The Missing Link

Introduction

By the early 1990s, most drought-triggered famines and food crises in Africa were not hard to predict. This is testimony to the success of famine early-warning systems (EWS), many of which were set up during the 1980s. In 1984–5, failure to prevent famine in the Sahel and Horn of Africa was widely attributed to a lack of early warning (EW). Since then, better prediction has been a major policy concern, for both donor agencies and national governments. Greatest attention has been paid to the establishment of purpose-built EWS. In the Sahel and Horn of Africa alone, more than eight new EWS were set up between 1985 and 1990. Considerable progress has been made, in improving data-collection methods, developing indicators, and making use of sophisticated information technology. By 1990 more information on the likelihood of famine was available to donors and governments than ever before.

Famine prevention, however, has remained an elusive goal. Better prediction has not led to corresponding improvements on the response side. There is clearly a missing link between the provision of EW information and the use of that information to trigger a timely preventive response. Anecdotal evidence abounds as to why EW information has not been influential. The EWS was not sufficiently vociferous; the information was inappropriate, late or untrustworthy; donors were ill-disposed to help a particular government; adequate resources were not available; institutional and logistical obstacles overwhelmed good intentions; the domestic political will to react was lacking, and so on. Intuitively such explanations make sense, but they do not amount to a systematic analysis of what happens to EW information once it enters the decision-making process and how it is used by policy-makers. How and when do they receive the information, and what kind of data are most influential and why? Most important, what are the missing links between prediction and prevention?

That is the subject of this book. It is based on an analysis of early warning and relief responses in 1990 and 1991 in five African countries: Ethiopia, Sudan, Chad, Mali and Kenya focusing on Turkana District in the north; and in a number of donor agencies. Our approach is to begin with the decision-making processes which determine the response to a threatened crisis. From there, it is possible to ascertain the relative importance of information and of other factors in triggering or inhibiting the response.

The agricultural year 1990–91 was a year of drought across much of the Sahel and Horn of Africa, with early indications that food crises would be widespread. In some regions, it was the combination of drought and conflict that threatened to cause the most severe cases of famine and suffering. Elsewhere, the drought was not necessarily as bad as in the mid-1980s. Nevertheless, conditions deteriorated to the point where relief aid was necessary, to protect livelihoods and often lives as well. A number of EWS in the region were being put to the test for the first time. Could they fulfil their ultimate objective of triggering an adequate and timely response to prevent acute food insecurity and/or famine developing? The results have been mixed. In some countries the EWS were remarkably ineffective. In others, they had much greater influence.

The reasons why EW information is, or is not, used fall into four broad categories: first, reasons to do with the EWS itself, and the information provided; second, reasons to do with the institutional context within which the EWS sits, and the institutional links to decision-makers; third, reasons to do with the broader political environment; and fourth, logistical obstacles to launching a timely and adequate response. Much has been written about the technical aspects of EW and numerous evaluations of different systems have been carried out. Most have focused on the internal workings of the EWS: the scope of indicators, accuracy of the data and timeliness of the warnings. These relate to the first category of reasons, concerned with the performance of the EWS itself. But few have looked at how EWS fit into the wider context. From the analysis presented in this book, the second and third categories of reasons emerge as the most important explanations of whether EW information is used, and of variations in performance between the different case-studies.

Most countries in the Sahel and Horn of Africa rely on the international relief system to provide resources to run relief operations in times of food crisis. National governments rarely have adequate resources or capacity to respond. Donor agencies and non-governmental organizations (NGOs), the key actors within the international relief system, have been particularly influential in developing famine prediction for the Sahel and Horn of Africa, both in funding and operational roles. But much less attention has been paid to developing the response side of the equation. There are two key reasons for the persistent failure to translate EW into timely response, illustrated in the case-studies in this book. First, the international relief system responds to famine once it is under way but is ill-equipped to respond to genuinely early warning, to intervene in time to prevent it. Second, it is not the severity of the crisis, but relations between international donors and national governments which tends to be the single most important determinant of the timing and scale of the international response. Thus, in the case of the southern African drought in 1992–3, as well as national capacity to respond, the desire of the donors to keep structural adjustment programmes on track and a determination to avoid further political unrest in the region combined to initiate a timely response by the international relief system (SADC, 1993). The same factors did not conspire to trigger a timely response to food crisis in the Sahel and Horn in 1991.

