Armed Conflict and Hunger--Calculating the Costs of Conflict

Conflict wastes lives, livelihoods, environmental resources, and materials. In certain cases, such as with the United States during World War II, war has been credited with reviving the economy, stimulating production, and reducing unemployment. In most cases, the products (and many of the people) are destroyed.

Warfare over the last 20 years has cost up to 1 million lives per year, and most of the casualties are civilians, not combatants. Estimates of deaths directly related to the fighting in wars run as low as 5 percent in the civil wars in Angola and Mozambique, although more than a million persons died in each of these conflicts. Lack of food and health care, plus the traumas of displacement, kill many more than actual combat does (Green, 1994).

The monetary costs of conflict are conventionally tallied in lost livelihoods, destroyed properties, and monetary needs for immediate relief and longer-term reconstruction. In addition, governments forgo investments in peacetime services and economic outputs and may suffer additional losses in high foreign exchange rates as they seek to contain bordering hostilities.

Green (1994) and Green and Mavie (1994) suggest the large scale of output forgone, expenditures diverted, and foreign exchange burdens indirectly related to the conflicts of southern Africa, but they also show the difficulties of precise calculation. For Mozambique, the cumulative loss of output due to the struggle between the government and Renamo insurgents probably exceeded US$20 billion from 1982 to 1992. Production losses were due to the deaths of some 1.5 million people and the removal of over half of the population from customary sources of livelihood (from 1.5 to 2 million were international refugees, 2 million were internally displaced into camps or resettlement schemes, 2 million were displaced but not into formal settlements, and more than 1 million were living in the vicinity of their ruined villages but were socioeconomically or psychologically displaced) (Green and Mavie, 1994, 78). The war also inflicted direct damage on markets, communications, public health services, and other infrastructure. Destruction of capital stock led to continuing losses of output with difficult-to-calculate impacts on income flow and multiplier effects. The conflict also drew enormous military investments from neighboring Tanzania and Zimbabwe. It is estimated that Tanzania invested $5 billion in military spending over the 30?year period 1961-91, which included Mozambique's prior liberation war against Portugal. This expenditure imposed high costs on the Tanzanian populace in terms of lost food security and health care due to the diversion of potential government resources away from agricultural and medical facilities and training. Such losses pale in comparison with Mozambique's devastation but still point to considerable regional effects beyond the immediate conflict (Green, 1994, 40).

Photo: Martin Lueders

After the fighting. Pathologist Isaac Moses' members of "skeleton crew, " exhuming remains for post-mortem of infant to determine cause of death; central Monrovia, Liberia, 1997.

Sivard (1996, 39) calculates world military versus social expenditures by considering investments in weapons as alternatives to health and nutrition expenditures. From 1960 to 1994, arms imports by developing countries totaled US$775 billion (in 1987 dollars). The enormity of waste in human lives is shown most dramatically not in monetary terms, but in limbs lost to antipersonnel land mines. An estimated 100 million antipersonnel land mines litter 69 countries; more than 10 million mines lie in Afghanistan, Angola, Egypt, and Iran. Cambodia is estimated to have one mine for every person in a population of 10 million, and one of every 236 persons is an amputee (Sivard, 1996, 15). Land mines that prevent farming and trade contribute to food insecurity following warfare and in turn to scarcities that contribute to continuing conflict potential. Clearing antipersonnel mines is both technically difficult and expensive-another cost of conflict.

Political analysts also speak of far-reaching political costs of conflict in developing countries. The most important of these is probably the undermining of the influence of the United Nations, which is seen as having failed to intervene effectively, and to a lesser extent, the International Federation of the Red Cross, which seems to have been powerless to deliver aid in the case of East Timor (Cranne, 1994).

In 1986, Willie Brandt chastised NATO countries for placing military and space program spending above foreign development assistance (Brandt, 1986). With the end of the Cold War, arms spending may finally be in decline, but emergency assistance for zones of armed conflict continues to hijack foreign assistance budgets that overall are shrinking in response to economic downturns and domestic pressures to cut budgets and "welfare" spending (Marchione, 1996).

