UN declared famine in Southern Somalia yesterday, after months of drought and hunger in the Horn of Africa. Looking at the map, large parts of Ethiopia and Kenya are not doing well either. It’s a serious and sad situation, but I find it somehow ironic that it was the Famine Early Warnings System Network that published the press release on behalf of UN, basing their resolution on the (arbitrary, I’d say) Integrated Phase Classification (IPC) scale, which argues that “famine exists when at least 20 percent of the population has extremely limited access to basic food requirements, global acute malnutrition exceeds 30 percent, and the death rate exceeds 2/10,000/day for the entire population.” Where did these percentages come from, why is it such a huge milestone that what most people have known for months is now finally recognised a real problem? And if there is such a network as Famine Early Warnings System, shouldn’t their role be to predict and prevent famines from happening, if it in fact is the case that last time a famine was declared was over 20 years ago (according to informative update on Somalia situation in Edward Carr’s blogpost Finally saying famine). Do we not learn anything in that time?
The same press release outlines causes for southern Somalia’s famine, saying it “is driven by a combination of factors. The total failure of the October‐December Deyr rains (secondary season) and the poor performance of the April‐June Gu rains (primary season) have resulted in crop failure, reduced labor demand, poor livestock body conditions, and excess animal mortality. The resulting decline in cereal availability and ongoing trade restrictions have subsequently pushed local cereal prices to record levels and substantially reduced household purchasing power in all livelihood zones. Large‐scale displacement and significant limitations on humanitarian access have further exacerbated the situation.”
This account rings true and takes into account many issues, but still, there is something missing. Causes are largely dealt with from an economic point of view, following in my opinion the popular and most widely-spread ‘entitlement approach’ view of famine as put forward by Indian economist Amartya Sen already in the 1980s. As Carr mentions the legal and logistical nightmare of getting food aid into the area, it’s a telling tale of the context of this famine – simple economics and blaming drought doesn’t explain why so many people have died and are dying every day in conflict-ridden Somalia, the uncontrollably large Dadaab refugee camp in Northern Kenya and surrounding areas.
I remembered I had written an essay on famines for my ‘Economic Development of Africa’ course a couple of years back, and digged it from the crates. While it is academic in nature and deals mostly with the Ethiopian famines of the 1970s, it could as well have been written now, using examples from Somalia and giving insight to the mechanisms at play today. While the UN and NGOs start shovelling food aid into the area as a work of patching-up started too late, I would agree with a comment Sanna Karlsson made in today’s (21.7.2011) Finnish newspaper HBL, about how conveniently, for these organisations, money starts pouring in again ensuring their continued necessity for existence. One might then wonder why not more is done to finally get to the root of the problem and solve the riddle of African famines, the only region in the world still experiencing them.
How useful is the ‘entitlement approach‘ in predicting and preventing famine?
(if you wish to use parts of text below please ask for permission first through the Contact form)
“Famine is conquerable. It has been eradicated from most of the world. But in some countries in Africa human suffering seems to be getting more rather than less common. This is a tragedy and a scandal” (de Waal, 1997: 1).
As the only continent still consistently setting the scene for famines, figures of excess mortality, starvation and food shortages often set the tone for famine writing by ‘outsiders’, often accompanied by now cliché images of starving African children. None of these factors need necessarily be causes or outcomes of a famine, as people experiencing famine, the ‘insiders’, often see it more “as a problem of destitution…and…as an intensification of ‘normal’ processes” (Devereux, 2001a: 119). The complexity of the concept, and the questions asked to try and understand it, have played an important role in the formulation of theories of and policy responses to famine.
While proponents from different areas of academia all attempt to provide explanations for famines, this essay will focus on the economic point of view. From the supply-side inclined Food Availability Decline (FAD) theorem that dominated famine discourse for long, to the demand-focused Entitlement Approach (EA) advanced by Amartya Sen in the early 1980s, focus has “shifted from expanding production of food to a much broader concept of food security that involves access to, and not just the supply of, food” (Fine, 1997: 619). With such differing views on the concept of famine though, and academic circles competing for the title of the ‘best theory’, it is important to examine the usefulness of EA, for both prevention and prediction require an all-inclusive understanding of the underlying causes, triggering factors, mechanisms, beneficiaries and victims of famine. As Sen himself puts it: “[i]n concentrating on entitlements, something of the total reality is obviously neglected in our approach, and the question is: how important are these neglected elements and how much difference is made by this neglect?” (1981: 50).
