Discussion with Chic Dambach, John Packer, Michael Lund and Lawrence Woocher (updated!)

The purpose of this post is to share a recent email exchange I initiated on the topic of conflict early warning and response.

———- Forwarded message ———-
Chic <chic@allianceforpeacebuilding.org>
Date: Sat, May 24, 2008 at 5:06 PM
Subject: RE: conflict early warning & forecasting
From my perspective, our prediction capability is very primitive, and that choosing a place to implement a violence prevention project is very much a guessing game.

I don’t believe in waiting for the predictive capacity to improve before taking action. Going to work in Guinea-Bissau was the right thing to do, and I believe the same about our project in Kyrgyzstan. However, in the long term, the ability to generate support for preventive work will be enhanced significantly if we can improve our ability to accurately predict where violence is most apt to happen. Furthermore, we can more effectively target limited resources where they are needed most if our prediction capacities improve.

There is a project underway at Brookings that has tremendous potential. I’m at home today and don’t have the names of the people doing it, but they showed me their model a few months ago, and it is very impressive.

As you know, there is more attention on violence prevention than ever before. That is a tremendously important development. However, it could lose credibility and support if we don’t improve our analytical skills as well as our application of peacebuilding skills.

Our top priority is to generate support for prevention projects, without waiting for better information, but I would love to see the academic and research community develop this capacity. Google and others could be very helpful.

———- Forwarded message ———-
John Packer <jpacker@hri.ca>
Date: Sat, May 24, 2008 at 6:09 PM
Subject: RE: conflict early warning & forecasting
Dear Chic and Patrick et al.,

Sorry to jump in, but to be clear: the problem is NOT early warning. This is nonsense. Did we not know what was going to happen in Rwanda? Could we not see the risks across Eastern Europe in the light of X-Yugoslavia? We read social behavior ALL THE TIME in our daily lives, in personal, social, corporate and other relations and we make calculations on which we rely (cf. anything from personal health care, the way we look after our children and plan their educations, and the way we invest economic resources including in the stock markets). As always, the problem is simply in EARLY and APPROPRIATE ACTION!

This is why I immediately share Chic’s observation that IPPP jumping into G-B made perfect sense AFTER a reasonable scan of the situation and some comparisons about relative chances of success, resource implications etc. Sensible engagement of knowledgeable people (like Michael, David Carment and Ben) was obviously useful, relatively cheap and fairly easily done (it was not like sending some one to the moon!). Taking on Tibet would have been stupid, as would have been some engagement with indigenous peoples in Canada – not that either situation is without risks of conflict or without merits in terms of human rights concerns and humanitarian values. But if we’re interested in addressing and preventing/curtailing/resolving relative risks of mass violence, then G-B was a pretty good case to select – perhaps not mathematically perfect, but pretty darn reasonable! And, importantly, it matched your capacities (see Mary Anderson’s cautions).

I do agree we need to improve our analytical skills, but this is not a matter of mathematics: the CAUSES for mass violence are pretty well known, certainly in terms of the relative risks.

MUCH more important is to develop the needed SKILLS for peacebuilding and to have LOTS and LOTS of people capable and actually engaged in such work. I am forever amazed at how juvenile or simply absent are these skills, even amongst the purported (and declared and congratulated) “peace-makers”.

Just to take one example in which I am now engaged: Who thinks Kirkuk (the mixed Iraqi city at the centre of the northern oil industry) is NOT a risk of violence for the city, province, country and region? You’d have to be plain stupid not to see it. Indeed, it’s the well-known possible tinder-box. But more than FIVE YEARS after the Iraq invasion, who has actually been doing anything – at all, much less something sensible and sustained? Answer: pretty much NO ONE!!!! About ten days ago, I facilitated one of the rare consultations among (exclusively!) Kirkuki local authorities, and we had a very good meeting with surprising consensus on most issues (the problem is absence of confidence due to recent bad experiences, despite the fact all these people grew up together, relatively live together, speak each other’s languages, have inter-marriages, etc. etc.). From an issue-based analysis, their problems are RESOLVABLE! They are “entitled to $1 Billion which was not spent from the Iraqi budget” but they’re unable to spend it due to insecurity and virtual absence of governance. The problem is partly in outside interferences (but that’s political reality): that’s a “herding cats” problem. The greater problem is the Kirkukis’ own lack of know-how in terms of their own governance – lack of simple/pure knowledge about alternative arrangements and about processes … absence of experience and the confidence which comes with it. ALMOST NO ONE IS HELPING THEM! And, frankly, not a lot of people really know HOW to help them! In their ignorance, many “helpers” are avoiding action, controlling the space (so keeping others out) and actually contributing to a worsening of the situation.


