Andy Catley is Research Director at the Feinstein International Center, Tufts University, Addis Ababa, Ethiopia. Adrian Cullis is Drought Risk Management/Emergencies Programme Coordinator at the Food and Agriculture Organization, Ethiopia.  

Introduction

For many years the deliberate removal of livestock from pastoralist areas of Africa during drought has been suggested as an appropriate drought response (e.g. Toulmin, 1995). Drought-related purchase of livestock and distribution of dried meat was used in pastoral areas of Mali in the 1980s (Oxby, 1989) and more recently, destocking was used in northern Kenya (Aklilu and Wekesa, 2002). Destocking and other types of livestock-related drought assistance fit well with the concept of saving lives and livelihoods. When viewed from a livelihoods perspective, destocking is a way to exchange some animals for cash, thereby giving pastoralists the cash they need to buy food, maintain a core herd and access the services they want (rather than the services aid agencies provide). This herd maintenance might involve purchase of fodder or veterinary care, thereby supporting local markets and service providers. Over time, two specific types of destocking have been recognized (LEGS, 2009). “Commercial destocking” refers to the purchase of animals by traders, and assumes that animals are still in a reasonable condition for sale and transport at relatively early stages of a drought. In contrast, “slaughter destocking” occurs later in drought, when animals are no longer in reasonable condition and therefore, not purchased by traders. As such animals are still fit for human consumption, slaughter destocking leads to the distribution of fresh or dried meat to selected households, and sale of hides and skins. Both approaches to destocking lead to cash transfers to pastoralist households during drought.

The emergence of yet another humanitarian crisis in the Horn of Africa in 2011 has led to a new round of reviews and evaluations of drought response programs. Within these programs, livelihoods support to pastoralist communities is now common, but accounts for only a small proportion of overall expenditure relative to food aid. For example, in Ethiopia in the drought in 2011 total aid including food aid was valued at US$800 million, while livelihood support totaled US$17.8 million (Anon, 2012) or 2.2 per cent. In addition, the distinction between food aid and livelihoods inputs is not always clear because food aid can be used by pastoralist households directly or indirectly to support livelihoods (Bush, 1995). In southern Ethiopia for example, some households reported using food aid as a supplementary feed for livestock during drought (Abebe et al., 2008).

In terms of livelihoods-based post drought recovery, the theory of drought cycle management indicates approaches such as restocking should only be needed if earlier livelihoods inputs were not implemented or where unsuccessful, as these earlier responses often aim to protect core livestock assets. Similarly, restocking can follow food aid provision as a stand-alone response that does not explicitly aim to protect assets. Various evaluations and studies on restocking in pastoralist areas are available, and provide lessons on issues such as the type and number of animals needed per household, and the extent to which restocked households reduce their reliance on food aid over time (Lotira, 2004; Wekesa, 2005). A longstanding issue with restocking is the relatively high aid cost per household, which relates to both the cost of the initial transfer of livestock and the cost of additional support such as food aid and other inputs until a viable, productive herd develops (LEGS, 2009).

Although the theory and practice of both food aid and livelihoods-based programming are now well known, relatively limited information is available on the relative costs and benefits of these two broad approaches. This paper provides a cost comparison of the different approaches by drawing on the impact assessment of the commercial destocking program in southern Ethiopia (Abebe et al., 2008) and using additional expenditure figures from the implementing agency, Save the Children US (SCUS). The destocking costs are compared to the cost of local and imported food aid, also by SCUS, followed by restocking of pastoralist households. Restocking costs plus additional food aid costs during herd growth were estimated from an evaluation of a restocking project in Ethiopia by Save the Children UK (Wekesa, 2005).

Commercial destocking in Ethiopia

In early 2006 commercial destocking was used in pastoralist areas of southern Ethiopia following weak rains in 2005.  The timeline of events is summarized in Figure 1, and shows that destocking took place relatively late in the drought, after substantial livestock deaths had been reported. Despite this, traders and pastoralists agreed prices for the sale of drought-affected cattle, and cattle purchases took place over about four weeks in February 2006. Private traders used their own capital for most of these purchases, and covered their own transport and other costs. The role of NGOs was first to liaise with government to ensure support for the approach, and then introduce the traders to the drought-affected areas. More details of the commercial destocking activities are provided in Box 1. 

