Economists (namely my business and govt lecturers) continue to blithely propagate the idea that market allocate resources efficiently. Of course, on a good day, they'll remind you that this assumes "you've got an optimal distribution of income", and promptly forget about it altogether. They'll start using phrases like "and this is why we must cut funding for social services." The argument is purely one of efficient allocation of goods.
I think that the major problem with all this is that economists believe that in an ideal world income is only influenced by how much someone wants that income. If it were true then incomes would be "optimally distributed" and we could leap to all the counterintuitive conclusions to which economists regularly leap. But any economist will happily tell you that incomes are influenced by all sorts of things. In fact, one of the core components of wage theory is that wages are influenced by both the willingness of people to work and the need employers have for their work. The willingness to work can be equated with how much someone wants income relative to how much they want spare time.
To make matters more ridiculous this is only looking at the combined willingness of all workers and the combined need for work of all employers. The need that employers have for each individual worker is very different. This half the explanation for why people incomes are different. The outcome of all this is that a person's income has very little at all to do with how much they want income.
I think that income serves too many purposes for it to be useful in telling us how much stuff people want to have. You can argue that if someone wants more stuff they'll simply go out and earn more income. So that when they go out and buy something we can safely say that the transaction is "optimal". Which merely means they wanted that thing more than anyone else did. But because income is also a way of motivating people to work it loses it's meaning as a measurement of desire for things.
To put it simply, income is only a useful proxy for utility when income is a function of and individual's demand for income. To the extent that it is a function of demand for income, then market's are "efficient". Since economists themselves show that income is a function of many other things, it seems unlikely that our free markets are going to be particularly efficient at all. And the old "efficient by not equitable" argument doesn't cut it. Economists mix up equity of income and equity of market outcome. It's true that given some distribution of income, a market will find the optimal-ish outcome. But who decided that markets should only be measured on their efficiency once the distribution is known. You can totally ignore inequity and still say that markets will not achieve an efficient outcome. Because the income distribution you're given not purely a function of utility. Given we don't really know what the utility distribution looks like - the income distribution is just a very lazy approximation - who's to say that the government doesn't have as good an idea of the optimal distribution as anyone.
The interesting thing about this is that you'd probably get a much more efficient distribution of resources if income was earned by queuing in a line for dollars. Everyone has the same number of hours in a day. The one small problem is that there wouldn't be anything to buy once you left the queue. Which leads you to the conclusion that communist-style rationing is probably far more efficient than capitalism. Capitalism actually trades efficiency for higher production.