Glen Whitman wonders why karaoke manufacturers record their own versions of hit songs rather than taking the originals and stripping out the vocals. The result is a nice summary of the law of copyright as it applies to covers and compulsory licensing. Glen’s conclusion:
Here’s your choice as a karaoke producer: You can use your own musicians and sound technicians to recreate the work, and then pay a few cents per song (multiplied by the number of copies made). Or you can use the original track and strip out the vocals; but in order to do so, you must obtain permission from the copyright owner, with all the transaction costs and probably higher price that would entail.
Is this system efficient? On the one hand, it’s clearly a waste of resources to hire musicians and sound technicians to reproduce works that already exist. In addition, the perceived quality will generally be lower than the original, since karaoke singers generally want something as close to the original as possible. On the other hand, extending the property rule to cover indirect duplication would create a hold-out problem: copyright owners could demand high prices for the right to create karaoke tracks. Real resources would be wasted on the negotiation process; worse, if negotiations ever broke down, some great songs might never get converted to karaoke form.
Extending the property rule could be justified on similar grounds to those that justify having copyright law at all: the property rule gives the copyright holder greater incentive to produce creative works in the first place. The added value of more created works outweighs, we hope, the loss from under-use of a nonrivalrous good. But how much added value are we really talking about here? As fun as karaoke is, it seems unlikely to me that demand for karaoke tracks would make the marginal difference between creation and non-creation for most songs. If I’m right, then anything impeding the duplication of songs for karaoke purposes would create an almost pure deadweight loss from under-use. The deadweight loss would take the form of drunken revelers unable to sing their favorite songs because permission has never been granted by the artists. The liability rule creates a reasonable second-best solution.
But could there not be a better solution that avoids the wasteful use of studio musicians? Perhaps copyright law needs a karaoke exception, which would allow the use of original recordings with stripped-out vocals for the same royalty that applies to new recordings.
While I was at Cato, we published a study arguing against compulsory licensing under copyright. I was and still am somewhat skeptical of this argument. The primary deadweight cost to using a property rule for song copyrights is not, as that study suggests, technological costs relating to finding and compensating rights holders. Rather, they are the deadweight costs from copyright holders setting inefficiently high prices in order to maximize revenue.
Copyright is a government-created monopoly. It’s not obvious to me, as a matter of principle, why a government-created right to prevent someone from using a particular work is inherently more libertarian than a government-created right to receive a fixed payment each time the work is used. Generally speaking, it seems to me that the ease with which musicians can create covers of other peoples’ songs has been good for our culture. Perhaps it would make sense to expand that right to include a compulsory license for karaoke machines.