This paper explores the nature of causality between corruption and FDI inflows in five Latin American countries. Unlike prior empirical studies that use correlation-based standard regressions, we employ dynamic cointegrated error-correction models to uncover the directions of Granger-causality between corruption and FDI inflows. The evidence is overwhelmingly supportive of causality that is either unidirectional running from corruption to FDI inflows without feedback, or bidirectional between the two variables. In addition, the results imply some challenges for policy-makers. Their attempts to control corruption as causality patterns between corruption and FDI inflows sometimes prove dynamic, i.e., changing their nature between the short- and the long-runs. Moreover, the presence of long lags in the causal effect of corruption on FDI inflows carries the risk that policymakers (given their usual short-run policy horizon) may mistakenly infer that corruption is not harmful and fighting it is rather futile.