This essay will be divided into four parts. The first part will focus on the entitlement approach by looking at its origins, theoretical propositions, empirical relevance and proposed preventative measures. The second part will deal with theoretical aspects, while the third part deals with the more empirical issues of EA. As will be seen, the approach is not without limitations, as its predictive power, distorted by theoretical problems, and its preventive power, deformed by its poor policy conclusions, will be mapped out. The fourth and concluding part will, apart from giving a summary of the main findings of this essay, present some thoughts for further study.
Until the 1980s, the dominant discourse on famine revolved around the Food Availability Decline (FAD) theorem. As its name suggests, FAD postulates that famines are caused by an aggregate decline in food supply, for example through drought, floods, war or in fact “anything which disrupts food production”, leading to crop failures and livestock deaths (Devereux, 1988: 270). Famine occurs when this disruption fails to meet the subsistence needs of many people over a prolonged period. The obvious and simple policy conclusion that follows is to channel food aid into the afflicted area (Devereux, 2001a). From this, it should be clear however that FAD is not a sufficient condition for famines to arise, as food can be acquired by other means than producing it oneself, for example through aid, and often not the only necessary condition, as EA proves.
Entitlements – endowments – e-mapping
Appealing to the limitations of FAD presented above, and the failure of the theory to take into account demand-side factors and food-distributional problems, Amartya Sen advanced the Entitlement Approach as an alternative. In his 1981 book titled Poverty and Famines, Sen points out that “[s]tarvation is the characteristic of some people not having enough food to eat. It is not the characteristic of there being not enough food to eat” (1981: 1).
In contrast to FAD, EA takes a distinctly microeconomic focus (Fine, 1997). At the core of Sen’s analysis of famine and starvation lies the structure and relations of ownership between individuals or households and food. “Sen shows that famines have occurred even with plenty of food in the afflicted country, because people have differential access to that food” (Devereux, 2001a: 131). These means of accessing food, ‘entitlement relations’ or ‘entitlement sets’, are divided into four broad categories (Sen, 1981: 2). Firstly, trade-based entitlements denote exchanges in the private ownership market, that is, gaining food through selling other products/services. Secondly, there are production-based entitlements relating to the ownership of factors of production, for example, owning land and thus being entitled to the produce of that land. Thirdly, own-labour entitlements suggest an ‘accepted’ ownership of wages earned, with which one is entitled to buy food. Fourthly, inheritance and transfer entitlements are connected with the receiving of endowments, i.e. assets and resources, in kind.
Through exchange entitlement mapping (e-mapping), individuals and households can convert their endowments into ‘entitlement sets’ (Devereux, 2001b). Holding institutional arrangements constant, Sen argues that e-mapping “will depend on the legal, political, economic and social characteristics of the society in question and the person’s position in it”, also taking into account social security provisions and taxation (1981: 46). A standard neoclassical utility function – with food commodities on the X-axis, non-food commodities on the Y-axis, a ‘budget line’ outlining the maximum food entitlements a person has and a commodity bundle that, below the line, adds to a person’s ‘starvation set’ – could be used to show e-mapping graphically.
Whether a person starves or not depends then on his/her endowments and e-mapping, caused by “a fall in the endowment bundle, or through an unfavourable shift in [e-mapping]” (Sen, 1981: 47). Sen separates two mechanisms here (1981: 50-51). First, a direct entitlement failure indicates a collapse of production-based entitlements, as droughts may destroy cash crops that farmers depend on. Second, a trade entitlement failure may arise through insufficient demand – resulting for example in laying off workers – so that “one can obtain less food through trade by exchanging one’s commodity [or own-labour] for food” (Sen, 1981: 51). Famine for Sen is thus “a phenomenon when a large number of people are not entitled to a minimum food requirement with subsequent widespread death” (Alemu, 2007: 98).
Considering its track record, from the ‘Great Famine’ of the late 19th century, killing one third of the population, to the mass starvation of previous and current centuries, Ethiopia could quite simply be crowned the ‘land of famine’ (Devereux, 2001a: 117). In the light of the above, out of four case studies presented in Poverty and Famines, the Ethiopian famine of 1972-74 in the Wollo, Tigrai and Harerghe regions deserves particular attention in this essay.
Sen found that while the Ethiopian famine was not induced by a food availability decline on the national level, food shortages did exist in the Wollo region, giving two reasons for this (1981). Firstly, farmers in the region experienced a direct entitlement failure, as severe droughts, due to poor rains in mid-1972 and early-1973, caused crop failure and stripped farmers of their production-based entitlements. Secondly, other classes in the area suffered of trade entitlement failure, due to “a collapse of income and purchasing power and of the ability of the Wollo population to attract food elsewhere in Ethiopia” (Sen, 1981: 94). Sen identified nomadic pastoralists as the main occupational group of victims during the famine, owing in large part, aside from livestock deaths, to the declining exchange entitlements of their animals in relation to purchased grains (1981).
