The other day, I posted on a paper by Doug Shackelford, Dan Shaviro, and Joel Slemrod that is a terrific framework for thinking about bank taxes. The authors looked at four ways to tax the banking business in the wake of the recent financial collapse—a transactions tax, a tax on bonuses paid to employees, and two levies on banks themselves.

They rejected the first two as either misguided or unworkable, but suggested there might be some benefit to taxing financial institutions. They looked at two basic models: A tax on “excess” profits similar to a plan developed by the International Monetary Fund, and a tax on certain bank liabilities. Several readers have asked for more about these two ideas, so here goes:

The IMF’s Financial Activities Tax (the FAT): The levy would be imposed on profits in excess of “normal” returns as well as employee pay above a base level. Former Joint Committee on Taxation chief Ed Kleinbard and Tim Edgar proposed a similar tax. The idea is that huge financial sector profits and bonuses are a proxy for inappropriate risk (the recent losses from which were borne to a troubling degree by taxpayers).

Such a tax may not be easy to implement. Determining profits that exceed normal is fraught with issues. Doing the same for compensation is a bigger challenge. As even the IMF notes, distinguishing excess pay from rewards for high productivity is an exercise in rough justice at best. Doug, Dan, and Joel raise yet another concern: If such a tax is imposed only on financial firms, these companies will find ways to redefine themselves to avoid the levy.

A Tax on Bank Liabilities: The second model is the one now being debated in Washington, as well as in Germany and the U.K. This tax would be imposed on all bank liabilities except for federally-insured deposits. A version proposed by the Obama Administration would tax big banks only.

Like the IMF’s plan, the goal is to both discourage risky behavior and recognize that occasional financial collapses are inevitable and that society, and not just imprudent banks, will pay the price. But there has been another issue in the U.S. debate: what to do with the money?

The original Obama bill would have raised about $90 billion over 10 years to repay the cost of the recent bailouts. Once enough was collected to do the job, the tax would expire. The Senate version was intended to finance a new bank bailout fund. But a tax whose major purpose is to finance a designated program has a big problem. If the levy also is enacted in part to discourage banks from taking inappropriate risk, wouldn’t they resume their bad behavior after the tax ended?

This design raises other questions as well. Should the tax be levied on only uninsured funds or on all deposits? Should the rate rise with the level of risk, and if so, how does government measure this? And since no tax can replace regulation (despite the fervent wish of some economists) how would a levy mesh with a new regime of bank rules?

Still, Doug, Dan, and Joel believe a bank liabilities tax may be the best option, especially if the goal is to discourage excessive risk. But Congress will have to be very careful in the way it designs the tax to avoid some troubling unintended consequences.