Human security report

http://www.humansecuritybrief.info/
http://www.humansecuritybrief.info/HSRP_Brief_2007.pdf

H U M A N S E C U R I T Y R E P O R T P R O J E C T
S I M O N F R A S E R U N I V E R S I T Y ,
C A N A D A H U M A N S E C U R I T Y B R I E F 2 0 0 7
F E A T U R I N G A S P E C I A L R E P O R T D Y I N G T O L O S E :
E X P L A I N I N G T H E D E C L I N E I N G L O B A L T E R R O R I S M ACKNOWLEDGEMENTS
In-House Team for the Human Security Brief 2007
Andrew Mack, director and editor-in-chief
Zoe Nielsen, deputy director
Tara Cooper, data research assistant
Ryan Cross, program assistant
Emina Dervisevic, program coordinator
Abhay Gupta, program officer
Robert Hartfiel, research manager, e-resources
Suzy Hainsworth, senior program manager
Mrinalini Menon, program officer
Irene Poetranto, e-resources research assistant
Christopher Rusko, e-resources research assistant
Consultants: Sebastian Merz and Amir Musin assisted in the preparation of data.
Editing and Proof Reading: Kirsten Craven, Craven Editorial; Jon Tinker, Panos Institute of Canada; and Ruth Wilson, West Coast Editorial Services.
Design and Layout: Susan Turner, Zuke Creative
Cover Design: Susan Turner, Zuke Creative
Cover Photo: Peter C. Brandt/Getty Images, New York City
This publication was made possible by the generous support given by the Department of Foreign Affairs and International Trade (Canada); the Norwegian Royal Ministry of Foreign Affairs; the Swedish International Development Cooperation Agency; the Swiss Agency for Development Cooperation; the Swiss Federal Department of Foreign Affairs; and the Department for International Development (United Kingdom).
We would like to especially thank our collaborators at the Uppsala Conflict Data Program (UCDP), Sweden and the International Peace Research Institute (PRIO), Oslo.
We would also like to thank André Gerolymatos, Don Hubert, and Tamir Moustafa for their comments on draft chapters.
ABOUT THE HUMAN SECURITY REPORT PROJECT
The Human Security Report Project’s (HSRP’s) mission is to undertake research on global and regional trends in political violence, and their causes and consequences, and to make this research accessible to policy and research communities, the media, educators, and the interested public.
HSRP was previously based at the Human Security Centre, Liu Institute for Global Issues at the University of British Columbia. In May 2007, HSRP relocated to the School for International Studies at Simon Fraser University.
The HSRP’s flagship publication is the Human Security Report, the first issue of which was published in October 2005 (http://www.humansecurityreport.info). The HSRP also publishes the Human Security Brief series (http://www.humansecuritybrief.info), and the forthcoming miniAtlas of Human Security, which will be copublished with the World Bank. The miniAtlas will be available in hard copy and online in English, French, and Spanish (http://www.miniatlasofhumansecurity.info).
The HSRP created and manages the Human Security Gateway (an online database of human security resources: http://www.humansecuritygateway.info/). It also produces the Afghanistan Conflict Monitor (a website highlighting new research and analysis on the conflict in Afghanistan: http://www.afghanconflictmonitor.org) and three online bulletins: Human Security Research (http://www.humansecurityresearch.info), Human Security News (http://www.humansecuritynews.info), and Afghanistan Security News (http://www.afghansecuritynews.info).
All of these publications and e-resources are available online free of charge.
The HSRP is funded by the governments of Canada, Norway, Sweden, Switzerland, and the United Kingdom.The distinction between human security and the more traditional concept of national security is an important one. While national security focuses on the defense of the state from external attack, human security is about protecting individuals and communities from any form of political violence. Human security and national security should be—and often are—mutually reinforcing. But secure states do not automatically mean secure peoples. Protecting citizens from foreign attacks may be a necessary condition for the security of individuals, but it is not a suffi cient one. Indeed, during the last 100 years far more people have been killed by their own governments than by foreign armies. All proponents of human security agree that its primary goal is the protection of individuals. But consensus breaks down over exactly what threats individuals should be protected from. Proponents of the "narrow" concept of human security, which underpins the HSRP’s research, focus on violent threats to individuals, while recognizing that these threats are strongly associated with poverty, lack of state capacity, and various forms of socio-economic and political inequity. Proponents of the "broad" concept of human security articulated in the UN Development Programme’s 1994 Human Development Report, and the Commission on Human Security’s 2003 report, Human Security Now, argue that the threat agenda should be broadened to include hunger, disease, and natural disasters, because these kill far more people than war, genocide, and terrorism combined. Although still subject to lively debate within the research community, the two approaches to human security are complementary rather than contradictory. WHAT IS HUMAN SECURITY? Human security is a relatively new concept, but one that is now widely used to describe the complex of interrelated threats associated with civil war, genocide, and the displacement of populations.
CONTENTS
Overview ............................................................................................................................1
Global Terrorism ...........................................................................................................1
Towards a New Peace in Africa? .................................................................................4
Global Trends in Human Insecurity: An Update .......................................................6
Conclusion ....................................................................................................................7
Chapter 1 Dying to Lose: Explaining the Decline in Global Terrorism ........................ 8
A Misleading Picture? ................................................................................................10
Iraq in 2007, a Dramatic Change ..............................................................................14
The Failure of al-Qaeda In Iraq—a Global Defeat for Islamist Terrorism ..............15
Islamist Terrorism: What the Statistics Tell Us ..........................................................19
Conclusion ..................................................................................................................20
Chapter 2 Towards A New Peace in Africa? ................................................................. 22
The Decline in Armed Conflicts ................................................................................22
Non-State Conflicts: A Long-Ignored Category of Political Violence ....................23
Organized Violence against Civilians .......................................................................24
What about Indirect Deaths? ....................................................................................25
Stops and Starts: Explaining Sub-Saharan Africa Conflict Trends .........................25
Conclusion ..................................................................................................................30
Chapter 3 Trends in Armed Conflict and Coups d’Éat ................................................. 32
State-Based Armed Conflict ......................................................................................32
State-Based Conflict Onsets and Terminations .......................................................34
Non-State Armed Conflict ........................................................................................36
Coups d’État ...............................................................................................................37
Chapter 4 Targeting Civilians ....................................................................................... 41
One-Sided Violence ...................................................................................................41
Human Rights Abuse .................................................................................................44
List of boxes
What is Human Security ..................................................................................................................II
Tracking Terrorism: A Complex and Contested Exercise ........................................................ 12-13
Why Most Terrorist Organizations Fail ......................................................................................... 16
Violent Strategies Losing Favour in the Middle East and North Africa ..................................... 21
The Rise and Decline of Coups d’État in Sub-Saharan Africa .................................................... 28
List of figures
Chapter 1
Figure 1.1 Global Fatalities From Terrorism, 1998-2006 .......................................................... 10
Figure 1.2 Global Fatalities From Terrorism, Excluding Iraq, 1998-2006 ................................ 14
Figure 1.3 Global Fatalities From Terrorism vs Fatalities From Terrorism in Iraq, January 2005 to September 2007 ............................................................................. 14
Figure 1.4 Pakistan: As Terrorist Attacks Increase,
So Does Opposition to Terrorism, 2004-2006 ......................................................... 19
Figure 1.5 Average Number of Fatalities From Jihadi Terrorism per Year, Before and After the Invasion of Iraq ...................................................................... 19
Figure 1.6 Global Fatalities from Islamist Terrorism vs Fatalities from Islamist Terrorism in Iraq, January 2005 to September 2007 ................................................................. 20
Chapter 2
Figure 2.1 Number of State-Based Armed Conflicts
in Sub-Saharan Africa, 1946-2006 ........................................................................... 22
Figure 2.2 Number of Reported Battle-Deaths from State-Based Armed Conflict in Sub-Saharan Africa, 1946-2006 ........................................................................... 23
Figure 2.3 Number of State-Based and Non-State Armed Conflicts
in Sub-Saharan Africa, 2002-2006 ........................................................................... 24
Figure 2.4 Number of Reported Battle-Deaths from State-Based and Non-State Armed
Conflict in Sub-Saharan Africa, 2002-2006 ............................................................. 24
Figure 2.5 Number of Campaigns of One-sided Violence
in Sub-Saharan Africa, 2002-2006 ............................................................................25
Figure 2.6 Number of Reported Deaths from One-Sided Violence
in Sub-Saharan Africa, 2002-2006 ............................................................................25
Figure 2.7 Average Number of Conflict Onsets and Terminations per Year in Sub-Saharan Africa, 1950-2005 ............................................................................26
Figure 2.8 A Dramatic Decline in Autocracies in Sub-Saharan Africa, 1946-2006 .................26
Figure 2.9 Number of State-based Armed Conflict Terminations
in Sub-Saharan Africa, 1950-2005 ............................................................................27
Figure 2.10 Average Number of Coups d’Etat and Attempted Coups d’Etat per Year
in Sub-Saharan Africa, 1946-2006 ............................................................................28
Chapter 3
Figure 3.1 Number of State-Based Armed Conflicts by Type, 1946-2006 ................................33
Figure 3.2 Number of Reported Battle-Deaths from State-Based Armed Conflict
by Type, 1946-2006 .....................................................................................................33
Figure 3.3 Average Number of State-Based Armed Conflict Onsets and Terminations
per Year, 1950-2005 ....................................................................................................34
Figure 3.4 Number of State-based Armed Conflict Terminations, 1950-2005 ........................35
Figure 3.5 Number of Non-State Armed Conflicts, 2002-2006 ................................................36
Figure 3.6 Number of Reported Battle-Deaths from
Non-State Armed Conflict, 2002-2006 .....................................................................37
Figure 3.7 Number of Coups d’État and Attempted Coups d’État, 1946-2006 .......................38
Figure 3.8 Average Number of Coups d’Etat and Attempted
Coups d’Etat per Year, 1946-2006 ..............................................................................39
Figure 3.9 Coups d’Etat and Attempted Coups d’Etat by Region, 1946-2006 .........................39
Chapter 4
Figure 4.1 Number of Campaigns of One-Sided Violence, 1989-2006 ....................................42
Figure 4.2 Number of Campaigns of One-Sided Violence per Region, 2002-2006 ................43
Figure 4.3 Regions with The Worst Human Rights Records, 1980-2006 .................................45
Figure 4.4 Regions with the Best Human Rights Records, 1980-2006 .....................................45
Figure 4.5 Political Terror Scale Scores for OECD vs LICUS Countries, 1980-2006 ...............46
Figure 4.6 Global Trend in Human Rights Abuse, 1980-2006 ..................................................46
Figure 4.7 The World’s Worst Human Rights Abusers, 2006 ....................................................47Giacomo Pirozzi / Panos Pictures. CHAD.
HUMAN SECURITY BRIEF 2007
1
OVERVIEW
human security brief 2007
This Brief focuses on three main issues. First, it challenges the expert consensus that the threat of terrorism—especially Islamist terrorism—is increasing. It tracks a remarkable but largely unnoticed decline in the incidence of terrorism around the world, including a sharp decrease in deadly assaults perpetrated by al-Qaeda’s loosely knit Islamist global terror network.
Second, it analyzes the marked decline in the number and deadliness of armed conflicts in sub-Saharan Africa that has taken place since the end of the 1990s. It attributes this decline—and the parallel but longer-term fall in coups d’état in the region—to a significant increase in international initiatives directed towards stopping ongoing political violence and preventing it from restarting.
Third, it updates the global trend data on armed conflicts, battle-deaths, coups d’état, and human rights abuses that were reported in the Human Security Report 2005 and Human Security Brief 2006. It finds that there has been little net change in recent years in the number of conflicts in which a government is one of the warring parties, but that other forms of political violence, including communal conflicts, have declined.
Global Terrorism
Chapter 1, "Dying to Lose: Explaining the Decline in Global Terrorism," presents a comprehensive review of the statistical data on global terrorism to the end of 2006—domestic as well as international—and reveals that the global death toll from terrorism has fallen.
This finding will surprise many. Since the 11 September 2001 al-Qaeda assault on the United States, the consensus among Western experts has been that the threat of terrorism, particularly Islamist terrorism, has been increasing. This was the view held by the 2006 and 2007 US National Intelligence Estimates, by a 2007 survey of 100 foreign policy and security experts published in a major US journal Foreign Policy, by a 2007 report on the terrorist threat to Europe from the director of the UK’s Security Service, and by a 2008 report from the official US National Counterterrorism Center (NCTC).1
A related concern is that the incidence of terrorism is increasing. For the period under review, this concern finds support in three major terrorism datasets, all produced in the United States: one by NCTC, one by the Memorial Institute for the Prevention of Terrorism (MIPT) in Oklahoma City, and one by the National Consortium for the Study of Terrorism and Responses to Terrorism (START) at the University of Maryland.
There is broad agreement among the compilers of these datasets that terrorism is intentional politically motivated violence perpetrated by non-state groups against civilians and/or noncombatants.
According to NCTC, the number of fatalities from all terrorist attacks, Islamist and non-Islamist, domestic as well as international, increased by 41 percent from the beginning of 2005 (the first year for which the agency has complete data) to
HUMAN SECURITY BRIEF 2007
2
the end of 2006. We use numbers of fatalities rather than numbers of attacks as our preferred measure, because the former are the best single indicator of the human costs of terrorism and because definitions of what constitutes an "attack" vary considerably.
