Last month I moved to Ethiopia to help set up the Agricultural Transformation Agency. It’s an initiative of the government to help increase the productivity and improve the livelihoods of smallholder farmers. I wanted to work in agricultural development because like many people I believe in something like the following:
1. Most of the poorest people in the world are farmers (and the majority of them are women)
2. Helping them grow more and linking them to markets reduces hunger and increase their income
3. The rest of society benefits from this because food becomes cheaper, people become healthier and work more effectively
4. Eventually, if agricultural productivity continues to increase, the sector will release workers who can move to cities and start an industrial revolution
The cause and effect are not always clear (is an agricultural revolution a cause of population growth or a response to it?), but it is a fairly accurate description of the trajectory of most currently developed countries. Even the city-states that grew rich without agriculture (Singapore, Hong Kong, the Gulf states) rely on large numbers of immigrants from places like South China and Pakistan that have experienced rapid agricultural growth and shed labour as a result.
This simplistic narrative has been challenged by Dani Rodrik and Margaret McMillan. They argue here that economic liberalisation in Africa and Latin America has benefitted the agricultural sector, but retarded economic growth because labour shifted from industry into agriculture and informal activities. By implication, promoting agriculture as an engine of economic development may be counter-productive. (The PDF is here).
Their argument is based on the recognition that economies develop in two ways: by productivity increasing within sectors, and by activity shifting from lower-productivity to higher-productivity sectors. In countries with low or zero growth, both technology and the sector mix are stagnant: this describes India in the 1970s, or parts of Africa now. Developed economies grow slowly because most of their population already work in the high-productivity sectors and need technological change to improve their standard of living. The highest rates of growth are in countries like China that are able to increase productivity in all sectors and shift labour from low- to high-productivity sectors. The best data I have show that in the late 1980s, 60% of Chinese worked in agriculture; now only 40% do and the proportion is falling fast.
This doesn’t mean that agriculture is always a low-productivity sector: in developed countries productivity is similar across sectors, mostly because farmers who can’t make a decent living get out and do something else. There is a set of countries, however, in which agricultural productivity is largely stagnant, while industrial and service productivity are growing fast. These are the emerging economies of Latin America, East Asia and Sub-Saharan Africa. They are at the bottom of the U-curve in the chart below.
Rodrik and McMillan trace the growth of economies between 1990 and 2005, a period of liberalisation almost everywhere. Most emerging markets implemented ‘Washington Consensus’ reforms that made them more open to trade and financial flows. The results in terms of growth were disappointing: most of Latin America and Africa stagnated in the 1990s. Were the reforms poorly implemented? On the contrary, they may have worked too well: as predicted by standard theory, opening to trade led to specialisation. The problem was that most Latin American and African countries turned out to have a comparative advantage in low-productivity activities, such as agriculture.
The results were most drastic in Argentina, which lost much of its manufacturing sector in the 1990s, even as the service sector and agriculture (led by soybeans) boomed. In Africa, many countries experienced slow growth, or even decline, in industrial sector employment. This is not just a ‘Dutch disease’ phenomenon: it occurred equally in resource-rich Nigeria and Zambia as in resource-poor Kenya and Malawi. The most common response to the decline in industrial employment in Latin America was a shift of workers into the informal sector and unemployment; in Africa, many went back to working on their family farm. In China, India and Thailand, on the other hand, there has been a steady shift of workers from agriculture into industry and services; even within cities, the growing formalisation of the retail sector has increased productivity and hence wages.
What do these results imply for development strategy? First, liberalisation is not a panacea: it may allocate resources more efficiently within sectors, but not necessarily across them. Second, islands of high productivity in one sector (like Zambia’s newly-efficient copper mines) do not drag the rest of the economy along with them. Third, restricting rural-urban migration is inimical to growth, because cities are hubs of higher-productivity activities.
However, this paper does not imply, in my view, that governments are wrong to pursue agricultural development – only that they are wrong to do so to the exclusion of other sectors. I would cite two economic and two non-economic reasons for why investment in agriculture is still justified in developed countries with large, poor rural populations.
One economic reason to invest in agriculture is that agriculture can provide the raw materials for higher-productivity activities. Brazil, for example, is now a world leader in sugar refining and vehicle ethanol technology – sectors it would have struggled to develop without plentiful, cheap raw sugar. Another is that increasing productivity in this low-productivity sector is generally the first step towards the growth of higher-productivity sectors. The Chinese economic miracle started with relaxing restrictions on private farmers in 1979.
The non-economic rationales for agriculture are that it feeds people and it is an egalitarian activity. Feeding people is attractive when global food prices are volatile, or transporting imported food is expensive: Ethiopia is much less likely to experience food-price riots than, say, Liberia or Egypt, because most of its food is home-grown. The final point is about distribution, rather than economic efficiency. Last year, Papua New Guinea earnt over a billion dollars from the export of gold – twice as much as it earnt from exporting coffee. Gold mines employ a lot of capital and several thousand highly skilled miners, a mix of Papua New Guineans and Australians earning up to $100,000. Coffee farms employ much less capital and are the primary livelihood for around 500,000 people in the Highlands, each earning a little under $1,000. In cash terms, gold is a more valuable activity than coffee. But coffee employs a hundred times more people, pays their children’s school fees and is much harder for politicians in the capital to profit from. Which industry is more conducive to social and economic stability?
Interestingly, Ethiopia is one of the few African countries in which labour shifted from lower- to higher-productivity sectors in the period 1990-2005 (Ghana and Mauritius are the others). This is despite a government policy of ‘Agriculture-Led Industrialisation’ for most of that period. I won’t speculate on the reasons, but I hope it bodes well for my work here.
