Monday night, when I wrote about tax clawbacks, I was afraid that the idea would be written off too quickly based on an oversimplistic view of the law. Two days later, it's like a movement. At least six members of Congress are on record as trying to craft some sort of tax clawback, Conor Clarke has Larry Tribe opining that a well-crafted clawback would be Constitutional, and widely read bloggers like Kevin Drum and Felix Salmon have considered the idea supportively. Tonight Bloomberg reports that

The senior members of the Senate Finance Committee from both parties proposed taxes totaling 70 percent on bonuses at AIG and other companies getting federal money during the U.S. financial meltdown. House Speaker Nancy Pelosi directed committees there to draft several alternatives and said her chamber may consider a bill as early as this week.

If we're going to do this, and it looks like we might, we had better get it right. Regardless of the legal technicalities, a tax clawback does represent a kind of escalation. It sits awkwardly with norms and ideals that are less a matter of law than we think but that are nevertheless an important part of American political culture. In our better moments, we dislike "collective punishment" and try not to change the rules of the game out from under people midstream. On balance, I think the benefits of a well designed tax clawback could exceed its costs. But a poorly designed clawback would set a corrosive precedent for no other purpose than to salve and misdirect public rage.

The main benefit of a tax clawback would not be to punish bankers for the looting they have already done, but to set a precedent. Many commentators (e.g. Surowiecki) have pointed out that during the credit bubble, market discipline failed not so much because shareholders expected to be bailed out, but because the employees who run financial firms could cash out short-term gains regardless of long-term costs to shareholders and taxpayers. The precedent of a tax clawback would put future employees of systemically important financial institutions in jeopardy. They would know that if their mistakes provoke a taxpayer bailout, their personal wealth would be on the line. Eliminating their sense of inviolability, making it impossible for bankers to simply walk away from the losses they impose on investors and taxpayers, would, I think, result in structural changes to financial institutions. Risk-takers would congregate in definitely-small-enough-to-fail boutiques and hedge funds. Managers of systemically important banks would lobby for regulation to prevent competition from forcing them into risky practices that might provoke a clawback of their personal net worth when things go bad.

The dumbest possible tax clawback would be a punitive one-off designed to recoup the AIG bonuses. The brazenness of those bonuses has galvanized public anger, and served usefully as a tipping point, but in the scheme of things recovering less than half a billion dollars of a multitrillion dollar bailout will not matter very much. In order to set a useful precedent, a tax clawback needs to be broadly and rationally targeted. That is, employees of any and all institutions whose weakness necessitates a public bailout must be subject to the clawback. The Paulson Treasury, as a matter of insidious policy, made it difficult to distinguish between failing and healthy banks by forcing solvent banks to suck up TARP money along with the zombies. A good clawback proposal would encourage healthy banks to return any public assistance they've received over a period of several months, and then claw back funds only from employees of banks that are unable to return the funds without violating capital or liquidity requirements. (The law would have to address wrinkles like how to let banks "return" noncash assistance such as asset guarantees.)

A good tax clawback would not have to be very punitive. While getting back the money is an important purpose of a clawback, establishing the principle that the people who run financial institutions will be made responsible for cleaning up their own messes is far more important. Levying a 100% tax on bonuses might be satisfying, but so draconian a law would only pass if it were uselessly addressed to a single scapegoat rather than applied to financial institutions broadly. I'd recommend a 50% tax on compensation above maybe $200,000 in any year during the four years prior to the public assistance. Since this tax would represent an unexpected expense to the people it would affect, I'd allow the liability to be spread out over a period of several years. In general, the law should be structured and justified as a means of having the parties responsible for a financial disaster bear part of the cost of the cleanup, not as punishment.

One might worry that if the tax is too mild, future bankers might not be discouraged from taking foolish risks at critical institutions. If a big bet can get you a $10M bonus this year, but you'd have to return $5M if things go wrong next year, it might still be worth taking the bet. I think there's less to this than meets the eye. Once a firm precedent is established that previous years' compensation is fair game to pay for a taxpayer bail out, bankers would have to keep in mind that tax rates can always change, and that legislators might be less reticent next time around, when the use of clawbacks would not be novel and controversial. The law might even establish a higher tax rate for future failures.

In order for an ex post tax to be Constitutional, it should apply broadly and have some legitimate purpose besides just punishing someone. Kevin Drum gets a bit sardonic about this:

So it looks like the answer here is simple: even though the purpose of this tax would pretty clearly be punitive with extreme prejudice, we need to carefully pretend that it's not. And we need to make sure the legislative history shows that it's not (it should be "manifestly regulatory and fiscal" Tribe says). Then everything is kosher! We can tax their socks off!

While a lot of us might want to be "punitive with extreme prejudice", this is too cynical a view. The requirements of the Constitution seem perfectly consistent with imposing a clawback that permanently alters the incentives of the people who run systemically important banks. A good law would be both retrospective and prospective. It would help defray the costs of the current crisis while firmly establishing the principle that the individuals who run critical financial institutions can be decompensated if they let those institutions melt down on their watch. The analogy to Superfund is quite close, I think. If we do this, we oughtn't conceptualize what we're doing as finding a loophole we can use to shaft the f@kers. We should craft a good law that lets us to recoup some of the cost of cleaning up existing messes, and that defines a framework for sharing the cost any future messes with the people most responsible for them.