『Abstract
　We examine the ‘resource curse’ using new data on historic resource stocks and an improved econometric methodology. The paper distinguishes between resource abundance (stocks) and extractive intensity (flows), focusing on relationships between resources and rule of law. Previously unavailable information on past resource stocks is estimated. We find that economically large initial natural resource stocks are associated with subsequent lower levels of rule of law and do not directly affect growth, while raw resource exports do not have a significant effect on rule of law when stocks are included in the analysis but do affect average growth rates. Sample size is maximized through the use of an EMis (expectation maximization with importance sampling) algorithm to replace missing data, minimizing the bias and inefficiency associated with listwise deletion, which commonly eliminates half or more of the available data in this setting.

Keywords: Cross-country comparisons; Multiple imputation; Resource curse; Resource extraction; Resource stocks; Rule of law』

1. Introduction and background
2. Empirical issues and data
3. Analysis and empirical investigation
　3.1. Missing data and multiple imputation
　3.2. Robustness to additional conditioning variables
4. Conclusion
Appendix A: Data Sources
Appendix B: Minerals included in the composite measure of mineral abundance
References