METHODS

We drew from the Uppsala Conflict Data Program/Peace Research Institute Oslo (UCDP/PRIO) Armed Conflict project database for conflicts causing an estimated total of at least 500 direct fatalities and reaching an intensity in which the violence was systemic and sustained. These included conflicts of any type with organized armed forces fighting within or between countries. Cases resulting in fewer than 500 deaths, whether caused by genocide, banditry, or terrorism, were excluded.

The Center for Systemic Peace (CSP) uses this database to specify dates of initiation, and termination of conflicts for the period 1946 to 2007.9 The CSP database is a unique source of detailed comparative measurement of the intensity of conflict (Table 1). It includes country-level information on 324 conflicts of any type in countries with populations more than 500 000, including a total of 41 countries experiencing conflict since the year 2000. The intensity scores used here represent a scale of the magnitude of violence on affected societies. Data in the CSP are collected from 16 journalistic and political science sources and scored by the Political Instability Task Force Episodes of the CSP on a 1 to 10 scale of intensity.Reference Marshall, Gurr and Harff10 The highest level of conflict is represented with a score of 10, while a score of 1 represents sporadic or limited violence (Table 2). The relative difference between score levels is intended to be equidistant in magnitude; thus a conflict with a score of 4 rates twice as intense as one with a score of 2. A conflict with a score of 10 would represent an intensity that could be found, for example, in the Japanese theater of World War II (WWII) during carpet and atomic bombing. The most intense scores in the past four decades included Vietnam and Rwanda, each with a score of 7. Since 1990, the only nations with scores of 6 or 7 were Afghanistan, Angola, Bosnia, El Salvador, Ethiopia, Iraq, Rwanda, and Sudan. An average conflict-to-intensity score was developed by multiplying the population in each intensity group by the total number of people in each area of current conflict.

Table 1. Components of the Conflict-Intensity Score

Table 2. Conflict-Intensity Levels

The population of each area in conflict was drawn from the US Census Bureau International Data Base.11 This source provided estimates based on census and other sources for each country and year since 1950 that have been determined from reviews of censuses, surveys, and vital statistics provided by National Statistics Offices, and data on international migration and refugee movements.

Only conflicts internal to China, including the Great Leap Forward and the Cultural Revolution, were excluded, as the large proportion of the population in the world present in China obscured trends external to that country. Where data were absent for years of conflict, a baseline growth rate of 3% per year was used.

The end of a conflict was defined as the year following the one in which hostilities were judged to have ceased. Data on population from countries at each rated level of conflict, including postconflict, were combined to create annual summaries of the number of people exposed to conflict by level of conflict intensity and year using Access software. This computation yielded estimated populations exposed to conflict by level of conflict intensity for each year from 1946 to 2007.

The data generated on populations exposed to conflict were then compared to an estimate of deaths due to direct conflict. The UCDP/PRIO Armed Conflict Data set is perhaps the most widely used source for estimates of direct battle deaths, drawn from a wide variety of informant and press sources.Reference Lacina and Gleditsch12 These data do not include estimates of deaths due indirectly to conflict, such as those that occur from conflict-related loss of public health infrastructure or systems protections (eg, water, food, sanitation, health, shelter, fuel).

Data on total populations living in territories at each rated level of conflict were combined to create annual summaries of the number of people exposed to conflict by level of intensity and year. This calculation yielded estimated populations exposed to conflict by level of conflict intensity for the years 1946 to 2007. From the year that a conflict ended, the next five years were defined as a postwar period. Each country was included during each of those five years to create a total population in postwar areas.

RESULTS

The number of people living in areas affected by conflict has risen (Figure 1), as has the size of the population in postconflict areas (Figure 2). This increase is due to both a shift toward lower intensity conflicts and the resolution of an increasing number of these conflicts. Together, these groups have constituted about 6% of the world's population during the last 15 years, up from an average of approximately 4% during the previous three decades.

Figure 1. Populations Living in Areas of Conflict by Level of Conflict Intensity, 1946-2007.

Figure 2. Populations Living in Areas of Conflict or Postconflict, 1956-2007.

An estimated 390 million people have lived in areas with some level of conflict in the year 1997 (Figure 3). In spite of an increase in the total population in these countries of about 30% in the following 10 years, the total population in areas with conflict declined to 370 million by 2007. The estimated mean intensity of conflict declined from 3.9 in 1997 to 3.4 in 2007. This decrease continued a long-term trend of declining intensity among existing conflicts, down from a mean value of 6.8 in the years immediately following WWII. The number of people living in areas where conflict had ended within the previous five years increased from 240 million in 1997 to 320 million in 2007 (Figure 2). The greatest decline occurred in areas of moderate conflict (levels 3 and 4), falling from 27% to 15% of all people living in conflict-affected areas in 2007. This left low- and high-level conflicts as an increasing proportion of all conflicts.

