Michael Lipton, Poverty Research Unit at Sussex University, March 2004
1. The narrative strikes back
In 1990-2003 the Green Revolution, irrigation, land redistribution, and staples yields all slowed down in Asia and, with a few exceptions, failed to reach Africa, where public and private investment in agriculture has dwindled. Globally (and in developing countries outside Asia), public-purpose and poverty-orientated agricultural research has stagnated or worse. In that context, following a sharp fall between the late 1970s and the late 1980s, aid to agriculture fell in real terms by about two-thirds to the late 1990s. i
Is this consistent with aid’s aim: to enhance low-income countries’ efforts to halve dollar poverty in 1990-2015? Staple foods - to which the dollar-poor devote over half their consumption - still come mainly from domestic agricultures. These are the main income source for the dollar-poor: 67 per cent of workers in China, 59 per cent in South Asia, and 60 per cent in sub-Saharan Africa report agriculture as their main income source, and for the poor proportions are even higher.ii Poverty incidence is considerably higher in rural than in urban areas, and highest among the rural majority whose main income source is either hired farm labour, or a very small farm.iii 70% of the dollar-poor are rural, and half are projected to be rural even in 2035; even their non-farm work and income derives mainly from demand by local small farmers.iv
This continuing dependence of the dollar-poor on small-scale farming for food, employment and income means that (a) DFID is right in recommitting aid to agriculture, (b) such aid, in the context of domestic programmes, should support small-farm, labour-intensive (i.e. wage-increasing) development; (c) DFID should press the World Bank to implement its stated intent to do the same, (d) PRSPs need specific small-farm content to succeed. In brief, an assault on poverty that neglects agriculture is foredoomed.
But why reject an opposite inference from the overlap of farming and poverty: that (e) poverty reduction means getting the poor out of agriculture? Because (as in E Asia) that works after, and due to, shared small-farm growth, not before it. Except in a tiny number of entrepot city-states (e.g. Singapore) and mineral economies unusually successful in both ‘poor pressure’ and management of economic surpluses (e.g. Botswana), almost all examples since 1700 of mass dollar-poverty reduction start with rises in income due to higher productivity in small family farms. The alternatives are not credible. (1) ‘Industrialisation first’ has huge capital costs per workplace, and has meant massively distorting subsidies and protection. That package dismally failed to address mass poverty in 1950-75. As Europe in 1740-1900 and Asia after 1980 show, it is fairly late in already successful, agriculture-led development processes before urban industrialisation does much for employment (and poverty reduction). (2) Rural non-farm growth - mainly trade, transport and construction - has in China and SE Asia been the consequence of expanded demand, based on growing productivity, by the nearby agricultural poor. (3) Large-scale, extensive, low-employment farming – as in parts of Latin America and Southern Africa – may achieve growth sometimes, but creates few workplaces, and is linked to persistent land and income inequality - making growth less effective in cutting dollar poverty.v (The proportion of farmland in small farms has risen in most of Asia; such farms remain efficient where land and water are scarce, and labour still plentiful.)
Where small family farms achieved rapid, employment-intensive staples growth in the 1970s and 1980s - in S and E China, in India outside the East-Central ‘poverty square’, in much of SE Asia - the baton of poverty reduction was later successfully passed to cash-crops, rural non-farm growth and urbanisation. Without initial employment-intensive, small-farm staples growth, such alternatives receive from agriculture, not the baton of success, but the rotten apple of failure. Migration or non-farm work, instead of being enabled by success, are forced on the poor as “distress diversification” out of failing staples sectors. That does not cure poverty, but embeds it.
The problems raised by Simon are real, and must be addressed in a strategy for farm development, and within it for aid and DFID. But the problems are just that: problems, to be solved, not arguments against a small-farm strategy as the centrepiece of effective initial poverty reduction.
2. Addressing Simon’s questions – and some others
John Mellor once drily remarked that, having learned the lesson of Asia’s Green Revolution lead areas – that mass poverty can be greatly reduced by rapid, science-based, and employment-intensive farm productivity growth – some donors, in slashing aid to agriculture, showed their determination not to make that mistake again! However, as Simon reminds us, times (science, prices, and much else) have changed. And some areas are less amenable to science-based productivity growth than the Green Revolution lead areas. Yet, even there, initially well-functioning roads, markets, credit, extension, etc. were (while helpful) not essential for mass, pro-poor farm-based progress. It has, however, three essential preconditions:
Disabling of farm incentives: Simon reminds us that open-market, real-terms world farm prices (not a problem-free concept) have been falling for forty years. Donor agencies should do much more to support the winding-down of farm subsidisation in rich countries; not only their farmers, but private-sector researchers too, respond to such subsidies by boosting overproduction, glutting world markets, and slashing developing-country farmers’ incentives. Yes, this makes farm-based poverty reduction harder. However, in huge areas of Asia, the Green Revolution has coped, boosting farm productivity amply fast enough to outweigh the fall in world farm prices. Reduction in domestic farm price extraction by developing-country governments - as they wound down their buy-cheap, sell-dear parastatals - has also softened the blow, to farmers, of falling world farm prices. There is, though, a case for directing much pro-poor agricultural investment (and aid) towards farmers, crops and areas least affected by global price trends: farm families poor enough to eat, or pay workers with, much of the extra coarse grains or root staples they grow, or living where transport costs from ports are high. That strategy, of course, can work only if the science, and its delivery systems, are there.