Despite the sometime misallocation, mismanagement, and politicization of development assistance, aid can support equitable and sustainable development. Foreign assistance has contributed to gains in child survival, life expectancy, and educational attainment in the developing world over the past several decades. For most of the poorer developing countries, aid remains an essential tool for assuring that everyone has access to food, basic education, primary health care, family planning services, clean water, and safe sanitation. In 1996, official development assistance from all sources totaled only US$58.2 billion, down 14.5 percent in real dollar terms from 1991 (OECD, 1998). Within this diminishing aid budget, an increasing proportion of both total assistance and food aid was directed to emergencies. In 1996 emergency assistance came to $5.5 billion, or 9.5 percent of all aid, compared with just 3.5 percent in 1987. In 1993 and 1994, emergency assistance peaked at 11 percent of all aid (Figure 1).

Assistance from the 21 members of the Organisation for Economic Cooperation and Development's Development Assistance Committee (OECD/DAC) accounted for 95 percent of all aid in 1996 (OECD, 1998). Although the total rose by 5 percent in real dollar terms from 1995 to 1996, it was still 9 percent lower than in 1992 in constant dollars.

The end of the Cold War deprived donors of a powerful political motive, and the share of the "peace dividend" devoted to development assistance was less than hoped (Lake, 1990). Meanwhile, former communist states in Central and Eastern Europe and the former Soviet Union, many of them in or close to conflict, received more than US$10 billion in aid in 1995 (the last year for which information is available), draining funds from poorer countries. Aid to Sub-Saharan Africa, where most of the poorest countries and several major conflict zones are located, fell from $18.9 billion in 1994 to $16.8 billion in 1996, a decline of 11 percent. Although the flow of private funds from DAC member countries to the developing world ($234 billion) far exceeded aid and other concessional flows from DAC member governments and multilateral organizations ($66 billion) in 1996, development assistance remains the most significant channel of funds for poor countries (OECD, 1998).

Food aid is also dropping precipitously; total tonnage from all donors declined 23 percent in 1996 and fell 57 percent between 1993 and 1996. On average in the mid?1990s, 35 percent of this greatly reduced tonnage went to meet emergency needs; in the 1970s, emergency relief accounted for only about 10 percent of all food aid (USAID, 1998). The WFP has seen an even more dramatic transformation of its activities: in 1996, emergency operations and protracted feeding of refugees and other displaced persons claimed 68 percent of WFP food aid tonnage, whereas a decade earlier, two-thirds of WFP assistance had gone into development efforts such as school feeding, maternal and child health projects, and food-for-work programs. Emergency operations in Rwanda and the former Yugoslavia and feeding of refugees and displaced persons in Afghanistan and Liberia alone claimed one of every three tons shipped by WFP in 1996 (WFP, 1998).

Although aid officials try to make relief function as development assistance, the "relief-to-development continuum" they talk about appears to be more wishful thinking than fact. The bulk of emergency food assistance is devoted to meeting basic human welfare needs.

Since the end of the World War II, the United States has tended to set global development assistance trends. In the 1990s, U.S. aid has led the pattern of growing slices of emergency aid coming out of a shrinking overall assistance pie. In fiscal 1996, U.S. bilateral emergency assistance came to US$1.35 billion (US$449 million in food aid, US$181 million in cash disaster assistance, and US$721 million in refugee aid) or 39 percent of the US$3.48 billion aid budget. Cash development assistance that same year totaled $1.68 billion, and the United States provided $445 million worth of development-oriented food aid. This represents almost a doubling of total U.S. emergency aid from US$786 million in fiscal 1989, although a slight decline from the high figure of US$2.2 billion in fiscal 1993. But overall aid resources fell 24 percent between fiscal 1985 and fiscal 1995. To meet emergency needs, U.S. policymakers diverted resources from global child survival initiatives and programs for sustainable development in Africa (Cohen, 1995; USAID, 1997).

Global aid from NGOs is similarly targeted more to disaster areas than to regions that are impoverished but peaceful. In the mid?1990s, such agencies as OXFAM United Kingdom and OXFAM Ireland, whose aim is long-term improvement of human conditions, were devoting 50 percent of their resources to areas of conflict; in Africa, the proportion is 70 percent (Cranne, 1994).

