The NME  (remember the NME?):
Sales of "old albums" have overtaken sales of "new albums" for the first time over the last six months in the US.
Sales of "old albums", which are classified as LPs that have been on sale for longer than 18 months, numbered at 76.6 million over the last six months, with sales of "new albums" numbering at 73.9 million, reports OC Weekly
Copyright laws are transparently no longer about rewarding artists or incentivising creation. The idea that there is, or ever has been, a musician who sat down to record and then thought "you know what? I would create this art, but my descendants will only get to reap the rewards for fifty years after I die, rather than my preferred seventy," is ridiculous.
What this news shows, though, is one very real reason why long copyright terms might be important: protectionism against the past.
Even with a copyright system which keeps pretty much every song recorded since World War II - and a number recorded before - out of the public domain, "old" albums are still outselling new ones in the US. "Old" is, in this case, defined pretty loosely, but it is hard to imagine what the music market would look like if copyright terms were reduced, even if just to the life of the artist.
Imagine being able to get every song ever recorded by Elvis, Jim Morrison, Elliott Smith or Notorious B.I.G. for free, legally. Would you spend as much on new artists? Would you, in fact, spend anything on new artists?
It's obviously not the case that these long terms directly help up-and-coming musicians. Almost all of the money on every Tupac album sold goes to Universal Music Group, and most of the rest goes to his heirs. While there is the argument that any extra income to record labels helps them take greater risks on new talent, it seems unlikely that that actually translates into them getting a cut of the sales.
But where it might help them is by boosting demand. If you are ambivalent between a Beatles and a Battles album, you are that much more likely to buy the latter if it doesn't cost a thousand times more.
Protectionism against the past, then: artificially raising the price of something you don't want to sell in order to make the thing you want to promote look better.
Of course, there's no guarantee that that actually works. All of the above assumes that people begin with a fixed amount of music that they want to consume, and that every "old" album they listen to is a "new" one they won't. But it is equally as likely that, under the current situation at least, people have a fixed amount of money they want to spend on music, and that making older albums free would increase, rather than decrease, what they spend on new artists.
How to tell the difference? All-you-can-eat subscriptions might provide the answer. Someone who pays £10 a month for Spotify has access to more music than they could reasonably listen to in their lifetime. If all they want is a fixed quantity of music, then they won't spend anything else. If they want to spend a certain amount of money, then they'll start spending more on albums and bands not on the service. Sadly, no-one seems to have done that study, though if anyone does know the answer, I'd be fascinated to see it.
This is all economist noodling, though. As ever, the model has been simplified, and in the real world people don't actually think of "music" as a vast homogenous mass which they purchase. A Radiohead fan won't stop buying their albums just because they could get Elvis for free, and someone who thinks hair metal is the pinnacle of generic perfection is unlikey to buy music from past 1979 no matter how much it costs. But I certainly would like some free Marvin Gaye.