A key empirical question is to what extent goods that are in short supply are substitutes or complements for things that are not in so short supply. For example, if electronic copies of music trade widely at a very low price, they may generate little revenue for musicians. On the other hand, it they are complementary to live music (and they are) then the "freebie" recordings raise the demand for live performances. So a perfectly reasonable business model in music is to give the recording away for free, and sell the live performances. In the free software industry, copies are cheap or free, but generate demand for (are complements of) services which are not.
What about books? One of the concerns of the publishing industry is that if free or cheap electronic copies become widely available revenue for authors will dry up. (Well that is a concern of authors, anyway.) In other words, are electronic copies substitutes for printed copies? In the long term, portable computers may replace printed books, so this is probably true. But it is far from true today. John Bennett draws our attention to an article by Tim O'Reilly that provides hard data on sales and downloads of Asterisk: The Future of Telephony, by Leif Madsen, Jared Smith, and Jim Van Meggelen, which was released for free download under a creative commons license. The article has quite a bit of information. Typically books sales spike when it is released then decline gradually - in our own research we find most sales occuring in the initial six months after release. The article has only two books to compare, but it seems as if the free download had little impact on sales, at worst causing sales to decline slightly faster. So at the moment at least it appears that elecronic copies are neither complements or substitutes for the printed version, but rather are neutral.