———- Forwarded message ———-
Patrick Meier <patrick.meier@tufts.edu>
Date: Sat, May 24, 2008 at 6:18 PM
Subject: Re: conflict early warning & forecasting

Dear Chic and John,

Many thanks for your replies.

I believe Lawrence is working with Joshua Epstein at Brookings on early warning models using agent-based modeling. Am I right Lawrence?

I tend to agree that early warning is not the problem. Even if highly reliable conflict forecasting systems existed, I doubt they would lead to effective and timely response. This is because policy makers do not draw on formal conflict early warning systems in the decision-making process. See this paper by Susanna Campbell and myself for supporting arguments/evidence. My main concerns with forecasting socio-political events stem (most recently) from the graduate course on complex systems I took at the Santa Fe Institute (SFI) in New Mexico. Even if we had timely, continuous, reliable and geo-referenced data, social systems still remain complex and predicting their behavior, in my opinion, is not possible–at least using econometrics. The two sources I most often draw on to support my perspective are:

Nassim Taleb (2007). The Black Swan: The Impact of the Highly Improbable. (Random House).

Charles Doran (1999). “Why Forecasts Fail: The Limits and Potential of Forecasting in International Relations and Economics.” International Studies Review, Vol. 1, No. 2, Prospects for International Relations: Conjectures about the Next Millennium (Summer, 1999), pp. 11-41

One of the most telling tests of how forecasting methods fare in the real world was run by Spyros Makridakis, “who spent part of his career managing competitions between forecasters who practice a ‘scientific method’ called econometrics […]. Simply put, he made people forecast in real life and then he judged their accuracy” (1). This led to the following lamentable conclusion “statistically sophisticated or complex methods do not provide more accurate forecasts than simpler ones” (2). And so, despite the fact that “billions of dollars have been invested in developing sophisticated data banks and early warnings, we have to note that even the most expensive systems have shown a striking inability to forecast political events,” not to mention galvanize any preventive measures (Rupesinghe 1988).

Other empirical studies demonstrate that experts, i.e., us (and our sophisticated systems and methodologies) are only marginally better than novices in our ability to accurately forecast political and economic events. Furthermore, these studies show that neither group’s forecasts are much better than random guessing. Of greater concern still is the empirical observation that experts nevertheless remain consistently overconfident of the accuracy of their own forecasts. This is compared to novices who tend to be more conservative vis-à-vis their forecasting abilities although they are equally (in)effective when it comes to accuracy. In addition, another study recently concluded that “out of close to a million papers published in politics, finance, and economics, there have only been a small number of checks on the predictive quality of such knowledge” (3)
———- Forwarded message ———-
Michael Lund <mslund@verizon.net>
Date: Sat, May 24, 2008 at 8:10 PM
Subject: RE: conflict early warning & forecasting

I think John puts his finger on the much larger problem. Also, notwithstanding the need to evoke some motivation to act by major countries, we need to be careful that we don’t paint ourselves in a corner by defining the problem we seek to avoid simply as particular moments when there is an outbreak of major violence leading to genocide or civil war. The overall rate of major armed conflicts has been in decline since the mid-1990’s, and the preceding amount was due in great part to the breakup of the Soviet system. I am wondering if the day for such major internal wars is largely over, for leaders’ strategies and pressure on them from international actors may be adjusting behavior to avoid major confrontations. Also, outmigration and remittances as a result of increased globalization are keeping many countries away from ever reaching the brink.

