Do European settlers account for 40% of all development ever outside Europe?

A new paper from Bill Easterly and Ross Levine:

Although a large literature argues that European settlement outside of Europe shaped institutional, educational, technological, cultural, and economic outcomes, researchers have been unable to directly assess these predictions because of an absence of data on colonial European settlement.
…we construct a new database on the European share of the population during the early stages of colonization and examine its association with the level of economic development today.
We find: (1) a strong and uniformly positive relationship between colonial European settlement and development, (2) a stronger relationship between colonial European settlement and economic development today than between development today and the proportion of the population of European descent today; and (3) no evidence that the positive relationship between colonial European settlement and economic development diminishes or becomes negative at very low levels of colonial European settlement, contradicting a large literature that focuses on the enduring adverse effects of small European settlements.
Our findings are most consistent with human capital playing a central role in the way that colonial European settlement affects development today.

The provocative post title comes from a tweet of Bill’s. He likes to provoke.

I think the wrong way to interpret their finding is that colonialism causes development.

I think the right way to think about it is to remember that modern economic growth is a product of ideas, skills, and the organization of society and firms. If you don’t have much of that stuff and add people that do–even colonists–growth is more likely.

What’s amazing is that there are other ways to get ideas and skills and social organization than colonialism. You would think, by now, these other paths would have swamped out the effects of early settlement. What I did not see in my quick reading of the paper is a sense of proportion–is the statistically significant result substantive? Less lazy readers than me: please chime in.