Understanding famine

A common theme running through this book is that a gap exists between theory and practice in famine prevention. Our understanding of famine as 'outsiders' has improved in leaps and bounds during the last 10 to 20 years. This should improve our ability to prevent it with appropriate policies and interventions. Instead, conventional and often inappropriate relief responses persist. Explanations of the causality of famine have shifted from a preoccupation with supply-side factors in the 1970s, towards a recognition of the pre-eminence of access or entitlement to food in the 1980s. Famine is no longer solely — or indeed even primarily — attributed to food availability decline (FAD), but increasingly to food entitlement decline (FED). Thus 'starvation is the characteristic of some people not having enough food to eat. It is not the characteristic of there not being enough food to eat' (Sen, 1981:1). This key distinction has fuelled much of the subsequent development of famine theory. The notion of entitlements has been refined in the light of recent famines in Africa. The health-crisis model of famine, for example, challenges Sen's assumption and the popular Western notion that famine mortality is caused primarily by starvation. Instead, the major killer is epidemic disease, either during or after periods of food shortage (de Waal, 1990; Dyson, 1993). The outbreak of mass communicable disease is often caused by economic and social collapse leading to large-scale migration and social disorder, itself a function of declining entitlements. A further refinement has been to move away from a narrow focus on exchange relations towards a wider bundle of entitlements incorporating different types of assets, including investments, stores and claims (Swift, 1989). In addition to buffering vulnerable people when famine threatens, however, this broader range of entitlements can also undermine food security, for example, via demands from kin for food, or various claims exacted by the state or by more powerful neighbours (Davies, 1995 (forthcoming)).

Linked to this is improved understanding of the strategies famine-prone people pursue to cope with famine. Interest in coping strategies arose in the mid-1980s as a means of trying to understand why it was that some people survived periods of dearth while others did not. Studies have revealed different stages of coping, from insurance, through to disposal of productive assets, to destitution as the crisis deepens and coping capacity runs out (Corbett, 1988). At the centre is a constant trade-off between short-term consumption and the preservation of longer-term capacity to produce: it is now widely recognized that households threatened by famine often choose to go without food rather than sell productive assets (de Waal, 1989; Devereux, 1993). Emphasizing indigenous strategies marks a shift towards a more people-focused understanding of famine, paying attention to the perceptions and priorities of famine victims.

Famines caused by conflict are increasingly common on the African continent, but are as yet poorly understood. There are at least three major ways in which war causes famine: first, the direct destruction of battle, and consumption of resources by armies; second, when famine is used as a weapon of war, for instance through restrictions on population movement, especially during sieges and for counter-insurgency purposes; and third, when state structures and warlords sustain themselves as predators of the poor, thereby creating famine (de Waal, 1993). Conflict, especially when it is prolonged, also serves to divert attention away from the pursuit of preventive policies. Furthermore, it tends to weaken those institutional structures not directly geared to sustaining and/or winning the war; and also creates a climate of suspicion and mistrust between governments and donors.

Early warning

There are many systems for predicting famine, from informal indigenous information systems through to formal EWS. This book is principally concerned with the functioning of formal EWS. An EWS can be defined as a system of data collection to monitor people's access to food, in order to provide timely notice when a food crisis threatens and thus to elicit appropriate response (Davies et al., 1991).

To some extent improved understanding of famine processes has fed into and informed the design of, and approach to, EWS. Inevitably there is a time-lag before theory is applied in practice, although the lag for EW has not been that long. The first major impetus to establish EWS in Africa came after the famines of the early 1970s in the Sahel, which the international community failed to recognize in time. EWS were set up to service donor and UN food aid institutions, and this remains their raison d'être. The earliest modern (and still one of the most influential) EWS, the FAO's Global Information and Early Warning System (GIEWS) is unequivocally supply-side oriented.

With the influence of entitlement theory of famine in the 1980s, most EWS began to incorporate indicators of effective demand for food, including price data and other socio-economic indicators, and a number responded to the work on coping strategies by trying to incorporate behavioural indicators of famine vulnerability. Measurement difficulties and scaling-up of local experiences make this one of the more challenging episodes in EWS development. Much of the most successful work has happened within small-scale, local-level information systems. Although most formal EWS operate at national and international level, after the mid-1980s a number of pioneering local-level systems were set up, often by NGOs.