The more serious problem may be a creeping societal degeneration in the poorest countries where the continuing lack of effective political processes and government are unable to avoid eventual state failure and chronically low development persists. Violence may occur but it is not necessarily going to be on a Rwanda-type scale. Bangladesh for instance.

———- Forwarded message ———-
John Packer <jpacker@hri.ca>
Date: Sun, May 25, 2008 at 9:55 PM
Subject: RE: conflict early warning & forecasting
Dear Michael, Chic, Patrick et al.,

I wholly agree with Michael. Indeed, I already think we’re playing into the doubters’ hands by accepting that we need better predictive capacities. Poppy cock! (And thanks Patrick for the supporting references and evidence.)

My point is that we’re basically quite capable of predicting the principal RISKS of violence creation and we also know what works generally in addressing them (i.a. cf. 60+ years of peace, stability and prosperity among NATO members).

As a matter of RISK, surely it is better to act to prevent what MIGHT occur even if we’re not certain, rather than to wait and see until it’s too late… especially when we KNOW the risks are mounting. For example, we know when our car is running a long time without a change of oil or on the same tires that we should not RISK non-attendance and so WE ACT to take care (change the oil and tires etc.). Indeed, it would be illogical to wait or carry on since we also know that the costs of the risk mount all the time … even towards catastrophic! I believe social science and plain old human experience have long taught us the risks associated with various kinds of socio-political illnesses and problems, and we avoid or withhold acting at our peril … with the prospective costs of the associated risks mounting all the time.

But separate from this is an entirely more constructive approach: instead of just thinking about how we can avoid our children becoming drug-dealers and delinquents, we actually take steps and invest in them becoming good citizens with self-esteem, creative opportunities and productive engagements. In other words, the preventive measures to avoid our children going astray are in fact to develop their positive potentials! The same goes for societies. Simply, it’s not just a matter or motivation of preventing violence that should merit our investments in building societies, but it’s the inherent value of creating peaceful, stable, creative and prosperous societies – of pursuing the civilizational capacity of our species (in global terms). It so happens that this IS also effective violence prevention (leaving aside some pathologies). So we do not need to wait for risks of violence to mount BEFORE we act to build peace. That’s just silly. Regrettably, it seems to be what we’ve been doing – and why we’re always so darn reactive and late and, therefore, facing poorer chances of success and often being taken over by “crisis managers” and frankly the military.

Michael is right hat the actual trend of violence is no longer in the direction of wars (in the classical sense, now fairly rare) or even civil wars, but more and more problems of disintegration and so-called failed and failing states. Waiting is especially a poor course of action vis-a-vis such situations or risks, because they are quite difficult to take hold of later (see Somalia). So such a course is also silly.

Chic, I would argue that, especially given the pausity of resources, we are very effective in our preventive action – when we engage. In CEE, I would argue the HCNM was/is about 100% successful: not a single case in which he has been involved has erupted into mass violence and only Macedonia deteriorated (and I argue we avoided civil war BECAUSE of the HCNM’s involvement). I would go further to say the same for efforts generally across the OSCE (with the exception of Chechnya… and Turlish Kurdistan, although arguments can also be made in that situation). And all this for very little money, comparatively speaking.

So I am dumb-founded at WHY there isn’t broad and abundant support – even enthusiasm – for more such institutions and engagements. What are we waiting for? Indeed, I would suggest the burden of proof should now be reversed and the doubters should prove that it is NOT worth the value of investing to so address the risks (even if small).

Chic, if you/they need arguments in terms of “interests”, how about mentioning Kabul, Kandahar, Kirkuk and Fallujah (and perhaps say “oil!” … or, more effectively, “gas prices!”). Without getting into the specificities of the cases (and there are many), just think of the social, political and ECONOMIC costs of these conflicts (now $2 TRILLION for Iraq alone, according to Joe Stiglitz).

