Is the world deteriorating?

Tikuisis, Peter ; Mandel, David R.

THE WORLD IS BECOMING A LESS CONFLICT-RIDDEN PLACE, AND IT HAS BEEN for decades. Wars and violent conflicts are on the decline, according to expert consensus. In particular, we note abundant evidence that suggests that the world is not deteriorating from Steven Pinker's conclusion of a long-term general decline in global violence (1) to Azar Gat's analysis pointing to a decline in war since the mid-1990s (2) to the 2013 Human Security Report proclaiming a substantial decline in conflicts since the early 1990s. (3)

It was for that reason that a systematic analysis of world crisis data from a global leader in crisis tracking left us baffled. We analyzed all of the International Crisis Group's monthly CrisisWatch statistics for the crises it has tracked from its inception in September 2003 to March 2014. Crisis Group's data paint a startlingly different picture: a world that has deteriorated substantially over the past decade. How could this be?

If Crisis Group was a mediocre crisis tracker, then we might grant the incongruity to incompetence. (4) But it is well resourced, drawing on multiple sources including over 100 analysts across five continents. In addition, it boasts a stellar cast. Past and current board trustees include a former UN High Commissioner; a former US under-secretary of state and ambassador to the UN, Russia, India, Israel, Jordan, El Salvador, and Nigeria; and a former UN deputy secretary-general and administrator of the UN Development Programme. Crisis Group also prides itself on "generally [being] regarded as the world's leading source of information, analysis and policy advice on preventing and resolving deadly conflict." (5) We can confidently rule out incompetence as a source of incongruence with expert consensus, which makes this puzzle all the more interesting.

First, the pieces: Crisis Group has tracked or is tracking 130 countries and regions (i.e., distinct conflict zones) over the 127-month period we examined. Crisis Group does not track every zone in every month. Some conflict zones were not on Crisis Group's radar when it started; some that were are not anymore.

We first verified whether there was a change in the number of conflict zones tracked over time. Contrary to the "improving world" thesis, we found that there was a slight, but statistically significant increase in the number of conflict zones that Crisis Group reported on during roughly the past decade.

We then reasoned that, while the number of conflict zones reported on had increased, Crisis Group might have been reporting more favorable conditions over time within those zones. However, that was not the case.

In each month, Crisis Group reports whether the conflict zones that it tracks have significantly deteriorated, improved, or on balance remained more or less unchanged. Deterioration means a significant increase in conflict; for example, "continued to worsen," "attacks significantly escalated," and "security further deteriorated" are some of the descriptions used for this assessment. Improvement means the opposite; for example, "comprehensive peace agreement," "new constitution signed," and "restore state authority." Unchanged means there was no change in conflict status since the last report.

Over the period we examined. Crisis Group issued 9,378 such assessments. Deteriorated situations outnumbered improvements in all but 3 of its 127 months of reporting history. And of the 130 conflict zones, 99 (76.2 percent) had a net surplus of deteriorations by March 2014 over the period they were tracked. Stated differently, over more than a decade of monthly reporting involving 130 conflict zones, the odds of deterioration rather than improvement or stasis were 3:1 based on Crisis Group's reporting.

What makes this puzzle more baffling is the fact that some of Crisis Group's leaders have issued statements in support of the improving world thesis. For example, in a 2012 opinion essay entitled "The Global March Towards Peace," Crisis Group's president emeritus, Gareth Evans, points out, "The bigger story has been concealed, as ever, by the media's daily preoccupation with current bloodshed: Over the last two decades, major wars and episodes of mass violence worldwide have become much less frequent and deadly." (6) Also in 2012, then-president Louise Arbour publicly expressed, "It may surprise some of you that there is actually some good news--namely that war may be on the retreat.... In fact over the past two decades the number of conflicts across the world has decreased." (7)

How could this puzzle be explained, and why was it not noticed before? We believe the answer to these questions lies in an important distinction that Nobel Prize-winning psychologist Daniel Kahneman and colleagues make between what they call the "inside view" versus the "outside view": two distinct information-processing styles that people can adopt. (8) The inside view is case based, nonstatistical, and has a narrative structure. It represents the story we might tell about a specific conflict zone at a given time or about that particular case as it develops over time.

