A while ago, in a comment thread on Stumbling and Mumbling, whilst arguing against someone who was using it as a point in favour of free markets, I asked what the value of economic efficiency was. Generally, efficiency, of itself, has no value, because it is purely instrumental, merely picking out the rate at which resources are converted into achievement of some end, meaning that the moral weight gets shifted mostly on to the end itself. Of course, because there are a plurality of morally important ends and a shortage of resources to achieve them with, it is important to be efficient, because we want to achieve as many of the morally important ends as much as possible given the resources available. That, however, is to do with the achievement of the ends themselves, not efficiency in and of itself, so efficiency per se does not have any intrinsic moral value.
There is however a sense in which efficiency is used by economists, if I remember first-year undergraduate economics with any accuracy, which does seem to have some independent moral weight to it. Economists call a distribution allocatively efficient when cost of the last unit sold is the same to the consumer and to the producer, when the marginal revenue of possessing whatever it is that was sold is equal to the marginal cost of producing it. Assuming - which I'll let slide for the time being - that marginal returns are diminishing and marginal costs increasing, at this point, neither party would benefit from any further trades, because neither would be prepared to meet the conditions necessary to make the trade beneficial to the other side. This is because for the producer, the price has to go up in order to meet costs, whereas for the consumer, the price has to go down in order to compensate for the loss of marginal revenue. For any additional trade to occur would be to make one party better off at the expense of the other, and that seems to have moral weight.
If I take some of your possessions, the mere fact that it makes me in some sense better off does not seem to mitigate or justify it. Granted, most people think that stealing in order to survive, at least at little cost and from those who are under no similar threat, is usually acceptable, but this is not that case. If redistributions were justified on the simple grounds that they made some people in some sense better off, then anyone stealing from anyone would presumably be acceptable, as long as the thief was in some sense better off afterwards, and usually, until they get caught, thieves are better off, or else they wouldn't steal. So, allocative efficiency looks like it points to a situation which redistributions from would be generally morally bad, because they would make one party better off at the expense of the other, and that's generally morally bad.
Allocative efficiency is a property, it is argued, of perfect markets. This, libertarians believe, stands as a moral argument in favour of laissez-faire economics. There are at least two problems here, the first being that there are no perfect markets, which constitute a kind of constraint-less fantasy world in which everyone has identically perfect information, there are no transaction costs, and all goods are undifferentiated. That's not totally decisive though, because some other markets might approximate pretty well to perfect markets, and be, in this respect, morally superior to anything else on offer. What is really decisive is to realise that the costs and benefits in question are dependent on the initial bundles that actors have, and what can be and is done with those bundles. Allocative efficiency can be achieved in a situation where one party has all the resources and the other starves to death, as long as the worse-off party has nothing the better-off party is prepared to exchange any of their resources for. The fact that under such a distribution it is never to both's advantage to make some kind of exchange does not obviously tell us anything about the moral properties of that distribution.
Granted, it's unlikely that there are many situations in which there are no mutually advantageous trades to be made, but many situations in which there are quite clearly lack moral respectability. It is presumably possible to rationally prefer slavery to starvation, so it is presumably possible to achieve allocative efficiency by enslaving the starving, as long as the party doing the enslaving prefers to have slaves over any other outcome. Initial bundles thus matter. Ronald Dworkin has reasonably convincingly argued that at least most, if not all, of the moral force of the idea of allocative efficiency comes from the thought that the costs to the two parties genuinely are equal, that is, represent equal commitments from each, which they could only do under, in a gross simplification, conditions of initial equality. This is because, proverbially, the worth of the ten pounds that keeps the starving man alive for another couple of days and the worth of the ten pounds which pays for my evening in the pub are not the same: the cost of the starving man losing the ten pounds is much greater than the cost to me of losing my ten pounds, because what they get converted into is of hugely different value.
Dworkin not only notes that costs, as comprehended by the market, are relative to initial bundles but to what can be and is done with resources. This is true not only in the simple sense that restricting the use of a particular resource means that is has less value to those who would have used it in that way, a loss which cannot be reflected in the costs to others, but also in the sense that the use of resources by others can increase or decrease the value to me of my resources. Planning regulations are one obvious example: building a skyscraper in the middle of a low-rise residential block imposes a number of costs on those living around it, and if we're interested in making sure that no costs are imposed without compensatory benefits, then that has to be taken into consideration, just as the costs to those who want to build skyscrapers should be. Because of collective action problems, it is often difficult to ensure that these costs are not ignored through the market: where costs are spread across a large group of people, they may find it difficult to coordinate action so as to recover their costs, the obvious solution to which, for Dworkin, is to buy insurance against such costs in the form of regulation.
So, the moral force of the idea of allocative efficiency depends on resources representing equal commitments from both parties to the bargain, and on all the costs, in the full moral sense, of uses on property being included in the assessment of the bargain. It is, politely, far from clear, given that a decent argument for the closed shop can be mounted from this position, that allocative efficiency then actually does tell in favour of laissez-faire economics.
Links added and slight editing, 22/11/055