Replicating Sachs and Warner: The 1997 Working Paper

This paper reports on my replication of the 1997 Sachs and Warner resource curse working paper. Exact pure replication is achieved. A test for statistical replication shows that some of the Sachs and Warner results attempting to determine the cause of the resource curse are not robust to the country sample. A test for scientific replication shows that findings of the resource curse are not sensitive to different measures of resource intensiveness, though they are subject to country sample and omitted variable bias. Reporting errors in the published paper show the value of researchers making both their data and code publicly available.