Measuring natural resources

Hypotheses about the resource curse hold that nat­ural resources have distinctive properties that hin­der development. A measure of natural resources should capture these properties and make it possi­ble place countries on a continuum from resource-poor to resource-rich. Arguably what is most dis­tinctive about the resources in question is that they derive from “natural” endowments that, because of their scarcity, can typically be sold for prices that far exceed the costs of extracting them.

For the purposes of this paper, I measure natu­ral resources using “rent” component of proceeds from mineral and fuel extraction. The economic concept of “rent” refers to excess profits after ac­counting for all relevant costs including a normal return on capital, and a “resource rent” is a rent that originates from access to a scarce natural re­source. The rent component thus depends on the quantity of a resource that is extracted, the cost of extracting it, and the international price at which it can be sold.

Resource rents are central to all major variants of the resource curse hypothesis. For example, Sachs and Warner emphasize how “wealth effects” caused by a resource boom undermine the com­petitiveness of the manufacturing sector. They as­sume for simplicity that the resource sector uses no capital or labor; so the boom consists entirely of rent.8 Meanwhile, “political” versions of the re­source curse assert the ability of either the state or rebel groups to capture “easy” rents from resource extraction.9

The data on resource rents that I use comes mostly from the World Bank’s Changing Wealth of Nations, which includes estimates for fourteen commodities.10 A major omission from the World Bank data set is diamonds, which feature promi­nently in discussions of the resource curse in Africa, particularly regarding the link between re­sources and violent conflict. Reliable information about diamond production is not readily avail­able, due to the historical secrecy of the dia­mond trade, wide variation in the value of gem-quality stones, and large-scale smuggling in con­flict zones. To address the omission, I calculated my own estimates of diamond rents, using data from several sources (and a fair amount of in­formed guesswork), which are reported in an ap­pendix. The estimates are very crude, but crude estimates are better than leaving diamonds out of the resource rent calculations.

Finally, the question arises of how to represent the value of resource rents while accounting for differences in country size. The two main candi­

1.      Sachs and Warner, “Natural Resources and Economic Growth,” 6–7.

2.      Ross, “Does Oil Hinder Democracy?”; Collier and Hoef­fler, “Greed and Grievance.”

3.      World Bank, The Changing Wealth of Nations: Measuring Sustainable Development in the New Millennium (Washington: World Bank, 2011). The commodities included are oil, natu­ral gas, coal, bauxite, copper, gold, iron ore, lead, nickel, phos­phate rock, silver, tin, zinc, and “forest rent.” (Since I focus on minerals and fuels, I exclude forest rent.)

 

Figure 1: Resource rents in sub-Saharan Africa, 1990–2008

 

dates are to express rents as a percentage of GDP or to express them per capita. Many critics of early studies of the resource curse pointed out that di­viding resource proceeds by GDP produces a mea­sure of “resource dependence” and not “resource abundance.”11 Imagine two countries with iden­tical natural endowments reaping identical nat­ural resource rents. One has succeeded in de­veloping an internationally competitive manufac­turing sector, and the other has failed to do so. Resource rents as a percentage of GDP will be lower for the successful industrializer, due to its larger GDP. The failed industrializer will appear more “resource-rich” by comparison, though its

11. Annika Kropf, “Resource Abundance vs. Resource De­pendence in Cross-Country Growth Regressions,” OPEC En­ergy Review 34, no. 2 (2010): 107–130; Michael Alexeev and Robert Conrad, “The Elusive Curse of Oil,” Review of Economics and Statistics 91, no. 3 (2009): 586–598.

resource endowments and rents are the same as its “resource-poor” counterpart. This kind of bias is minimal in comparing African countries, most of which have similarly low levels of industrial development. So while I express resource rents as a percentage of GDP, using rents per capita would yield almost identical results.12

Figure 1 presents regional trends in resource rents from 1990 to 2008 (the latest year for which complete data were available), adjusted for infla­tion. The solid line is for the thirteen mineral and fuel commodities in the World Bank data, plus my estimates for diamonds. The dotted line excludes fuel rents. One striking feature is that the solid line is con­

12. To be more precise, the correlation between (the natural logarithm of) rents per capita and (the logistic transformation of) rents as a proportion of GDP is 0.96.

4 sistently far above the dotted line, which high­lights the fact that the bulk of Africa’s resource rents are derived from fuel (especially oil). For most years, fuel rents alone are two to four times greater than mineral rents.

Another striking feature is that Africa has ex­perienced a sharp boom in resource rents since about the year 2000. The trend in rents during the 1990s was downward. However, by 2008 total rents were about four times higher than in 1990, and about eight times higher than their low point in the late 1990s. The increase affected mining and fuel.

 

Country: 
Sectors: