(quoting myself here...)
OK it's not quite as simple as that, because you have to factor in time horizons and the early adoption phase, in which coin production costs are marginal (thus not allowing for efficient distribution). With BTC the design was such that a great deal of coins were produced during the early phase for a very low cost (the first 1,000,000 BTC were mined for a cost of what it would take to mine less than 1 BTC today, roughly speaking).
More efficient distribution would thus require (aside from finding a way to make mining a net positive process) a shorter early adoption phase in which production costs are trivial, so that we could arrive at the point where mining produces tangible positive effects more quickly. However, the early adoption phase should still last long enough and be lucrative enough for the coin to gain popularity.
Efficiency in the context of this thread is the rate (relative to the maximum possible rate) of adoption of the currency as a currency and not as an investment asset. Very few people in normal times (before sovereign debt crisis safe haven demand) considered holding dollars long-term (in your mattress, i.e paying no interest) as an investment. A hallmark feature of a currency is you rarely concern yourself with its exchange value.
I am asserting that the challenge is to design an upstart crypto-currency which has enough near-term investment appeal to drive excitement and investment adoption, while simultaneously driving synergistic adoption for spending which eventually subsumes the investment gains.
Some here are challenging my assertion.
Realize that A comes before B. B is the goal, where at which time A cashes out without ponzi crash, because B supplies demand. This is accomplished by raising transactions to a very high level early with PC mining.