Overcrowding and poor sanitation characterised the relief camps set up by local initiative, but having received aid from abroad two months too late, the size of the camp had already decreased by three-fourths (Sen, 1981). All in all, an estimated 50 000 – 200 000 people died during the famine (Sen, 1981: 86).
What should be done to prevent entitlement failures from happening then? While recognising that “[t]here is no ‘magic bullet’ to deal with the entrenched problem of hunger in the world”, Sen gives two options, the first of which, increasing purchasing power, is of less interest in this essay (Sen, 2002).
Sen places great emphasis on the creation of a conducive democracy and a free press. To tackle the rooted problem of hunger, in Sen’s words,
“it requires political leadership in encouraging democratic governments in the world, including support for multiparty elections, open public discussions, elimination of press censorship, and also economic support for independent news media and rapid dissemination of information and analysis” (2002)
Devereux provides a rationale for this line of thinking, because “[i]f market agents…have good information about future food availability, then their decisions will be rational and prices will be accurate” (2001a: 130). For prevention and prediction, a free press could in theory channel the problems facing famine victims to a wider audience, sparking a need within a democratic government to act against such human misery.
Since its conception in the beginning of the 1980s, the EA has been subject to immense scrutiny however. Focus has particularly been on its representation of famine, the theoretical groundings of the approach, and the economic framework it operates in.
The concept of famine
For any study of famine to be valuable, the subject in question needs to be properly assessed. As said in the Introduction, ‘famine’ is a multi-layered concept with almost as many definitions as there are studies of it. Especially the time scope of famine is contested. While the EA considers famine to be a discrete event, a one-off occurrence of entitlement failure, Alemu warns of the dangers of such a positioning, “not only for its explanation but also for designing policies to deny famine a future” (2007: 95). This is because an events-based approach to famine fails to take account of the specific historical background, with its social, political and economic realities often forming the basis for an understanding of individuals and communities facing famine. More in line with an ‘insiders’ account then, famine in academia is nowadays more often defined as a process, implying extension and intensification of normal practices (Watts, 1991). Sen points out that, “[f]or entitlement in a multi-period setting the initial formation of the problem would require serious modification and extension” (1986: 10 in Devereux, 2001b: 249). This confirms that EA is not suited for a long-term analysis of famine, hindering its usefulness for prevention and prediction
The causes of famine
Linked to the previous discussion, there is disagreement over the causes. While FAD takes a supply-side approach, and EA takes a more demand-side approach, the truth may lie somewhere in between. Though incorporating it as a ‘special case’ of EA, Sen did not only seek to refute FAD (Fine, 1997). He further ignores or neglects views about poor infrastructure, which often make it hard for food to come to an area in times of famine, and about the institutional setting in Africa, more often associated with high risks than transparency and accountability (Alemu, 2007). In focusing on entitlement collapses and market failures brought about by economic ‘shocks’, EA is only capable of accounting for some proximate triggering causes of famines, without dwelling much on underlying social and political factors, that this essay will further deal with later.
The straightforward way in which EA presents famine results in an oversimplification of the process, where even Sen himself recognises some limitations. First, he notes that people may not always consume as much food as their entitlements would allow, e.g. due to fixed food habits or a choice to starve over selling productive assets (Sen, 1981: 50). Devereux finds that these latter ‘coping strategies’ – of protecting assets for the future and letting the weakest and most dependent family members die by denying them food – often characterise famine responses, and may in fact be rational decisions, partly explained by power relations within a household (2001b: 249-250). This is problematic for EA, which seemingly makes an “implicit assumption that a food shortage triggers an automatic behavioural response…dominated by the search for food”, and can only explain deviances of this in terms of the ‘passivity’ of victims (Devereux, 2001b: 249; Fine, 1997).
Furthermore, while Sen often views famine mortality as triggered by an entitlement collapse followed by hunger, starvation and ultimately death, he also has reservations about the cause of death during famines, as starvation may be ‘triumphed’ by epidemics (1981). Disassociating starvation from famine however risks missing that, starvation being a prolonged process of hunger, famine occurs due to changing entitlements and vulnerabilities of people over time (Alemu, 2007). To be able to prevent mass deaths from occurring, it is important to know who is deprived, who is dying and why. For this purpose, EA seems useless, as it “seeks to analyse how famines happen; it does not claim to explain why they happen” (Devereux, 2001b: 131).
