In this case, observing the actual spending practices of people can only set a lower bound on willingness to pay. If you're obtaining something for free, it certainly doesn't preclude you from being willing to pay for it as well. The fact that people spent significantly more on music (per capita) in the past[1] -- when it wasn't nearly as easy to obtain music for free -- suggests that how much people are currently paying is probably a poor measure of their willingness to pay.

[1] http://www.businessinsider.com/these-charts-explain-the-real...