In the Philippines, data generated by Bangsamoro Conflict Monitoring System from 2011 to 2014 has started to yield important insights useful in the government’s efforts toward peace and development in the Bangsamoro.

What can we do?

Acknowledging the potential of these systems, The Asia Foundation has been working with the International Development Research Centre, the World Bank, and the Korean Trust Fund to improve their quality and to set up similar initiatives in other countries. This involves three things:

First, we are encouraging mutual learning among the NVMS, DSW, and BCMS to help them strengthen their systems through technical assistance and a series of workshops and exchanges.

Second, we are exporting best practices elsewhere. The project is producing a methodology toolkit, which will act as a “how to” guide for organizations interested in building their own violence monitoring systems. And we are producing research papers to highlight the analytical uses of the data derived from the three systems. And over the coming months, we will be piloting similar systems in Nepal and Myanmar.

Third, we are seeking to build a set of minimum protocols for violence monitoring systems to ensure that data can be compared across different countries, in much the way it can with the cross-country datasets. With donors and governments seeking to find better ways to measure progress against SDG 16 (on levels of violence), there is much scope to plug data and analysis from these systems into global debates and monitoring efforts. We need better violence monitoring systems in more countries to capture progress and challenges. In the end, peace might begin with good data.

Thanit Herabat is a program assistant for The Asia Foundation’s conflict and development team, Patrick Barron is regional director for conflict and development, and Bryony Lau is a senior program officer, also in the same team. All are based in Thailand. They can be reached at [email protected], [email protected], and [email protected]. The views and opinions expressed here are those of the authors and not those of The Asia Foundation or its funders.

• Previous in blog« Local Pathways to Disability-Inclusive Governance in Indonesia

• Next in blogThe Rise of Smart Phones in Cambodia Challenges Social Norms »

Contact

The Latest Across Asia