Center for Financial Innovation and Stability

Notes from the Vault

Larry D. Wall
March 2015

The Dodd-Frank Act shifted the base for assessing Federal Deposit Insurance Corporation (FDIC) assessments from total deposits to a new base of total assets less capital. The change was intended, by advocates such as the Independent Community Bankers of America, to provide a “fairer” allocation of deposit insurance premiums by better aligning insurance premiums with asset shares. However, it has had the effect of imposing insurance premiums on all marginal asset acquisition decisions by insured banks, given that capital is relatively fixed in the short run. This Notes from the Vault post takes a closer look at the merits of the change in the FDIC assessment base.

Insurance premiums and fund
Congress created the deposit insurance fund (DIF) through the Banking Act of 1933 and the Banking Act of 1935.¹ Given that Congress created the DIF, Congress could in theory have funded and administered it in any number of ways, including using taxpayer-financed appropriations. In practice, legislated changes affecting the DIF and the way the FDIC has managed it are consistent with the fund being used as an accounting mechanism designed to have the banking industry pay, over time, for the costs of insuring deposits and resolving failed banks.

The DIF increases to the extent its revenues exceed its expenses. The revenues consist largely of bank insurance premium assessments and interest earned on the fund's Treasury portfolio.² The expenses are mostly the FDIC's operating expenses and the increase (or decrease) in the FDIC's expected cost of resolving failed banks.³

The Dodd-Frank Act set a minimum designated reserve ratio for the DIF equal to 1.35 percent of estimated insured deposits by 2020.⁴ The FDIC Board has set a longer-run target ratio of 2.0 percent to provide greater flexibility in resolving failed banks during a crisis. If the FDIC were to need additional funding, the agency has the authority to borrow $100 billion from Treasury and $100 billion from the Federal Financing Bank.⁵ However, Diana Ellis, director of the FDIC Division of Insurance and Research, observes that the banking industry is obligated by law to repay funds borrowed from Treasury over a period of seven years. Thus, both the FDIC's target for the DIF and the borrowing authority provided by Congress are consistent with the goal of having the banking industry fund the cost of deposit insurance.

Given this overall goal, a reasonable additional strategy is to have each bank pay assessments equal to its expected cost to the deposit insurance fund. Such a strategy fairly spreads the cost of insurance across banks and also aligns each bank's incentive to increase deposits with the cost of providing additional insurance. The FDIC has long based insurance assessment on each bank's deposits, which roughly scales the base to the magnitude of potential losses. However, prior to the passage of the FDIC Improvement Act in 1991, the assessment rate was set by law to be a fixed rate independent of the riskiness of the bank. Since then, the FDIC has based insurance premiums on increasingly more sophisticated measures of the risk of losses to the DIF.

These goals provide a base for evaluating the question of whether larger banks have caused a greater proportion of the FDIC's losses and, hence, whether the shift in assessments to the larger banks advanced the goal of having banks' assessments roughly approximate their expected cost to the fund.

The table below provides a comparison of FDIC losses with the estimated pre-Dodd-Frank assessment base (deposits) and the estimated post-Dodd-Frank assessment base (assets less capital) by six size categories. The FDIC losses are taken over the period from 2000 through February 28, 2015, for all FDIC–insured institutions.⁶ The deposit and asset figures are for the fourth quarter of 2014 for all FDIC–insured institutions.

What stands out in the table is that the banks that are plausibly in the too-big-to-fail (TBTF) category (over $100 billion in assets) account for over 60 percent of the assessment base (old or new) but accounted for none of the losses to the FDIC. In other words, the large banks bore a hugely disproportionate share of the cost of realized deposit insurance losses in the 2000s under either assessment base. Each of the other size categories are subsidized by the largest banks, with each of these categories' exposing the DIF to losses of two to three times their respective shares of the assessment base.⁷

Also, in the table observe that the change in the assessment base does not make a very large difference in the largest banks' share of that base. The increase in the assessment base of the large banks only increases by 1.55 percentage points as a result of the move to assets less capital versus a base of deposits only.⁸ Thus, the change in the assessment base does not currently seem to be having a major effect on the distribution of deposit insurance costs.

