Last night the Treasury and Barney Frank, head of the House Financial Services Committee, unveiled a draft of legislation intended to address the regulation of “too big to fail” banks and financial institutions.

The main provision of the jointly-written legislation would be to impose the costs of bailouts of a systemically-important firms on other big firms. Regulators would assess companies with more than $10 billion in assets with fees to go into a pool to finance bailouts of other “too big to fail” banks.

The idea is that any firms large enough to warrant an implicit bailout should bear the costs of any similar bank that is bailed out because regulators consider it too big to fail, thereby sparing taxpayers the bill for rescuing companies that took on too much risk because they knew the feds had their back.

The New York Times is reporting that the assessments for the bailout fund would only be levied after the collapse of a systemically important firm.

Initial thoughts:

1. This plan would give big banks incentives very similar to those that Fannie and Freddie faced. Once firms reach the arbitrary $10 billion cutoff, they know that they have socialized risks. Bank managers will not care if those risks are shared with other large financial institutions or the general public. In other words, the regulators will create an identifiable group of firms that are designated to receive bailouts in case of a collapse.

2. The provision that the assessments for the bailout fund will only be made after a firm collapses seems like a large drawback. Consider that the recent collapses happened because the industry as a whole was on the verge of a collapse. If this measure had been in place when Bear Stearns was going under, would regulators really have considered going to other banks and asking them to draw on their own weak balance sheets to fund the bailout? Wouldn’t it be more likely that banks would get them to suspend or postpone the assessments?

3. The overall solution is much more palatable for the financial industry than a Glass-Steagall II or firm size-restriction measure would be.

More on this topic soon.