MIPT’s data show global fatalities from terrorism increasing more than fourfold from 1998 to 2006, with the steepest increase coming after the US-led invasion of Iraq in 2003. START records a 75 percent increase in 2004—the latest year for which it has released data.
The rising terrorist fatality toll revealed by the MIPT and NCTC datasets to the end of 2006, coupled with the bleak assessments of Western security analysts and intelligence agencies, appears to provide compelling support for the claim that both the incidence and the threat of global terrorism have indeed increased.
In fact, the data are open to a very different interpretation—one that strongly challenges the prevailing consensus. A more critical evaluation of the evidence indicates that the global toll from terrorist violence is not increasing but decreasing.
The reason that all three datasets show global fatality tolls from terrorism rising so steeply after 2003 is that each of them counts as acts of "terrorism" a very large percentage of deadly assaults against civilians by non-state armed groups in Iraq. This counting method, and the fact that for most of the period since the invasion in 2003 the conflict in Iraq has been the world’s deadliest, leads to some rather startling findings. MIPT claims that in 2006 an extraordinary 79 percent of global fatalities from terrorism were in Iraq.
This counting approach is unusual because the intentional killing of civilians in wartime is not normally described as "terrorism," but as a "war crime" or "crime against humanity." It is problematic because neither MIPT nor START is consistent in its counting practices. They both count as victims of terrorism a large percentage of the civilians intentionally killed by non-state armed groups in Iraq’s civil war, but they generally count only a small percentage of the civilians similarly killed by non-state armed groups in Africa’s civil wars.
MIPT, for example, records more than 2,000 deaths from terrorism in Iraq in 2004, but not a single death from terrorism in the Sudan, despite the fact that hundreds if not thousands of civilians were being slaughtered by insurgent groups and militias in Darfur during that year. The same was true in the Democratic Republic of the Congo in 1999 and in Uganda in 2002. In each case, large numbers of civilians were killed by insurgents, yet in each case MIPT did not record a single death from terrorism.
This failure to report the deaths of civilians intentionally killed by non-state armed groups as "terrorism" is repeated in country after country throughout Africa. It is not clear why this should be the case, but the dataset compilers may have been influenced by the way the US State Department categorizes what it calls "foreign terrorist organizations."
For a terror group to be so labelled by the State Department, it must "threaten the security of US nationals, or the national security ... of the United States."2 This highly US-centric definition excludes very large numbers of non-state groups in Africa and elsewhere that are guilty of perpetrating deliberate politically motivated violence against civilians—i.e., terrorism.
Whatever the reason for these coding decisions, the effect is the same. The death tolls attributed by MIPT and START to "terrorism"—civilians intentionally killed by non-state armed groups—in Africa’s civil wars is tiny compared with the reported death toll from terrorism in Iraq’s civil war.
Given these inconsistencies in counting civilian fatalities between Iraq and Africa, and given that deadly assaults on civilians by non-state armed groups in civil wars are not normally described as terrorism, there is a defensible case for removing the Iraq data from the global terrorism counts.
When this is done and the trend lines are redrawn, a dramatically different picture emerges. Absent Iraq, both START and MIPT now show net declines of more than 40 percent in fatalities from terrorism since 2001. NCTC’s fatality trend data still show an increase from 2005 to the end of 2006, but it is much less steep than when Iraq deaths are included.
In other words, the claim that the incidence of global terrorism is increasing is dependent upon accepting the unusual argument that violence intentionally perpetrated by non-state armed groups against civilians in Iraq should be treated as terrorism, but that similar violence elsewhere should not be counted as terrorism. If this argument is rejected, there has been a net decline in the global terrorism toll between 2001 and 2006.
In 2007, a Dramatic Change
Even if the unusual practice of counting the intentional killing of civilians in civil wars as terrorism is accepted, and even if the Iraq "terrorism" data are included, there has still been a substantial decline in the global terrorism toll—but this is a very recent development.
When this Brief was being prepared, complete annual data were only available from MIPT and NCTC to 2006 (2004 in the case of START). But in December 2007, NCTC released H U M A N S E C U R I T Y B R I E F 2 0 0 7 3
new figures that revealed that global terrorism fatalities had declined by 40 percent between July and September of that year—driven by a 55 percent decline in the terrorism death toll in Iraq.
The new data showed that while the war in Iraq had been driving the global terrorism death toll up from 2003 to 2006, it was now driving it down. (New data from NCTC released as this Brief went to press revealed that there had been a further 20 percent decline in the Iraq death toll from October to December 2007.)
This decline in what NCTC and MIPT count as fatalities from terrorism was part of an overall decrease in political violence in Iraq in 2007. This decrease was driven by several factors:
° The ceasefire observed by the Shia Mahdi Army since August 2007.
° The decision by former Sunni insurgents to ally with US forces against their former comrades in the Islamist "al-Qaeda In Iraq" organization.
° The "surge" of 30,000 extra troops deployed from the US in the first half of 2007, coupled with a new US counterinsurgency approach that places greater emphasis on defending the population.
° "Ethnic cleansing" in Baghdad—which has meant fewer mixed neighbourhoods, more "defensible space," and less disputed territory to fight about.
Of these factors, by far the most significant for the future of the Islamist or "jihadi" terror campaigns around the world has been the extraordinary revolt of Sunni Muslims against al-Qaeda In Iraq.
The reduction in casualties brought about by these changes has been dramatic, but in the absence of any major progress on the political front, the risk of a resumption of major conflict remains very real. But whatever happens, it is highly unlikely that al-Qaeda In Iraq’s fortunes will be reversed.
The Decline in Islamist Terrorism
The major concern in the West is not with local terrorist organizations fighting over local issues, but with the global campaigns of al-Qaeda and its loosely knit affiliates around the world. Here too, according to NCTC, there has been a remarkable, though uneven, recent decline in fatalities from Islamist terrorism. And here too the fall has been driven primarily by changes in Iraq.
The recent decline in Islamist terrorism reverses a steep increase that started after the invasion and occupation of Iraq in 2003. NCTC’s data indicate that the global fatality toll from Islamist attacks more than tripled between the beginning of 2005 and July 2007. But then came a dramatic change. Between July and September 2007 Iraqi deaths from Islamist violence dropped by more than 65 percent. This in turn drove a decline of 47 percent in the global toll.
NCTC’s was not the only quantitative data showing a decline in Islamist terrorism. The US-based Intelcenter think-tank published a study in mid-2007 that examined the 63 "most significant" attacks launched by al-Qaeda and its affiliates over a period of nearly 10 years.
It is highly unlikely that al-Qaeda In Iraq’s fortunes will be reversed.
The study included the major attacks most associated with Islamist terrorism—those in Bali, London, Madrid, Amman, and Jakarta, as well as 9/11 in the US. It did not include the Islamist violence in Iraq’s civil war that has been the main driver of the global Islamist death toll recorded by NCTC, nor did it include deadly assaults on civilians by insurgents in Afghanistan or other civil wars. Intelcenter found that by mid-2007 the number of Islamist attacks around the world had declined by 65 percent from the high point in 2004, and that fatalities from such attacks had declined by more than 90 percent.
In other words, the fall in Islamist terrorist violence has been remarkable whether or not the intentional killing of civilians in Iraq is counted. But just what has driven this decline—and what it means for the future—is less clear.
Does the Decline in the Incidence of Islamist Terrorism Mean That the Threat Has Diminished?
The decline in the fatality toll from Islamist terror operations does not necessarily mean that the threat has diminished. It is conceivably the case that al-Qaeda, or an affiliated group, may launch another attack on the scale of the 9/11 assault on the US—or one that is even more devastating. Should such an attack occur, it could dramatically reverse the downward trend in fatalities.
However, there are several reasons for believing that the recent decline in Islamist terrorism does in fact mean that the threat is diminishing.
First, counterterrorism efforts, although still plagued by a multitude of problems, are more widespread, more coordiH U M A N S E C U R I T Y B R I E F 2 0 0 7 4
nated, and more effective today than they were prior to 9/11. Part of the reason we are seeing fewer terrorist attacks is that a greater number are being prevented from occurring before they can even be launched.
Second, there is growing evidence of bitter doctrinal infighting within, and defections from, the now largely decentralized global Islamist network. Such developments are a classic sign of organizational crisis and incipient breakdown. Given that the Islamists have failed to achieve any of their strategic goals, and given the humiliating recent defeats experienced by al-Qaeda In Iraq, this development is not surprising.
Islamist terror groups confront a fundamental impasse––one largely of their own making.
Third, is the extraordinary drop in support for Islamist terror organizations in the Muslim world over the past five years—a decline that is driven by the increasingly popular rejection of the terrorists’ indiscriminate violence (that mostly targets fellow Muslims), their extremist ideology, and their harshly repressive policies.
Chapter 1 of this Brief focuses on this third issue. It argues that the more the Islamists attempt to impose their values and policies, and the more violence they perpetrate against their coreligionists, the more they lose support. The evidence for this is now overwhelming.
A Pew poll in July 2007, for example, revealed that Muslim support for terrorist violence against civilians had declined by half or more over five years in all of the four countries polled: Afghanistan, Pakistan, Saudi Arabia, and Iraq.
By late 2007 in Afghanistan just 1 percent of Afghans "strongly supported" the presence of the Taliban and foreign jihadi fighters in their country. In Pakistan, support for Islamist political parties has collapsed—dropping by more than four-fifths between the 2002 and the 2008 national elections. And in the North-West Frontier Province where al-Qaeda has its strongest presence in Pakistan, support for Osama bin Laden dropped from 70 percent in August 2007 to 4 percent in January 2008.
A December 2007 poll in Saudi Arabia found that Osama bin Laden’s fellow countrymen had "dramatically turned against him, against al-Qaeda, and against terrorism in general." And in Iraq, where the Islamists have suffered their greatest recent strategic setback, a major poll also released in December 2007 found that 100 percent of Iraqis—Sunnis as well as Shia—found al-Qaeda In Iraq’s attacks on civilians to be "unacceptable."
This pattern has been repeated in country after country in the Muslim world. Its strategic implications are critically important because historical evidence suggests that terrorist campaigns that lose public support will, sooner or later, be either abandoned or defeated. Without popular support, the Islamists cannot hope to create a successful political revolution—lacking any serious conventional military capacity, they cannot hope to defeat incumbent regimes by force of arms.
As Muslim publics increasingly reject Islamist policies and terror tactics, they are more likely to cooperate with official counterterror campaigns. This is precisely what happened in Iraq, where Sunni insurgents became so alienated from their former al-Qaeda In Iraq allies that they joined with the US in an anti-Islamist alliance to defeat them.
Even where terror organizations have a modest degree of support, their campaigns are still mostly notable for their failure rate. A 2005 study in International Security that examined 42 terrorist campaigns waged by 28 terror organizations of all types over a period of 5 years found that they failed to achieve their limited policy goals 93 percent of the time.
More than six years after 9/11, Islamist terror groups confront a fundamental impasse—one largely of their own making. Their indiscriminately violent terror tactics and harshly repressive policies have dramatically eroded their popular support in the Muslim world, sparked deep divisions within the global Islamist movement, and catalyzed increasingly effective counterterror campaigns around the world.
These and other changes examined in Chapter 1 of this Brief suggest that although the threat posed by al-Qaeda and its affiliates is still serious and far from being eliminated, the prognosis for this loosely knit global terror network is now bleak.
Towards a New Peace in Africa?
Commenting on sub-Saharan Africa’s security situation in January 1999, the Economist reported that "from north to south, east to west, large swathes of the continent are at war, but almost all efforts at pacification have come to naught."3 At the time, this assessment was unsurprising. In 1999 sub-Saharan Africa was the world’s most war-afflicted region, with a battle-death toll that was greater than the rest of the world’s combined. H U M A N S E C U R I T Y B R I E F 2 0 0 7 5
But as a guide to the future, the Economist’s pessimism was misplaced. Between 1999 and 2006 (the most recent year for which we have complete data), sub-Saharan Africa’s security landscape was transformed. The number of armed conflicts being fought in the region fell by more than half. The number of people being killed dropped even more steeply—by 2006 the annual battle-death toll was just 2 percent of that of 1999.
This was not the only positive change. Conflicts are conventionally defined as armed confrontations between a government and another government, or between a government and an insurgent group. But this definition—one used by almost all of the major conflict datasets—completely ignores communal and other "non-state" conflicts, those in which a government is not one of the warring parties.
To address this omission, the Uppsala Conflict Data Program (UCDP) created a new non-state conflict dataset for the Human Security Report Project (HSRP) that has revealed just how important this hitherto uncounted category of conflict is—particularly in sub-Saharan Africa.
In 2002 there were 26 of these non-state conflicts in sub-Saharan Africa—twice as many as those in which a government was one of the warring parties. But in this category, too, there has been a sharp decline in both conflict numbers and death tolls since 2002. Between 2002 and 2006 non-state conflict numbers had dropped by more than half across the region, and their death tolls had fallen by some 70 percent.
Violent campaigns waged against defenseless civilians by governments or rebel groups constitute a third type of political violence. Once again, we find a positive change taking place in the new millennium. Between 2002 and 2006 the number of campaigns of "one-sided violence" against civilians fell by two-thirds, and their death tolls fell by more than 80 percent.
Sub-Sahara African governments have also been highly vulnerable to coups d’état in the past. Indeed, the region has suffered almost half of the world’s coups since 1946. But here, too, there has been a positive change. The average number of coups per year in the new millennium has been some 40 percent lower than in the 1980s—the peak decade for coups in the region.
