From a utilitarian perspective, devising an optimal copyright regime requires a simple balancing of copyright’s benefits against its costs in order to maximize the resulting social welfare. While simple in theory, this balance has proven surprisingly difficult to achieve in practice. In part, this difficulty arises because the data available is inevitably limited, and often observational, and so the conclusions are fraught with ambiguity and uncertainty. In part, it arises because there are vested political interests that do not want an optimal copyright regime. Rather, they seek a copyright regime that maximizes their own welfare, whatever that may be. As a result, whatever data is available, and whatever uncertainty or ambiguity remains, they will twist in order to ensure that copyright serves them, rather than society as a whole. Whether due to the first problem or the second, there are also surprisingly few empirical studies of the costs and benefits associated with broadening or narrowing copyright. This study aims to help fill that gap by examining how file sharing has affected the creation and dissemination of new original music.

The first is familiar: Lower returns for creating music will lead to fewer new authors and artists, as lower returns lead individuals to devote their talents elsewhere in the economy. The second may be less familiar, but is no less real: Lower returns for any given work means that those individuals who do become authors or artists will, on average, each create more works.

when wages are low, an initial wage increase leads individuals to devote more time to working, as the higher wage leads individuals to substitute work for leisure (the “substitution” effect). As wages continue to rise, however, higher wages generate higher income, which leads in turn to increased demand for a variety of goods including leisure (the “income” effect). Once wages increase to the point where the desire for leisure from the income effect outweighs the desire for work from the substitution effect, further wage increases actually lead to less time spent on work. Economists refer to this as the backward-bending labor supply curve.

Economic theory thus suggests that we should expect broader copyright to lead to more works from having more new authors, but at the potential cost of fewer works from our most popular existing authors. Narrowing copyright, on the other hand, whether through a formal statutory amendment or through an informal technological development such as the rise of file sharing, should have the opposite effect: fewer works from having fewer new authors or artists entering the field, but more works from our most popular existing authors.

Whether file sharing will lead to more or fewer works, on balance, then, depends on the relative magnitude of these two marginal effects.

While the sharp decline in revenue that paralleled the rise of file sharing is associated with a decrease in the number of new artists in the study’s sample of the top fifty each year, ceteris paribus, it is also associated with an increase in the average number of hits by each new artist who does make it, ceteris paribus. Moreover, because the latter effect outweighed the former, the decline in revenue since file sharing began was not associated with any decline in music output. To the contrary, the decline in revenue was associated with an increase in the number of new hit songs that appeared annually in the study’s sample of the top fifty, ceteris paribus.

This study therefore concludes that file sharing has not reduced the creation of new original music. It may have led to fewer works as a result of fewer new artists entering the field, but it was also associated with an increase in output by those artists who chose, despite the lower returns, to devote their talents to making music. Given that file sharing undeniably promotes the broad dissemination of existing works, this conclusion suggests that file sharing is both fully consonant with copyright’s constitutionally-delimited purposes and welfare enhancing. While the study has limitations, particularly those associated with the use of observational data generally, it nonetheless suggests that file sharing is not, for the music industry, at least, the problem that the copyright industries have claimed.

While the data represents the longest and most detailed study of file sharing’s impact on music output that I have seen, it still covers only twenty-nine years and one country.