Democracy And Natural Resources

In the popular imagination, democracy is the foundation for prosperity. Strong, trustworthy political and public institutions are a precondition for sustained economic growth and development. The empirical evidence however, is decidedly mixed – not a few studies find little to no evidence of a link, and some fund links going the other way i.e. income creates democracies but not the other way around.

This new working paper from the IMF partly explains why (abstract):

Income and Democracy: Lipset's Law Revisited
Hoeffler, Anke, and Bates, Robert H. & Fayad, Ghada

Summary: We revisit Lipset‘s law, which posits a positive and significant relationship between income and democracy. Using dynamic and heterogeneous panel data estimation techniques, we find a significant and negative relationship between income and democracy: higher/lower incomes per capita hinder/trigger democratization. Decomposing overall income per capita into its resource and non-resource components, we find that the coefficient on the latter is positive and significant while that on the former is significant but negative, indicating that the role of resource income is central to the result.

To explain the results in plain English:

1. On an overall basis, higher per capita income levels slow down democratisation (!).
2. Lower income levels on the other hand speed up democratisation (!). This is of course completely counter-intuitive.
3. But when resource and non-resource incomes are disaggregated, the results are different.
4. When income is primarily resource based, the original finding stays the same.
5. But with non-resource based income, the findings are reversed i.e. higher incomes result in pressure for more democracy.

So the bottom line here is that when an economy is based on extraction of natural resources such as oil, political development tends to be slower. This makes sense, as resource extraction industries are usually monopolistic or oligopolistic and tend to foster rent-seeking behaviour in the surrounding environment.

This result also fits nicely into the literature on FDI determinants, where foreign investors in primary industries (mining and agriculture) tend to be insensitive to the state of political development, unlike with investors in downstream activities where rule of law, property rights, and a stable political environment carry greater weight.

Natural resource extraction is a take it or leave it business, as you can’t exactly pack up and go elsewhere. Of course, there remain counter-examples (*cough*China*cough*), but it seems in this case there’s some basis to the popular notion.

A couple of notes on the methodology:

1. Countries that remained fully authoritarian or fully democratic were dropped from the sample tested (no variation in the dependent variable). As such Australia and Saudi Arabia among others were dropped from the sample, and as was the US.
2. Granger causality tests showed bi-directional causality i.e. democracy affects income, but income also affects democratisation. A classic chicken and egg problem.

Technical Notes

Hoeffler, Anke ; Bates, Robert H. ; Fayad, Ghada, "Income and Democracy: Lipset's Law Revisited", IMF Working Paper No. 12/295, December 2012