Explaining Changing Patterns of Political Violence in sub-Saharan Africa
In the aftermath of the Cold War, sub-Saharan Africa underwent a period of wrenching political change. In 1989 there were 36 autocratic regimes in the region; in 2000 there were just four. But few of the dictatorial regimes were replaced by inclusive democracies—a type of government that has a relatively low risk of succumbing to armed conflict. Most of the new governments were neither fully autocratic nor fully democratic, but a volatile mix of the two.
These mixed regimes—political scientists call them "anocracies"—are associated with much higher risks of armed conflict than either autocracies or democracies. From 1988 to 1999 the number of anocracies increased fifteenfold in sub-Saharan Africa, from two to 30, a change that helped drive the steep increase in the number of state-based conflicts (those involving a government as one of the warring parties) that had started in the 1990s. The average number of new conflicts starting each year in the 1990s was double that in the 1980s.
This upsurge of conflict onsets (some were old conflicts that had restarted) ensured that sub-Saharan Africa was the world’s most violent region in the 1990s. What prevented the total number of conflicts being fought each year from being even higher was that the average number of conflict terminations each year also increased dramatically during the decade.
The doubling of new conflicts starting in the 1990s indicates that whatever conflict prevention initiatives were being attempted during this period were having a negligible impact. This was bad news for policy-makers at the UN and elsewhere, where the idea that "prevention is better than cure" has become widely accepted—though much less widely practiced.
The number of armed conflicts in the region fell by more than half.
But the fact that the average number of conflicts ending each year in the 1990s was more than twice that of the 1980s, and that a much greater percentage of these terminations was made up of negotiated settlements, was good news. It indicated that what the UN calls "peacemaking"—initiatives designed to bring conflicts to an end—was meeting with growing success.
In the new millennium, the security situation in sub-Saharan Africa underwent a second remarkable change. Not only did new outbreaks of warfare sharply drop but there were appreciably more old conflicts stopping than there were new conflicts starting. As a consequence of these changes, conflict numbers shrank from 13 in 2000, to only seven in 2006.
For policy-makers, it is critically important to understand why both the number and the deadliness of conflicts of all H U M A N S E C U R I T Y B R I E F 2 0 0 7 6
types have dropped so dramatically throughout sub-Saharan Africa since 1999.
The most robust finding on the causes of war by researchers is that the higher the per capita income a country enjoys, the lower its risk of armed conflict. This is why most wars take place in very poor countries. The evidence for this war-poverty association is overwhelming.
But while economic growth is clearly associated with reductions in the incidence of both conflicts and coups over the long term, it cannot explain the steep recent decline in the number of armed conflicts in sub-Saharan Africa. Since 2000, income levels have not risen high enough, or over a long enough period, to explain the reduction in new outbreaks of conflict. This is also true of all the other "structural" explanations that researchers have focused on—those stressing the impact of other slow-changing risk factors, such as "youth bulges," population size, dependence on primary commodities, and so on. The explanation for the radical improvement in sub-Saharan Africa’s security climate in the new millennium must lie elsewhere.
The HSRP’s research suggests that the drivers of this remarkable decline in armed conflicts in sub-Saharan Africa are to be found not in long-term structural change, but in the post-Cold War surge of policy initiatives designed to stop wars ("peacemaking") or prevent them from starting again ("postconflict peacebuilding"). Relatively little effort has been put into conflict prevention.
These peacemaking and postconflict peacebuilding initiatives—which include third party mediation efforts to end ongoing conflicts, humanitarian missions, and peacekeeping and peacebuilding operations, have all greatly increased in number since 1990.
Non-state conflict numbers have undergone a marked decline.
The success rate of individual initiatives has often been modest, and critics have been quick to point to the failures, particularly of UN peace operations. The criticisms are accurate enough but too often miss the important truth that the net effect of this upsurge of security-related initiatives has been highly positive, notwithstanding the lack of success in individual cases.
As we noted in the 2005 Human Security Report, if these sorts of results can be achieved by peacemaking and peacebuilding missions that too often suffer from inadequate resources, ineffective coordination, inappropriate mandates, and lack of political support, then the potential for improvement—via better designed, resourced, and implemented policy initiatives—is clearly very large.
Global Trends in Human Insecurity: An Update
The 2005 Human Security Report found that armed conflicts involving a government as one of the warring parties had declined by more than 40 percent around the world from the end of the Cold War to 2003. The most severe conflicts—and the number of genocides—had declined by some 80 percent. Coup d’état numbers were down by 60 percent from the high point in 1963.
The Report also found that the number of battle-deaths in state-based armed conflicts had declined even more steeply than the conflict numbers—though over a much longer period. The average number of battle-deaths per conflict per year—the best measure of the deadliness of warfare—had fallen from 38,000 in 1950, to just 600 in 2002. By 2006 the estimated global battle-death toll had declined further, but only very slightly. Moreover, uncertainty about fatality numbers in Iraq and elsewhere has meant that while we can be confident about the long-term trends, no conclusions should be drawn from minor year-to-year variations in reported death counts.
The 2006 Human Security Brief found that the global decline in state-based conflict numbers had more or less levelled out, and this Brief reveals that this situation changed little between 2005 and 2006. But at the regional level there have been major changes since 2002. While sub-Saharan Africa has witnessed a major decline in political violence, two other regions—Central and South Asia, and the Middle East and North Africa—have seen sharp increases in both conflict and fatality numbers.
As reported in Chapter 2, the HSRP now has five years of data on "non-state" conflicts—those fought between communal or rebel groups or warlords, but in which the government is not a warring party.
Five years is long enough to detect trends, and we can now report that worldwide, non-state conflict numbers have undergone a marked and consistent decline since data were first collected in 2002. In fact, they declined by a third—from 36 to 24—between 2002 and 2006. Reported battle-deaths from these conflicts declined by 60 percent over the same period. Much of this global reduction in the number of non-state conflicts and associated fatalities has been driven by the improvements in sub-Saharan Africa.H U M A N S E C U R I T Y B R I E F 2 0 0 7 7
When we look at the combined global total of state-based and non-state conflicts, we find that there has been an 18 percent decline—from 68 in 2002, to 56 in 2006.
A third type of political violence involves deadly campaigns against defenceless civilians. Campaigns of "one-sided violence"—which can be perpetrated by either governments or non-state armed groups—often take place during civil wars, with sub-Saharan Africa having the lion’s share.
Campaigns of one-sided violence, like armed conflicts, have to result in at least 25 fatalities a year to be counted as such. In 1989 there were 19 such campaigns, but their number grew unevenly throughout the 1990s—lending support to the view that targeting civilians had become an increasingly prevalent element of the post-Cold War security landscape. In 2004, the peak year, 38 campaigns of one-sided violence were being perpetrated around the world, but since then there has been a sharp drop. In 2006 there were just 26—a net decline of 32 percent.
Political violence also involves deadly campaigns against civilians.
This Human Security Brief also provides new data to update global and regional trends in core human rights abuse—primarily imprisonment and physical violence. But measuring such abuse is both difficult and controversial. It is difficult because there is no single, accessible, and objective yardstick to measure human rights violations. It is controversial because many human rights organizations reject the very idea of quantifying abuses of rights, on both methodological and moral grounds.
The UN’s Human Rights Council has no mandate to collect comprehensive data on human rights abuse, and any attempt to secure one would almost certainly be frustrated by member states. So, the international community finds itself confronting a critically important human security issue with no official data to determine whether or not its policies are having any impact.
The HSRP relies on the Political Terror Scale (PTS), a composite index compiled annually by researchers at the University of North Carolina, Asheville. The PTS uses data on core human rights violations in individual countries around the world that are drawn from the annual reports of Amnesty International and the US State Department.
The three regions that have had the worst human rights records between 1980 and 2006 are sub-Saharan Africa, the Middle East and North Africa, and Central and South Asia. The Americas, Europe, and East and Southeast Asia and Oceania score best. But the starkest differences in levels of rights violations turn out to be those between rich and poor countries.
There are a number of important methodological challenges associated with the rights violation data that are reviewed in Chapter 4, but reporting practices today are more extensive and consistent than they were in the past, and dataset compilers are more sensitive to coding challenges and political biases that may have distorted the trend data during the Cold War years.
Conclusion
The long-term statistical data on human insecurity reviewed in this Brief—on terror attacks, and on the global and regional incidence of wars, coups, and core human rights abuses—can be useful to policy-makers in several ways.
First, it can help them better understand the drivers of political violence of all types.
Second, it can provide insights into which policies may be most useful in forestalling such violence, in halting violence that cannot be forestalled, and in preventing violence that has been halted from starting up again.
Such information and analysis is a necessary—though not sufficient—condition for the "evidence-based policy" that increasing numbers of international organizations and donor governments are demanding.
The recent changes reported in this Brief provide grounds for modest optimism—not least because the evidence clearly indicates that efforts to stop violent conflicts and to prevent them from starting again can be remarkably effective. But few of the "root cause" drivers of warfare and deadly assaults against civilians—from poverty to group inequality—have improved, and some have worsened. Given this, and with 56 armed conflicts still being waged around the world, there are certainly no grounds for complacency.H U M A N S E C U R I T Y B R I E F 2 0 0 7 8 Khalid Mohammed / Associated Press. IRAQ. Dying to Lose: Explaining the Decline in Global Terrorism In October 2003, then US secretary of defense, Donald Rumsfeld, noted in a confi dential memo to senior administra tion offi cials, "We lack metrics to know if we are winning or losing the global war on terror.4 "Today there are "metrics"— notably three datasets—one from an offi cial US government agency and two others that are funded by the US Department of Homeland Security.5 This Brief provides the fi rst critical assessment of their fi ndings. Each of the three datasets tracks the global incidence and human costs of all forms of terrorism—domestic and international, religious and secular. However, notwithstanding the mass of data that is now available, determining whether terrorism is increasing or decreasing around the world remains a complex and controversial task. In part this is because attempts to measure a phenomenon, the very meaning of which is subject to intense—and often highly politicized— debate are bound to be contested. The United Nations (UN) has consistently failed to reach an agreed defi nition of terrorism in part because, as the well-known cliché puts it, "One man’s terrorist is another man’s freedom fi ghter." For the purposes of this chapter, terrorism is defi ned as a tactic—"the intentional use of violence for political ends by non-state actors against civilians." This defi nition is broadly compatible with those adopted by the three datasets discussed here. In what follows we provide a brief overview of how security experts view the global terrorist threat. We then subject the claims associated with this assessment to a critical test, drawing on the statistics from the three datasets. We show how the statistical information that these and other datasets provide can be read in very different ways and that a close examination of the data, together with other research fi ndings, reveals a picture that is very much at odds with the mainstream consensus. The Expert Consensus More than six years after al-Qaeda’s September 11 assault on the United States, expert opinion in the West holds that the threat of global terrorism is growing. There are few dis senting voices. In August 2007 a nonpartisan survey of 100 leading US foreign policy and security experts by the Center on American Progress and the US journal Foreign Policy reported that 84 percent of those polled rejected the assertion that the United States was winning the war on terror. The central focus of this "war" is, of course, Islamist terrorism.6 This pessimistic assessment was in line with the fi ndings of the 2006 US National Intelligence Estimate, which reported that "activists identifying themselves as jihadists ... are increasing in both number and geographic dispersion."7 Similar sentiments were reiterated in the July 2007 National Intelligence Estimate.8 In November 2007 the director of the UK’s Security Service claimed that in the previous 12 months C H A P T E R 1H U M A N S E C U R I T Y B R I E F 2 0 0 7 9
there had been "an increase in [terrorist] attack planning across the continent."9
The consensus view of the various Western intelligence agencies is in turn supported by statistics from the three datasets referred to in this chapter. The National Counterterrorism Center (NCTC), the official US agency charged with tracking the incidence of terrorism around the world, has data that show the number of terrorist attacks—and the fatalities they cause—have increased steeply worldwide from 2005 to 2006—the last year for which the agency has complete data. Similarly, the US-based Memorial Institute for the Prevention of Terrorism (MIPT), which has statistics on international and domestic terrorism going back to 1998, shows fatalities from terrorism worldwide increasing sharply from 2003—as does the relatively new National Consortium for the Study of Terrorism and Responses to Terrorism (START) dataset from the University of Maryland.10
In all three datasets it is clear that the sharp increase in "terrorism" fatalities around the world has been driven by the rapidly rising civilian death toll in Iraq that followed the US-led invasion in 2003. But describing the intentional killing of civilians in civil wars as "terrorism" is both unusual and somewhat controversial. It also has the effect of greatly inflating the global terrorism toll.
A recent statistical study revealed a huge jump in Islamist terror incidents worldwide after 2003.
The MIPT, NCTC, and START datasets all include fatalities from domestic as well as international terrorism. However, the concern driving the US-led "Global War on Terror" (GWOT) is not local terror groups fighting over local issues, but the threat—and especially the threat to the West—from Islamist terrorists associated with the global campaigns of al-Qaeda and its affiliates.
This concern is understandable. Islamist groups around the world are well-organized and well-funded; their members are resolutely committed to their cause; their networks have a global reach; they communicate, inform, and propagandize via hundreds of Islamist websites; and they have launched major attacks on six continents. A recent statistical study that is discussed in detail later in this chapter revealed a huge jump in Islamist terror incidents worldwide after 2003—an increase again driven by events in Iraq.
