Liebowitz sees decline, blames downloading. I disagree.

Stan Liebowitz has posted an updated and revised examination of the impact of MP3 downloads on record sales.

The bottom line: MP3 downloading is harming sales. No other explanations that have been put forward to explain the recent decline hold up under analysis.
Nevertheless, how far should we be willing to go to protect the record industry from a 20-25%
decline in business? Are recent proposals, such as the Berman bill going too far? My own view is that
such proposals are going too far. Should we switch to a non-market alternative as suggested by Lessig,
Natenal, and Romer? I believe that switching to a non-market alternative should be an absolutely
last-resort policy [...] Allowing record companies to protect their wares using digital right management technology seems to be a far more reasonable alternative.

I've just managed to read this through in detail; it is a thorough piece of work, as before, and measured in its conclusions, but it is still missing a few key measures.

Stan notes that the 'amount of time spent on activities' survey he uses to dismiss displacement of music by DVD or other activities, is methodologically flawed.

The librarying of DVDs rather than albums is not really tested. Seeing per capita DVD sales (and rentals) on the same composite chart showing the different music formats would be very interesting.

The cell-phone/text messaging/chat displacement should not be underestimated either. The amount spent on cellphone connectivity by the text-message generation is enough to displace a fair bit of prerecorded entertainment spending.

These quibbles aside, there is a key methodological problem at the core of the project. It works from the assumption that the RIAA records sold represent the totality of music consumption, and that any online music downloaded necessarily is an infringement of RIAA copyrights.

In fact, there is a great deal of music available online that is from artists outside the RIAA orbit, and freely made available, sometimes to promote CD sales, sometimes just for pleasure. This does reduce RIAA profits, but by displacement, not by theft.

A more subtle point is that network theory predicts that as interconnections between customers grow, the chance of a big cascade hit grows at first, but then as they are exposed to a more diverse range of advice and opinions, it falls off. The hugely increased connectivity that the net and cellphones enable may have caused the phase transition on the customer base.

This is discussed in Duncan Watts' classic paper.
In either of these two scenarios, the independent labels and musicians do better.