I've already had this debate with Tim in private (we're friends). Some of his analysis of transactional use is quite misleading. For instance: http://letstalkbitcoin.com/blog/post/a-marginal-economy-vers...

Here he compares bitcoin to m-pesa which is a deeply misleading comparison. The former has a floating price, the latter is a money substitute and its price is fixed against the currency which is substitutes for. This is why I've suggested to Tim that it is total purchasing power transferred that should be measured, and I think I have convinced him of that.