Land Reform is not a Panacea

Justice in the world of the second best

Land reform can often have mixed results. In a world where there are systematic frictions to transacting land, it’s perfectly plausible for breaking up the estates of the landlords and redistributing to those who work it to make everyone, or almost everyone, worse off. Farms generally have increasing returns to plot size. A simple reason are new technologies which may require high fixed costs to use. If the land is half the efficient size for one tractor, you cannot buy half a tractor. It is conceivable that people will share capital implements, but increasing the number of negotiations which must be done will lead to many failures. Empirically, there is considerable “slack” in developing countries, primarily arising from indivisible goods.

We should also expect “land to the landless” style reforms to be relatively worse the more modern agriculture gets. In the pre-modern economy, we should expect people increasing the productivity of land to be dependent on local knowledge. In the modern economy, however, increasing productivity comes from ideas spread from elsewhere. These ideas are taken up first by bigger farms, and only later diffuse to small farmers. Larger farms can be more risk-taking in a world of very poor insurance markets, as is ofthe the case in the developing world.

Adamopoulos and Restuccia (2014) summarize the facts. Farms in poor countries are much smaller, and they are much less productive, by a factor of 34 and 16 respectively. Despite the fact that we should expect farms to get smaller as outside opportunities decrease, farms are so much smaller in poor countries that it becomes extremely inefficient. (Intuition: the farm owner is a manager, and could work anywhere. As wages rise elsewhere, people should reallocate and the land managed by one manager should increase). This misallocation persists largely because of government policy to subsidize smallholders, or tax large farmers. Many developing countries have swiftly increasing progressive taxes, essentially exempting smallholders while penalizing the successful.

And now onto some historical examples. Some particular crops might be more productive, but require large fixed costs. Sugar is the archetypal example of this. On the island of Hispaniola, Haiti grows coffee and the Dominican Republic grows sugar. The roots of this lie in the aftermath of the slave revolt — the old owners were driven out, and the land redistributed. This redistribution was persistent — land inequality was substantially lower even in 1950. There were no large plantations whatsoever, and plots were quite small. Sugar is something of an industrialized good, requiring considerable exertion to break, boil, and refine the sugar. Plantations can grow coffee and sugar, but small plots can only grow coffee. When the market changed to be less favorable to coffee, Haiti fell far, far behind.

In Taiwan, the KMT engaged in substantial land reform in several phases. Oliver Kim and Jen-Kuan Wang show that, while the second phase (which redistributed the formerly Japanese-owned lands) boosted yields, the third phase (which broke up estates and redistributed) substantially lowered productivity. In the Philippines, the Marcos government redistributed land in 1988, breaking up large estates and then putting restrictions on whether the land could be sold. Adamopoulos and Restuccia (2020) estimate that the reforms reduced agricultural productivity by 17%.

I would not claim that land reform cannot be good. I think entrenched aristocracies are often politically poisonous, and form an anti-progress bloc. I would simply caution us against assuming that it will increase agricultural yields, and that redistribution is necessarily good. It also suggests that parties, like the EFF, which are agitating for land expropriation without redistribution, are quite likely to arrive at outcomes that make everyone but a lucky few recipients of land worse off.

I believe the ideal mechanism for righting wrong in land distribution is, allowing that the land was not cultivated by slaves, for the land everywhere to be auctioned, with the prior owners having to pay a price equal to the unimproved value of the land, and everyone else paying whatever the auction mechanism discovers the price to be, including improvements. It would be, in essence, an enormous tax upon unjustly earned gains.