Yes, great we’ve got Milt doing a fabulous service: Milt, you’re surely going to heaven! But surely that’s absurd … one modest retired American businessman backing an initiative for a whole country?! Problem is we only seem to have one Milt – and, very regrettably, he just ain’t rich enough!!!! Still, aside from the evident value of Milt’s support for the G-B people, the greater utility is in the demonstration effect: if MILT can do this with a few experts, why on earth can’t the Governments of the world (or even a few) get together to back seriously such work in various obvious places?

One problem, I honestly fear, is that IF a number of Govts actually do come forward (or if Soros, Gates or someone were to offer big money), in truth we do NOT have many qualified people really available to do the work. Simply, we’ve never invested in it and so we might well fail due to all the problems of un-qualified people (and perhaps adventurers) running hither and thither in willy-nilly “actions” leading to nothing or worse (again, see Mary Anderson).

So, Lawrence, if you’re reading this, we need to build institutional capacities – tools, skills, experience and confidence! When the call comes, we need to be able to succeed.

———- Forwarded message ———-
Lawrence Woocher <lwoocher@usip.org>
Date: Fri, May 30, 2008 at 10:52 AM
Subject: Re: conflict early warning & forecasting


Thanks for the stimulating exchange. I’d like to jump in with a few thoughts. Apologies for chiming in a few days late.
We in this community are all persuaded that more vigorous and effective conflict prevention is possible and needed. Yet we observe relatively little serious preventive effort by our governments and intergovernmental orgs., at least relative to the opportunities for useful preventive action. And relatively little support for preventive action from private donors. This leaves puzzled (and frustrated).
However, I think it is counterproductive to frame the question as what is “the problem” that limits the extent of support for conflict prevention. First, there are almost certainly lots of specific factors that contribute to the problem. Any one of these need not be a sufficient explanation of the failure to engage in prevention. For example, in retrospect we may see that there were adequate warning signs in some place and time to have triggered actions to forestall major violence (i.e., “early warning was not the problem”). Nevertheless, these signs may have been ambiguous or accompanied by conflicting signals, limiting their impact.
Second, what is more important going forward is to understand which of the factors that make a difference in the extent and quality of preventive action are subject to influence–by the expert community, activists, or sympathetic officials. I believe some of the factors that limit the investment in prevention relate to basic cognitive processes; e.g., psychologists have found people generally exhibit an  “aversion to certain losses,” meaning they are willing to risk much larger future costs to avoid a small but certain loss now. This suggests that as long as the risk of major violence is uncertain, decision makers are likely to be reluctant to take costly preventive actions–even when these present costs are clearly smart investments. This might be an important factor in explaining the meager investment in conflict prevention, but it’s not obvious what we or others could do to lessen its negative impact.
In contrast, to Chic’s point, if we can generate more accurate political forecasts and/or increase policymakers’ confidence in the tools of conflict prevention, perhaps they will be more likely to act preventively. The question, thus, is not whether forecasting tools or preventive techniques are adequate, but whether improving them would make a difference in policymakers’ (or donors) actions.
In sum, I’d urge us to think collectively about what we think are the factors that both make a difference and are amenable to influence. More accurate forecasts? Stronger preventive capacities? Better knowledge of what kinds of preventive strategies to use in different contexts? A wider community of skilled practitioners? Better arguments about the merits of prevention?
Lastly, Patrick is right that USIP is providing modest support to Joshua Epstein and his team at Brookings to develop an early warning application based on his agent-based model of civil violence. Josh sent me a first draft recently, so it’s moving forward. I’ll be happy to keep you informed of progress.


One response to “Discussion with Chic Dambach, John Packer, Michael Lund and Lawrence Woocher (updated!)

  1. Patrick Meier mentioned the finding from a paper of mine with Scott Armstrong that experts using their unaided judgment to predict decisions in conflicts are not better than novices or chance. Please don’t miss the good news story! We have developed two methods that do provide accurate forecasts for conflicts: structured analogies and simulated interactions. The key papers are available at conflictforecasting.com under “Papers”. They are Green (2002 and 2005) and Green & Armstrong (2007).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s