This proliferation of emergency needs and the consequent squeeze on development assistance are likely to continue, although reductions in development aid result in the elimination of programs that might help households and communities in resource-poor countries become more food secure and less prone to conflict. The U.S. Agency for International Development (USAID) recognizes that "sustainable development that creates chains of enterprise, respects the environment, and enlarges the range of freedom and opportunity over generations should be pursued as the principal antidote to social disarray" (Atwood, 1994), but USAID also faces shrinking budgets.

Within shrinking budgets, international agricultural research is suffering cuts and is increasingly forced to compete with programs supporting the complementary goals of environmental protection, health and nutritional welfare, livelihood security, and infrastructure development. In 1993, the share of agriculture in OECD aid fell to less than 8 percent, down from 12 percent just four years earlier (Randel and German, 1996). Cuts in agricultural research come at a time when all countries are facing relative scarcities of land and water, and problems of food insecurity and environmental degradation are growing. In many developing countries, the gap between "resource-poor" and "high-potential" areas is growing, adding to perceptions by those left behind that they are unfairly deprived. Population growth without access to technologies to intensify production or protect the environment ripens the potential for environmental destruction and conflict. Conflict and its aftermath thus contribute to shortfalls in food measurable in terms of costs in production forgone.

Conflict accounts for downturns or lower than expected values in agriculture, gross domestic product, and trade. This is because conflict directly and indirectly reduces land and water, plant and animal breeding stocks, human resources, and financial capital to invest in agriculture, environmental protection, and human well-being. Food projections through the year 2000 and into the 21st century suggested anecdotally that Mozambique, in the absence of conflict, could make southern Africa self-sufficient in rice. However, the recent floods in southern Africa have been serious setbacks. Cambodia and Myanmar could brighten the outlook for Asian rice in the absence of conflict-related production and marketing disruptions. Beyond anecdotal evidence, it is possible to estimate roughly the extent of food production losses due to conflict by examining food production trends in war-torn countries. Of special interest are the extent of food production declines possibly due to war in southern Africa, where the aggregate trend in food production per capita is downward.

For Sub-Saharan Africa, production scenarios are calculated with and without conflict as a historical factor from 1970 to 1993. Two methods have been used to compare actual and "peace-adjusted" food production in individual countries and in the region as a whole. The first method in Table 3 adjusts mean food production trends in war-torn Sub-Saharan African countries. Columns 3 and 4 show annual mean food production per capita in war and non-war years for Sub-Saharan countries that have had conflicts, and column 5 shows the difference (3- 4) as a percentage of non-war production for each country. The annual impact of these differences on food production in the region as a whole (column 7) is calculated by weighting the contribution of each country's population to the region as a whole.

The second method in Table 4 calculates the differences in mean growth in food production during war and non-war years and their contribution to regional food production trends. Columns 3 and 4 show growth during years of war (beginning one year before war) and relative peace, and column 5 calculates the difference (3- 4) for each country. Annual impact of these differences on food production in the region as a whole (column 7) is again calculated by weighting the contribution of each country's population to the region as a whole.

Tables 5 and 6 show the total impact of internal wars on food production levels for Sub-Saharan Africa as a whole. Using the mean food production method, Table 5 subtracts the total impact on food production for the region (Column 2, derived from the mean weighted country impacts in Table 3) from the actual food production level observed for each year (Column 1) to reach a peace-adjusted level of food production (Column 3). This is the level that shows what food production might have been in the absence of conflict in the region. The figure for food from peace (column 4) shows how much additional food the region might have produced in the absence of war as a percentage of its actual production.

Using the growth method, Table 6 "peace-adjusts" regional food production by recalculating food production each year using the adjusted annual growth figures shown in parentheses in column 3 (these figures are derived from the adjustments to annual growth for each country at war shown in column 7 of Table 4). Again, "food from peace" represents as a percentage of actual production the additional food the region might have produced had peace prevailed. The growth method, therefore, accumulates war's effects on food production because hypothetical gains in production in one year become the basis for calculating gains in the next year, and so forth. In contrast, the mean food production method treats the effects of war each year without regard to effects in previous years, allowing one to better compare the effects from year to year. Actual and peace-adjusted per capita food production by these two methods are plotted in Figures 2 and 3.

These methods, admittedly rough estimates of country and regional departures from historical food production trends, suggest the extent to which armed conflict has interfered with food production in Sub-Saharan Africa over the period 1970-93 and the additional quantities of "food from peace" that might have been available had wars been absent.