The original effect of the change in the assessment base was larger, with nondeposit liability funding accounting for almost 25 percent of deposit and nondeposit funding in mid-2010 when the Dodd-Frank Act was passed. Now that this funding accounts for only 14.4 percent of total funding, the different between assets and deposit measures is much smaller. Indeed, this change is what we should expect, because if the government makes some funding sources more expensive (nondeposit liabilities, in this case), banks will obtain less funds from those sources.

Too-big-to-fail subsidies
The table highlights the fact that the change in the assessment base increased the gap between the cost the TBTF banks imposed on the deposit insurance fund during the financial crisis and their share of the insurance premiums. However, one could argue that the change in the assessment base is nevertheless justified by the substantial subsidies the TBTF banks received during the crisis in the form of underpriced loans, guarantees, and equity investment. I concur that the TBTF banks obtained substantial subsidies during the crisis, but the argument that these subsidies justify a change in the assessment base for deposit insurance depends upon who bore the risk of higher losses. The change in the assessment base primarily benefits smaller banks, hence, in order to justify the change, it should be the case that small banks were most at risk.

The banking industry, including TBTF banks, obtained support from various Federal Reserve credit and liquidity programs, Treasury support through various Bank Investment Programs and the Capital Purchase Program, and FDIC support through the Temporary Liquidity Guarantee Program. Each of these programs provided funds or guarantees on better terms than the users could have received from the market. The TBTF banks also benefited from market participants' expectations that further assistance would be provided if necessary.⁹ Moreover, these benefits were not provided for "free" even though the support programs ultimately earned more than they lost. The FDIC, Federal Reserve, and Treasury were taking the substantial risk that more banks would fail with large losses to the support programs.

Small banks would not have been at risk, however, from the potential of higher losses to the Federal Reserve's and Treasury's support programs. The support provided by the Federal Reserve and Treasury was not insured against loss by the FDIC.¹⁰ Rather, any losses taken by the Federal Reserve or Treasury would ultimately have been borne by taxpayers.

In contrast, the DIF was exposed to the potential for large losses to the Temporary Liquidity Guarantee Program from the failure of a TBTF bank. However, small banks need not have been exposed to more than their fair share of these losses from this program. In general, the FDIC is required to use a resolution method that imposes the least cost on the DIF, which implies that the uninsured deposits and nondeposit liabilities can avoid losses only if doing so reduces the FDIC's resolution costs. There is an exception to this general rule, often referred to as the systemic risk exception, which the FDIC had to invoke in order to undertake the Temporary Liquidity Guarantee Program.

However, the provisions creating the systemic risk exception also provide that the losses from invoking the exception shall be repaid from a special assessment that may be imposed on depositories (banks) or depository holding companies or both. Further, the assessment terms are to take account of the types of entities that benefited from the action taken. In other words, if the FDIC's Temporary Liquidity Guarantee Program (or any other exception to least cost resolution) results in losses and the TBTF banks and bank holding companies benefited disproportionately, the FDIC already had ample authority to impose disproportionately high insurance assessments on them.

Thus, the benefit that TBTF banks received from the support programs in the recent crisis need not have resulted in higher insurance premiums for small banks. Nevertheless, one could argue that the change in the FDIC assessment base was more about providing funds for future bank failures. The change in the assessment base would be appropriate to the extent it better aligns TBTF insurance premiums with the likely costs of resolving failed TBTF banks in the future.

TBTF banks could pose a larger share of future failures—indeed, given the statistics in the table—their share could not be smaller. However, in response to new authority granted by the Dodd-Frank Act, the FDIC has developed resolution methods that are designed to have holding company creditors bear all of the losses by the failure of a systemically important bank group. The resolution method, called Single Point of Entry, has not yet been implemented and I have raised concerns about both large bank resolutions in general and the procedure required to invoke the FDIC’s new authority. However, the clear intent of current efforts is to shift more of the risk of failure of TBTF banking organizations onto private market participants in a way that completely protects the DIF.

Thus, the change in the assessment base has been justified on the grounds that asset base is a "fairer" allocation of insurance premiums. However, a large fraction of the TBTF subsidies incurred in the recent crisis was provided by taxpayers. Further, the FDIC had the authority to shift more of the cost onto TBTF institutions had the DIF suffered losses from the Temporary Liquidity Guarantee Program. Finally, work is ongoing to reduce the risk that the failure of a systemic bank will impose large costs on the FDIC.