In contrast, the outside view is category based, statistical, and can be put into a narrative structure only with concerted mental effort. Our analysis of Crisis Group's indicators is eminently based on an outside view approach. The inferences we drew from its data came from looking across all time points and all reported conflict zones. We were not concerned with the narratives, as accurate as they may be, that accompanied specific reports about individual conflicts, or with the narrative trajectories of any particular conflict over time, but only with what the big picture revealed based on an across-the-board aggregation of its front-page indicators.

Therein lies the key to the puzzle we uncovered. The inside view is the default explanatory system for humans. People are naturally adept at telling (and listening to) stories and comparably poor at pulling out inferences from large impersonal databases. It is why intelligence analysts value good narrative over statistical prediction--or, as former CIA veteran Sherman Kent lamented, a preponderance of "poets" and a paucity of "mathematicians." (9)

It is also why people tend to be moved to action more by a single evocative story of personal suffering than by a cold statistic about mass casualties of genocide and other forms of collective violence--what Paul Slovic calls "psychological numbing." (10) The inside view, our cognitive default, actively and automatically promotes it since we emotionally connect with a specific case of suffering more than with even a staggering, but seemingly impersonal statistic.

As good as Crisis Group's conflict analyses might be, they adopt an inside view approach. While they may report on many conflicts, those analyses are still case focused. The problem is that the inside view, although psychologically compelling, is prone to systematic biases such as base rate neglect and overconfidence in forecasting. (11)

Switching to the outside view can offer useful consistency checks on inside view thinking. In our study, it revealed a glaring inconsistency between a world leader in conflict tracking and expert consensus on the decline in violent conflict since the 1990s. That inconsistency was not explicit in any Crisis Group report, but could be discerned only by looking holistically and quantitatively across its full reporting history. Indeed, as we have also noted, that inconsistency is evident even within Crisis Group since some of its leaders have noted the improving world trend while presiding over an organization whose aggregate indicator analyses show an opposing deteriorating trend.

That revelation is important. Policy advisers, such as Crisis Group, that rely primarily on an inside view approach will benefit from periodic review of what their reporting shows when it is subjected to an outside view analysis. Anomalies or inconsistencies that are revealed might point to the need for procedural reforms.

For instance, we examined when conflict zones disappeared from Crisis Group's radar. We expected that reporting would stop only after a zone had stopped deteriorating for at least a few months prior, but we found no consistent pattern. In some cases, deterioration was reported just before conflict reporting had stopped, by as little as one month. When we looked across all zones where reporting had stopped for six months or more, we found 36.1 percent ended with a surplus of deteriorations to improvements compared to only 13.9 percent where improvements exceeded deteriorations. That suggests that situations that had changed prior to when reporting stopped were more likely to be worse than when they began.

It is important to interrogate the data on any topic from both the inside and the outside views and to triangulate the results from those alternative perspectives. If they do not add up, there is a good chance that overreliance on the inside view is responsible.

There is also the possibility that improvements, which likely are not reported as frequently by the media as deteriorations, fall below the detection threshold or acceptance criteria for consideration. Although media bias, reflecting the fact that bad news sells, should not be an overriding concern given that Crisis Group conducts its own analyses, we acknowledge that even the best analysts may be subtly influenced by the preponderance of negative reports.

There is another potentially biasing factor in reporting by Crisis Group that ought to be noted; namely, Crisis Group's advocacy mission, which seeks to prevent violence and stop violent conflicts that are already in play. To influence the international community to take conflict-mitigating action, it is more effective to emphasize the negative than the positive. Indeed, it is a well-established fact that negative events are more likely to capture people's attention and motivate their subsequent action than positive events. (12) Thus, Crisis Group's focus on deterioration may be partly strategic.

