There is a view in the literature that curbing corruption is concurrently growth augmenting. We present evidence that such is not always the case: independent of its indirect effects, a drop in corruption is growth augmenting only if there has been a persistent decline in corruption in the past.