Conclusion
The change in the FDIC's assessment base shifted more of the cost of bank failures from smaller banks to the very largest banks. This change occurred despite the fact that the DIF incurred no losses from the failure of TBTF banks. To be sure, the TBTF banks received substantial subsidies during the financial crisis, but those subsidies largely came from the taxpayers and any losses to the FDIC's special program could have been recovered by disproportionately high assessments on the largest banks. Thus, the real beneficiaries of the change in assessment are smaller banks that now pay an even smaller share of the deposit insurance costs their failures impose on the FDIC.

Nevertheless, the change in the assessment base might not be a very important issue if all it did was to increase slightly the largest banks' share of the FDIC's insurance assessments. However, the change in the assessment base also had the unintended effect of substantially increasing the cost of using the Federal Reserve's primary tool for normalizing monetary policy, interest on excess reserves, as I previously have discussed.¹¹

In part because of these higher costs, the Federal Reserve is contemplating the use of overnight reverse repurchase agreements with nonbank financial firms, especially money funds. Among the concerns expressed by members of the Federal Open Market Committee is that "that in times of financial stress, the facility's counterparties could shift investments toward the facility and away from financial and nonfinancial corporations, possibly causing disruptions in funding that could magnify the stress." Another concern of some of the participants is that "a relatively large ON RRP facility had the potential to expand the Federal Reserve's role in financial intermediation and reshape the financial industry in ways that were difficult to anticipate."¹²

Thus, the changes to the FDIC assessment base benefit smaller banks rather than the taxpayers who are actually bearing the cost of TBTF guarantees. Moreover, an unintended consequence of the change may be to make the financial system less stable, possibly resulting in pressure for additional taxpayer support of the financial system in a future crisis.

Larry D. Wall is the executive director of the Center for Financial Innovation and Stability at the Atlanta Fed. The author thanks Paula Tkac and Warren Weber for helpful comments on the paper and Pam Frisbee for research assistance. The views expressed here are the author's and not necessarily those of the Federal Reserve Bank of Atlanta or the Federal Reserve System. If you wish to comment on this post, please e-mail atl.nftv.mailbox@atl.frb.org.

_______________________________________

¹ See the FDIC’s A Brief History of Deposit Insurance in the United States.

² The FDIC provides an Annual Report that summarizes its financial activity.

³ The FDIC’s supervisory and other activities are paid for by deposit insurance assessments. Note that the Financial Section of the 2014 Annual Report shows negative numbers for insurance losses in 2013 and 2014. These negative numbers are primarily due to a more favorable reassessment of the expected losses from failures during the crisis.

⁴ Alternatively, the FDIC can set the 2020 target for the fund “as a comparable percentage of the new assessment base, average consolidated total assets minus average tangible equity.”

⁵ This authority is discussed in the first footnote of the Deposit Insurance Fund’s 2013 Financial Statements.

⁶ This period includes the losses from the recent crisis but misses the elevated losses between 1981 and 1993. However, the structure of the banking industry was rather different over most of the 1981–93 period, as state laws banned interstate acquisition and in many cases limited intrastate branching at the start of this period.

⁷ Joseph H. Neely of Neely and Associates has a different take on the cost distribution implications of having small and very large banks in the same fund.

⁸ Of course, the large banks’ share of the actual assessments also depends upon the rate they are being charged. However, it appears likely that the rate paid by the largest banks is in the same range as that paid by smaller banks. The FDIC Quarterly for the Third Quarter of 2014 reports that banks accounting for over 60 percent of the assessment base paid a rate between 5.0 and 7.5 basis points. Further, over 90 percent of the banks paid between 2.5 and 10 basis points.

⁹ See the Government Accountability Office report Large Bank Holding Companies: Expectations of Government Support.

¹⁰ The Federal Reserve’s claims were backed by collateral provided by the bank. However, the FDIC is not legally responsible to make good if the collateral provided to the Federal Reserve is insufficient.

¹¹See also the Shadow Financial Regulatory Committee.

¹² This is an example of an “unknown, unknown” discussed by Mark Fisher that can arise from the Federal Reserve operating with a very large balance sheet.