Theoretical limitations of entitlements
Apart from conceptual difficulties, Sen’s analytical framework may be questioned, as he himself notes that “there can be ambiguities in the specification of entitlements” (Sen, 1981). The ‘fuzziness’ results from problems relating to explaining socially-determined entitlements and property rights within communities, as Sen only focuses on one unit of analysis, the individual or representative household (Devereux, 2001b). Due to this ‘methodological individualism’, the EA is “analytically weakest in precisely those socioeconomic contexts for which it was designed, namely, famine-prone communities…where common property and open access regimes for resources dominate private property and market-based exchange” (Devereux, 2001b: 255).
Though recognising that entitlement relations may be more complex than the four sets he explores, Sen nevertheless ignores ‘extra-entitlement transfers’ that occur outside a society’s legal frameworks (1981: 49). This is a significant omittance many EA critics have used as evidence of the approach being generally un-applicable to famines in Africa. Essentially, it comes down to the lack of understanding of “the historical political economy by which certain sorts of entitlements come to be socially distributed” (Watts, 1991: 21). Also, ignoring extra-legal entitlements, especially during war, further ignores the very political nature of most current famines. The “disruptive effects” and “unruly practices”, such as those of looting and brigading, make EA inefficient in preventing and predicting ‘war famines’ (Devereux, 2001b). This is because of its very focus on economic factors, entitlement and market failures in otherwise perfectly functioning markets, lending little room for famine explorations in conflict situations within the household and in the wider community.
In fact, Fine argues the EA is not a theory of famine causation at all, but merely an approach (1997). While Sen also argues that EA is “a general framework for analysing famines rather than one particular hypothesis about their causation”, attempting to refute FAD theory implicitly places EA in an equivalent position (Sen, 1981: 162 in Devereux, 2001: 247). According to Fine, Sen “seems to overlook the fact that differences in method are as contested as approaches to famines themselves”, highlighting the importance of a clear terminology if there is an aim to extend or criticise a theory (1997: 623). In short, while EA explains some economic factors of famine, it can not encapsulate the complexity and diversity of livelihoods found in sub-Saharan Africa, and their proneness to famine.
With a poor theoretical background, or as Fine has argued, no theory at all, it is difficult to provide helpful guidelines for famine-prone countries to follow. Indeed, Sen’s conclusions for famine-prevention – especially of fostering democracy and a free press – as well as empirical causalities are subject to much criticism.
Keen “suggests a need to investigate precisely how particular democracies function and which groups they may fail to protect” (1994: 212). Drawing on examples from the South-western Sudanese famine, from 1983 to 1989, Keen shows that in more than one respect, state, central and local elites did little to stop famine from progressing and even benefitted from a prolonged starvation of the peoples of Darfur (1994). During this time period the government in Khartoum saw the prospect of defeating rebels and accessing valuable resources in the affected area, the central government “manipulate[d] civil society for its own purposes”, and some local groups reaped economic benefits – all through exploiting the vulnerability and political powerlessness of those, often poor groups of society, subjected to famine (Keen, 1994: 211.212). As such, the umbrella term ‘democracy’, which Sen so freely uses as a necessary factor for the prevention of famines, is unsatisfying and misleading.
On the matter of free press, starving children continue to pose in newspaper images – how helpful these images are, can be contested. Appealing to the rise of humanitarianism in the ‘North’ and the focus on a ‘right to food’, news stories serve to fuel the dedication ‘humanitarian internationals’ have to relieving famine through food aid (de Waal, 1997). While humanitarianism does not prevent famine, and can sometimes act as a restriction on other efforts to relieve it, de Waal still finds free press important in providing a platform for famine victims to voice their concerns (1997: 5). Nevertheless, even a reporter at one of Britain’s largest daily newspapers, Guardian, has become aware of the transition from ‘foreign reporting’ to ‘disaster journalism’, describing it as a
“quick search for the most tragic story, the most dramatic images, shown without context, scant explanation of why the disaster happened, and with no background information on how that country or community became so troubled in the first place” (George, 2009).
In this context, it is clear that a free press alone cannot act as a preventative mechanism, as the causes of famines are often overlooked, nor can it be useful for predicting famines, as news by their very nature appear when something has already happened.
