This paper investigates the dynamic causal link between exports and economic growth using both linear and nonlinear Granger causality tests. We use annual South African data on real exports and real gross domestic product from 1911-2011. The linear Granger causality result shows no evidence of significant causality between exports and GDP. The relevant VAR is unstable, which undermines our confidence in the causality result identified by the linear Granger causality test. Accordingly we turn to the nonlinear methods to evaluate Granger causality between exports and GDP. First, we use Hiemstra and Jones (1994) nonlinear Granger causality test and find a unidirectional causality from GDP to exports. However, using a more powerful and less biased nonlinear test, the Diks and Panchenko (2006) test, we find evidence of significant bi-directional causality. These results highlight the risk of misleading conclusions based on the standard linear Granger causality tests which neither accounts for structural breaks nor uncover nonlinearities in the dynamic relationship between exports and GDP.