Ronald Coase famously advised economists to "look out the window" every so often. It's advice I (try to) take to heart. Here's an example. On Monday afternoon, I was standing behind our building waiting for a few people and "enjoying" the smell of the dumpsters behind the cafeteria next door. An obvious externality that calls for a Pigovian solution if a Coasean solution is unavailable, right?
Not obviously. I can't dig up the cite at the moment, but I recall David Friedman being insistent about making sure the accounting is complete. It's probably rare that an activity produces only positive or only negative externalities. While walking from my car to my office, I noticed something else: the air smelled like bacon. The negative externality of the dumpster smell has to be balanced against the positive externality of the bacon smell. Perhaps that's the implicit Coasean bargain: put up with a slight stench if the wind's right and you're standing right behind your building in the afternoon, and you'll get the pleasant smell of tasty bacon in the morning.
The upshot for public policy is this: we can't make policy designed to fix externalities without taking all the relevant externalities into consideration. For more, here's John Nye's essay in Regulation on "The Pigou Problem." And, while Googling David Friedman and externalities, I was reminded that the Concise Encyclopedia of Economics article on externalities was written by none other than our very own Bryan Caplan.