The fact that the loose Islamist terror network inspired by Osama bin Laden has metastasized in recent years creating quasi-independent "homegrown" or "self-starter" Islamist terror nodules in Europe and elsewhere has been a cause for further concern.
The expert consensus in the West is that the threat of global terrorism is growing.
Some US commentators even believe that the West confronts an existential "Islamo-fascist" terrorist threat as grave as the dangers posed by Nazi Germany.11 Many more believe it is simply a matter of time before an Islamist terror organization gains access to, and uses, weapons of mass destruction (WMD).12
Although there are some notable dissenters,13 the expert consensus in the West is that the threat of global terrorism—and of Islamist terrorism in particular—is growing.
What the Data Reveal
In this section we review the data on the incidence of all types of terrorism around the world. Later, we address the particular challenges involved in determining whether Islamist terrorism is increasing or decreasing.
In tracking terrorism from year to year we rely primarily on fatality counts, rather than the number of attacks. This is partly because fatalities are the best measure of the human cost of terrorism, but it is also because the definition of terrorist "attack" can differ from dataset to dataset. For example, it is possible to count 100 coordinated bombings in a single city in a single day as one terrorist incident—or as 100. Yet, regardless of how the data compilers decide to count incidents in a case like this, the fatality toll will remain essentially the same.
According to NCTC, the number of fatalities from all terrorist attacks, Islamist and non-Islamist, domestic as well as international, increased by 41 percent from the beginning of 2005 to the end of 2006. NCTC recorded 14,618 fatalities in 2005; 20,573 in 2006.14
MIPT’s dataset shows global fatalities from terrorism increasing from 2,172 in 1998, to 12,070 in 2006, an increase of some 450 percent. Most of this increase takes place after the invasion and occupation of Iraq in 2003.15 The START dataset shows fatalities rising by 75 percent in 2004 alone—2004 is the last year for which the START team has released data.H U M A N S E C U R I T Y B R I E F 2 0 0 7 10
The escalation of the global fatality toll is clearly revealed in Figure 1.1 above. The peak in 2001 is caused by al-Qaeda’s September 11 attacks on the US.
The rising post-2003 fatality toll revealed by all three datasets, coupled with the bleak assessments of US intelligence agencies, appears to provide compelling evidence for the claim that the global terrorist threat has indeed increased significantly. However, as we will see, the data are open to quite different interpretations.
A Misleading Picture?
The reason that the NCTC, MIPT, and START global fatality tolls rise so dramatically after 2003 is because all three datasets are counting a large percentage of all civilian fatalities from intentional violence in Iraq’s civil war as deaths from "terrorism." For example, NCTC’s estimate for fatalities from terrorism in Iraq in 2006 is 13,343. This is nearly 80 percent of the total Iraqi civilian fatality toll of 16,657 for that year as estimated by the independent US organization, icasualties.org.16
In 2006 Iraq’s share of global deaths from terrorism—as recorded by NCTC and MIPT—was startlingly high. According to NCTC, in 2006 some 64 percent of terrorist fatalities worldwide were in Iraq. MIPT’s data indicate that Iraq’s share was an extraordinary 79 percent.17
Since the concept of terrorism remains contested, the counting rules used by NCTC, MIPT, and START are as legitimate as any others. But they are unusual because counting the intentional killing of civilians in civil wars as "terrorism," as all three datasets do, is a sharp departure from customary practice. As Ohio State University’s John Mueller has noted: "When terrorism becomes really extensive in an area we generally no longer call it terrorism, but rather war or insurgency."19 Moreover, as a July 2007 US Congressional Research Service report noted, NCTC’s Iraq data are, "largely the product of sectarian violence, rampant criminal activity, and home-grown insurgency—[and therefore] grossly distort the global terrorism picture."20
Over the past 30 years, civil wars in the Democratic Republic of the Congo (DRC), Angola, Liberia, Sierra Leone, Uganda, Bosnia, Guatemala, and elsewhere have, like the war in Iraq, been notorious for the number of civilians killed.But although much of the slaughter in all these cases was intentional, politically motivated, and perpetrated by non-state groups—and thus constituted "terrorism" as conceived by MIPT, NCTC, and START—it was almost never described as such.
Accounts of the human costs of these conflicts typically refer to "death tolls"—a term that usually includes both combatants and civilians. Insofar as the intentional killing of civilians in wartime has been the focus of specific attention, it has traditionally been described as a "war crime" or "crime against humanity," or even "genocide"—but not "terrorism."21
However, the departure from traditional practice is not the only reason for concern. What makes MIPT and START’s fatality counting practices particularly problematic is that they are not applied consistently. To be more specific, while both institutions count a large percentage of all violent civilian deaths in Iraq’s civil war as terrorism, they code extraordinarily few of the thousands of violent civilian deaths in Africa’s many civil wars since 1998 this way. (NCTC does not cover the years in which the sub-Sahara African conflicts noted below were taking place and therefore its data are not considered here.)
We know that the politically motivated killing of civilians by non-state armed groups has been seriously undercounted in Africa by MIPT and START because we can compare their terrorism fatality data with statistics compiled by Uppsala University’s Conflict Data Program (UCDP).22 UCDP does not use the term "terrorism," but the UCDP dataset on "one-sided violence" includes fatality data on intentional politically motivated violence perpetrated against civilians by non-state armed groups. This is very close to the definition of terrorism used by MIPT and START.
Comparing UCDP’s data on Africa’s civil war fatalities with those of MIPT and START is instructive. Take the case of Sudan. In 2004 UCDP, whose estimates are always conservative,
Figure 1.1 Global FatalitiesFrom Terrorism, 1998-20061999199820002001200220032004200520062500020,00015,00010,0005,0000YearNumber of FatalitiesSTARTMIPTNCTC
Data Sources: MIPT; NCTC; START.18
All three terrorism datasets show a sharp worldwide increase in fatalities following the invasion of Iraq in 2003.H U M A N S E C U R I T Y B R I E F 2 0 0 7 11
counted 723 civilian deaths perpetrated by the Janjaweed and other non-state armed groups. Yet, MIPT recorded zero deaths from terrorism in Sudan in 2004; START counted just 17.
MIPT defines terrorism as politically motivated "violence, or the threat of violence, calculated to create an atmosphere of fear and alarm ... [and] generally directed against civilian targets."23 Given this definition, how could MIPT, which reported 2,471 deaths from terrorism in Iraq’s civil war in 2004, record no terrorism deaths at all from Sudan’s civil wars in the same year?
In the DRC in 1999 large numbers of civilians were being deliberately targeted by rebel groups.
Sudan is not the only African country where MIPT and START appear to be using different fatality counting rules from those they use in Iraq. In the DRC in 1999—where large numbers of civilians were being deliberately targeted by rebel groups in a vicious civil war—the same pattern is evident. MIPT again found that there had been no fatalities from terrorism; START counted seven; UCDP 624. In Uganda in 2002, MIPT’s terrorism count was again zero, START’s was 107, while UCDP’s was 1,109.
Perhaps the most telling comparison is that between MIPT’s estimate of terrorism’s share of all deaths—combatants as well as civilians—in Iraq in 2006 with its estimate of terrorism’s share of all deaths in sub-Saharan Africa in 1999. We chose 1999 because it has the highest death toll from armed violence in that region of any year from 1946 to 2006, and because Africa’s wars at this time, particularly in the DRC and Angola, were notorious for their attacks on civilians. We chose 2006 for Iraq because this was the year that that country experienced its highest death toll since 1998 according to NCTC and MIPT.
If MIPT’s coding practices had been the same in sub-Saharan Africa as they are in Iraq, then we would expect that terrorism’s share of all fatalities in Africa in 1999 would have been significantly greater.
In fact, while MIPT’s data indicates that 48.4 percent of all fatalities in Iraq in 2006 were due to terrorism, in sub-Saharan Africa for 1999 it finds that just 0.06 percent of fatalities were due to terrorism. If the intentional killing of civilians is not counted as terrorism in Africa’s civil wars, it should not be counted in Iraq’s civil war either.24
It is not clear why MIPT and START use different coding practices in Iraq and in sub-Saharan Africa, but a review of their terrorism fatality counts in different countries around the world suggests one possible explanation.25 In countries where intentional political violence against civilians is widely viewed as terrorism in the US and by the international community—in southern Thailand, Indonesia, the Philippines, Spain, and Israel, for example—it is counted as terrorism by MIPT and START. Similarly, in sub-Saharan Africa, MIPT and START recorded all of the fatalities from the al-Qaeda attacks in 1998 and 2002 in Kenya, and in Tanzania in 1998, as terrorism.
In civil wars where some insurgents are widely identified in the US as "terrorists"—in Iraq, Afghanistan, and Colombia, for example––MIPT and START also tend to count civilian fatalities from insurgent violence as terrorism. However, in civil wars in which intentional violence against civilians by rebels is not widely identified as terrorism in the US and elsewhere, MIPT and START ignore, or seriously undercount, civilian fatalities from political violence. The sub-Saharan African wars noted above are cases in point.
These coding decisions suggest that MIPT and START researchers may have been influenced by the US State Department’s criteria for determining what constitutes a "foreign terrorist organizations", in particular the requirement that such organizations must, "... threaten the security of US nationals, or the national security ... of the United States."26 Clearly this highly US-centric definition excludes many of the non-state groups in Africa and elsewhere that are guilty of perpetrating intentional, politically motivated violence against civilians—i.e., actions that fit the broadly consensual definition of terrorism that the MIPT, NCTC and START datasets use.
Treating civil war deaths in sub-Saharan Africa differently from those in Iraq, distorts the trend data.
