Chair:

Simon Maxwell, ODI

Speakers:

Stephen Devereux, Institute for Development Studies

Ed Clay, Overseas Development Institute

Louise Bohn, University of East Anglia

John Seaman, Save the Children Fund (UK)

Jonathan Kydd, Imperial College

Megan Vaughan, University of Oxford

Discussants:

Graham Stegmann, Department for International Development

Richard Dowden

The wide-ranging discussion following the presentations highlighted the problems caused by failures in the regions early warning systems; the conflict between donors, governments and NGOs objectives; the deteriorating governance and accountability of key institutions; the failures of agricultural input and output markets post-liberalisation; and the underlying vulnerability of many households as a result of HIV/AIDS. Further detail on these issues is given in the topic-based summary below. This summary cannot reflect the full breadth and diversity of the discussion, nor any consensus; rather it highlights some of the main issues raised.

Key questions were asked about the failure to learn from the lessons of past food security crises in the region and elsewhere, in terms of establishing effective early warning systems and effective food reserve management. The need to improve the governance and accountability of all agencies (government, IFIs, donors, UN institutions, and private sector actors) was noted, as was the role that effective civil society organisations can play in contributing to this. The absence of a sound agricultural growth model for the region was noted, and was attributed to failures in agricultural service delivery, the unreliability of output markets and lack of economic integration, which have not been addressed by liberalisation. It was emphasised that there is an urgent need to address the mounting vulnerability of the population in the region, who are rapidly losing the ability to cope in the face of HIV/AIDS and market failures. This is likely to involve a combination of effective safety nets (for asset building as well as immediate consumption) and market interventions.

It was questioned why so much relevant research-based policy analysis exists relating to all these issues but is not being used to implement practical changes. A number of participants highlighted the value of continuing the expert informed dialogue over the coming months.

1. Rural poverty

Key issues discussed:

livelihoods, diversification, decentralisation

Recent research evidence suggests that the current situation in rural Malawi, where three quarters of households corresponded to the poorest households in Tanzania and Uganda, is due to an erosion of Malawis assets over the last 10 years. (The food shock of 2001 is relatively small compared to 1991.) For example, people have been selling their assets to pay for caring for and treating HIV/Aids victims. Livestock are no longer used as a buffer against bad times because security at the village level, both personal and commercial, has dramatically reduced as a result of economic and political liberalisation. There has also been an erosion of common pool assets such as woodland and wild resources. This combination of factors means that rural Malawi finds it difficult to respond to shocks, more so than before.

It has been found that those people better off in rural areas in Malawi (and other countries touched by the current crisis) have off farm income sources/diverse livelihoods. Generally livelihoods are not built on rural economy alone, nevertheless agriculture is an important motor of rural economy. The rural growth linkage model argues that in poor rural economies agriculture is the best, and sometimes only, engine of growth, but it is difficult to see how agriculture can grow in countries like Malawi, given land constraints and current technologies and ineffective services, and it is not clear that sufficient farm-non-farm linkages exist. More research is needed.

Decentralisation is being promoted throughout the region. A side effect of this is that district authorities are desperate to earn tax revenue. There have been dramatic increases in taxes, and these have serious disincentive effects for local trade and the growth of rural non-farm enterprises.

2. Liberalisation

Key issues discussed:

Institutions and food reserves

The discussion first focussed on strategic grain reserves. The successful operation of conventional SGRs is hampered by short-term considerations relating to financing, and ideological concerns regarding non-intervention. Either these need to be resolved, allowing a longer-term, more strategic management perspective (e.g. planning over a 5-10 year horizon rather than 1-2 as at present), or SGRs operations should be limited to distribution of immediate food relief, not market manipulation. The significance of long-term fluctuations in food supplies is well known but it appears that it has been overlooked in the region in recent years. For example, the IMF work from a short-term financial not a medium-term food security perspective and gave the right advice re: the Malawi SGR from the former perspective.

If SGRs continue to be used for longer-term strategic food stocks management, then they need to be taken out of individual governments regular budgets and not be expected to be self-financing. And a regional perspective is needed.

SGRs could take advantage of the more sophisticated financing instruments available regionally these days, e.g. via the SAFEX futures market, including call options, which allow rights to maize every year to be secured. No longer simply a question of buying and selling grain in real time. Zambia has been recommended to have a strategic reserve of dollars instead of maize, although this has yet to materialise.

Questions of accountability were raised. There is a major role for private players in the regional grain market, but SGR operations must be accountable to and funded by public institutions. The political nature of decisions about stock sizes was recognised. Should reserves be controlled by the UN/donors?

Better to split the issue into short and long run considerations? For example short run mechanisms to deal with occasional problems that did not interfere with longer run management. Would this disempower national governments? The issue is do we want to stop bouts of famine, or look at food reserves as part of a longer-term development process?

