I think in Piketty's defence, he set out to gather historical evidence and try to put inequality in a historical perspective.

Due to the kind of records available, in order to construct some of the long-run measures he is interested in, he has to use market values, because that is what there is data for. (Historical tax records in the UK, for instance regarding Estate tax work in this way, I understand.)

There have been some good discussions over at CrookedTimber. I'll try to reprise some of my thoughts from that discussion over here in some other comments.