Here are some quotes from the paper I found significant.

File sharing proponents have also suggested that music sales decreased as consumers substituted into other types of entertainment goods, such as prerecorded movies and video games. However, Michel (2005) concludes there is no evidence that CEX [the study that the author is using for data] households' decrease in music purchases corresponds to an increase in spending on movie tickets, prerecorded movies or video games.

He also raises the possibility that some of the decrease may be due to legal music downloads, but believes that this is less true for the years covered by his study than in more recent years.

From his conclusion:

Our micro-level data test results suggest that file sharing may have reduced album sales (between 1999 and 2003) by as much as 13 percent for some music consumers. At minimum, our test results indicate that the relationship between computer ownership and music purchases (in the CEX) weakened after Internet file sharing became a viable option for music purchasers. No similar negative change exists in the data prior to the initiation of the first file-sharing service.

After reading his paper, I'm inclined to agree. The data does suggest that there may have been a 13% drop in CD purchases among computer owners who spent more music than the median consumer, but there may be alternate explanations to account for at least some of the drop. However, it seems clear that filesharing did reduce music sales among that population.