As expected, in 13 of 14 countries food production was lower in war years. Drops were as low as 3.4 percent in Kenya and as high as 44.5 percent in Angola (Table 3). The mean decline in annual production was 12.3 percent. These decreases were paralleled by shortfalls in food production growth rates, which were observed in all countries except Nigeria (Table 4). Growth declined as little as 0.3 percent in Chad to as much as 8.1 percent in Liberia; the mean was 2.9 percent.

Such variation is expected given the widely differing scope and scale of the conflicts. The calculations also reflect expected differences in the scale of impacts within and across countries. The impact of Angola's war on the country and region is very large, twice as great as that of Zimbabwe, which experienced less disruption in food production (Table 3). The Ethiopian war had a medium impact on the country's annual production compared with other countries in the region, but its regional impact is magnified to 1 percent for each year of war because of its size. The massive food assistance requirements of both Angola and Ethiopia during and immediately following their war years verify the large-scale disruptions. The calculations also show the aggregate effect of large numbers of wars on both food production and growth during and immediately following the Cold War years.

Adjusting regional food production for peace reverses neither the general downward production trends for the region nor the obvious effects of the major droughts of 1974, 1984, and 1992. But the adjustment does dampen the effects. By the mean food production method the greatest gain would have been nearly 6 percent in 1991 (Figure 2). The costs of conflict are more graphically apparent by the growth method, which shows production gaps widening with each decade: from zero to 1 percent in the 1970s; 1.3 to 3.5 percent in the 1980s; and 3.9 to 5.3 percent in the 1990s (Figure 3).

In interpreting these data certain caveats are in order. The FAO country-level data (like other statistical sources) are not very reliable. Even in ordinary times, but especially in times of crisis, data collection and reporting techniques lack accuracy, miss variations within countries, and often fail to take into account significant contributions of the informal economy, especially activity in parallel markets.

Moreover, the impact of conflict on food production is complex and hard to separate from other factors. The rough calculations presented here lump together synergisms of armed conflict, bad weather, human illness, or volatile commodity prices on food production, distribution, and consumption. In specific cases such as Ethiopia, Sudan, and most recently Rwanda, food crises due to drought and agricultural and relief mismanagement preceded and triggered violent overthrow of their governments, only to be followed by even greater food shortfalls in the years of conflict that followed.

The calculations also aggregate country-level data, so they do not identify within-country regional shortfalls and food insecurity that result from conflict or contribute to social discontent. The central issue in food crisis is not aggregate food production but who has food and who does not. In any food crisis situation, some benefit rather than suffer from selective regional downturns in food and income.

The ways in which governments or others intervene to prevent or redress food insecurity may be more important for peace than the original shortfall or impending crisis. During the 1992 drought, the Somalian government exhibited little capacity to respond, and people remained mired in famine and conflict. The Hutu government in Rwanda responded selectively, excluding many who subsequently joined the political opposition and participated in violent actions. By contrast, more stable governments in Botswana and Zimbabwe successfully weathered drought-related food shortfalls; their governments had early warning and timely response systems in place, and with donor assistance, managed to stave off crisis, famine, and civil unrest.

Finally, the data also raise questions that cannot be answered without additional year-by-year scrutiny of the food production data; for example, why does food production appear higher in Chad and Uganda during war years?

Such caveats notwithstanding, the data in these tables indicate that stagnating food production and declining growth are closely related to conflict, as either cause or effect. But quantitative data show only that they are closely linked. Attention to additional historical data by country and year can help separate out initial or sequential cause and effect and offer lessons for future conflict prevention.

In 1999-2000, Ellen Messer was a Visiting Associate Professor at Tufts School of Nutrition Science and Policy and a Visiting Scholar in Anthropology at the Massachusetts Institute of Technology.

Marc J. Cohen is special assistant to the director general of the International Food Policy Research Institute and the former editor of Bread for the World's World Hunger Report.

Jashinta D'Costa is an independent consultant. She has been a consultant to the United Nations Development Program, Save the Children-US., and National Council for Agricultural Education, as well as served as research associate with Bread for the World Institute.

This article is adapted from Food from Peace: Breaking the Link between Conflict and Hunger, Washington: International Food Policy Research Institute, 1998, by the same authors.