David Gordon, in his book Essential Rothbard, writes approvingly of Rothbard's dismissing the idea that positive externalities are a social problem:
A and B decide to pay for the building of a dam for their use; C benefits though he did not pay. ... This is the problem of the Free Rider. Yet it is difficult to understand what the hullabaloo is all about.
Am I to be specially taxed because I enjoy the sight of my neighbor's garden without paying for it? A's and B's purchase of a good reveals that they are willing to pay for it; if it indirectly benefits C, no one is the loser. (28)
But isn't it obvious that the "social problem" arises not when the dam benefits C but precisely when the dam is not produced in the first place, because it is non-excludable, and people like C will free ride on it? If it were possible to offer to sell the dam's services to C at the price sufficient to cover costs, then C would agree to pay.
The positive externality is then a "problem" not because it is enjoyed by C, but because it is not enjoyed by anyone including C.
I mean, even the most primitive undergraduate micro textbook will argue that goods with positive externalities are "underproduced."
No one is saying that Rothbard ought to be taxed for enjoying the neighbor's garden. The "tax" is a (bad) solution to the problem. But the problem itself remains: since, in building his garden, my neighbor fails to count my pleasure in seeing it, there are fewer gardens than there would be if such benefits were internalized. Some predictable -- even by economists -- increase in human happiness is therefore unrealized.
Gordon does not help it when he does not present the opposing view as strongly as possible. Why would he battle a straw man? This is a pity, because Rothbard's argument has merit and deserves to be developed further. Perhaps Gordon would reply that Rothbard rejected the idea of efficiency in economics as "operationally meaningless." (29-30) Therefore, such underproduction cannot be condemned as inefficient.
Now I, too, consider neoclassical efficiency to be a naive and hopeless concept. But, first, economic efficiency is not thereby rendered a meaningless notion; it is fully tractable when applied to the market process within the Austrian tradition, as I show in my SAtK, I, 10-14.
Second, even within conventional econ, we can still have a little model in which an excludable dam would yield better results than a non-excludable one.
At the same time, I think that we are uniquely "helpless" before most externalities, whether positive or negative, and should almost never involve the state in "fixing" them, as in the longer run this will do more harm than good. Externalities must simply be endured.
Externalities then are a metaphysical complaint, a sort of economic problem of evil.
But for all that, there is evil here, and a world without externalities would be happier than the present world.
In other words, externalities are of the same type of evil as:
- entropy, or
- the fact that no factory is 100% efficient and in particular that all production emits some waste, or
- that human bodies are fragile, or
- that some people are unjust, or
- scarcity of the factors of production
is an evil. It's part of the design of this universe; some of these evils can be mitigated with time via economic progress but never eliminated; nor does the state play any role in their mitigation.
For example, technological improvement might help us to internalize some externalities, analogously to how it helped us, say, to hail an Uber taxi easily.
In another example of transaction costs, Rothbard writes: "What is so terrible about transaction costs? On what basis are they considered the ultimate evil, so that their minimization must override all other considerations of choice, freedom, and justice?" Gordon comments: "If one responds that reducing these costs has some, but not overriding importance, Rothbard's question compels one to specify how much, and why, they are to count." (34) Well, they are costs, and hopefully sooner or later, solutions that increase welfare by diminishing these particular costs, too, will be found. They count no more and no less than other costs that also constrain human happiness.
Note though that if such solutions are found, they will be found by entrepreneurs not by economists. Most economists are terrible entrepreneurs, and vice versa. Economists should not aspire to drive the market. So, I agree that this "approach" to welfare economics is absurd: economists can contribute little to diminishing transaction costs.
For these reasons, I agree that externalities are not a "social" problem or physical evils; but they are still metaphysical evils, an ultimate and irreparable defect in the nature of the world.