Another Blog Post With a Conclusion Listed as a Question?

This post by William Easterly at Aidwatch asks, "Was the poverty of Africa determined in 1000 BC?" Easterly cites an article by Diego Comin, Erick Gong, and himself:

78 percent of the difference in income today between sub-Saharan Africa and Western Europe is explained by technology differences that already existed in 1500 AD – even BEFORE the slave trade and colonialism.

Moreover, these technological differences had already appeared by 1000 BC. The state of technology in 1000 BC has a strong correlation with technology 2500 years later, in 1500 AD.

Whenever looking at a "strong correlation", it is important to ask yourself, "Could some third variable explain both factors?" My immediate thought is there are probably two confounding variables. Firstly, Africa is largely landlocked, with a dearth of navigable rivers, making trade and the diffusion and adoption of technology much more difficult. Secondly, Africa has conditions that are ideal for the spread of tropical diseases such as malaria.

When the authors do not control for these factors, they find that "78 percent of the difference in income today between sub-Saharan Africa and Western Europe is explained by technology differences that already existed in 1500 AD".

However, when the authors add independent variables for "distance to equator", "distance to equator squared", and "landlocked", the regression coefficient for "migration-adjusted technology level in 1500 AD" drops to 1.770 (significant at the 5% level). With this coefficient, only 42% of the difference is attributable to technology differences in 1500 AD.

Lastly, when the authors perform a regression analysis using "landlocked" and "tropical dummy" as independent variables, they find that the regression coefficient for "migration-adjusted technology level in 1500 AD" is 2.590 (significant at the 1% level). This coefficient explains 62% of the difference in income between Europe and Africa in 2002.

I also have a fairly major quibble with the "landlocked" variable. If you look at Wikipedia's list of landlocked countries, you will notice that DR Congo is not considered landlocked because it has 40 km of coastline. However, looking at a map, you will see that DR Congo is effectively landlocked. Furthermore, the "landlocked" countries of Western Europe seem to all be outliers in terms of GDP per capita on a purchasing power parity basis: Luxembourg (#3), Liechtenstein (#1), San Marino (#18), Switzerland (#19), and Vatican City (not ranked). In fact, if you were to only look at Western Europe, you would say that being landlocked is in fact an advantage. The only non-landlocked country in Western Europe with a higher rank than Switzerland is Ireland at #16 (and I have a feeling that they may have fallen a bit in the standings since 2009).