Speakers:
Stefan Dercon, University of Oxford
John Farrington, Overseas Development Institute
Chair:
Matthew Greenslade, DFID

John Farrington began by pointing out the different perspectives of those concerned with agricultural growth, and those with social protection. The former come from a background in biological sciences or agricultural economics, and focus on the agriculture sector; the latter from social welfare concerns, and focus on the individual household. There are clear links between the sector and household, above all in flows of labour.

Assets matter: they are closely linked to levels of vulnerability.

Unlike production, the aims of protection are to reduce the variance of both household incomes, preferably while there is a rising trend, and of costs to households, preferably within a falling trend.
Threats to the poor can be seen either as shocks, unforeseen surprises that arise at short notice; or as stresses over the medium to long term. The latter are often less well considered in policy thinking than the former.

Current approaches to social protection include: the World Bank's social risk management that emphasises ways that people can recover from shocks; risk reduction; and efforts to reduce vulnerability, largely by building up assets.

Agricultural policy needs to take into account the potential to reduce risks and vulnerability. Investments in rural infrastructure, improving or clarifying the access of the poor to assets not formally owned (forest, grazing land, water), through legislation and regulation such tenure and inheritance rules, minimum wages, standards, etc,; through delivery of services - including those that promote farming systems that do not entail high levels of risk, and the provision of insurance.

Social protection measures that can promote growth include cash transfers, subsidised or free food sourced locally, employment creation schemes, subsidised or free farm inputs, and financial services.

Above all, avoid policy that exacerbates problems: imported food aid that arrives late and swamps recovering markets, the promotion of farming methods and systems involving high costs and where profitability can be highly variable.

In conclusion, he noted the importance of taking a broader view of risk that includes stresses as well as shocks; the need to consider both household and production activities together; to mainstream ideas about protection into production sector policy-making, and to find ways to reduce vulnerability.
Stefan Dercon  Stefan Dercon began by stating that while clearly both agricultural growth and social protection are important for poverty reduction, it would be naïve to expect too much from either.

Agricultural growth matters, but historically reduced poverty takes place as people leave farming. The importance of agriculture varies by context. In resource-rich countries, the challenge is one of managing wealth: agriculture's role might then be a source of subsequent diversification, or a way to distribute wealth. In coastal economies where the opportunity to trade is strong, the role of agriculture may be to allow the transfer of resources out of the sector to allow other sectors to grow. In landlocked countries without good resources, agriculture may be the only option but growth miracles are unlikely to be possible.
Agriculture's potential for growth may be limited by technical innovations often yielding finite increments, rather than steady progression. In addition, agriculture faces limits of capital, the sheer quality of resources and remoteness that limit possibilities in 'poor areas', and high risks.

Much agriculture, especially in the poorer and more marginal areas, confronts substantial risks. As is well known, households manage risk both by ex ante and ex post means for example, by diversification in the former case, and coping in the latter. But these strategies come at high cost to average incomes and welfare. For example, when ICRISAT carried out surveys of villages in southern India in the 1970s and 1980s, they found that the poorer households earned 25-50% per unit value of asset, since they could not risk investing in higher-yielding, but riskier alternatives. Coping does not always smooth consumption: shocks can be severe, and the household barely gets by. So, for example, the droughts of the early 1980s in Zimbabwe hit the nutrition of children under-five so hard that surveys carried out 16 years later showed that they were physically smaller than previous or subsequent cohorts, and earned less.

The point is clear: risk represents for the poor a trap one that prevents them climbing out of their poverty, and one that periodically knocks them back down.

What, then, to do? Will growth help? Yes, but what will be the source of such growth? Does it help to redistribute income? Yes, very much so: this is a case where reducing inequality actually raises efficiency. But in thinking of possible transfers, bear in mind that those trapped by risk are usually deeply poor; while transfers usually do best at raising the marginally poor over the poverty threshold. So the transfers have to be substantial enough to lift people out of the poverty trap. Targeting transfers may help, but a crucial point is that the transfers need to be credible and dependable if those facing risk are to be able to invest in higher-yielding activities, safe in the knowledge that the transfer will protect them from the impact of risks.
Matthew Greenslade

Matthew Greenslade (Chair)  Discussion raised the following issues:
# Does reducing the risks faced by the poorest constitute a point for optimism? Perhaps, but some of the losses reported to risk aversion seem too large.
# What about the impact of risks on schooling? Emergency payment of fees would help. Mexico's Progresa really should have additional school fee payments linked to a neutral indicator, such as the weather.
# Cash transfers: in India the main transfer to the poor is heavily subsidised food, through a system that is so inefficient that every rupee's worth of food costs four or five rupees to deliver. The system survives since it suits wealthy farmers to produce additional food at the taxpayer's expense. Far better would be a cash transfer, such as through universal old-age pensions.
# The non-working poor clearly do not benefits from measures designed to help production: they also need dependable transfers.
# Policy-makers need more evidence of the value of social protection in terms of the potential to stimulate production; and, for the non-working poor, measure of welfare.
# Transfers have to be dependable: field studies in Ethiopia show that food aid allocation is, from the point of the view of the recipients, effectively random, since there is no correspondence in deliveries to hazards faced.
# The poor will need help in demanding their right to entitlements once these are in place.