A recent Economist article poses the question Does freer farm trade help poor people? Given the ideological slant of the Economist, it is unsurprising that the article concludes yes. The interesting aspect of the article discusses two World Bank research papers that indicate that the way in which agricultural trade is regulated has major consequences.

The links between trade, food prices and poverty reduction are more subtle. Different types of reform have diverse effects on prices. When countries cut their tariffs on farm goods, their consumers pay lower prices. In contrast, when farm subsidies are slashed, world food prices rise. The lavishness of farm subsidies means that the net effect of fully freeing trade would be to raise prices, by an average of 5.5% for primary farm products and 1.3% for processed goods, according to the World Bank.

These effects are still much smaller than recent food-price spikes, but would they, on balance, help or hurt the poor? In crude terms, food-exporting countries gain in the short term whereas net importers lose. Farmers are better off; those who buy their food fare worse. Although most of the world’s poor live in rural areas, they are not, by and large, net food sellers. A forthcoming study* of nine poor countries by M. Ataman Aksoy and Aylin Isik-Dikmelik, two economists at the World Bank, shows that even in very rural countries, such as Bangladesh and Zambia, only one-fifth of households sell more food than they buy. That suggests the losers may outnumber winners.

But things are not so simple. The authors point out that net food buyers tend to be richer than net sellers, so high food prices, on average, transfer income from richer to poorer households. And prices are not the only route through which poverty is affected. Higher farm income boosts demand for rural labour, increasing wages for landless peasants and others who buy rather than grow their food. Several studies show this income effect can outweigh the initial price effect. Finally, the farm sector itself can grow. Decades of underinvestment in agriculture have left many poor countries reliant on imports: over time that can change.

The World Bank has often argued that the balance of all these factors is likely to be positive. Although freer farm trade—and higher prices—may raise poverty rates in some countries, it will reduce them in more. One much-cited piece of evidence is a study† by Thomas Hertel, Roman Keeney, Maros Ivanic and Alan Winters. This analysis simulated the effect of getting rid of all subsidies and barriers on global prices and trade volumes. It then mapped these results on to detailed household statistics in 15 countries, which between them covered 1 billion people. Fully free trade in farm goods would reduce poverty in 13 countries while raising it in two.

But lately the bank seems to be taking a different line. Robert Zoellick, the bank’s president, claims that the food-price crisis will throw 100m people below the poverty line, undoing seven years of progress. His figure comes from extrapolating the results of a different study** by Mr Ivanic and Will Martin, another World Bank economist. This study analyses the effects of more expensive staple foods on poverty by examining household surveys in nine countries. In seven cases, higher food prices meant more poverty. (Dani Rodrik, a blogging Harvard economist, was one of the first to highlight the tension between these studies.)

In fact, the bank’s results are not as contradictory as they seem. The two studies are based on different sets of countries: only Peru, Zambia and Vietnam appear in both. And the gloomy analysis measures only the effect of pricier staple foods, whereas the other examines freer trade in all farm goods. Such trade brings broader benefits: even if higher prices for staples exacerbate poverty in some countries, at least in the short term, the effect may be outweighed by increased demand for other farm exports, such as processed goods, as rich countries cut tariffs.

These subtleties suggest two conclusions. First, the bank, and others, should beware sweeping generalisations about the impact of food prices on the poor. Second, the nature of trade reform matters. Removing rich-country subsidies on staple goods, the focus of much debate in the Doha round, may be less useful in the fight against poverty than cutting tariffs would be. The food-price crisis has not hurt the case for freer farm trade. But it has shown how important it is to get it right.

These papers only assess trade rather than agricultural practices. I would add that the ecological fit of agriculture to the place in which it is practiced will also have substantial impacts on the potential for a regions ability to escape from poverty. Increases in agricultural production that damages other ecosystem services that are important for local people’s livelihood, such a fisheries, fuelwood, flood regulation, or water quality, can do more damage than good.