Common sense suggests that file sharing should have a fairly substantial negative impact on music sales. A paper by Felix Oberholzer-Gee and Koleman Strumpf published in The Journal of Political Economy argues that any such effect is negligible. On the grounds that wishing for ponies does not make actual ponies appear, I have been personally skeptical of this result. Stan Liebowitz - with whom I disagree on many things copyright, but who is a meticulous and honest researcher - has gone through the paper carefully. You can find the fruit of his labor here. He examines the data that is available and finds little merit in the Oberholzer-Gee and Strumpf conclusions. "File sharing has no impact on sales" is not a terribly good argument against copyright because the evidence suggests it isn't true.