"Too big to fail"

We've heard this phrase a few times, most recently with Fannie Mae, Freddie Mac, and AIG: it is "Too big to fail." I heard it yesterday on a financial show talking about how important the banking sector is to the economy - the commentator said that if a shoe factory fails, it fails and someone else will make shoes, but if the banking sector fails (as the credit freeze demonstrates) then it creates a lot of collateral damage.

I don't have a problem with logic that declares something "too big to fail" as such, but it occurs to me that anytime we use this phrase, there are two implications which we cannot ignore.

The first is that if something is too big too fail, that it must be regulated. I'm not a fan of excessive regulation, and I believe in markets, but markets only work because the risk of failure keeps investors and businesses prudent. I.e., excessive speculation and risk taking are curbed by the possibility of losses. Take away the possibility of failure, and you are creating incentives for reckless behavior - writing bad insurance policies, loaning to people who are not creditworthy, etc. So if we are going to label an entity as being too big to fail, we must compensate for this by replacing the market-based constraints on risk taking with formal regulatory constraints. Otherwise, nothing will prevent the conditions that led to the near-failure in the first place.

The second implication is that if something is too big to fail, then there has been a marketplace breakdown that has concentrated too much market share in that entity. One of the great things about a marketplace is that their distributed nature make them resilient to individual failures - in fact, those failures are a necessary and integral part of the functioning of a marketplace. Risk taking is rewarded when wisely taken; innovation necessarily involves risks. And failure checks excessive risk taking and weeds out bad ideas and weak execution. Without failure, there can be no innovation, no learning. A marketplace that does not have enough diversity of players to suffer a periodic failure of one or more of those players is therefore not a functioning one.

A concentrated market may not rise to the level of illegal monopoly, but I would argue that it's effects can be just as bad. Therefore, per my regulatory argument above, we have a choice in these situations. We can fix the marketplace by finding mechanisms to create the distributed failure-tolerant environment I describe above that is an integral aspect of a functioning market. Or we can decide for one reason or another that we are OK with the market concentration and instead choose to replace the risk of failure with a regulatory regime.

What is not a viable option, though, is to choose not to choose. If something is too big to fail, we cannot rescue it and then do nothing to either fix the market concentration or regulate it. Otherwise, we are simply inviting more of the same problems.