On a positive note, our further outside view analysis offers some support for Crisis Group's Watchlist reporting, which "alerts readers to situations where, in the coming month, there is a particular risk of new or significantly escalated conflict, or a particular conflict resolution opportunity." We explored whether more reports of deteriorations versus improvements would follow risk alerts, and whether more reports of improvements versus deteriorations would follow resolution opportunities. We found thirty-six deteriorated situations versus six improved situations in the month following a "Conflict Risk Alert," and twelve improved situations versus seven deteriorated situations in the month following a "Conflict Resolution Opportunity." Hence, Crisis Group's forecasting odds are 6:1 in favor of risk alerts leading to deterioration and 2:1 in favor of resolution opportunities leading to improvements. Of course, we also note that the forecasts and subsequent indicators are not from independent sources. They are in fact from the same source--Crisis Group. Nevertheless, the analysis indicates the sort of consistency check that monitoring organizations could self-apply.

So, is the world deteriorating? It might appear so from a conflict-specific inside view but, as pointed out earlier, this perspective can distort the long-term and conflict-general outside view. According to scholarly consensus, the outside view points to a gradual de-escalation in global conflict prior to and during the period that Crisis Group has been tracking conflicts. The high quality of Crisis Group's work notwithstanding, we do not believe that its aggregate indicator findings seriously challenge that conclusion.

[c] Her Majesty the Queen in Right of Canada, as represented by the Minister of National Defence, 2015.

Notes

Peter Tikuisis is a senior scientist for the Socio-Cognitive Systems Section at Defence Research and Development Canada, Toronto Research Centre. He is also adjunct research professor at the Norman Paterson School of International Affairs, Carleton University, Ottawa. He has published research in bubble physics, thermal physiology, soldier performance, and most recently on terrorism and state fragility. David R. Mandel is a senior scientist for the Socio-Cognitive Systems Section at Defence Research and Development Canada, Toronto Research Centre. He is also adjunct professor of psychology at York University, Toronto. He has published research in human thinking and reasoning, judgment and decisionmaking, and the psychological bases of conflict and collective violence.

(1.) Steven Pinker, The Better Angels of Our Nature: Why Violence Has Declined (New York: Viking Penguin Press, 2011).

(2.) Azar Gat, "Is War Declining--and Why?" Journal of Peace Research 50, no. 2 (2012): 149-157.

(3.) Human Security Research Group, "Human Security Report 2013: The Decline in Global Violence: Evidence, Explanation, and Contestation," Human Security Report Project, Simon Fraser University, Vancouver, Canada, www.hsr group.org/human-security-reports/2013/text.aspx, accessed 23 June 2014.

(4.) International Crisis Group, "Crisis Watch: A Monthly Bulletin on Current and Potential Conflicts Around the World," Brussels (headquarters), Belgium, www.crisisgroup.org/en/publication-type/crisiswatch.aspx, accessed 30 April 2014.

(5.) See www.crisisgroup.org/en/about.aspx, accessed 25 November 2014.

(6.) Gareth Evans, "The Global March Toward Peace," Project Syndicate, 2012, www.gevans.org/opeds/oped136.html.

(7.) Louise Arbour, "Crisis and Conflict: Global Challenges in 2012," International Crisis Group speech, 2012, www.crisisgroup.org/en/publication-type/speeches /2012/crisis-and-conflict-global-challenges-in-2012.aspx.

(8.) See, for example, Daniel Kahneman and Dan Lovallo, "Timid Choices and Bold Forecasts: A Cognitive Perspective on Risk Taking," Management Science 39, no. 1 (1993): 17-31.

(9.) Sherman Kent, "Words of Estimative Probability," 1964, Central Intelligence Agency, https://www.cia.gov/library/center-for-the-study-of-intelligence/csi-publications /books-and-monographs/sherman-kent-and-the-board-of-national-estimates -collected-essays/6words.html.

(10.) Paul Slovic, "If I Look at the Mass I Will Never Act: Psychic Numbing and Genocide," Judgment and Decision Making 2, no. 2 (2007): 79-95.

(11.) Kahneman and Lovallo, "Timid Choices and Bold Forecasts."

(12.) Shelley E. Taylor, "Asymmetrical Effects of Positive and Negative Events: The Mobilization-minimization Hypothesis," Psychological Bulletin 110, no. 1 (1991): 67-85.