Whatever the reason for treating civilian deaths in civil wars in sub-Saharan Africa differently to those in Iraq, the practice distorts the trend data. Had civilian fatalities from intentional violence in sub-Saharan Africa been counted the same way as civilian fatalities in Iraq were counted, the MIPT and START trend data would have revealed a far higher global death toll from terrorism from 1998 onwards––and the sharp post-2003 increase in fatalities caused by Iraq’s civil war would have been much less significant.H U M A N S E C U R I T Y B R I E F 2 0 0 7 12 Just three research institutions—all based in the US—track the incidence of terrorism around the world and publish their fi ndings annually: ° The National Counterterrorism Center (NCTC), an offi cial US agency created in August 2004, collects and records data on terrorism as part of its mandate. It has published statistics on international and domestic terrorism since 2005. NCTC was created in response to criticism about the inadequacy of the US State Department’s annual Patterns of Global Terrorism reports.27 ° The Memorial Institute for the Prevention of Terrorism (MIPT) was created in memory of the Oklahoma bombing in 1995. MIPT has data from 1998 to the end of 2006.28 Prior to 1998, MIPT has data on international terrorism only. ° The relatively new National Consortium for the Study of Terrorism and Responses to Terrorism (START) at the University of Maryland has two datasets. GDT1 1970-1997 was created from the Pinkerton Global Information dataset. Data for the GDT2 dataset, which is the one referenced here, were collected by the START team. Thus far, START has only published its fi ndings to the end of 2004.29 Both START and MIPT are funded by the US Department of Homeland Security. All three datasets collect information on both domestic as well as international terrorism. Previously, datasets like "Iterate" and the State Department’s "Patterns of Global Terrorism" focused exclusively on international terrorism— which over the past 30 years has killed, on average, fewer than 500 people a year.30 The current, more inclusive, approach to data collection refl ects the belief that the distinction between "domestic" and "international" often obscures more than it reveals. In Europe, for example, many Islamist terrorists were European nationals, but were inspired by organizations like al-Qaeda and often had links with overseas terror groups. Despite the huge challenges involved in compiling terrorism datasets, MIPT, NCTC, and START have made a major contribution to our understanding of the changing incidence of terrorism around the world. Without the systematic and timely collection of global and regional data on terrorist attacks and fatalities, there is no way of determining whether or not the incidence of terrorism is increasing or decreasing—information that is essential for evaluating the changing nature of the terrorist threat and the success, or failure, of counterterrorism policies. We address the issue of whether or not it is appropriate to categorize the intentional killing of civilians in wartime as terrorism elsewhere this chapter. But this is by no means the only controversial issue that dataset compilers working in this area have to address. Here we review three other major challenges. Access to Reliable Data All three datasets rely on media and other reports in compiling statistics on terrorist attacks and fatalities. However, deaths often go unreported in civil wars and hence are not recorded. Even when deaths are reported, it is often diffi cult to determine whether the victim was a civilian or a combatant in civilian clothes. This matters because killing combatants does not normally count as an act of terrorism. Terrorists do not always claim responsibility for their actions. So even when it is clear that the victim is a civilian, it may not be possible to determine the identity of the perpetrator. Knowing the identity of the perpetrator is important—the intentional killing of civilians by non-state armed groups will be counted as terrorism, the intentional killing of civilians by government forces will not. A similar problem arises when researchers try to de termine whether violence was perpetrated with political or criminal intent—again being able to make this distinction is crucial because purely criminal violence does not count as terrorism. Should Terrorism Counts Include Only Civilian Deaths? Most analysts agree that one of the defi ning characteristics of terrorism is that it involves attacks on civilians. Yet, at the same time—and somewhat paradoxically—few in the West would dissent from the claim that the al-Qaeda attack on the USS Cole was an act of terrorism—even though the Cole was an on-duty warship. MIPT, NCTC, and START address this issue differently. TRACKING TERRORISM: A COMPLEX AND CONTESTED EXERCISE Confl icting defi nitions, inadequate data, and inconsistent coding rules greatly complicate efforts to measure the incidence and intensity of terrorism around the world.H U M A N S E C U R I T Y B R I E F 2 0 0 7 13 It is important to note here that there is nothing in the defi nitions of terrorism adopted by MIPT or START that suggests that fatalities from intentional violence against civilians in Africa’s civil wars should not be included in the global terrorism count. Global Terrorism Trends without Iraq What happens if we remove Iraqi deaths from the global ter rorism count to determine what the underlying trends are? There is a defensible case for doing this since, as noted above, the intentional killing of civilians in civil wars has not tradi tionally been described as terrorism. In addition, by counting civilian deaths from intentional violence in Iraq’s civil war as terrorism, MIPT and START are not only at odds with tra ditional practice but also with their own coding practices in Africa’s civil wars. In Figure 1.2 the fatality trend lines from Figure 1.1 are redrawn with the Iraq death toll omitted. A radically different picture now emerges. The huge increases in the global terrorism death toll following the invasion of Iraq that were so dramatically evident in the MIPT, NCTC, and START trend lines in Figure 1.1 have disappeared completely. Now neither of the two datasets that record fatalities back to 1998 shows any substantive increase—indeed both show a net decline in fatalities, from 2001 in the case of MIPT, and from 1998 in the case of START.36 Clearly, if the hitherto unusual practice NCTC uses the term "noncombatant" rather than "civilian" in its discussion of what constitutes terrorism. "Noncombatants," according to NCTC, include "military personnel outside a war zone or warlike setting." Since the Cole was neither in a war zone, nor a warlike setting, the al-Qaeda attack was clearly an act of terrorism for NCTC. MIPT acknowledges that terrorism is "generally directed against civilian targets," but goes on to state that when attacks on military or police forces are carried out "in order to make a political statement," they should be designated as terrorist acts.31 It is, however, often impossible to know the intent of perpetrators, so it is quite unclear how coders could make such determinations with any degree of confi dence. But there is no doubt that MIPT assumes that a great deal of violence against the police and the military is intended to "make a political statement," since in 2005 military and police deaths constituted more than a third of all fatalities in the MIPT database. START recognizes that opinions differ as to whether attacks on the military or police should be counted as terrorism and does not stipulate whether they should or should not be included. Rather, START leaves it to users of the dataset to create their own defi nitions. They can do this by using fi lters to exclude (or include) particular categories of victims, including the police and the military.32 At this stage of its development, however, START’s dataset is far from user-friendly and contains many anomalies. The Difference between "Terrorism" and "Insurgent" or "Sectarian" Violence Both MIPT and NCTC make a distinction between "ter rorism," on the one hand, and "insurgent" and "sectarian" violence, on the other. Yet, while both defi ne "terrorism," neither defi nes "insurgent" or "sectarian" violence in a way that is helpful in distinguishing terrorism from the latter two forms of violence. MIPT, for example, defi nes terrorism as "a tactic," while "insurgency" is described as "a political-military strategy."33 But this distinction does not tell us whether a particular attack on civilians should be coded as a case of terrorist violence or as a case of insurgent violence. NCTC notes that "in the cases of Iraq and Afghanistan, it is particularly diffi cult ... to distinguish terrorism from the numerous other forms of violence, including crime and sectarian violence."34 However, while NCTC defi nes terrorism, it does not defi ne either insurgency or sectarian violence. Yet, without clear and unambiguous coding rules—which in turn require clear defi nitions—data cannot be coded consistently. And consistency is critical. As Alan B. Krueger and David Laitin noted in their infl uential 2004 critique of the US State Department’s Patterns of Global Terrorism data: "Time-series analysis, which seeks to discern trends in given phenomena over time, requires a consistent approach to collecting data."35 Researchers at MIPT, NCTC, and START are acutely aware of the diffi culties of working with contested defi nitions and insuffi cient and often inaccurate information. All three datasets are seen as "works in progress," with data constantly being revised as new information becomes available. As the discussion above clearly indicates, the challenges involved in tracking terrorism are very real. Nevertheless, the data that MIPT, NCTC, and START provide, when used with due caution, can be illuminating. They reveal surprising and important fi ndings about current terrorism trends and the factors that drive them.H U M A N S E C U R I T Y B R I E F 2 0 0 7 14
of counting civilian fatalities from intentional violence in wartime as terrorism is rejected, then the trend data in Figure 1.2 pose a major challenge to the expert view that the global terrorist threat is increasing.
Iraq in 2007, a Dramatic Change
Thus far we have only reviewed the global fatality data to the end of 2006—this being the last year for which NCTC and MIPT have complete annual statistics.37 However, in December 2007 NCTC released new fatality data covering the period from the beginning of that year to the end of September.
The new data reveal a dramatic decline in terrorist fatalities from March to September 2007. The decline in Iraq for this six-month period is 61 percent; the worldwide decline is 46 percent. And, as Figure 1.3 makes clear, the civilian fatalities in Iraq that had driven the global terrorist toll sharply up from 2005 to 2006 were now driving it sharply down.
If NCTC’s practice—which is shared by MIPT and START—of counting the deliberate killing of noncombatants in civil wars as terrorism is accepted, then the steep reduction in such killings in 2007 poses an additional challenge to the expert consensus that the global terrorist threat is worsening. If the intentional killing of civilians in Iraq is not counted as terrorism, then the evidence still suggests there has been a decline in terrorist fatalities—although in this case the decline starts earlier and is more modest.
NCTC was not the only organization to record a drop in deadly assaults on civilians in Iraq in 2007. In September 2007 General David Petraeus, commander of the Multi-National Force-Iraq, testified to the US Congress that there had been an unprecedented decline in violence in Iraq. Civilian deaths, he claimed, had declined by 45 percent Iraq-wide since the high point of sectarian violence in December 2006.38
Petraeus’s claim was strongly disputed by opponents of the war in the US.39 But the declining trend he reported was virtually identical to that reported by the UK-based Iraq Body Count (IBC), a strongly antiwar organization that has kept a careful record of civilian fatalities from organized violence for several years.40 IBC’s data show that civilian deaths dropped by some 69 percent from the middle of 2006 to November 2007.41
In the months that followed Petraeus’s presentation, the death toll kept dropping. In mid-November, the US military reported that civilian fatalities were 60 percent lower than in June, while the weekly count of armed attacks across the country had shrunk from 1,600 to 575.42 In December the military claimed that fewer weapons were entering the country from Iran, while the number of foreign fighters entering Iraq from Syria was down by 25 to 30 percent.43
The Iraqi Interior Ministry, which uses a different methodology for counting civilian deaths than that used by the US military, reported in early December 2007 that 538 Iraqi civilians had been killed in November, two-thirds fewer than the August toll. This was the lowest monthly civilian death toll reported by the ministry since February 2006.44
These various estimates are based on quite different counting methods, some more thorough than others, but all the data reveal a similar trend in declining fatalities—military as well as civilian.
Figure 1.2 Global Fatalities From Terrorism,Excluding Iraq, 1998-200619991998200020012002200320042005200625,00020,00015,00010,0005,0000Number of FatalitiesYearSTARTMIPTNCTC
Data Sources: MIPT; NCTC; START.
Absent Iraq, there has been no major increase in fatalities from terrorism since 2001.
Figure 1.3 Global Fatalities From Terrorism vs Fatalities From Terrorism in Iraq, January 2005 to September 2007Jan-05Apr-05Jul-05Oct-05Jan-06Apr-06Jul-06Oct-06Jan-07Apr-07Jul-073,0002,5002,0001,5001,0005000Global FatalitiesFatalities in IraqNumber of FatalitiesYear
Data Source: NCTC.
Violence against civilians in Iraq drives the global terrorism toll.H U M A N S E C U R I T Y B R I E F 2 0 0 7 15
By the end of 2007, with tens of thousands of Iraqi refugees beginning to return to Baghdad, there was no longer any doubt that the security situation in the country had undergone a major change. But what had driven the change remained the subject of lively debate.
Why Civilian Fatalities in Iraq Declined in 2007
Because it is clear that NCTC is counting such a large percentage of all civilians killed in Iraq by insurgent or sectarian violence as victims of terrorism, any explanation of the decline in the civilian death toll, in general, will also be a major part of the explanation of the decline in the fatality toll from terrorism as NCTC defines it.
The steep decline in the Iraqi and global terrorism tolls in 2007 was driven by a series of major changes in the Iraqi security environment during the year. First, was the much-vaunted "surge"—the deployment of nearly 30,000 extra US troops to Iraq in the first half of 2007.45 Second, was a major shift in US military strategy on the ground. In 2007 providing security for the population had, for the first time, become a top priority for the US military—a radical change from past practice. The additional troops provided by the surge greatly facilitated this new policy.
The third change was the security effect of forced population movements—the good news about declining civilian deaths in Baghdad was due in part to the bad news about "ethnic cleansing." In Baghdad sectarian violence had continued to drive people from their homes throughout the surge buildup in the first half of 2007. Areas controlled by Shia expanded in the north of the city, while Sunnis, who were mostly on the losing side, consolidated in the south.46
The sharply redrawn sectarian boundaries that were the consequence of ethnic cleansing created more "defensible space" for both communities, while far fewer vulnerable mixed neighbourhoods meant that there was less territory to fight about. This, plus the heightened local security provided by the US, increased the costs of sectarian violence while reducing its benefits, which in turn pushed down the civilian death toll.
The fourth major change in the Iraqi security environment was the announcement in April 2007 by Shiite militia leader Moqtada al-Sadr that the Mahdi Army, his powerful but deeply factionalized militia, would observe a unilateral ceasefire.47 In mid-November 2007 the US military reported that the Mahdi Army’s ceasefire had been "a significant factor behind the recent drop in attacks in Baghdad."48 As this Brief went to press there was major fighting ongoing between the Shia-led government forces and Mahdi Army militias. This will not necessarily have caused an increase in fatalities from terrorism, however. Combat fatalities (including civilians inadvertently caught in the crossfire) are not counted as terrorism by any of the datasets under review.
Finally, and of critical importance for understanding the challenges that Islamist terror organizations confront elsewhere in the Muslim world, was the surprising alliance formed between the US military and its former Sunni insurgent enemies against the Islamist terrorists of al-Qaeda In Iraq.
The Failure of al-Qaeda In Iraq—a Global Defeat for Islamist Terrorism
In July 2005 Ayman al-Zawahiri, al-Qaeda’s main strategist and number two to Osama bin Laden, described Iraq as the location of "the greatest battle of Islam in this era."49 This battle was being fought on Iraqi soil by a foreign-led group of Sunni militants who had become known as al-Qaeda In Iraq (AQI).
In 2005 and early 2006, AQI was pursuing a nationwide terror campaign against Iraq’s "apostate" Shia community. AQI suicide attacks against Shia mosques and other civilian targets were intended to provoke Shia revenge attacks against Sunni communities that would lead to a Sunni-Shia civil war. The resulting turmoil would, it was believed, precipitate the withdrawal of the US and its allies.
The good news about declining deaths in Iraq was due in part to the bad news about "ethnic cleansing."
