We introduce a notion of causal independence based on virtual intervention, which is a fundamental concept of the theory of causal networks. Causal independence allows for de ning a measure for the strength of a causal e ect. We call this information ow and compare it with known information ow measures such as the transfer entropy. Contents