Writing for The Guardian, Saskia Sassen argues that at least part of what we label `migration’, specifically rural-urban migration, is being driven by the rise of large-scale land deals in developing countries:
Migrating to the cities is one major option. When politicians drone on mindlessly about¬†more than half the world’s population becoming urbanised, they rarely bring up the diverse ways in which people are being pushed off their land. Where else can they go but cities?
These new migrations are a key marker of our epoch. (Another is the¬†migration¬†of¬†high-level professionals.) They co-exist with older migrations. The rapid changes at point of origin also explain why most migrations are to cities. And they explain new types of migration, from rural areas to the global north (notably from several sub-Saharan countries to Europe).
In effect, expulsions are being rebranded as migrations, a phenomenon that will not cease anytime soon, given the ongoing search for land for crops, mining and water by governments and firms from a growing number of countries.
The causality of Sassen argument is not implausible – large-scale evictions are on the rise, and it’s perfectly possible that some of these people then make their way to urban areas. But is this having any meaningful effect on rural-urban migration? As is frequently the case on the Poverty Matters blog, we have someone desperately asserting something without much in the way of evidence to back up their claims.
It’s hard to read too much into cross-country scatter plots, but indulge me for a moment by taking a look at the following figure, which combines data from the Land Matrix database of large-scale land deals and World Bank data on urbanisation rates and arable land:
On the Y axis we have the average rate of urbanisation a country faced between 2000-2011 (the period covered by the land matrix data) and on the X axis we have the total number of hectares reported sold divided by the ¬†(average estimated) arable land in that country. The left graph uses data on all countries that I could get World Bank Data for (using wbopendata) and the right graph restricts the sample to countries where there has been at least one reported land deal during the period covered. As can be seen, there’s very little in the way of any convincing relationship in any direction between land grabs and urbanisation. The latter is my best back-of-the-envelope proxy for rural-urban migration, although obviously it will also be correlated with within-city growth.
Now, you should interpret this with some caution, as some of the numbers look a bit high (I encourage readers to look into the data themselves, as this is a bit slapdash). I should note that there still doesn’t appear to be a relationship between the two when¬†I use levels instead of changes, or if I use the number of people living in slums as an outcome, nor when I use the total number of land deals a country has seen, the total number of hectares lost, log hectares, or, as above, hectares a percentage of arable land. Nor is there a relationship when I disaggregate the land grab and urbanisation data by year, or when I use lagged values of land grabs. ¬†There is also no relationship when I use a sample including countries which never have a land deal, and restricting it to only countries which have land deals.
Now, this isn’t identified in any reasonable sense – so you should take this purely as descriptive (lack of) evidence. But I think it is safe for us to call most rural-urban “migration” what it is: “migration.”
There are a number of reasons why this is the case. First, it’s not clear that the next-best opportunity for a displaced farmer is the nearest city. Indeed, despite the fact that many land¬†seizures are detrimental to local communities, they sometimes provide job opportunities for recently-displaced land owners. This is not necessarily welfare-enhancing, but would go some way to mitigate the effect that Sassen purports in her article.
Furthermore, there’s growing evidence¬†(also¬†here,¬†here¬†and here)¬†that tenure insecurity actually leads to less, not more, ¬†internal migration and migration abroad, as weak property rights not only force landowners to maintain a physical presence to guard their ownership, but also reduce land liquidity, making it harder to sell and get the hell out of dodge. So we’d actually already expect to see more migration from areas with stronger property rights, which will also be the areas which are going to be shielded from large scale land grabs. This doesn’t necessarily speak to Sassen’s main hypothesis (and might even be hiding any evidence of it), but just shows that the reality might be a little bit more complicated than “migration is just another word for land grabbing.”