In late 2006, an impact assessment of the commercial destocking was conducted, focusing on potential livelihoods impacts in households where cattle were sold (Abebe et al., 2008). The assessment comprised two main activities, being interviews with the private traders who purchased cattle, and the use of participatory methods with a random sample of 114 households that sold cattle. This sample was thought to represent 20% of all 570 households involved in the project.

Interviews with traders indicated that they had purchased at least 20,000 cattle, resulting in cash transfers to pastoral households valued at approximately US$ 1.01 million. These figures were used to estimate a basic benefit-cost ratio for the approach, by assuming a benefit of US$ 1.01 million and comparing this to the aid costs incurred by the NGO implementer, Save the Children. The result was a benefit cost of 41:1, while also noting that the activity was implemented late and an earlier response would probably have achieved an even greater benefit-cost (Abebe et al., 2008).

The participatory impact assessment with pastoralist households showed that while 28 per cent of income from destocking was spent on human food, 37 per cent of income was used to protect livestock through motorized transport of animals to distant grazing areas, livestock feed, and veterinary care  (Abebe et al., 2009).

 The comparative model

We developed a simple comparative model to compare two scenarios. The “food aid plus restocking” scenario involved provision of food aid, substantial livestock herd depletion, followed by restocking. The “commercial destocking” scenario involved a timely commercial destocking program at the onset of drought, without the need for food aid or restocking. Each scenario is explained in further detail below.

Food aid plus restocking scenario – after the failure of the long rainy season in 2005, a pastoralist household faces destitution as the next rainy season, in late 2005, also starts to fail. Grain prices are rising and as livestock lose body condition, their market value starts to fall. A few weeks pass and stories of livestock deaths and drought prompt UN agencies and NGOs to conduct drought assessments. Some weeks later, bags of emergency food aid are appearing in towns and villages, and the household distributions begin. By this time, the pastoralist has slaughtered his calves to try to protect their mothers during the drought, but now the cows are also dying. After another two months, his 20 core breeding stock are decimated. A month later, some light rain falls and a “drought recovery” program is designed with the objective of “returning destitute pastoralists back to a sustainable livelihood”. The pastoralist is restocked with 30 sheep and goats, and given more food aid.

Commercial destocking scenario – after the failure of the long rains in 2005, humanitarian agencies are carefully tracking rainfall in late 2005. As drought begins, they contact livestock traders and support events where the traders meet pastoralist leaders and discuss if and how cattle and other livestock might be purchased. A commercial destocking program quickly evolves within days, with the traders using their own cash and trucks to buy livestock from drought-affected areas. The pastoralist gets involved and sells two young bulls. The prices are not high relative to the best market values in a good year, but with the income he knows that he can buy enough maize to feed his family of six people for two months, while also spending additional income to protect his best breeding cows. His decision is based on the fact that he can sell a young bull for Ethiopian birr 440 (US$50, at 2005-6 drought prices), and with that money, can buy 200kg of maize (cost Ethiopian birr 160/100kg). Also, he knows that each person will eat around 0.5kg of maize a day and so for two months, he will need 180kg of maize for the family. From the sale of the other bull, he can buy some livestock feed and transport some cows to a better grazing area. At the end of the drought, the core breeding stock are alive, the household has not received food aid, and a restocking program is not needed. The main assumptions and variables used in the model are presented in Box 2.

Cost comparison

Using the scenarios and assumptions above, costs are summarized in Table 1.  The results in Table 1 indicate that for support that meets basic food security objectives over two months, and before restocking, the provision of local food aid cost 17 times the cost of commercial destocking (b/a, Table 1), whereas the provision of imported food aid cost 29 times the cost of commercial destocking (d/a, Table 1). If the costs of restocking are included in the analysis, local food aid plus restocking cost 125 times commercial destocking (c/a, Table 1), whereas imported food aid plus restocking cost 137 times commercial destocking (e/a, Table 1).

