The Proposed “Financial Crisis Responsibility Fee”

Editor’s Note: In testimony before the U.S. Senate Finance Committee, Douglas Elliott discusses the proposed “Financial Crisis Responsibility Fee,” arguing that it is a reasonable solution to recovering taxpayer losses on the TARP and ensuring that these losses fall on the financial industry, not the public.

Thank you, Chairman Baucus, Ranking Member Grassley, and members of the Committee, for your kind invitation to discuss this important, and fairly misunderstood, topic. My name is Douglas Elliott and I am a Fellow at the Brookings Institution. The views expressed here are my own and do not necessarily reflect those of the Brookings Institution, which does not normally take “house” positions on policy issues.

Since I have not previously testified in front of this committee, let me briefly explain my background. I was an investment banker for almost two decades, primarily at J.P. Morgan. As such, I provided a wide range of advice and capital raising services to a client base of financial institutions. At Brookings, I have used the expertise developed as an investment banker to focus on financial institutions, markets, and their regulation, publishing several dozen papers on the financial crisis and regulatory reform[1]. The research is actually a continuation of work that I did from 2003-6 when I founded a small think tank called the Center On Federal Financial Institutions[2], for which I worked as a volunteer until we ran out of money and I had to return to Wall Street. COFFI focused on the federal government’s own lending and insurance activities, which were already of very considerable size even prior to the financial crisis. Finally, to save everyone the usual Washington guessing game, let me explain that I am a political independent and a moderate.

I am strongly supportive of the overall idea of recovering the losses on the TARP from the financial industry so that taxpayers do not bear the ultimate cost. Repayment is self-evidently the right public policy unless there is a compelling reason not to charge a fee. It is difficult to envision a public policy rationale for why it would be more appropriate for taxpayers to bear the cost than for the industry that benefitted directly from the government actions, absent some substantial negative effect of imposing the fee. As I will explain, I firmly believe that the financial industry and the economy could absorb the relatively small fee required without significant negative effects.

I am less wedded to a particular way of imposing the fee, although I like the general approach the Administration has proposed, since it would help in a modest way to achieve other public policy objectives at the same time. It is structured to fall most heavily on the firms that represent the most risk to taxpayers going forward and that represented the greatest danger during the crisis. It is also tilted away from the core lending activities that are generally supported by deposits and towards the riskier activities that are less central to economic growth. These are worthwhile goals, although I view them as clearly secondary to the objective of repaying the taxpayers.

Opponents legitimately point out that imposing a fee on a bank, as with any other business, is likely to cause at least some of the cost to be passed through to the bank’s customers. In particular, they suggest that loans could become significantly more expensive and harder to obtain. They are right directionally, but a review of the facts shows that any effect should be quite small, given the immense size of the financial industry. Also, it is worth noting that taxpayers as a group still deserve their money back even if some taxpayers ultimately bear a portion of the cost in their capacity as bank customers, especially since those who end up bearing the greatest burden are also likely to have benefitted the most from the rescue of the financial system. For example, a business that ends up paying slightly more for a loan going forward might have been paying much more for loans now if our financial system had melted down completely.

The Administration is proposing to collect $9 billion a year for a minimum of 10 years. This compares to a core earnings power for the U.S. banking industry of approximately $200 billion a year, after-tax[3]. Non-interest expense, of which compensation is the main item, is at least $300 billion, after-tax[4]. Thus, the fee would be less than 2% of after-tax income plus non-interest expense, which appears extremely reasonable given the scale of aid we taxpayers have provided the industry.

Comparing it on another dimension, banks and thrifts reported $13 trillion of assets to the FDIC, which does not count considerable investment banking and other non-bank assets. Thus, the industry could cover the $9 billion fee by charging less than an additional 0.1% on each dollar of assets, on a pre-tax basis, assuming the fee is not tax-deductible. In practice, the industry might pass along half of this to customers, or approximately 0.05% per dollar of assets, and absorb the other half by taking a 1% hit to income plus non-interest expense. For comparison, the Fed would never bother with an interest rate move this small, because the effect on the overall economy would be minor.

I notice that industry lobbyists are magnifying the reported effect by aggregating the 10 years of fees to a total $90 billion hit to capital and indicating that this $90 billion could have supported $900 billion of lending at a 10:1 ratio. They could just as easily have indicated that the industry could support an additional $900 billion of lending by voluntarily cutting non-interest expense by 3-4% for each of the next 10 years, principally by cutting compensation. In addition, $9 billion a year for 10 years has a lower value in today’s dollars than $90 billion. Discounting at a 12.5% cost of equity would bring that figure down to under $50 billion, making the necessary adjustments even smaller.

Opponents have also argued that the TARP money infused into the banks as additional capital is likely to be repaid at a net profit to the taxpayers and therefore it is unfair for the financial industry to bear the cost of the TARP. They are factually correct about the repayments, although the figures are sometimes exaggerated by looking only at the returns to taxpayers on the banks that have already repaid their aid. Our problem was never going to be the banks that were strong enough to repay quickly; the losses will come over time from the weak ones that never make their full payments. Nonetheless, I do expect that the narrowly defined category of capital infusions from TARP into the banks will yield a profit for taxpayers.

However, I think it borders on the churlish for the banks to focus on such a narrow measure when the total aid provided by the taxpayers was of far greater value to them than the relatively small cost of the proposed fee. The banks, and to a lesser extent other financial institutions, were major contributors to the financial crisis that led to our most severe recession since the Great Depression. Despite this, they were massively aided by taxpayers, with the government taking unprecedented actions to assist. The aid that was provided was generally priced well below what the private market would have charged for taking on the same risk. This was appropriate, but it means that merely paying off the aid under the terms required does not come close to fully compensating taxpayers for the risks they successfully took to restore the economy.

