@mzapata I am unsure how to forecast. Would Fulani Herder-Farmer violence resolve this question / be considered as systematic killings? Although there is not a clear and overarching political agenda behind the violene afaik, the killings are not unrelated and def. a result of greed and grievance. I'd argue that Fulani attacks should resolve this question, especially in regard to their ethnic dimensions, the reprisal attacks etc. Did the EWP discuss this issue yet? (Need to discuss this in 2017 was lower, as the death toll was too).

Cec338245b186a865fdc48963d77e3af?s=80&d=identicon
mzapata
made a comment:

Hi @JS,

We did discuss the issue of herder-farmer violence in Nigeria and determined that it did not qualify as a campaign of violence by an armed group. The experts we consulted said that much of the violence entails micro-level conflicts between herders and farmers within and between communities. Violence across the middle belt is being perpetrated by many groups with a variety of motivations for various purposes (for example, land disputes, banditry, ethnic grievances, etc.) and though many of these groups target civilians, we don't see evidence that they are working in coordination as part of a campaign against civilians. Though many perpetrators share an ethnicity, we thus far have not seen evidence of large-scale coordination of an ethnic-based campaign. Hope that helps, happy to discuss further!

Fc513a593698e997bb89e8cefda61caf?s=80&d=identicon
JS
made a comment:

Hi @mzapata, thanks a lot for that reply! This really helps me forecasting and structure my thinking. I reduced my forecast by 50 percentage point as a result (It's at 30% now).

I am (sadly) no country expert & just quickly mapped the ACLED-Data a week ago: https://twitter.com/JensStappenbeck/status/993830669007622144

Fulani already killed ~850 civilians during the first four months of this year according to ACLED, so the violence in general is quite massive and way higher than last year. Although most attacks are located in Northern Benue, I cannot deduct from the data that the attacks are part of a campaign against civilians or coordinated. I'd say that obv. there are political, ethnic and - maybe increasing - religious connotations to the violence. I also would suspect that the rise in violence could hint towards greater coordination.

This doesn't have to be the case though (and very well might not be). I agree that there is a difference between the ethnic connotations to violence that I was thinking about & the large-scale coordination of an ethnic-based campaign that the EWP would for example be looking for. In the Philippines, where there is political backing at the highest level for violence against drugusers for example, the case is much clearer (although, interesting enough, I did not assume that the EWP would include the Philippines at first as I for example related it to drug-relad violence in Latin America). In South Sudan however I was surprised that killings of civilians by all opposition groups have been somewhat combined to resolve the question, while their coordination is a bit questionable as well. If killings by Ex-Fulani-militias would rise again in the CAR (which they do, but I assume not to the extent that's relevant here), I'd think they would resolve this question, although coordination between them is highly questionable as well as they also do attack each other sometimes, which the Fulani don't. It's very difficult to judge these events (and to gain insight into how the EWP might judge them). Right now I don't have any suggestion for a better approach/better definitions than you though. Are there any plans to revive the EWP blog? I think having a place to discuss approaches / guidelines for resolvement of these questions would def. be of value, although I am not sure how big the interest for such debates might be. But if there is a place for debates like this, it's at the EWP right now.

If I come accross a report that provides any detail into Fulani-groups & points in either direction of (non-)organization, I'll just tag you here.

Cec338245b186a865fdc48963d77e3af?s=80&d=identicon
mzapata
made a comment:

Hi @JS,

You hit on one of the major challenges for us in question resolution. While I know that for GJO forecasters sticking to one data source (i.e. ACLED) and basing our resolutions solely on its information might be a simpler and cleaner way to resolve questions, we fear that in many cases it does not provide a complete picture. When we resolve, we start with ACLED (for the countries it covers) or a similar data source when available, then filter by interaction type to isolate violence against civilians. If the numbers are high enough we then look into potential perpetrator groups, conducting qualitative research into who is likely working together (for example, in Darfur, we count Janjaweed as part of the state-led campaign even though they are not officially a state security force, and in the Philippines we count vigilante groups operating within the Double Barrel campaign against drug users as part of the campaign). When warranted, we take an additional step of contacting a country expert to discuss (as we did for Nigeria).

But the challenges you highlight of resolving Nigeria, CAR, and South Sudan are issues that we struggle with regularly and reevaluate as new information becomes available. Though perhaps less satisfying for you, we do ascribe to a certain status quo bias in our project in general: so in the case of South Sudan, we determined an onset of mass killing in 2013 by Machar supporters -- at one point we knew enough to call it a campaign and since then we have not seen significant evidence to the contrary. Compare to Nigeria, where we have not yet seen evidence that the violence is a part of a campaign. These determinations I think you have seen in our annual “State of the World” blog posts. (Also, as a side note, seeing intercommunal violence as perpetrated solely by Fulani feeds a dangerous narrative gaining traction across the Sahel these days...happy to discuss that one further.) It is an imperfect science, with spotty information of varied precision, and we do reevaluate determinations as new information becomes available (also note that ACLED updates as they receive more information, so the analysis for the most recent round of determinations was done immediately after the April 1 closing date).

To that end, yes please comment/tag me if anything comes up! Though it would be easier to switch this conversation to email I think it's useful for other forecasters to see/weigh in.

Fc513a593698e997bb89e8cefda61caf?s=80&d=identicon
JS
made a comment:

Hi @mzapata, thanks again for the explanation! Off-Topic: I'll make sure to heed the advice in regard to the Fulani. Highlighting only them is prob. as useful as calling every islamist group in Mali as being Daesh. I was trying to quickly make a point of the scale of violence vis á vis BH. This could & should have been done better.