David Friedman recently had an interesting externality exercise post:
If actions I might take would provide benefits for other people which I am not in a position to charge them for, I have too little incentive to take them. If my actions would imposes costs on other people which I am not required to reimburse them for, I have too much incentive to take them. So let government subsidize or mandate the production of positive externalities, tax or ban the production of negative externalities, thus making us all better off.
I offer the following challenge to readers. List all the positive and negative externalities from educating children. For a second challenge, pick some other public policy commonly defended on externality grounds, and try to list the externalities with the wrong sign–the ones that are an argument for subsidizing what we now tax, or taxing what we now subsidize.
He discusses examples from population and education; I naturally wonder if we can come up with energy and environment examples. Any ideas? How about the wind power subsidies we were discussing last week?
My starting point for such ideas is an idea that is implicit in his laying out the externality accounting problem: yes, living in society means that our actions and outcomes are frequently interdependent. But just because your action affects my welfare does not imply that every such interaction needs to be “internalized”. The “externality” can be sufficiently small that at the margin, if it were internalized, it wouldn’t change the outcome. In such a case, a government policy to internalize an externality is only redistributive. It doesn’t increase efficiency if it doesn’t change the outcome.
A concrete, simple example: my neighbor has a nice front garden. I derive value from her garden, but I don’t pay for the plants. If there’s a (Pareto-relevant) externality, the fact that I don’t pay induces her to plant less and have a smaller garden than she would otherwise. But consider the fact that she probably loves gardening and is the one who gets the most aggregate benefit from her garden, and I have a smaller benefit. If that is the case, it’s possible that at the margin, my marginal benefit from her garden is low enough that if I pay her for my value, it won’t change the size of her garden. In that case, the externality is not Pareto-relevant, if I pay her she won’t plant more plants, and my payment to her does not change the outcome, it just redistributes resources from me to her.
Just because an externality exists, that doesn’t mean it’s relevant. If it’s not relevant and we implement a policy to internalize it, that policy is redistributive.
How do we identify which externalities are relevant and which are irrelevant? That’s tricky. In the unreal world of fully defined property rights and zero transaction costs, relevant externalities are the ones that private parties go out of their way to internalize through side payments. But ill-defined property rights and positive transaction costs obscure that relationship. So in the real world, what do we do? I think we focus on identifying situations in which transaction costs are high enough to cause this problem, and recast the role of public policy as the reduction of transaction costs.