For Whom Is Gas a Curse? Heterogeneous Effects of Natural Gas Dependence on Human Development: A Combined Approach of Double Machine Learning with Causal Forests

The relationship between natural gas wealth and socioeconomic development has yielded conflicting results in the “resource curse” literature. This study reframes the issue, shifting the focus from whether gas constitutes a curse to determining for whom it becomes one. Using a panel of 65 gas-producing countries with complete data through 2021, we combine two-way fixed effects with Double Machine Learning and causal forests to estimate the average and heterogeneous causal effects of gas rents on the Human Development Index. The average treatment effect is virtually zero, consistent with previous conflicting evidence. Causal forest estimates reveal substantial heterogeneity since 55% of country–year observations show negative effects. A clustering of the estimated effects identifies four country regimes, ranging from gas-dependent economies experiencing negative effects to diversified producers with good governance that benefit from gas revenues. The largest negative effects are concentrated in the Middle East and North Africa, while OECD countries experience positive effects. Exposure to external trade is the characteristic most closely tied to less favorable effects, whereas strong governance goes with more favorable ones once openness and income are held constant. Thus, the “curse” and the advantages of gas can coexist, depending on each country’s institutional and economic characteristics.