The point was made that market liberalisation in general has undermined accountability. But how to improve accountability? A return to the state intervention of the 1970s and 1980s? What is different this time, that means increased state intervention will improve governance and accountability?

A note on SGR management in Malawi in 2001-02: it was suggested that the grain reported to be missing by the media was in fact used in 2001 (the first import of maize did not occur until November 2001). There was no secrecy: government budgets are published and donors have to be informed. The National Food Reserve Authority transferred the maize to ADMARC, for people in rural areas to buy as shortages were already being felt.

3. Early Warning:

Key issues discussed:

Data, crop figures, micro/macro level data, variables used as early warning signals, performance tables and best practice.

Many early warning systems do not produce data that is useful for monitoring food security:

• Most early warning systems look at maize production or rainfall. As a result there is not enough reliable information on crops other than maize, which can be very important for food security.
• Many early warning systems, such as FEWS-NET, focus on macro data and do very little ground-truthing. People are not getting out into the villages. Based on available macro data, Malawi is one of the top 10 performing economies in the developing world, but in terms of livelihoods on the ground this should be questioned.
• Early warning systems may not collect information on food prices, so difficult to monitor entitlement (access) rather than food availability. The SCF household food economy approach has much to offer in this context.

Also, early warning systems often operate far below professional best practice. At present many governments in Southern African do not have the capacity for a well-functioning statistics system. Liberalisation has led to contracts for data collection and analysis being put up on a 2/3 years basis, so there is little institutional learning. Such turnover is too high to help Africa. It has been very difficult to implement minimum standards for best practice, despite the technical information existing. Donors are as much to blame as anyone.

In addition, there are conflicts of interest. In one case, funding for the early warning system came from a donor also promoting increased production of food security crops. The early warning system data showed an increase in production of these crops that had not happened on the ground, resulting in fundamental misinterpretation of the current food security situation.

NGOs often have micro level data that is relevant. But because it is small-scale and "soft" data, it is not believed by decision makers. Joint missions comprised of all stakeholders may improve ownership and result in appropriate action.

Early warning systems are better developed and maintained in countries with regular food security problems, e.g. Ethiopia.

4. Safety nets and social protection

Key issues discussed:

Attention to safety nets, types of safety nets, government intervention and land ownership

Intervention at which stage in the marketing chain? The objective should be to make farming profitable and enable the necessary transactions (input procurement, output sales).

In this context, safety nets should not just focus on Aids orphans (which is most prevalent today) but also on access to inputs, and dealing with climatic variability.

A range of safety net options were discussed:

• Food subsidies. ADMARCs use of cross-subsidies in Malawi in the 1970s ensured adequate food production but rates of child malnutrition indicated it did not solve problems of access. But at what level between import or export parity? There is a large literature on different countries experience, e.g. Indonesia. Price stability rather than price levels is the key. In the absence of reliable food markets, there exists tension between setting prices high enough for farmers to invest and low enough so that farmers dont believe they will starve if they dont grow maize price setting must be made from both these points of view.
• Fair price shops. In the past, they only worked where there are the resources to control them otherwise they become a mechanism for targeting a limited number of (usually privileged) people. How will it be done differently if done again? Like school feeding? The 1994 famine in Bangladesh proved fair price shops do not work.
• Access to inputs needs to be restored but not by going back to the old institutions. Starter packs of seed and fertiliser have been tried in Malawi and in two years of good rains are estimated to have increased maize production by 25%. May be more cost-effective than fertiliser subsidies, especially in countries with porous borders, or than importing food, but what are the longer-term sustainability issues? Impact on private sector input suppliers; relevance to all farmers. Is fertiliser-for-work a better option?
• Technology. There may be potential for improving food security through technological changes, for example the proposal by the CG Centres Future Harvest programme to promote a "wide food band" of a range of drought-tolerant crops (e.g. cassava) and fertilityenhancing crops (e.g. legumes).
• Different types of safety net are needed for different farmers. For example, due to land pressure and other resource constraints, farmers in Southern Malawi have very little potential for increasing yields in response to any agriculturally-based incentive. Interventions to support off-farm activity may be more appropriate, especially as most of these farmers are net purchasers of grain. Existing safety nets fail to take this into account
• Off-farm options. In one pilot project in Malawi, farmers were offered vouchers redeemable against cash purchases or a starter pack all bar one made purchases rather than taking starter packs. A majority of these bought soap, which deteriorates very little over time and can be used flexibly, either to sell for cash or for domestic use.
• School feeding programmes: these have the potential to build human capital at the same time (by encouraging families to keep children in school), and thus contribute towards a movement out of agriculture over the longer-term. There was disagreement about this as it has not worked elsewhere.