Discussion

The scenarios we developed for the cost comparison involved assumptions and estimates, leading to an imperfect analysis. However, the scale of the difference in costs between commercial destocking and food aid plus restocking indicates that minor errors in the model are unlikely to alter the overall result viz. that a food aid plus restocking approach is vastly more expensive than early response with commercial destocking. Also, we did not include in the analysis the potential economic benefits of commercial destocking such as support to local markets and services when people use the income from destocking to buy the items and services they need (Abebe et al., 2008). In contrast, food aid may also lead to livelihood benefits and assist asset protection, but in the absence of evidence that quantifies these benefits, the use of food aid for asset protection is difficult to justify.  It is also evident that during severe drought, high levels of livestock losses occur during food aid provision. At present, food aid dominates drought response whereas our analysis indicates that the balance of funding needs to shift radically towards early drought response and approaches such as commercial destocking. The destocking experience in Ethiopia indicates that pastoralists will sell animals during drought if a reasonable price is offered and when cash payments are made promptly by buyers.  Also, livestock prices in the 2011 drought were substantially higher than 2006 as more traders were operating in the area. There is still a role for food aid during drought, but relative to livelihoods-based responses pastoralists in Ethiopia saw this role as benefiting mainly poorer households with few or no livestock (Abebe et al., 2008).

The further use and adaption of the simple modeling method used in our analysis would help donors and aid agencies better understand the economic implications of late response and asset depletion in pastoralist areas relative to enabling asset protection during drought.  However, the analysis of food aid benefits and comparison with other types of support is now further complicated by the emergence of safety net programs in pastoralist areas of Kenya and Ethiopia (Devereux and Tibbo, 2012). Although these programs focus on cash transfers, food transfers are also used in some areas and may be preferred when for example, local markets are disrupted or not functioning, or when cereal price inflation is high. Food or cash transfers may be expanded during drought in communities that are already receiving a safety net, thereby hindering analysis of specific drought-related support compared to regular support. Although safety net programs may have objectives related to food security, asset protection and asset building, analysis of cash transfers in pastoralist areas indicates that the level of transfer is too low to achieve meaningful asset building, and almost by definition, very poor or destitute pastoralists have no assets to protect (Catley and Napier, 2011). It follows that food aid transfers of an equivalent value to cash transfers are likely to achieve comparable impacts, being mainly food security impacts.  A more rigorous analysis could use the type of model described in this paper, and compare the cost of safety net provision of food or cash, with the cost of commercial destocking during drought.   For example, assuming that a safety net is continued to our model pastoralist household of two adults and four children during two months of a drought, how does the cost compare to the cost of commercial destocking at US$4.3 per household?

At first sight, the benefits of commercial destocking point to a need to strengthen livestock marketing in pastoralist areas, as a robust trade and active private sector enables commercial destocking. In the case study in southern Ethiopia in 2006, a growing formal livestock export trade was an important factor for encouraging private traders to invest in destocking (Abebe et al, 2008). However, the links between livestock trade, strengthening of pastoralist livelihoods and drought response are complex, not least because ongoing commercialization of livestock systems in pastoral areas benefits wealthier herders far more than more vulnerable, poorer herders. In high-export areas of southern Ethiopia and Somali areas, commercialization is seen by some analysts as a key driver of pastoralist destitution and explains the apparent contradiction between increasing numbers of destitute pastoralists with simultaneous growth in livestock exports (Catley and Aklilu, 2012). Critically, commercialization results in larger herds for the wealthy and smaller herds for the poor, and an increasing asset gap. For drought recovery programs, these trends indicate that a pastoralist household in a commercializing system needs far greater livestock holdings than the “minimum herd size” used by some restocking projects, and clearly, not all households can achieve a commercial herd size. It follows that approaches such as restocking and safety nets need to recognize that not all pastoralist households with very low or no livestock can, or want to, return to pastoralism.  For NGOs and donors working in high-export pastoralist areas, the challenge is to integrate these issues into long-term development strategies, and structure strategies according to pastoralist wealth groups and commercialization trends. A hard reality seems to be that while commercialization and population growth trends have been evident for decades and contribute to vulnerability, among aid actors drought is the more noticeable event and so responses are framed narrowly around dealing with drought. However, drought recovery and development programs need to consider whether a return to pastoralism is really viable for many poorer households given the competition they will face from a commercializing, wealthy and well-connected sector within pastoral areas. A further consideration, given this context, is how to reshape the use of food aid to complement coherent, long-term poverty reduction strategies in these areas (Barrett and Maxwell, 2005).