Food-production and food-supply forecasts are in many ways still the best-developed EW capability. A number of EWS continue to be food-supply driven. However, many have now developed multi-indicator models, incorporating a wider range of socio-economic indicators, which enable them to be sensitive to less-dramatic changes in food situations than famine. The complexity of famine processes implies that multi-indicator local-level EWS are most likely to be able to: detect deterioration in food security sufficiently early to launch a timely response; monitor and be sensitive to complex famine processes within different groups in a population; and identify appropriate public action in line with local people's priorities. Nevertheless, the original rationale for setting up EWS, to service food-aid donor agencies, has changed very little during the last 20 years. Their reference point, particularly for the recommendations they make, continues to be large-scale famine catastrophes.

Least progress has occurred in conflict situations, despite growing demands for EWS to enter this arena. The problems here are overwhelming, not least because of political sensitivities which can undermine the 'technical neutrality' most EWS strive to preserve.

Famine prevention: the relief response

Response can be defined as additional resources channelled to famine-prone people in order to assist them in withstanding the effects of declining access to food. For the international relief system, this means resources over and above normal development aid. In practice, response usually means saving lives or providing food to prevent starvation. Much less often, it means saving livelihoods, or providing food and other resources to protect future capacity to subsist, as well as to ensure current consumption.

The conceptual advances in our understanding of famine processes have major implications for preventive public action. First, they imply that famine policy cannot be limited to preventing large-scale excess mortality due to starvation, but instead must be viewed as halting the progress of a downward spiral of increasing vulnerability, leading to economic and social disintegration, destitution and eventually death from one of a number of causes, of which the outbreak and transmission of communicable disease is particularly serious. This spiral is illustrated in Figure 1.1. For famines to be prevented, intervention on the spiral must occur well before the point of death has been reached, as indicated.

Second, intervention must be early enough to protect livelihoods before lives are threatened, in accordance with people's own priorities (cutting back on consumption to preserve productive assets). This requires a more developmental approach to relief early enough in the spiral to prevent destitution and to reinforce existing capacity to cope.

Third, the preventive process is institutionally and politically more complex than a narrowly defined humanitarian relief operation which aims to prevent death by starvation. It implies that a range of interventions is required, including improved health care and the protection of water sources. This suggests a heavy institutional load, both for donor agencies and national governments, and a restructuring of bureaucratic procedures which tend to separate famine and related emergencies from more general development activities.

The current international relief system falls far short of this ideal. Its approach is founded on a simplistic and reductionist model of famine — of death by starvation, to which the most logical response is to provide food aid. Of course there are other factors feeding these assumptions: at least until recently, readily available food surpluses in the North, whereas money and other relief resources have been more difficult to obtain. The principal characteristics of the international relief system are that it is geared to saving lives, not livelihoods; that it mainly provides food aid, on the premise that starvation is the main cause of famine mortality; and that it delivers inadequate quantities of relief too late — after the start of the hungry season.

Figure 1.1 shows how far down the famine spiral people must sink before they usually receive international relief assistance. This is supported by evidence from the case-studies in this book. What is conspicuously clear is that our improved knowledge has not percolated through to famine-prevention practice. And this cannot be blamed on lack of information; famine EW is where greatest progress has been made. Instead, the problems are to do with inherent institutional and political weaknesses in how the international relief system functions. These themes are explored in Chapter 3.

The Sahel and Horn of Africa: conditions ripe for famine

Coupled with this poorly functioning international relief system is a constellation of forces in Africa which conspire to create conditions ripe for famine. What makes the Horn of Africa, in particular, and to a lesser but significant extent the Sahel, so vulnerable to famine when other countries, such as India, are succeeding in preventing it?3 First is the fact that the underlying scale and intensity of food insecurity on the African continent are increasing. More people are vulnerable to famine now than 20 years ago, due both to deteriorating trends and rising incidence and severity of short-term shocks. Nearly half the population of sub-Saharan Africa — an estimated 235 million people — face conditions of chronic food insecurity. A third of all children under five are chronically undernourished. Population growth has resulted in the absolute number of hungry children rising from 18 million in 1975 to 30 million in 1990 (ACC/SCN, 1992).