Home -> Law Blog Directory -> Legal Commentary Blogs -> Economics of Contempt

Legal Commentary

: Economics of Contempt

Debating the Optimal Bank Size

By Anonymous

Everyone seems to agree that financial regulations should strive to prevent banks from ever becoming "too big to fail" in the first place. First, you have to identify the point at which a bank becomes "too big to fail." Felix Salmon proposes limits based on bank assets:

To get specific, I think that maybe $300 billion in assets would be a reasonable cap on bank size -- there's very little evidence that banks get any economies of scale beyond that in any case.

If the government is serious about limiting the size of banks, then a limit of $300 billion in assets is way too high.

When Continental Illinois was on the verge of failure in 1984, the FDIC determined that the bank was too big to fail, even though it held only $84 billion in assets (adjusted for inflation). Instead of immediately taking Continental into receivership, the FDIC provided a temporary bailout?it injected $2 billion into Continental, and guaranteed all of the bank's depositors and general creditors against loss. The FDIC's decision to protect all of Continental's depositors and general creditors was extremely controversial at the time, and led to a prolonged debate about the "too big to fail" problem.

So if a bank with only $84 billion in assets is too big to fail, capping bank size at $300 billion doesn't make much sense.

In my experience, the definition of "too big to fail" fluctuates depending on the state of the financial markets. When financial markets are fragile and investors are on edge, banks that in normal times would be considered too small to pose systemic risks suddenly become too big to fail. When times are good and financial markets are relatively stable, only the biggest banks are considered too big to fail. For this reason, any express regulatory limit on bank size needs to allow for at least some flexibility.

Full post as published by Economics of Contempt on March 31, 2009 (boomark / email).