Key Players in Economic Development

Institutions and the resource curse: New insights from causal machine learning

“…Lastly, we include proxies for the transportation infrastructure. The reason is that a lot of the transportation infrastructure in Africa is mining-related (e.g., [ 46 ]), which can create positive spillovers for local economic development (e.g., [ 47 ]) or spread local conflict (e.g., [ 48 ]). Indeed, adding a large set of covariates is no panacea for a potential omitted variable bias through other, unobserved confounding variables.…”

“…STUVA requires the absence of spillovers. The evidence on spillovers is mixed: [ 17 ] find no evidence for spillovers from mining, while [ 47 ] find evidence for spillovers from changes in mineral prices among districts connected by roads or ethnolinguistic linkages.…”

“…We follow a large and growing literature in using nighttime light intensity to proxy for economic development at the subnational level and in particular the recent work using nighttime light intensity to study the effect of mining activities on economic development at the subnational level include (e.g. [ 17 , 47 ]). These data are based on daily measures from the Operation Linescan System of the DMSP.…”

“…As an outcome variable, we use the natural logarithm of the average nighttime light pixel value in a given district and year. To avoid losing observations with zero reported nighttime lights, many scholars have added 0.01 before taking logs (e.g., [ 17 , 47 , 51 , 53 ]). We use this transformation as well in order to ensure the comparability of our results to those in the previous literature.…”

“…For this purpose, we use information on mining activity from the SNL Metals & Mining database and information on the world market prices of minerals from various sources (see Table A1 in Online Appendix in S1 Appendix ). Hence, we rely on the same type of data for natural resources as [ 16 , 47 ].…”