Re-visiting the Ethiopian famine 1972-74
Seeing that there are a number of problems related with EA, it may be useful to quickly re-examine Sen’s Ethiopian case study. Alemu argues the ‘normal level’ of food production Sen uses as a milestone for refuting FAD should be examined, as it is found that starvation levels were high and calorie intakes were low already before 1972 (2007: 118). Devereux also points to a “possible theoretical contradiction in Sen’s analysis” (1988: 275). Since Sen argued pastoralists in Wollo faced a direct entitlement failure due to loss of cattle, and an exchange entitlement failure resulting from it, then how can a fall in market price for livestock, indicating excess supply, be explained? According to Devereux, this could be due to a failure to separate market from total food demand (1988). Furthermore, in contrast to Sen, even price rises in Wollo may have made a significant contribution. Drought-induced falls in output accompanied by withholding of surpluses, perhaps as a coping strategy for future survival, along with increased market dependence of food producers helped spread price rises to neighbouring regions, suggesting that “certain famines produce predictable food ‘price ripples’, which both are caused by the crisis and contribute to it” (ideas by Seaman & Holt in Devereux, 1988: 275-276). If EA fails to pay attention to the supply-side, its predictive power is severely limited.
While EA has shed light on proximate triggers of some famines, as Devereux argues, “Sen’s approach is significantly weakened, both conceptually and empirically, by its methodological individualism and by its privileging of economic aspects of famine above sociopolitical determinants” (2001b: 245). Largely dismissing the supply-side further leads to a dismissal of many factors relevant for famine analysis, while the strict legal entitlements makes explaining community ownership and war-time violations problematic through EA. Furthermore, the focus on single events instead of a prolonged process of economic as well as social and political vulnerability and proneness to famine has undermined EA’s preventive and predictive qualities. It is also interesting to note that even by effectively ignoring the political embeddedness of famines, Sen’s preventative responses rest on two very political aspects, democracy and a free press, which, it should be noted, are rarely completely free from vested state and corporate interests.
As has become clear in this essay, more than one factor affects how, why and who is affected, as “[t]here is no single theory which explains every aspect of every recorded famine” (Devereux, 1988: 282). While the outlook for preventing and predicting famine is somewhat dampened by this pessimistic expression, one should not be completely disheartened. However, for more valuable debates, economists need to start looking outside the market, forming other bases for analysis than EA and, rather than exogenising variables clearly related to famines, recognise the importance of social and political factors in determining famine causes and outcomes. It could therefore be argued that the elements Sen neglects in his analysis are of major importance and distort the real reasons underlying prolonged processes of starvation and destitution. Famines have been prevented, predicted and abolished in other parts of the world; the key lies in understanding the specific characteristics that makes it so persistent in Africa.
Alemu, G. (2007) ‘Revisiting the Entitlement Approach to Famine: Taking a closer look at the supply factor – critical survey of the literature’ in Eastern Africa Social Science Research Review. Vol. 13, No. 2. June.
Devereux, S. (1988) ‘Entitlements, availability and famine – A revisionist view of Wollo 1972-74’ in Food Policy. August.
Devereux, S. (2001a) ‘Famine in Africa’ in Food Security in Sub-Saharan Africa. Eds. Devereux, S. & Maxwell, S. ITDG Publishing: London.
Devereux, S. (2001b) ’Sen’s Entitlement Approach: Critiques and Counter-critiques’ in Oxford Development Studies. 29, 3.
George, S. (2009) ‘What is Development Journalism?’ in Guardian Online. Accessed 10.12.2009 at http://www.guardian.co.uk/journalismcompetition/professional-what-is-development-journalism
Fine, B. (1997) ‘Entitlement Failure?’ in Development and Change. 28.
Keen, D. (1994) ‘Chapter 6 Discussion and Conclusions’ in The Benefits of Famine: A Political Economy of Famine and Relief in Southwestern Sudan, 1983-89. Princeton University Press: Princeton.
Sen, A. (1981) Poverty and Famines – An Essay on Entitlement and Deprivation. ILO, Oxford University Press: New York.
Sen, A. (1986) Quote from ‘Food, economics and entitlements’ in Lloyds Bank Review. 160.
Sen, A. (2002) ‘Why Half the Planet is Hungry’ in Observer of London. June 16. Accessed 12.12.2009 at http://www.globalpolicy.org/component/content/article/211/44284.html
Waal, A. de (1997) ‘Introduction’ in Famine crimes: Politics and the Disaster Relief Industry in Africa. James Currey.
Watts, M. (1991) ‘Entitlements or Empowerment? Famine and Starvation in Africa’ in Review of African Political Economy. No. 51. July.