Co-blogger Eric Posner writes:

Ilya, citing a column by Steven Landsburg, asks why a bailout is necessary. In Landsburg’s words, “Just because the banks disappear doesn't mean the lenders will. Borrowers will still want to borrow and lenders will still want to lend. The only question is whether they'll be able to find each other.” That’s the question and the problem with answering the question is that no one knows the answer. History suggests, however, that the probability they won’t “find each other” is not zero; let’s call it one percent. If there is a one percent chance that the current financial meltdown causes a catastrophic outcome—not just people out of work and the reduction in the standard of living, but predictable bursts of xenophobia, beggar-thy-neighbor policies, global political instability, and all the rest—how much should the government do to prevent that from happening and at what cost?

Let's assume there is a small chance that, absent a bailout, there will be a "catastrophic outcome" of the sort Eric describes. That possibility is not by itself sufficient to justify the bailout. One also has to consider the risks of the bailout itself. Some of these are highly likely to happen, such as hundreds of billions of dollars of expenses borne by taxpayers, the creation of moral hazard for industries who will expect to be bailed out in the future, and the economic inefficiency caused by government control of a large part of the economy. In addition, there is a small but probably nonzero risk that the bailout itself might lead to a "catastrophic outcome" thatt goes beyond the onerous costs described above.

For example, despite the two year sunset provision in the draft legislation, that legislation might end up getting extended, leading to permanent government control of the finance industry, and an eventual monopolization of finance by government; a future administration could claim that the government cannot manage its massive portfolio of finance firms efficiently unless it has a monopoly of lending, and future, economically illiterate public opinion, might buy the claim. That in turn might lead to a severe permanent decline in our standard of living, as we spiral into quasi-socialism. By transferring enormous resources from successful businesses to failing ones, the bailout could also damage the economy enough to cause a deep recession, which in turn could lead to "predictable bursts of xenophobia, beggar-thy-neighbor policies, global political instability, and all the rest."

I don't consider these scenarios to be likely. But if we are going to justify the bailout by weighing low-probability catastrophic risks, we have to consider both sides of the equation. We can't defend the bailout by claiming that it might stop a low-probability catastrophic event without considering the possibility that it might itself become the cause of such a catastrophe.

There are low-probability catastrophic risks on both sides. When you consider the fact that the bailout also has severe non-catastrophic costs that have a very high probability of occurring, the case for it still hasn't been made.