However, these provocations, plus the militants’ efforts to impose their extremist ideology on the local Sunni populace in al-Anbar province and elsewhere, and their savage attacks on anyone who challenged them, had generated growing Sunni anger, not just in al-Anbar but throughout Iraq.50
In September 2006 a nationwide opinion poll revealed that the terror tactics of AQI were rejected by large majorities of Sunnis, as well as overwhelming majorities of Shia and Kurds.51A year later, anti-al-Qaeda sentiments in Iraq had grown even more intense. An ABC News/BBC/NHK poll revealed that 100 percent of those surveyed—Sunni and Shia alike—found AQI attacks on Iraqi civilians "unacceptable"; 98 percent rejected the militants’ attempts to gain control over areas in which they operated; and 97 percent opposed their attempts to recruit foreign fighters and bring them to Iraq.52H U M A N S E C U R I T Y B R I E F 2 0 0 7 16 Despite a huge surge of new research since al-Qaeda’s 9-11 strikes on the US in 2001, remarkably few studies have analyzed why the overwhelming majority of terrorist campaigns, non-Islamist as well as those associated with al-Qaeda and its affi liates, fail to achieve their strategic objectives.53 Most research has focused on why terror campaigns start, rather than why they end. There are many possible explanations for the failure of terror campaigns. Prominent among them, as Audrey Kurth Cronin has argued, are doctrinal infi ghting, lack of effective operational control, and lack of unity—all very evident in the case of al-Qaeda and its affi liates.54 However, the historical evidence also suggests that terror campaigns that lose public support will eventually be abandoned, even if the terrorists themselves remain undefeated.55 As Cronin puts it, "Terrorist groups generally cannot survive without either active or passive support from the surrounding population."56 The recent history of terrorism in Europe is an instructive case in point. In the 1960s and 1970s, and through into the 1980s in some cases, there was an upsurge of urban guerrilla/terrorist activity in Europe—the Baader Meinhof Gang/Red Army Faction launched attacks in Germany, the Red Brigades in Italy, Action Direct in France, the Fighting Communist Cells in Belgium, the Revolutionary People’s Struggle in Greece, and the Angry Brigade in the UK. However, the radical neo-Marxist political agendas of these small essentially middle-class organizations, like the maximalist goals of Islamist groups, had zero appeal to the citizens that the radicals hoped to mobilize. Insofar as the militants had any coherent strategy, it was to use violence to provoke indiscriminate state repression, which they hoped would in turn radicalize their potential support base. But, like the Islamist terror campaigns in Egypt and Algeria, and al-Qaeda Iraq, the violence of the neo-Marxist groups succeeded only in alienating them still further from society, while catalyzing—and creating public support for—tough offi cial antiterror policies. Only a small percentage of the active members of these organizations was ever captured, killed, or imprisoned. The rest simply gave up on strategies that—as individual members of these organizations increasingly came to realize—had no chance of succeeding, while putting them at great personal risk. By contrast, a number of terror campaigns employed by national liberation movements against colonial powers– –against the British in Cyprus and Yemen and the French in Algeria, for example—achieved real success. However, here the strategic circumstances were completely different. In an era when anticolonialist sentiments were growing rapidly in both the developed and developing world, the nationalist rebels, unlike Europe’s neo-Marxist radicals or today’s jihadi terrorists, had widespread popular support. In such a strategic context it is not surprising that terrorism proved to be an effective tactic. The anticolonial nationalists had time—and history—on their side. These successes have few parallels in the current era, however. Today’s terrorists are not fi ghting European powers with few vested interests in clinging to an outmoded colonial system.57 They are confronting incumbent national governments that have an existential interest in avoiding defeat.58 Since the armed forces of these governments are almost always far larger, as well as better armed and trained than are the terrorists, it is not surprising that the latter so rarely prevail. Just how infrequently terrorist organizations achieve their goals in the current era was revealed in a rare quantitative study published in International Security in 2006. In an analysis of the successes and failures of 28 terrorist organizations in 42 campaigns over a fi ve-year period, Max Abrahms found that terrorists failed to achieve their stated policy goals in 93 percent of cases—a remarkably high failure rate.59 Although the defi ning characteristic of terrorism is the use of political violence against civilians, Abrahms noted that terror groups that mostly attacked civilians had a success rate of zero. In the rare cases where terrorism succeeded, the militants had limited policy objectives and attacked military targets more than they did civilians.60 A case in point is the suicide bombing of the US Marines’ barracks in Lebanon in October 1983 that left some 300 US Marines and French paratroopers dead. The terrorists’ objective was limited—to achieve the withdrawal of a small number of foreign troops from Lebanon—and the target was military.61 US and French forces pulled out of the country early in 1984. This case is misleadingly cited by Osama bin Laden and others as evidence that terrorism succeeds. WHY MOST TERRORIST ORGANIZATIONS FAIL While many believe that terrorism is an effective tactic for achieving political objectives, the evidence suggests otherwise.H U M A N S E C U R I T Y B R I E F 2 0 0 7 17
The growing revulsion felt towards AQI was to become a major strategic liability for the militants, not least because it paved the way for the unprecedented US-Sunni security collaboration that had gathered pace in the last half of 2006 and accelerated through 2007. Sunni insurgents, who had previously been killing Americans, were now working alongside them in a campaign to hunt down and kill their former AQI allies.
Throughout 2007, as part of the extraordinary process that had become known as the "Sunni Awakening," tens of thousands of mostly young Sunni men, many of them former insurgents, flocked to join anti-al-Qaeda "concerned local citizens" militia groups—large numbers of which are armed and funded by the US.
The Islamists’ failure in Iraq is neither accidental nor unique.
The US military’s new Sunni allies—there are now more than 90,000 of them—provided priceless intelligence on the identity and location of AQI fighters with some of the most valuable information coming from AQI defectors who had joined the new militias.62 Previously, US counterinsurgency operations had lacked reliable information on who—and where—the militants were. The predictable result was that many innocent civilians were arrested without good reason and interned—or were killed or injured in offensive sweep operations by Coalition forces. Unsurprisingly, this increased popular hostility towards the occupation, while generating more volunteers for the insurgency.
In the late summer of 2007, the combined efforts of the concerned local citizens groups and US forces had dealt a series of crushing blows to AQI in most of its urban strongholds in the country—a dramatic reversal of the terror group’s fortunes in a relatively short period of time. The new US-Sunni alliance was also an important factor in the nationwide decline in civilian—and thus terrorist—fatalities as counted by NCTC.63
By November 2007 it had become evident that an equally remarkable—though much less widespread or publicized—movement was underway to create Shia "concerned citizens" auxiliary police forces. According to the US military, some 15,000 volunteers had joined 24 all-Shiite groups, while a further 18 mixed Sunni/Shia groups had also been formed.64 The protection offered by both Sunni and Shia concerned citizens militias to the local communities in which they operated was yet another factor driving the civilian death toll down in 2007.
At the end of December 2007, General Abdul Kareem Khalaf, a spokesman for the Iraqi Interior Ministry, claimed that 75 percent of AQI’s networks and safe havens had been destroyed.65 AQI activity was now concentrated in Iraq’s north where the organization was under growing pressure from US forces.
While the decline in civilian casualties in Iraq has been widely welcomed, the security situation in the country is far from stable. The alliance between the US military and former Sunni insurgents in al-Anbar and elsewhere is not a collaboration based on shared values. It was, and remains, an initiative based on common opposition to a common foe—"my enemy’s enemy is my friend."66
If AQI is completely crushed, the rationale for the US-Sunni cooperation disappears. There are real concerns in Washington that, should this happen, the former insurgents, now re-armed and trained by the US, will again turn their guns against the Americans. The Shia-dominated Iraqi government, on the other hand, worries that the 90,000-plus US-armed and trained Sunni militia is undermining a sectarian balance of power that has come to favour the Shia majority.
While Iraq’s security future remains uncertain in many respects, by early 2008 one thing was very clear: AQI, while far from being completely crushed, had suffered a stunning defeat—politically as well as militarily. Hated by both the Shia and Kurdish communities and having deeply alienated its former Sunni allies, there appeared little prospect that Osama bin Laden’s Iraqi affiliate would be able to make a comeback.
The Islamists’ failure in Iraq is neither accidental nor unique. Throughout the Muslim world there have been similar reactions against the extremist ideology and the indiscriminate violence that have become one of the hallmarks of Islamist campaigns.
The Sources of Islamist Political Failure
AQI’s failure in Iraq parallels earlier failures of violent Islamist movements in the Muslim world—notably in Egypt and Algeria. In all three cases, growing revulsion at the policies and the indiscriminate violence of the militants generated a popular backlash and effective campaigns of often ruthless official repression.67 Similar negative reactions to Islamist political agendas are now evident throughout the Muslim world. Indeed, evidence that large and growing majorities of Muslims reject the Islamists’ harsh and repressive ideology is overwhelming.H U M A N S E C U R I T Y B R I E F 2 0 0 7 18
Most Muslims (79 percent according to Gallup) share the militants’ belief in the importance of sharia law.68 But the way this belief is interpreted by mainstream believers is dramatically different from the extremist policies and practices that al-Qaeda and other Islamist groups seek to impose wherever they have the opportunity.69
It is also clear from opinion polls that most Muslims embrace a wide range of other views that the extremists reject. A major Gallup poll of 10 Muslim countries conducted between August and October 2005 found that, notwithstanding the strongly anti-American sentiments that hold sway in much of the Islamic world, there was also widespread support for the very liberal values that the Islamists reject:
The vast majority of those surveyed support freedoms of speech, religion and assembly—as well as a woman’s right to vote, drive and work outside the home. The majority of opinion in every nation surveyed, save Saudi Arabia, also believes it is appropriate for women to serve at the highest levels of government ... A mean of 60% in the ten countries said they would want religious leaders to play no direct role in drafting a country’s constitution.70
Similarly, an ABC News/BBC World Service poll taken in Afghanistan and released in December 2007 found that large majorities of Afghans, men as well as women, supported women’s rights to be educated, to vote, to work outside the home, and to hold government office.71
The liberal values revealed by the Gallup and ABC/BBC polls are completely antithetical to those of the Taliban, al-Qaeda, and other Islamist extremists—a fact that has clear strategic implications. As the 2006 US National Intelligence Estimate put it: "The jihadis greatest vulnerability is that their ultimate political solution—an ultra-conservative interpretation of sharia-based governance spanning the Muslim world—is unpopular with the vast majority of Muslims."72
The reality is that al-Qaeda’s extraordinarily harsh pan-Islamist ideology and the policies that are associated with it appeal to only a tiny—and shrinking—minority of Muslims around the world. And the more the Islamists attempt to impose it, the more rapidly they lose support.
Muslims around the world are not only deeply opposed to the Islamists’ ideology and policies, they also strongly reject their use of suicide attacks and other deadly assaults on civilians. This is not surprising—the majority of victims of jihadi/Islamist violence have been fellow Muslims.
A Pew poll released in July 2007 revealed that "in Lebanon, Bangladesh, Pakistan and Indonesia, the proportion of Muslims who view suicide bombing and other attacks against civilians as being often or sometimes justified has declined by half or more over the past five years. Wide majorities say such attacks are, at most, rarely acceptable."73
A December 2007 poll conducted in Saudi Arabia revealed that Osama bin Laden’s fellow countrymen had "dramatically turned against him, his organization of al-Qaeda, Saudi fighters in Iraq, and terrorism itself."74
In Afghanistan, where the government is locked in a confrontation with a resurgent Taliban supported by foreign jihadis, MIPT found that terror attacks on civilians increased from 28 in the first quarter of 2005, to 123 in the second quarter of 2006.75 However, as the level of violence rose, so too did popular antipathy towards the Taliban and their foreign jihadi allies.
The ABC News/BBC poll referred to earlier found that by late 2007, just 1 percent of Afghans expressed "strong support" for the presence of the Taliban and jihadi fighters in the country.76
In Pakistan—a country widely believed to be harbouring Osama bin Laden and his deputy al-Zawahiri and providing a home base for al-Qaeda and Taliban militants—the trend against the Islamists is even more pronounced.
As the level of violence rose, so too did popular antipathy towards the Taliban and their foreign jihadi allies.
Figure 1.4 reveals that the percentage of Pakistanis believing that acts of terrorist violence against civilians are "never justified" rose from 35 percent in 2004, to 69 percent in 2006.77 The small minority supporting terrorism shrank significantly over the same period. It is no accident that this decline in support coincided with a tenfold increase in terrorism over the same period—from five terrorist attacks in the first quarter of 2004, to 50 in the last quarter of 2006.78 As the attacks increased, opposition to them almost doubled.
In August 2007, 33 percent of Pakistanis supported al-Qaeda; 38 percent supported the Taliban. By January 2008 al-Qaeda’s support had dropped to 18 percent; the Taliban’s to 19 percent.79 When asked if they would vote for al-Qaeda, just 1 percent of Pakistanis polled answered in the affirmative. The Taliban had the support of 3 percent of those polled.80H U M A N S E C U R I T Y B R I E F 2 0 0 7 19
In the North-West Frontier Province, where al-Qaeda has a strong presence, the percentage of those with a favourable opinion of Osama bin Laden had dropped from 70 percent in August 2007, to just 4 percent in January 2008—an extraordinary decline over such a short period.81 Bin Laden’s support level halved nationwide over the same period.
The reason for this sea change in public opinion in Pakistan, according to Terror Free Tomorrow, was "increased terrorist attacks by al-Qaeda and the Taliban, and the assassination of former Prime Minister Benazir Bhutto."82
Further evidence of the decline in support for the Islamists comes from the 2008 election results, where Islamist parties gained just 2 percent of the national vote, a fivefold decline from the level of support that they had enjoyed in 2002.
The Islamists’ minimal popular support and their lack of conventional military capability means that any attempt to mount a popular armed uprising in Pakistan would be doomed to failure. Their lack of broad-based support within the military precludes a successful military coup.83 The Islamists have a strong presence in the northwest of the country, but no way of leveraging it to achieve state power.
This pattern—the lack of Islamist conventional military capacity, an absence of broad-based support within the military, and minimal popular support—is evident in all other countries where Islamist terror groups are active.
Islamist Terrorism: What the Statistics Tell Us
Case studies can inform us about Islamist terror campaigns in particular countries, while public opinion surveys can reveal levels of support for Islamist ideology and tactics in the Muslim world. But neither can tell us whether or not the incidence of Islamist terrorism is increasing or decreasing—which is perhaps the most important objective measure of the threat. For this we need to turn again to the datasets.
Given the intense concern that Islamist terrorism generates around the world, there is surprisingly little accessible long-term statistical data on its scope or incidence. One notable exception is found in the March 2007 study by Peter Bergen and Paul Cruikshank that extracted data on the incidence of jihadi terrorism from the MIPT dataset.85
The authors compared MIPT’s jihadi attack and fatality numbers around the world before and after the invasion of Iraq in 2003. Their findings provided strong support for the expert consensus that the threat of Islamist terrorism is increasing.
The average annual global fatality toll from jihadi terrorism for the postinvasion period (March 2003 to September 2006) was 237 percent higher than in the pre-invasion period (September 2001 to March 2003). The average yearly total of jihadi attacks increased by more than 600 percent.86
Figure 1.5 clearly shows the sharp increase in the fatality rate from jihadi attacks following the invasion of Iraq. It also suggests that the wars in Iraq, and to a much lesser degree Afghanistan, have been the major drivers of the worldwide increase in jihadi attacks.
The Bergen/Cruikshank study reveals a major increase in the average number of jihadi fatalities for the pre- and post-
Figure 1.4 Pakistan: As Terrorist Attacks Increase,So Does Opposition to Terrorism, 2004-20062004Q12004Q22004Q32004Q42005Q12005Q22005Q32005Q42006Q12006Q22006Q32006Q49080706050403020100Number of Terrorist IncidentsOpposition to Terrorism (Percent of Survey Respondents)Incidents and OppositionYear
Data Sources: MIPT; PEW Global Attitudes Project.84
As terrorist incidents increase, so does opposition to terrorism.