Besides, we should focus more broadly than solely on the direct benefits of the government’s extraordinary actions. At the point the government acted so forcefully, trillions of dollars of value had been destroyed on securities and loans, much of it in the hands of the institutions which would be paying this fee. Failure of the government to act in the extraordinary manner that it did would have allowed a further meltdown that would have destroyed trillions of dollars more in value. The industry should be extremely grateful for this aid, instead of minimizing the nature of the help in order to avoid a relatively trivial fee. Nor do I think it is in their long-term interests to be perceived as refusing to make the taxpayers whole.

Let me be clear. I do believe the industry has good arguments on a few of their specific complaints, particularly the unfairness of asking the banks to pay for the auto rescue. Similarly, in a narrow sense, they have a reasonable argument for postponing the imposition of the fee until we know the size of the ultimate losses better. However, I think these arguments are dwarfed in importance by the over-riding requirement that taxpayers be repaid in total and that it be demonstrated quickly that this will indeed occur. The proposed fee is so far below the value the government provided to the financial industry that there should be no quibbling about exactly how this modest figure was calculated.

I should also note that some opponents of the fee make the argument that their particular financial institutions benefitted from the government’s actions only indirectly, in the same manner as anyone else living in this country. This notion is misleading. Any large holder of financial instruments is far better off today than they would have been if the government had refused to put the taxpayers’ money at risk. It is true that everyone in the economy benefited indirectly, but those with major financial investments benefitted much more significantly than those who were only affected by the general economy. It would be reasonable to have a portion of the fee spread much more widely across financial institutions, but we have to draw the line somewhere and the approach the Administration suggests is reasonable and has other advantages.

There has also been a movement to have Fannie Mae and Freddie Mac foot their share of the bill. This would be great in theory, but is impractical. Whatever the technicalities of how the government accounts for Fannie and Freddie, they are effectively government-owned. The current market value of the private shareholdings in the two firms is about $2 billion, reflecting the fact that they are likely to prove totally worthless. The fact that they have value at all is a combination of option value, since a miraculous recovery could occur, plus the possibility that they may be paid something for their nuisance value. If we reconstitute them as public companies in the future by selling them back to the market, the expected cost of any future fees would just be subtracted by potential buyers from the price they would otherwise pay us. So, including Fannie and Freddie in the fee would have virtually no net benefit to taxpayers.

I will not say a lot about the specifics of the fee, both because I feel less strongly about them than I do about the broad concept and because I suspect that there will be considerable further changes before final legislation is finished. Given that we are going to charge a fee, it is reasonable to try to assist with another important public policy goal, which is to provide disincentives for institutions to act in ways that present the greatest risks to the public. The Administration’s first step in this direction was to limit the financial institutions subject to this fee to those that have at least $50 billion in assets. This is sensible, since institutions smaller than this represent a considerably lower risk of a taxpayer bailout than do the largest institutions. The exact cut-off, however, is arbitrary and could be moved without great harm.

The second step was to suggest using “risk-weighted assets” as the base figure from which to calculate the fee. The size of assets makes sense to use as the base, because it is a good measure of the size of a financial institution and size directly relates to both capacity to pay and the total level of risk represented by an institution. Using the risk-weighted approach is consistent with the main method for calculating the risk of an institution for determining minimum capital requirements. The primary risk for a bank is that the value of its assets declines. Since certain of those assets are riskier than others, risk-weighting simply uses a higher multiplier for risky assets than for non-risky ones. Thus, a bank with risky assets will look bigger for purposes of this fee than an otherwise identical bank with lower risk assets. This does have the effect of modestly tilting the fee towards commercial lending, including to small businesses, and away from mortgage lending or holding government debt. However, the total effect on lending is likely to be small anyway, so the implications of that tilt are not serious and do have the benefit of reflecting risk levels.

Third, the portion of assets funded by equity and insured deposits would be excluded. This is intended to provide incentives for banks to support their activities with these more stable sources of funding and, to some extent, to counter any likely effects on lending. In practice, deposit money tends to be used to support loans, particularly at the smaller banks, so excluding deposits from the fee should reduce any impact on lending. It is also currently true that deposits are subject to FDIC insurance premiums and therefore this approach avoids charging twice, however the FDIC seems likely to move to charging based on assets rather than deposits, in which case this point would become moot.

Overall, the structure of the proposed fee seems broadly reasonable, but may need tweaking to avoid technical difficulties, such as have been raised in regard to repurchase agreements. More importantly, whatever the method, we should put into place a fee that recovers for the taxpayers the losses they are currently bearing from the TARP. These losses should fall on the financial industry, not the public.

Thank you for the opportunity to express my views on this important topic.

[1] My Brookings papers can be found at
[2] My COFFI papers can be found at
[3] FDIC figures show that the institutions reporting to it earned approximately $280 billion in 2009 prior to taxes and excluding loan losses. Adding in earnings for the industry from investment banking and other activities that are not conducted in entities reporting to the FDIC, subtracting a normal level of loan losses, and adjusting for taxes, the result is approximately $200 billion a year as a running rate.
[4] FDIC figures show non-interest expense of $385 billion in 2009. Adding an estimate of investment banking and non-banking expenses and subtracting taxes implies a figure north of $300 billion, although it is hard to be precise.