Art Carden and Steve Horwitz have an absolutely essential article up at the Library of Economics and Liberty about market failure. Why is this article absolutely essential? Because, although market failure is a real thing, it is too often naively assumed that the proper response to any and all market failures is to override the market. It’s not. From the article:
Externalities, public goods, asymmetric information, and market power provide necessary—but insufficient—conditions for intervention to be justified. They certainly are not talismans that provide interventionists with carte blanche to tinker with the members of a society as if they were pieces on a chessboard. Too often, critics of markets think that merely invoking these terms destroys the case for free markets.
So market failure is when, because of externalities, public goods, etc., the market fails to arrive a the best solution. The problem is that imposing some kind of non-market solution raises new risks, for example that the government intervention will fail as or even more spectacularly than the market would.
The article does a great job of going into depth on each of the varieties of market failure, and I’ll just add my $0.02 to their conclusion. There are twin extremes that need to be avoided. The first is the extreme of assuming that the market is always right. The second is the extreme of assuming that, whenever the market fails, government is always right. Unfortunately–while the first is easily recognized as an extreme by most folks–the second frequently goes unchallenged. It ought to be.
There is no perfect ideological answer to real-world problems. Sometimes market failures really do require intervention. Sometimes, however, they don’t. One of the biggest differences between liberals and conservatives, in my experience and according to psychological research, is that conservatives can accept when the optimal solution is just unreachable and a second-best strategy is the only thing we can hope for.