Figure 1.5 Average Number of FatalitiesFrom Jihadi Terrorism per Year, Before and After the Invasion of Iraq1,8001,6001,4001,2001,0008006004002000Fatalities worldwideFatalities worldwide,excluding IraqFatalities worldwide,excluding Iraqand AfghanistanBefore Invasion After InvasionNumber of Fatalities
Data Source: Peter Bergen and Paul Cruickshank.87
The global death toll from Islamist terrorism increases dramatically after the invasion of Iraq.H U M A N S E C U R I T Y B R I E F 2 0 0 7 20
war periods. But it does not tell us what happens within those periods, and it only extends to September 2006.
NCTC has data from the beginning of 2005 to September 2007 and its data can also be disaggregated to track fatalities perpetrated by different terror groups, including "Sunni Islamic extremists," a category that includes al-Qaeda and its affiliates around the world, and is largely synonymous with Islamist terrorism.
As Figure 1.6 shows, Islamist terrorism was indeed increasing around the world from 2005. But in July 2007 things began to change. Between July and the end of September, the global fatality toll from Islamist terrorism halved—going from 727 deaths in July, to 385 in September. Much of this decline was driven by the even steeper decline in Islamist terrorism fatalities in Iraq; these dropped by 73 percent from August to September.
Global fatalities from Islamist terrorism over this period declined less than those in Iraq primarily because fatalities in Afghanistan increased while fatalities in Iraq decreased.88
Given recent events in Iraq in late 2007—especially the major setbacks experienced by AQI—there is little reason to assume that the decline in Islamist terrorism in that country that started in September 2007 will be reversed.
NCTC is not the only research institution whose statistics reveal a decline in Islamist terror attacks and fatalities.89 The Intelcenter, a US think-tank based in Alexandria, Virginia, that focuses on Islamist terrorism, recently examined "the 63 most significant attacks executed by al-Qaeda, regional arms and affiliate groups over the past nearly 10 years."90 These include the attacks in Bali, London, Madrid, Amman, and Jakarta, as well as the September 11 attacks in the US—i.e., those most associated with Islamist terror. The survey did not include jihadi/Islamist attacks in Iraq and Afghanistan and other "insurgency theatres."
Intelcenter found that by August 2007, the number of Islamist attacks and fatalities, and the average number of fatalities per attack, had all declined from a high point in 2004. Attacks were down by 65 percent—from 20 to seven. Fatalities decreased by 92 percent, from 739 in 2004, to 56 in August 2007.91 The average number of individuals killed per attack went from 67 to six over the same period.
Intelcenter’s data, while not as current as those of NCTC, also present a picture that is sharply at odds with the consensus view that the Islamist terror threat is increasing.
Conclusion
This chapter has argued that there is little objective evidence to support the claim that the threat of terrorism is increasing around the world—at least as measured by fatalities from terrorist attacks. It has shown that the big increases in the global terrorist toll to the end of 2006 that were recorded by NCTC and MIPT were the result of counting a large percentage of the civilian deaths from insurgent and sectarian violence in postinvasion Iraq as "terrorism." We have argued that there are defensible grounds for rejecting this counting approach.
If the Iraq fatalities are removed from the global terrorism data, there is no evidence of any substantial increase in the fatality toll since data on both domestic and international terrorism began to be collected in 1998. Indeed, the two datasets that have statistics going back to 1998 both reveal a decline in deaths from terrorism since 2001.
There is little evidence that the threat of terrorism is increasing.
However, even if we accept that it is appropriate to count civilian deaths from political violence in civil wars as terrorism, the latest statistics from NCTC—the only dataset that has usable data for 2007—still show a decline in the global death toll from terrorism. We have shown that the extraordinary 46 percent drop in fatalities worldwide from all forms of terrorism that NCTC’s data reveal for the period of March to September 2007 was driven almost entirely by the 61 percent decline in deadly assaults on civilians in Iraq.
Figure 1.6 Global Fatalities from IslamistTerrorism vs Fatalities from Islamist Terrorism in Iraq, January 2005 to September 2007Jan-05Apr-05Jul-05Oct-05Jan-06Apr-06Jul-06Oct-06Jan-07Apr-07Jul-079008007006005004003002001000GlobalIraqNumber of FatalitiesYear
Data Source: NCTC.
Once again global fatalities from terrorism are driven by fatalities in Iraq.H U M A N S E C U R I T Y B R I E F 2 0 0 7 21 The change described here is one of a net—not uniform— decline. As terror attacks have declined in Iraq and elsewhere, they have increased in Afghanistan and Pakistan. In the West, as we noted earlier, the main focus of concern has not been local terror groups fi ghting over local issues, but the transnational Islamist terror organizations that are the central target of the $140-billion-a-year global war on terror.92 The GWOT, as it is often called, has had some signifi cant tactical successes. Al-Qaeda’s global terror campaign has been disrupted by an assault on its fi nancial networks, by the loss of its sanctuaries in Afghanistan, and by the death or capture of individuals in key leadership positions. In the Muslim world, however, Washington’s antiterror efforts have been widely interpreted as being directed against Islam. This fact, plus the intense and widespread Muslim opposition to the US invasion and occupation of Iraq, undoubtedly helped swell support for the Islamist cause—offsetting many of the initial tactical gains. However, any inital support the Islamists secured by capitalizing on the widespread anti-American sentiment in the Muslim world has been largely negated by their violent attempts to impose their harsh ideology and policies on their coreligionists. The response has been widespread public revulsion and a dramatic decline in popular support for the terrorists. The evidence suggests that in 2008 the international community confronts a terrorist threat that is both serious and far from being eliminated, but that is in no sense comparable to the dangers posed by Fascism as some alarmist commentators have claimed. The threats to individuals that al-Qaeda and its affl iates pose are real, but they need to be put in perspective. Like organized crime, terrorism can kill individual citizens, but its perpetrators lack both the public support and the military capacity needed to defeat governments. In the long term, perhaps sooner, Islamist terror organizations will join the overwhelming majority of other terrorist groups that have failed to achieve their objectives—from the anarchists of the nineteenth century, to the neo-Marxists of the twentieth. Their members will be killed, captured, or—most likely—will simply abandon a struggle that lacks popular support and that history suggests is doomed to failure. The failure of Islamist terror groups in the Middle East and North Africa to prevail either militarily or politically has been associated with a remarkable, but little-noticed, shift in grassroots strategies to effect political change in the region. A new study from the Center for International Development and Confl ict Management at the University of Maryland has revealed a sharp decline over the past quarter of a century in the percentage of organizations pursuing violent strategies—including terrorism—to effect political change across the Middle East and North Africa.93 The researchers used the Minorities at Risk database to examine violent versus nonviolent strategies employed by 102 political organizations that represent 29 different ethnic/ national/religious groups throughout the Middle East and North Africa. The researchers found that the percentage of organizations employing violent political strategies dropped almost fourfold between 1985 and 2004—from 54 percent to 14 percent.94 This extraordinary decline is associated with the defeat of violent political movements in Egypt, Algeria, and elsewhere. The percentage of organizations using nonviolent protest politics increased more than threefold from 1985 to 2004, while the number using electoral politics more than doubled in the same period.95 This study provides further support for the thesis that there has been a long-term shift—albeit with signifi cant reverses from time to time—away from reliance on terrorist tactics to effect political change in the Middle East and North Africa. This shift is associated with, and indeed part of, the uneven decline in armed confl ict numbers in this region since the early 1980s. VIOLENT STRATEGIES LOSING FAVOUR IN THE MIDDLE EAST AND NORTH AFRICA For almost a quarter of a century a profound, but little-noticed, change has been underway in the Middle East and North Africa away from reliance on violent strategies to effect political change.H U M A N S E C U R I T Y B R I E F 2 0 0 7 22 Simon Maina / AFP / Getty Images. KENYA. Towards A New Peace in Africa? Recent news from sub-Saharan Africa has not been good. Since the end of 2007, spiralling intercommunal violence in Kenya has killed more than 1,000 people and displaced well over a quarter of a million. Somalia, still without a functioning government, has become the battleground of a bitter low-level proxy war between Eritrea and Ethiopia. The growing violence in Darfur has spilled over to envelop neighbouring Chad, and the Central African Republic, while in southern Sudan, the 2005 peace agreement that stopped a civil war that has cost two million lives is in grave risk of breaking down. In the Democratic Republic of the Congo (DRC), elevated levels of disease and malnutrition caused by almost a decade of political violence have been killing an average 40,000-plus people a month—half of them chil dren—since 2003.96 The current horror stories are real enough, but behind the headlines is another very different and far less depressing reality—one that gets little media coverage. Not withstand ing the current violence, sub-Saharan Africa is dramatically more secure than it was less than 10 years ago. Twenty-three of the region’s states, some half of the total, were embroiled in state-based confl ict at sometime during the 1990s—a decade that saw confl icts erupting across the continent at double the rate of the 1980s. However, in the new millen nium something remarkable happened. The Decline in Armed Confl icts Between 1999 and the end of 2006, the number of state-based armed confl icts in sub-Saharan Africa—those in which a gov ernment is one of the warring parties—declined sharply, as Figure 2.1 indicates. In 1999 there were 16 such confl icts in the region, the highest number since 1946; in 2006 there were just seven—a drop of 56 percent. The overwhelming majority of these confl icts were fought within, not between, states—a pattern that has been fairly constant throughout the world for more than 30 years. This has been a net decline, of course. C H A P T E R 2 Figure 2.1 Number of State-Based Armed Confl icts in Sub-Saharan Africa, 1946-2006 Data Sources: UCDP/PRIO and UCDP/Human Security Report Project Dataset.97 In sub-Saharan Africa confl ict numbers plunged in the new millennium.H U M A N S E C U R I T Y B R I E F 2 0 0 7 23
Conflicts have continued to break out since 2000, but at some 60 percent of the rate of the 1990s. And more importantly, conflicts are now ending at more than twice the rate of the 1960s, 1970s, and 1980s.
In 2006 the centre of gravity of political violence in sub-Saharan Africa was the arc of instability in northeast and Central Africa with conflicts in Burundi, the Central African Republic, Chad, Somalia, Uganda, and Ethiopia. But while a source of great human suffering, these conflicts are far less deadly than those of the late 1990s. (Although the war in Darfur is directly linked with the conflicts in Chad and the Central African Republic, it is not included in the sub-Saharan Africa conflict totals because Sudan is part of the Middle East and North African region.)
Between 1999 and 2006, most of the major conflicts in West Africa (Sierra Leone, Liberia, Cote d’Ivoire) and Central Africa (notably Angola and the DRC), as well as the war between Ethiopia and Eritrea, either came to an end, or the violence sharply de-escalated leading to major reductions in fatality tolls.
As Figure 2.2 shows, the battle-death toll from state-based conflicts in sub-Saharan Africa increased more than tenfold between 1961 and 1986—reflecting in part the increase in conflict numbers. It then declined sharply to the mid-1990s, before rising dramatically again from 1997 to 1999. The estimated fatality toll in 1999 was in the vicinity of 99,000, the highest in the region since the end of World War II. Just two wars accounted for more than three-quarters of the battle-deaths that year. An estimated 48,000 people were killed in the DRC and 30,000 in the war between Ethiopia and Eritrea. The tolls in Angola and Sierra Leone were 10,000 and 7,000, respectively.
Conflicts have continued to break out since 2000, but at some 60 percent of the rate of the 1990s.
Then things changed. By 2001 the Ethiopia/Eritrea and Sierra Leone conflicts were over. By 2002 the major fighting was over in Angola and the DRC, and the fatality count for the region as a whole had shrunk to less than 5,000. The level of violence continued to drop, and by 2005 the fatality estimate for sub-Saharan Africa was just 1, 851, the lowest in 45 years and less than 2 percent of the 1999 fatality count.
Non-State Conflicts: A Long-Ignored Category of Political Violence
Most quantitative analyses of Africa’s wars focus on state-based conflicts. They do not include non-state conflicts, those that are fought between various armed groups—rebels fighting rebels, warlords fighting warlords, and between different ethnic and religious communities. These conflicts, whose distinguishing feature is that none of the warring parties is a government, are completely ignored in most of the major conflict databases.99 This is a serious omission since in some years there are more non-state conflicts being fought around the world than state-based conflicts.
Until recently there was no comprehensive annually updated dataset on non-state conflicts. However, in 2004 the Uppsala Conflict Data Program (UCDP) created a new dataset for the Human Security Report Project. The trend data on sub-Saharan Africa’s non-state conflicts, shown in Figure 2.3, are instructive.
Two things are apparent from Figure 2.3. First, there were two or more times as many of the rarely reported non-state conflicts in sub-Saharan Africa as state-based conflicts in every year save 2006. Second, the number of both non-state and state-based conflicts declined throughout the period: the former by 54 percent, the latter by 46 percent. The
Figure 2.2 Number of Reported Battle-Deathsfrom State-Based Armed Conflictin Sub-Saharan Africa, 1946-2006121086420Number of Deaths (tens of thousands)Year1961195119561946196619711976198119861991199620012006
Data Sources: Lacina & Gleditsch andUCDP/Human Security Report Project Dataset.98
Just two wars—those in the Democratic Republic of the Congo and between Ethiopia and Eritrea— drove most of the dramatic increase in fatalities from 1997 to 1999.H U M A N S E C U R I T Y B R I E F 2 0 0 7 24
overall decline was from 39 conflicts (state-based and non-state) in 2002, to 19 in 2006.
