Fixing bankers' compensation is all the rage these days. The latest salvo in the ongoing tussle is the December joint proposal by France and Britain to impose a windfall tax on the bonuses of bankers, along with the crisis "responsibility fee" suggested by the United States. These proposals have raised eyebrows across the financial and political spectrum, mainly because they are often viewed as a populist measure aimed at placating taxpayers upset over governmental support for a financial system many regard as the primary villians of the global economic crisis. At the very least, it is argued, such a tax is justified since the unprecedented profits of banks would not have been possible in the absence of government largesse (as Mohamed El-Erian has argued).
Beyond such natural justice arguments, there are, at least, three economic arguments behind imposing a contingent windfall tax on firms in the financial sector.
First, such a tax would serve to (partially) offset the part of fiscal deficits that were incurred for the purposes of government-financed bailout packages. This goes beyond the government getting its fair share of the upside---presumably, such an upside should have been priced into the interest on bailout loans, as well as any options or warrants received as part of an original rescue package. This is effectively a premium payment for future government assistance, which may be much more difficult to collect (both economically and politically) once supernormal profits erode and the memories of the crisis fade. This ex ante insurance argument has been made by, among others, Doug Diamond and Anil Kashyap.
Second, the tax would act as a mechanism that could dampen the moral hazard that is a direct result of these bailouts. If firms expect that their future profits from risk-taking activity will be accompanied by a commensurate tax should such risky activity lead to the need for a bailout, then they are more likely to factor such a cost into their investment decisions, which in turn reduces the chances of moral hazard. As David Stockman has pointed out, this also has the nice side effect of limiting the size of a bloated financial sector that has increasingly been directed toward non-social welfare enhancing activity.
Third, to the extent that tax revenues are ultimately rebated to taxpayers (either directly or indirectly by paying down on the fiscal deficit), the tax would serve to reduce the deadweight losses that have arisen in the financial sector as a consequence of (one would hope temporarily) greater market concentration.
Of course, financial sector firms will resist such a move. The most likely concern, from the perspective of a given financial center, is that such a move would lead to an exodus of staff from, say, London to New York. Even if governments were to fail to coordinate on the imposition of a windfall tax, such a threat is, ultimately, not credible. The tax will be levied on profits from the previous financial year, and would need to be paid regardless of whether the firm relocates in the following year or not. But since it is also meant to be an extraordinary tax, the cost will have been sunk, and so any relocation decision would be based on weighing the future likelihood of another extraordinary tax versus the costs of relocating today; this calculus is far different from a choice between paying the tax versus moving.
In practical terms, the windfall tax should be levied on banks, and not on individuals directly. While this does open up the possibility of accounting mischief to sidestep the tax, it allows financial firms the maximum flexibility to distribute the costs of the tax in a manner most consistent with internal firm goals. Besides, with the spotlight shining so brightly on banker bonuses, it seems unlikely that the financial glitterati would be able to nonetheless richly reward themselves at the expense of their shareholders. The tax should also be contingent on profitability; the objective of such a tax is, after all, to reduce the monopoly rents accruing to the net beneficiaries of an implicitly government-supported financial sector, not to punish firms that are continuing to struggle simply because they happen to be in the financial sector. Finally, if the tax revenue is in fact not directly rebated (perhaps as part of a stimulus package), the proceeds should be placed in a special fund designed for funding future bailouts, as and when needed.
Update: Brian Bell and John van Reenen break down the nature of banker compensation in much greater detail, and in particular make an impassioned case in support of clawback agreements.