Knowledge Problem

He Gets Vaccinated So He Can Get Kissed

Lynne Kiesling

Over at Marginal Revolution Alex reveals that he has gotten a flu shot, and uses that fact to draw out a brief analysis of the external benefits that he does not entirely get to enjoy from his shot: his shot reduces contagion, reducing illness in others, which has only a small benefit that redounds to him. He closes by asking for suggestions on how to encourage more people to get flu shots, short of flat-out coercion and mandatory flu shots.

Let’s turn Alex’s point on its head. Alex gets a flu shot because he wants to get kissed.

The primary reason Alex gets a flu shot is to reduce his (and his family’s) probability of illness. A second reason may be to reduce the illness probability facing his friends. A third reason may be to reduce the illness probability facing his students.

This is the crucial point to remember whenever anyone starts talking about tax or subsidy policies to “internalize externalities”: at the margin, the greatest beneficiary of Alex’s action is Alex. If the marginal benefit to him of getting the shot is larger than the marginal cost, then he’ll do it, even if we don’t have some elaborate scheme of subsidies to compensate him for the benefit he generates for us. If his desire to prevent illness in himself and in his close circle is high enough, he’ll do it, and any payment we make to him at the margin could be inframarginal; it might not change his behavior.

That case is what Buchanan & Stubblebine (Economica, 1961) classified as a Pareto-irrelevant, or just irrelevant, externality. In that case, a system of payments provides nothing more than a wealth transfer without changing the actual amount of the beneficial behavior that occurs.

Note the important policy implication of this insight: not all externalities need to be eliminated to achieve efficiency and the optimal amount of the behavior in question. In an efficient equilibrium, externalities still exist, it’s just that they are inframarginal.

So let’s sharpen Alex’s question. What is a good policy for ensuring that Pareto-relevant externalities are reduced or eliminated? How do we determine if an externality like the one he’s describing is relevant or irrelevant?