Avoiding the Lewis path

In the best of all worlds, improvements in agriculture productivity leads to surpluses that allow capital accumulation and the development of industry, which then provides better inputs for agriculture. This is a virtuous circles that eventually leads to agriculture using only a tiny fraction of the workforce and representing a minuscule portion of GDP. This so-called Lewis path to growth has happened in many western economies, but does not seem to take off in Africa, in particular.

Bruno Dorin, Jean-Charles Hourcade and Michel Benoit-Cattin show that the Lewis path is not the unique equilibrium path in a growth model. A particular concern is the so-called Lewis trap that would result from a lack of additional agricultural land, where agriculture keeps growing in the labor force for little gain in output. But why insist on farming where land is no good? We have a global economy now and can produce goods where the comparative advantage is highest. Many areas of Africa are simply no good for agriculture, so we should stop insisting that they should go through all the motions of the Lewis path. Go straight to manufacturing and import food (my previous rant on this). This would also imply that other areas would specialize in agriculture, which is good even though the authors complain that this would lead to urban poverty there. People will move where the jobs are, for example to charter cities.