Overall, the experience in Ethiopia point to three main conclusions. First, in a context of drought occurring in a wider context of a dynamic livestock export trade, commercial destocking is a very effective way to support pastoralist livelihoods. Second, the wider use of commercial destocking and aid investments in livestock export markets needs to take account of wealth differentiation in pastoralist areas and how wealthier households tend to benefit more from exports than poorer households. Third, far more evidence is needed to show the impact of livelihoods-based drought responses, and especially, impact assessments and economic analysis involving comparison of different approaches with food aid. Such assessments need to take account of the repeated cycles of food aid, provided over decades, and the long term costs of food aid and attempts to rebuild pastoralist herds.        

References

Abebe, D., Cullis, A., Catley, A., Aklilu, Y., Mekonnen, G. and Ghebrechirstos, Y. (2008).Livelihoods impact and benefit-cost estimation of a commercial de-stocking relief intervention in Moyale district, southern Ethiopia. Disasters 32(2), 167-189.

Aklilu, Y. and M. Wekesa (2002). Drought, livestock and livelihoods: Lessons from the 1999-2001 emergency response in the pastoral sector in Kenya. Humanitarian Practice Network Paper 40. Overseas Development Institute, London.

Anon (2012). Disaster Risk Management Agriculture Task Force briefing paper 10. Ministry of Agriculture, Addis Ababa.

Barrett, C.B. and Maxwell, D. (2005). Food Aid After Fifty Years: Recasting its role. Routledge, London and New York.

Bush, J. (1995). The role of food aid in drought and recovery: Oxfam’s North Turkana (Kenya) Drought Relief Programme, 1992-4. Disasters 19(3), 247-259.

Catley, A. and Aklilu, Y. (2012). Moving Up or Moving Out? Commercialization, Growth and Destitution in Pastoralist Areas. In: Catley, A., Lind, J. and Scoones, I. (eds.), Pastoralism and Development in Africa: Dynamic Change at the Margins. Earthscan, Oxford, in press.

Catley, A. and Napier, A. (2010). Rapid Review of the Cash-for-Work and Natural Resource Management Components of the RAIN Project. Feinstein International Center, Tufts University, Addis Ababa.

Devereux, S. and Tibbo, K. (2012). Social protection for pastoralists. In: Catley, A., Lind, J. and Scoones, I. (eds.), Pastoralism and Development in Africa: Dynamic Change at the Margins. Earthscan, Oxford, in press.

LEGS (2009). Livestock Emergency Guidelines and Standards. Practical Action Publishing, Rugby.

Lotira, R. (2004). Rebuilding herds by re-enforcing gargar/irb among the Somali pastoralists of Kenya: evaluation of experimental restocking program in Wajir and Mandera Districts of Kenya. African Union/Interafrican Bureau for Animal Resources, Nairobi and Feinstein International Center, Tufts University, Nairobi.

Oxby, C. (1989). African Livestock Keepers in Recurrent Crisis: Policy Issues Arising from the NGO Response. International Institute for Environment and Development, London.

Toulmin, C. (1995). Tracking Through Drought: Options for Destocking and Restocking. In: I. Scoones (ed.) Living With Uncertainty: New Directions in Pastoral Development in Africa. Intermediate Technology Publications, London.

Wekesa, M. (2005). Terminal evaluation of the restocking/rehabilitation programme for the internally displaced persons in Fik Zone of the Somali Region of Ethiopia. Save the Children UK, Addis Ababa and Acacia Consultants, Nairobi.

Acknowledgements

This study was funded by the United States Agency for International Development under a Cooperative Agreement to Tufts University for the Pastoralist Livelihoods Initiative Policy project in Ethiopia. The views expressed in the paper do not necessarily reflect of the views of USAID.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>