Non-UN peacekeepers: New data and research

Our new data on non-UN peacekeeping shows that regional missions are equally effective as the UN in lowering violence against civilians by governments, but not by rebels.

Corinne Bara, 1 April 2020

In the past two decades, regional organizations and coalitions of states have deployed more peace operations than the UN. And still most quantitative peacekeeping research studies only the effect of UN missions. This is because detailed data on non-UN missions has not been available so far.

With detailed I primarily mean dynamic data – data that tells us how many non-UN peacekeepers have been deployed in any month in a conflict. Monthly data is crucial as most conflict researchers now study conflict dynamics with monthly (some even daily) data on violence. This detailed violence data is matched with monthly data on UN deployments, while the best available data on non-UN missions by SIPRI so far has been annual. Moreover, the SIPRI data is not easily available at the moment: For years before 2000, one has to consult the peacekeeping tables in the printed SIPRI Handbooks. For years after 2000, SIPRI theoretically has an online platform on peacekeeping data, but this has been offline for several years now.

To overcome this lack of data, I have collected and published the UN and Non-UN Peacekeeping Dataset. The data reports the approximate monthly number of troops, police, and observers deployed by the UN, regional organizations, and coalitions of states to civil conflicts globally between 1993 and 2016. These data are compiled and consolidated from the annual numbers provided by SIPRI.

But how do we get from annual SIPRI numbers to a monthly dataset? The solution is as simple as effective: I use linear interpolation between known annual data points to generate the monthly estimates. I also extrapolate from the first available data point to the beginning of missions. While this may seem like a crude procedure, it works well for two reasons.

First, the most plausible theoretical assumption is that increases or decreases in deployment levels are gradual. While we can expect some missions that build up quickly to increase their presence in waves, most personnel changes are likely to be more incremental. Second, I can show empirically that a linear interpolation proxies the monthly personnel levels very well. We can know this because for the UN part of the data we have true monthly data to compare to. The correlation between the true monthly and our interpolated monthly numbers in UN missions is a staggering 0.97.

So, when faced with the choice of not studying non-UN peacekeeping at all, or using an interpolated version, the latter is a better choice. This especially given how little we currently know about the effectiveness of regional peacekeeping missions.

In an article in International Peacekeeping, published together with my co-author Lisa Hultman, we do exactly that. We study whether regional peacekeeping operations are equally effective in mitigating violence against civilians by governments as the UN. What we find is that at comparable mission sizes and compositions (UN missions are on average larger and have more police and observers alongside military troops), then regional operations are equally effective in mitigating violence against civilians by governments, but only UN troops and police curb civilian targeting by non-state actors.

This is just a starting point. There is much to do in terms of studying the impact of non-UN missions, and one important research avenue is to study how effectively UN and non-UN missions work together. There is a lot of cooperation but also a division of labor between different peacekeeping organizations, and we know little about how well this works. We hope our new data both enables and encourages this new strand of peacekeeping research!