Over at EconLog, commenter Thaomas took friendly objection to David Henderson’s insistence that there ain’t no such thing as a free lunch (“TANSTAAFL”) . Here’s Thaomas’s comment in full:
Actually, out of equilibrium, There Is Such a Thing As a Free Lunch. Any improvement in policy is a free lunch, unless one means that it is very unlikely that literally everyone will benefit from a better policy. A revenue neutral move to lower corporate taxes would have been a free lunch. Free trade would be a free lunch. Higher immigration would be a free lunch.
I think that economists’ main job should be to look for ways of creating free lunches and preventing the destruction of free lunches.
I agree with what I take to be the spirit of Thaomas’s comment – namely, that opportunities for gains from trade are everywhere and that “economists’ main job should be to look for ways” to clear the way for these gains actually to be gotten. But I disagree that there exist free lunches.
The examples of free lunches that Thaomas mentions are examples of sources of gains from trade. And I am wholly on board with making trade and immigration freer, and with lowering corporate taxes (in budget-neutral ways). Economically, policy moves in these directions would surely create net gains. But as Jon Murphy points out in a response to Thaomas, these gains would not be free.
The point of my post here – which consists below of sharing two comments that I left at EconLog – will strike many readers as pedantic. But I plead that it’s not. (It is, however, a post meant largely for my fellow economists.) Getting language right is important, for if we don’t we are easily misled.
Here’s my first comment; it’s in direct response to Thaomas’s comment quoted above:
I agree with Jon Murphy. Thaomas here seems to use “free lunch” in a way that is common but is mistaken – namely, to indicate unambiguous net gain in ‘social’ welfare. But as Jon points out, the fact that there’s a gain doesn’t mean that the action that was undertaken to bring that gain about is costless. Something was sacrificed, and the (subjective) value of that which was sacrificed is the cost. The fact that the cost is less than the benefit does not mean that this benefit is free.
By the way, this same conclusion holds even for changes that economists call “Pareto-improving changes” – at least insofar as such changes result from conscious human choices. (“Pareto-improving changes” are changes that leave at least one person better off while leaving no one worse off.)
Change brought about by human choice requires foregoing something; this truth holds even for Pareto-improving change. The value of that which is foregone is a cost. This cost might be minuscule relative to the gain; but it is nevertheless a cost. Because it is impossible – literally impossible – to choose without incurring a cost, it is impossible to choose in ways that bring about change without that change having a cost.
Here’s a simple demonstration of the cost that attends a Pareto-improving change:
Sam values Sara’s apple more than Sam values his (Sam’s) peach, while Sara values Sam’s peach more than Sara values her (Sara’s) apple. Sara and Sam thus agree to exchange. Sam gives Sara his peach in exchange for Sara giving Sam her apple. Both parties are made better off while no one is made worse off. Still, costs are involved. The cost to Sam of securing his gain is the subjective value that he anticipated he would have gotten from eating his peach had he not transferred it to Sara. The cost to Sara of securing her gain is the subjective value that she anticipated that she would have gotten from eating her apple had she not transferred it to Sam.
My second comment was sparked by another commenter (“artifex”) writing, in response to my first comment, this: “Unlike your example of exchanging an apple for a peach, the standard examples of comparative advantage leave all parties involved with at least as much of all goods as before.” However, my second comment (just below, and modified slightly) goes well beyond a reply to artifex:
My example [in my first comment, above] of the exchange of a peach for an apple is not intended as an example of the operation of comparative advantage. A standard example of the latter, such as this one , involves production; my example involved no production and wasn’t meant to do so.
My peach-apple example was meant to show the simple point that economic change brought about by human choice involves cost even if that change makes at least one person better off without making anyone worse off. Each party in my example experienced gain, but each party nevertheless incurred a cost to achieve his and her gain.
Nothing in my example is anything beyond second-chapter ECON 101. Yet it is common practice for economists to draw from it a conclusion that I think is mistaken. (I say “I,” but I really mean “I who have learned much from Jim Buchanan  and scholars in the LSE subject-cost tradition .”) The mistaken conclusion is that the gain that arises from an exchange of the sort that I put as an example involves no cost to society.
The problem with such a claim about “no cost to society” is that society is not a sentient creature that incurs or experiences costs (or benefits, or anything else). The claim “no cost to society” would be correct only if some third creature, not Sara or Sam but “Society,” had a capacity to experience gains and losses and experienced these gains and losses by being somehow connected to Sam’s and Sara’s actions (or to each of their utility functions) – and then, this “Society,” experiences “gain” when the actions of Sara and Sam result in net gains to Sara and Sam, and experiences “loss” when the actions of Sara and Sam result in net losses to those two persons. This “Society’s” gains and losses arise for it independently of any choices that it makes, for it – “Society” – makes no choices. In this example, “Society” is merely an entity that experiences the net consequences of the choices and actions of others.
In reality, of course, no such creature, “Society,” exists (although people, including most economists, regularly talk and write as if there is such a creature). In reality, costs and benefits are experienced only by individuals.
In my simple example, each party obviously experienced a cost – namely, the sacrifice of the anticipated subjective value of continuing to possess that which he or she gave up as a means of gaining the opportunity to experience the subjective value of possessing that which he or she received in return. That this sacrifice is a cost can be seen by recognizing that each party would prefer to get what the other has without having to give up that which he or she has. Sam would prefer to have both the peach and the apple rather than only the apple. Similarly, of course, for Sara.
The fact that each party, at the moment of choice, experiences a gain in subjective utility without any other person suffering a loss in subjective utility means neither that these gains are free lunches nor that no costs are incurred in society in order to make these gains a reality.
In summary, it is misleading to talk of “costs to” and “benefits for” society. Yet I don’t go so far as some people and reject the concept of society as meaningless. Society is real, and the word “society” is meaningful. But society is made up exclusively of individuals; it is not a creature independent of the individuals who comprise it. Society, as such, experiences nothing. But obviously the people who are in society experience much.
In my example of Sam and Sara exchanging an apple and a peach, costs are incurred in society but not by society. The gains, although net, are not free lunches.