Simon points out that much past farm investment in developing countries was supported by huge input subsidies, which are not – or should not be – affordable as a cost of a small-farm strategy. However, such subsidies were, and often still are, far outweighed by concealed taxation of farm sales (mainly by foreign exchange manipulation), and by steering public support for health, education, roads, etc. away from farming areas. While the case for reducing price biases against farmers has won acceptance by most developing-country governments, many - especially in Africa - have responded to fiscal stress by winding down their already inadequate support for rural and agricultural infrastructures, often those that provide public goods and persuade young or educated people to stay in rural areas and help them prosper.
The smaller farmer in developing countries tends to be
much more labour-intensive in methods and crops, reflecting national comparative
advantage. So she should gain most from globalisation. Yet in practice,
as Simon says, radical changes in supply chains in SE Asia, Latin America
and Eastern and Southern Africa - supermarkets, export horticulture, grades
and standards - often seem like ‘three ugly sisters’ of globalisation,
because they are managed in ways that exclude small farmers. However,
there are plenty of success stories in avoiding such ‘intermediation
failures’ between small farmers and new market institutions.vi
Aid resources should be used to spread and develop them.
I believe Simon is mistaken about economies of scale. Hardly any of the thousands of studies find these in developing-country farming.vii Small farmers find it cheaper and easier to supervise family and hired labour, and large farmers to borrow, invest and manage capital. The shift to larger-scale farming happens rather late in the process of poverty reduction, as rural labour becomes scarcer and capital acquisition more feasible: for instance, agricultural censuses show the proportion of cropland in small farms in India rising from the 1960s to the present. In middle-income countries with very unequal land distribution (i.e. high proportions of land in big farms) such as S Africa and Brazil, poverty is much higher than would be expected from average income, and it is essential to get more farmland (and water) into the hands of the poor. DFID’s current agricultural policy review, plus its earlier Land Policy Paper, place it well to provide support for consensual, science-backed processes of land redistribution.
Simon rightly warns that the donor community’s increasingly management-centred, macro-orientated approaches to aid and development policy (including liberalisation), in part via their impact on PRSPs, neglect ground-level sector realities in general and agriculture in particular. Paradoxically, such approaches, though inspired by donors’ thrust towards poverty reduction, are implemented in a way that has underpinned the drain of resources from agriculture, the main source of income for the poor, and one where shifting land and control to poor farm users is generally efficient as well as equitable. In redefining its programme for aid to small-farm, employment-intensive and science-based growth in rural Africa, DFID’s new policy initiative can help to put this right.
3. The demographic window: scenarios, science and agriculture
To project the role of agriculture in reducing poverty (and to infer appropriate aid responses) means projecting (a) the poor’s access to employment-based entitlements to food, and (b) the numbers, especially of the poor, seeking these. As for (b), recent fertility reductions mean, in 2000-2030, big rises in the ratio of people of prime working age (15-64) to others (normally dependants) in many very poor regions. In 2000, there were 101 people of prime working age for every 100 dependants in Ethiopia; the projection for 2030 is 139. For Nigeria the rise is from 101 to 150; for Bangladesh from 126 to 181; and for India from 140 to 172.viii This is a ‘window of opportunity’ for growthix - and for poverty reduction. We know that this is speeded up by fertility reduction, mainly via delayed rises in the ratio of workers to dependants.x
The process depends, however, on those extra workers finding employment or self-employment, so they have enough extra income to afford adequate food. Almost invariably since 1700, this poverty reduction process has been initiated by employment-intensive, small-farm agricultural growth, first for staple foods, then for cash crops – later releasing demand-driven rural non-farm and urban expansion. The process increasingly needs support by crop and water science, often by appropriate rural institutions and infrastructures, sometimes by land and water redistribution. There has been a collapse of aid, and to some extent domestic investment and infrastructures, reaching poor farmers and stimulating the science that serves them. DfID’s initiative will help restore the right priorities. While we must allow for the risks and options that Simon discusses, focus on the soil, water, seed, and land needs of the rural poor is the central priority.
i Unless otherwise stated, sources and evidence are in IFAD, Rural Poverty Report 2001, Rome, 2001.
ii FAOSTAT (updated Feb 2004); data for 2000. These regions together contain 87 per cent of the dollar-poor (UN, Human Development Report 2003, OUP, New York, 2003, p. 41; data for 1999).
iii Below 1 hectare of irrigated, or 2 of rainfed, land. K. Sundaram and S. Tendulkar, ‘The working poor in India’, ILO, 2002; R. Eastwood and M.Lipton, ‘Rural-urban dimensions of inequality change’, WIDER, Helsinki, 2000.
iv M. Ravallion, ‘Urbanisation and poverty’, mimeo, World Bank, 2000; P. Hazell and C. Ramasamy, Green Revolution Revisited, Baltimore: Johns Hopkins, 1991.
v A. de Janvry and E. Sadoulet, ‘Growth, poverty and inequality in Latin America’; Review of Income and Wealth, 46 (2000), 267-87; R. Eastwood and M. Lipton, ‘Pro-poor growth and pro-growth poverty reduction’, Asian Development Review, 18 (2000), 22-58.
vi See IFAD 2001, chapter 5.
vii H. Binswanger et al. in T.Srinivasan and J. Behrman (eds.), Handbook of Development Economics vol IIIB, North Holland, 1995; IFAD 2001, ch. 3.
ix In 1962-90 this contributed 1.7%/year to growth of income per person in E/SE Asia. D. Bloom and J. Williamson, Demographic transitions, human resource development, and economic miracles in emerging Asia’ HIID Working Paper, Harvard, Cambridge (Mass.), 1977.
x R. Eastwood and M. Lipton, ‘Impact of changes in human fertility on poverty’, Journal of Development Studies, 31 (2000): 1-30.
Hosted by the Natural Resources Institute, University of Greenwich
Last Updated: July 20, 2004