A few words about urbanization and growth

From a paper I am writing for the World Bank:

Every affluent country in the world is urbanized. Among OECD countries, 77 percent of people live in urban areas, and among World Bank-designated high-income countries, 78 percent of people live in urban areas. The poorest two countries in the OECD, Turkey and Mexico, are 67 percent and 76 percent urbanized, respectively. The least urbanized affluent country, Portugal, is 55 or 59 percent urbanized, depending on source. At the same time, the world’s lowest income countries are generally not urbanized: in 2004, the urbanization rate among the World Bank’s designated low income countries was 31 percent. All of the countries with urbanization rates of less than 20 percent, Burkina Faso, Burundi, Cambodia, Ethiopia, Malawi, Nepal, Papua New Guinea, and Uganda, are low-income countries, all with Gross National Incomes Per Capita of less than $660, and most with GNIs that are substantially lower than that.[1] The correlation between urbanization and PPP Per Capita GDP in 2000 was .70. In short, urbanization accompanies affluence.

That urbanization accompanies affluence does not, however, mean that urbanization causes affluence. First, it is worth noting that Latin America and the Caribbean are 77 percent urbanized, and the countries in that region are certainly not among the World’s richest (nor are they in general, among the poorest). There are also very poor countries in Africa--Cameroon, Mauritania, and Senegal--that are at least 50 percent urbanized. All of these countries have per capita GNIs of less than $1010.[2] Hence affluence does not necessarily follow urbanization.

One of the most interesting questions in development economics, then, is whether urbanization causes affluence, or whether affluence causes urbanization. Knowing the direction of causation is important, because it will dictate whether policy should encourage urbanization or be neutral with respect to urbanization. Discussion below will outline arguments for both directions.

Principles set forth by Richard Freeman, however, suggests that the evidence is already sufficient to know that policy should not discourage urbanization. Freeman’s three rules of econometrics are: (1) it had better be there in the ordinary-least-squares regression; (2) it had better still be there in the econometrically-sophisticated high-tech instrument procedures; (3) it had better still be there for small technical tweaks to the econometrically-sophisticated procedures. That urbanization has a deleterious effect on affluence is not there in scatter-plots and correlations, and, as we shall see below, is not there in the OLS regressions in the literature.

The fact that there is no evidence that urbanization inhibits development is in itself important. On my visit to Bangladesh in 2004, some officials were convinced that urbanization was shown scientifically to be “immoral,” and wanted to know where to find the work that would prove it to be so. I disappointed them when I was not able to point them to any such work.

---

[1] Afghanistan is likely within this group as well, but there is no good current data on urbanization in the country.

[2] Gabon, a sub-Saharan that is 84 percent urbanized, has a per capita GNI of more than $5000, a very high number for the region. But the primary reason for the high number is the presence of oil.