Figure 3. Populations Living in Areas of Conflict and Battle Deaths, 1951-2006.

The size of the postconflict population was largest in the years after WWII. In the 1950s, few conflicts ended and few people entered the postconflict group. The end of postcolonial wars and wars of national liberation in the 1970s brought another peak in postconflict populations.

A third peak occurred in the late 1990s, with the end of the Cold War-related hostilities. The continued resolution of conflicts since that peak has created the highest level of postwar populations since WWII. The number of people now living in areas where war ended in the past five years is almost as large as the number living in areas of conflict at any level (Figure 2). This period is only the second time since the end of WWII that the ratio of people in postconflict to conflict areas has been as high. Combined with the declining proportion of high-intensity conflicts, the declining number of new conflicts, and the resolution of some hostilities of long duration, the relative importance of the intensity of low-level conflict and postconflict has never been greater in the past 50 years.

DISCUSSION

Trends over time are subject to large fluctuations, and the long-term trend toward decline has been interrupted several times during the past 50 years. The picture of conflict magnitude and intensity is thus not a simple one. Still, several major trends emerge from these data. The size of populations in postconflict environments, relative to those in areas of current hostilities, is now higher than it has been at any time since the end of WWII.

This research identifies three key trends. First, there is a long-term increase in the number of people living in conflict-affected areas, defined to include both conflict and postconflict areas. Second, the proportion of conflict-affected people living in areas of lesser conflict intensity has risen dramatically, while moderate-level conflict has declined. Third, there is an unprecedented rise in the size of postconflict populations relative to conflict-affected populations. To our knowledge, only the second of these trends has been described previously.8

The decline in widely used estimates of direct-conflict deaths contrasts dramatically with the increase in the number of people living in conflict or postconflict areas. Accounting for deaths in conflicts is notoriously difficult, as many deaths may not be counted. Other researchers have described this deficit as “the shrinking costs of war.” This research, depending on a scale that focuses on populations exposed to risk rather than counts of deaths, may provide a more comprehensive estimate of the impact of war beyond those deaths that can be accounted for. It is thus a broader measure of the human costs of conflict. These costs may not be shrinking if more people are exposed to a living environment that includes exposure to conflict. Livelihoods, access to essential goods, and access to health services may all suffer in the lead up to, during, and after a war.

Vulnerability to ill health and compromised well-being is high in this population. It has been reported that people in fragile states represent 15% of all those living in lower- and middle-income countries. They represent one-third of the world's poor, one-half of all children dying before age 5 years, and more than one-third of all maternal deaths.Reference Chakrabarti14 Twenty-two of the 34 countries farthest from reaching the Millennium Development Goals (MDGs) are in, or are emerging from, conflict.15 Yet while funding for emergency health programs in high conflict and disaster areas has risen rapidly, very little funding is available for areas where conflict is ending or has recently ended.Reference Heller16Reference Mlambo, Kamara and Nyende17Reference Jones, Hilborne and Anthony18Reference Guha-Sapir, van Panhuis, Degomme and Teran1920 This analysis adds further strength and urgency to the need to redress this gap through programs focusing on training health workers, establishing disease surveillance, strengthening national planning, providing basic primary care, and transitioning from emergency international humanitarian health programs to national health systems in countries emerging from conflict.

Thus, the rise in the size of postconflict populations, relative to the population of people in areas of current conflict, has important implications for humanitarian and development strategies. Funding for development assistance doubled in the first decade of the 21st century, while funding for humanitarian assistance tripled, reaching $18 billion US dollars in 2008.21 Health programs are the largest part of this assistance, following food aid, at $1.5 billion.22

Health has become a pivotal element in determining success or failure of the postconflict period. Given the large and growing population living in these areas, investments in strengthening health systems and human resources for health should become priority investments. Strategic planning specific to country, region, and culture is a crucial first step in the overall process of recovery. Such skills were rarely a priority during the last three decades of program-centric approaches to humanitarian emergencies. Postconflict-specific research, education, and training, inclusive to the entire humanitarian, donor, and military communities, must address short-term relief and strategic planning skills, application, monitoring, evaluation, and measures of effectiveness for long-term development.

Renewed attention to opportunities to stabilize and rebuild societies will be important components of foreign and international policy in the years ahead. Low-level conflict and postconflict countries constitute the major areas failing to meet international development targets summarized in the MDGs. In the years ahead, conflict trends should be analyzed more effectively with the humanitarian measures introduced here, if more effective choices with assistance resources are to be made.