When we look at the non-state conflict battle-death tolls we find a similar pattern of decline. As Figure 2.4 indicates, there were some 4,600 non-state battle-deaths in 2002; in 2006 there were just over 1,300—a decline of some 70 percent. The combined fatality toll from state-based and non-state conflict dropped by almost two-thirds from 2002 to 2006.
Organized Violence against Civilians
A third category of political violence focuses on deadly assaults on civilians—by both governments and armed non-state actors. UCDP calls this phenomenon "one-sided violence." The rationale for having this separate category is that killing defenseless civilians is fundamentally different from armed conflict and should therefore be treated as such. This is already standard practice with genocides—an extreme form of one-sided violence.100 For a campaign of one-sided violence to be recorded, 25 or more civilians must be killed by a government or organized non-state group within a calendar year.
Most, but not all, campaigns of violence against civilians take place in the context of civil wars. This was certainly the case with the world’s worst single case of one-sided violence since World War II—the genocide in Rwanda in 1994, when an estimated 500, 000 to 800,000 civilians—Tutsis and moderate Hutus—were slaughtered in the space of little more than three months.101 If the higher estimate in Rwanda is correct, then the death toll is greater than the total number of people killed in all the wars being waged around the world in 1950—the deadliest year for battle-deaths since the end of World War II.
Since 1994 there has been nothing remotely as horrific as the Rwandan genocide in sub-Saharan Africa—or indeed anywhere else in the world. The recent violence in Kenya, while garnering a huge amount of media attention, has accounted for little more than a tiny fraction of the Rwandan death toll. However, campaigns of one-sided violence, by governments as well as rebels, continue to kill the innocent across sub-Saharan Africa. After fluctuating in the 1990s, the number of campaigns of one-sided violence peaked in 2002. But as Figure 2.5 reveals, this was followed by a sharp though uneven decline: between 2002 and 2006 the number of campaigns of one-sided violence dropped by 67 percent. The steep decrease in organized violence against civilians in sub-Saharan Africa between 2002 and 2006 parallels similar declines in both state-based and non-state conflicts over the same period.
Like the fatality count from state-based and non-state conflicts, the death toll from organized violence against civilians also declined sharply from 2002 to 2006. Given the big decline in the number of campaigns of one-sided violence in this period this is not surprising. Figure 2.6 shows the extent of the decline in fatalities from one-sided violence: The death toll in 2006 was just over one-sixth of that in 2002. As noted in Chapter 4, the one-sided violence fatality statistics are the most prone to error. While we are confident that
Figure 2.3 Number of State-Based and Non-State Armed Conflicts in Sub-Saharan Africa, 2002-200645403530252015105020022003200420052006State-Based ConflictsNon-State ConflictsNumber of ConflictsYear
Data Source: UCDP/Human Security Report Project Dataset.
The number of non-state conflicts in sub-Saharan Africa dropped by more than 50 percent between 2002 and 2006.
Figure 2.4 Number of Reported Battle-Deaths from State-Based and Non-State Armed Conflict in Sub-Saharan Africa, 2002-200610,0009,0008,0007,0006,0005,0004,0003,0002,0001,000020022003200420052006State-Based ConflictNon-State ConflictNumber of DeathsYear
Data Source: UCDP/Human Security Report Project Dataset.
Battle-deaths from non-state conflict declined by 70 percent between 2002 and 2006.H U M A N S E C U R I T Y B R I E F 2 0 0 7 25
the downward trend from 2002 to 2006 is correct, the absolute numbers for any particular year may be subject to quite a wide margin of error.
What about Indirect Deaths?
Battle-deaths and deaths from one-sided violence are only a relatively small part of the human cost of Africa’s wars. The number of "indirect deaths"—fatalities caused by war-exacerbated disease and malnutrition—is many times greater than the number of deaths that occur as a direct consequence of violence in most poor-country wars.
There is very little in the way of reliable statistics on indirect deaths for sub-Saharan Africa—or anywhere else in the world for that matter. Indeed, there is only one country, the DRC, for which there are nationwide estimates of indirect deaths over time. These estimates derive from a series of mortality surveys undertaken by the International Rescue Committee (IRC) since 1999.
The most recent IRC survey, released in January 2008, estimates that there have been some 5.4 million "excess" or indirect deaths in the DRC since 1998.102 The overwhelming majority of these deaths were from war-exacerbated disease and malnutrition. This extraordinary figure raises an obvious question. Since we do not have comparable data for any other sub-Saharan African country, how do we know that indirect deaths have not been rising in the region while violent deaths have been declining?
The short answer is that we cannot be absolutely sure, but that it is unlikely for a number of reasons.103 The key drivers of indirect deaths are the intensity and scope of the violence, the number of displaced people, the state of local health services before and after the conflict, and access to humanitarian assistance. We know that since 1999:
° There has been a major decline in the scope and intensity of conflicts.
° Refugee numbers have shrunk substantially.
° The share of global humanitarian assistance going to Africa doubled between 1999 and 2006—from 23 percent to 46 percent.104
The combined effect of these factors suggests that indirect deaths in the region have very likely declined, along with conflict and fatality numbers, and numbers of campaigns of one-sided violence and their resulting fatality tolls since 1999.
The next Human Security Report will examine the "hidden costs of war," focusing in particular on the factors that drive indirect deaths.
Stops and Starts: Explaining Sub-Saharan Africa Conflict Trends
The security situation in sub-Saharan Africa in the 1990s was extraordinarily volatile. During this decade, there were twice as many state-based conflict onsets—including old conflicts that had restarted—as in the 1980s, but there was also an even larger increase in the number of conflicts that ended.105
Understanding trends in onsets and terminations is important for policy-makers. A net increase in conflict numbers, for example, could be due to more onsets or to
Figure 2.5 Number of Campaigns of One-sidedViolence in Sub-Saharan Africa, 2002-2006161412108642020022003200420052006Number of CampaignsYear
Data Source: UCDP/Human Security Report Project Dataset.
Between 2002 and 2006 the number of campaigns of organized violence against civilians fell by two-thirds.
Figure 2.6 Number of Reported Deaths from One-Sided Violence in Sub-Saharan Africa, 2002-20063,5003,0002,5002,0001,5001,000500020022003200420052006Number of DeathsYear
Data Source: UCDP/Human Security Report Project Dataset.
Fatalities from one-sided violence declined by more than 80 percent from 2002 to 2006.H U M A N S E C U R I T Y B R I E F 2 0 0 7 26
fewer terminations. The first would suggest that conflict prevention policies were having little effect; the second that efforts directed at stopping conflicts—"peacemaking" in UN-speak—were ineffective.
As Figure 2.7 shows, in the 1990s the average number of conflicts starting each year was twice that of the previous decade. This unprecedented increase suggests that any conflict prevention efforts that were being tried in this period were having a negligible impact. However, the average number of terminations per year was more than twice that of the 1980s and a growing percentage of these terminations were negotiated settlements, which suggests that peacemaking efforts were meeting with increasing success.
While this latter development is encouraging, the fact that in the new millennium the average number of conflict onsets per year remains higher than in every decade since World War II, save the 1990s, is a source of continuing concern. It reflects the ongoing political instability in the region and the continuing failure of conflict prevention policies to have much impact.
Why the Sharp Increase in Conflict Numbers in the 1990s?
The increase in new state-based conflicts in sub-Saharan Africa in the 1990s was not unique to the region and was clearly related to the end of the Cold War. Regimes and rebel groups that had long been propped up by the assistance given by one or the other of the two superpowers suddenly found that this support—political as well as economic—had disappeared. The result in many cases, not just in sub-Saharan Africa, was regime change and ongoing political instability.
However, in sub-Saharan Africa something else was happening: The countries of the region, to a greater degree than in other parts of the world, were undergoing profound and wrenching political change. In 1988 nearly 90 percent of sub-Saharan African states had autocratic governments. By 2006 there were just two autocracies in the region, while the number of democracies had increased sixfold, from three to 18.
Had the only change been a decrease in autocracies and an increase in democracies, it would likely have enhanced regional security, since democracies tend to experience fewer armed conflicts than do autocracies. But these were not the only changes.
Figure 2.8 uses data from the Polity IV Project at the Center for Systemic Peace in Virginia. This dataset tracks not only trends in the number of autocracies and democracies but also trends in "anocracies"—a third regime type, one that is neither fully democratic nor fully autocratic, but a mix of both systems.
The increase in the number of anocracies in sub-Saharan Africa between 1988 and 2000 is startling—far greater than in any other region of the world. In 1988 there were two anocracies and 37 autocracies in sub-Saharan Africa. By 2000 there were just four autocracies, but 30 anocracies. This change is an important part of the explanation for the sharp increase in conflict numbers in the 1990s. As Monty Marshall and Benjamin R. Cole point out in their "Global Report on Conflict, Governance and State Fragility, 2008":
Figure 2.7 Average Number ofConflict Onsets and Terminations per Yearin Sub-Saharan Africa, 1950-2005OnsetsTerminations4.543.532.521.51.501950–591960–691970–791980–892000–051990–99Onsets and TerminationsYear
Data Source: UCDP/Human Security Report Project Dataset.
The rate at which conflicts were starting and ending in the 1990s was twice that of previous decades.
Figure 2.8 A Dramatic Decline in Autocraciesin Sub-Saharan Africa, 1946-200619611951195619461966197119761981198619911996200120064035302520151050AnocraciesDemocraciesAutocraciesNumber of RegimesYear
Data Source: Polity IV.106
The number of autocracies declined dramatically following the end of the Cold War.H U M A N S E C U R I T Y B R I E F 2 0 0 7 27
Anocracies have been much more vulnerable to new outbreaks of armed societal conflict; they have been about six times more likely than democracies and two and one-half times as likely as autocracies to experience new outbreaks of societal wars.107
Given this, and given that the number of anocracies in sub-Saharan Africa increased fifteenfold between 1988 and 2000, it is not surprising that there were twice as many new state-based armed conflicts in the 1990s as in the 1980s.
Other explanations for the sharp increase in war onsets in the 1990s are much less compelling. The quantitative literature on the causes of armed conflict stresses "structural" variables like income per capita, demographic factors such as disproportionate numbers of young unemployed males, or dependence on primary commodities. What all these factors have in common is that, unlike political regimes, they change very slowly.
The association between income per capita and conflicts is the most robust finding in the quantitative literature. Yet, average incomes have to change a great deal to bring about a significant change in the risk of new conflict onsets. The changes in average per capita incomes in sub-Saharan African countries in the early 1990s were not nearly sufficient to explain the doubling of conflict onsets in that decade.
How Conflicts End
Understanding why more conflicts have been ending—and not restarting—since the end of the Cold War requires a more detailed examination of the different ways in which conflicts were coming to an end in this period.
The pattern of war terminations in sub-Saharan Africa has changed substantially over the past two decades—as it has in the rest of the world. The number of state-based conflicts that terminate in victories has decreased sharply, while the number ending in negotiated settlements has risen.
These changes, as Figure 2.9 indicates, are striking. From 1950 to 1989, 28 state-based conflicts in sub-Saharan Africa ended in victories, six in negotiated settlements. Then from 1990 to 1999, there were six victories and 12 negotiated settlements. The third type of termination recorded in Figure 2.9 is labelled "Other." This is the category that includes conflicts that peter out without either a victory or a peace agreement—or where the death toll falls below the threshold of 25 for a full calendar year. As was the case with negotiated settlements, there was an explosion of "Other" terminations in the 1990s.
As Figure 2.9 reveals, in the new millennium, the number of conflicts ending in victories continued to shrink while the number ending in negotiated settlements continued to grow. Between 2000 and 2005 just one state-based conflict
Figure 2.9 Number of State-based Armed Conflict Terminations in Sub-Saharan Africa, 1950-2005
VICTORIES
NEGOTIATED SETTLEMENTS
OTHER
TOTAL TERMINATIONS
Years
Total No.
Number Restarted in Under 5 Years
Percent Restarted in Under 5 Years
Total No.
Number Restarted in Under 5 Years
Percent Restarted in under 5 years
Total No.
Number Restarted in Under 5 Years
Percent Restarted in Under 5 Years
Total No.
Number Restarted in Under 5 Years
Percent Restarted in Under 5 Years
1950-59
1
0
0.0
0
0
0.0
2
0
0.0
3
0
0.0
1960-69
9
0
0.0
1
0
0.0
2
2
100.0
12
2
16.7
1970-79
9
4
44.4
1
0
0.0
4
0
0.0
14
4
28.6
1980-89
9
1
11.1
4
0
0.0
4
1
25.0
17
2
11.8
1990-99
6
1
16.7
12
8
66.7
20
14
70.0
38
23
60.5
Total 1950-1999
34
6
17.6
18
8
44.4
32
17
53.1
84
31
36.9
2000-2005
1*
0
0.0
10*
0
0.0
10*
6
60.0
21*
6
28.6
Total 1950-2005
35
6
17.1
28
8
28.6
42
23
54.8
105
37
35.2
Data Source: UCDP/Human Security Report Project Dataset.
Since the end of the Cold War, the number of conflicts ending in negotiated settlements has increased; the number ending in victories has decreased.
*Includes terminations for which it is too early to determine failure rate over the five-year period.