Wednesday, January 25, 2012

externalities ⊄ market failures

This non-technical paper has a really great presentation of four ways in which externalities are often corrected without government intervention. (Oftentimes, since externalities mean that the full cost and benefit of a decision is not felt by the decisionmaker, he may make the wrong decision, and that this can lead to market failure, in which society would be better off if in sum if he made a different decision.) Even the jargon is explained along the way, so I highly recommend it to non-economists wanting to understand the topic.* To summarize:

  1. Simple Coasian internalization: Coase's theorem says that if property rights are clearly enough defined, and transaction costs are low enough, externalities will not lead to market failure. That is, if someone wants to do something (like paint their house chartreuse) that has a side-effect on someone else (the homeowners on the block whose property values drop), as long as it's easy enough for homeowners to get together and negotiate, they will either agree to pay off the chartreuse-lover not to paint his home or to pay off the other homeowners to compensate for the loss in property value. Which of these outcomes is chosen is simply which one leads to the greatest total good. It's important to note here that externalities are symmetric: you may be hurt by my smoking, but I would be hurt by your smoking ban. So, whatever the outcome, someone is going to be harmed, and the optimal outcome leads to the person being harmed who least minds it, regardless of how money changes hands in order to get everyone to agree to that outcome (because side payments are zero-sum and therefore irrelevant to the utilitarian policy analyst).
  2. Complex Coasian internalization: But, sometimes it's not so easy for parties to negotiate, or it's not even clear who owns the entities that are being affected by the externality. Lighthouses can't negotiate with passing ships about whether to provide their services, and no one owns coastal safety. Therefore, you might expect a shortage of lighthouses. But, if different groups with different motivations can combine, suddenly it might be privately worthwhile to provide that public good after all because the affected party essentially is the affecting party (and therefore transaction costs don't matter either). For example, if harbor owners also run lighthouses, now they have an incentive to provide the right number of them, because they will be able to charge harbor fees in the amount that reflects how much boats value having a safe place to go.
  3. Informal mechanisms: Basically, even without governments, property rights can be effectively defined and enforced and transactions handled through social institutions and norms. That is, the above conditions for Coase's theorem to hold don't need to come from the government. For example, no one may own the space in a common area, and it may be impossible to negotiate with everyone who uses it, but nonetheless, social norms can dictate that smokers stay in one corner away from the non-smokers, and that trespasses against this norm are punished with evil glares.
  4. The externality just may not exist at the relevant margin: This is a bit of a facetious way to argue that externalities don't always cause market failures, and I almost dismissed it as such, but actually there are plenty of contexts in which its important to remember... Basically, while some actions do cause externalities when those actions are taken at certain levels, that doesn't mean it happens at the relevant levels. It may be true that libraries are things with wonderful positive externalities, but if there were privately run libraries in every city already, it may not be true that the government needs to worry about a library shortage. Additional libraries don't add any extra positive externality on top of what the old libraries already provide. (I think #4 is worth including because of the phenomenon in which voters consider whether things are good at all but not how much of those things are optimal given inevitable tradeoffs...)

*I'm less convinced by the later argument that there doesn't seem to be an unoptimal amount of cybersecurity, and I have more to say about that, but one blog post at a time...

No comments: