Why the Time is Ripe for Fixing FDICIA

Our rationale for questioning the exception under FDICIA is clear. One hundred percent protection has proven a high-cost policy. The lack of market discipline under such a regime suggests that the current TBTF exception could encourage excessive risk taking in the future.

The need to fix this aspect of FDICIA now, however, may appear counterintuitive. The banking industry is enjoying record profits, banks have been able to broaden the activities from which they can generate revenue and the largest U.S. banks have grown to the size that some analysts believe is required to compete internationally. Yet, it is these very trends that make enactment of our reform timely.

More TBTF Banks
More large banks controlling a greater share of uninsured deposits have resulted from recent bank consolidation. Hence, regulators may view more banks as requiring TBTF treatment under FDICIA. Consequently, more uninsured depositors will assume that their bank is TBTF and their deposits will be protected.

No written list of the TBTF institutions exists, but previous empirical work used the 11 largest institutions highlighted by the Comptroller of the Currency in his 1984 testimony. Those banks controlled 23 percent of all assets at the end of 1983 and had an average size, in 1997 dollars, of about $94 billion. The 11 largest banks as of the end of 1997 held 35 percent of all assets and had an average size of nearly $157 billion. The growing number of large banks also means that the TBTF list may have expanded past the 11-bank list. The smallest of the 11 TBTF banks in 1983 had assets of just over $38 billion in 1997 dollars. Using $38 billion as a cutoff suggests that regulators may view depositors at the nation's 21 largest banks as benefiting from implied deposit insurance (see Table below).

The top 21 banks that regulators may view as TBTF have increased their share of uninsured deposits from 30 percent in 1991 to 38 percent in 1997. (The top 50 banks increased their share from 44 percent in 1991 to 55 percent in 1997.) At the same time, total uninsured deposits have increased. Consequently, the taxpayers' exposure at TBTF banks has increased significantly since FDICIA.*

More Bank Powers
Banks are not only growing in size, they are growing in powers, that is, in their ability to offer nonbank financial services such as the sale and underwriting of insurance and securities. Both the Office of the Comptroller of the Currency and the Board of Governors of the Federal Reserve System have expanded the types and/or the amount of nonbank financial activities in which banking organizations can participate. Both regulators also support legislation that expands their nonbank financial and commercial activities (they disagree as to the organizational structure in which such powers will be exercised and the specific type and amount of new powers made available). Expanding powers would increase the role of banking organizations in the financial system, making it more likely that regulators will consider the failure of a large institution as destabilizing. For these reasons, we view banking deregulation before deposit insurance reform as "putting the cart before the horse," a view long held by researchers at the Federal Reserve Bank of Minneapolis.⁷

More Bank Profits
The most extensive reforms to the bank regulatory system have occurred after or during banking crises. Although some of these reforms were well-conceived, good times present a superior time for action for at least two reasons. First, we should put depositors and other creditors at risk now precisely to avoid a future debacle. Second, a reform may put institutions in a weaker financial position by increasing their cost of funds. We should, therefore, enact reforms when banks have more resources to absorb the higher costs.

TBTF Banks Have Increased in Size and May Have Increased in Number

In 1984 congressional testimony, the Comptroller of the Currency suggested that the 11 largest banks were TBTF. As of year-end 1997, the banks in the right column exceeded the minimum size required—$38.2 billion—to make the Comptroller's TBTF list.

Assets as of 12/83 (billions of $)				Assets as of 12/97 (billions of $)
1983 Ranking	1983 Dollars	1997 Dollars		1997 Ranking	1997 Dollars
1 Citibank	$113.4	$182.1		1 Chase Manhattan	$297.0
2 Bank of America	109.6	178.0		2 Citibank	262.5
3 Chase Manhattan	79.5	127.6		3 Bank of America	236.9
4 Manufacturers Hanover	58.0	93.1		4 NationsBank	200.6
5 Morgan Guaranty	56.2	90.2		5 Morgan Guaranty	196.7
6 Chemical	49.3	79.1		6 First Union	124.9
7 Continental	40.6	65.2		7 Bankers Trust	110.0
8 Bankers Trust	40.1	64.4		8 Wells Fargo	89.1
9 Security Pacific	36.0	57.9		9 PNC	69.7
10 First Chicago	35.5	57.0		10 KeyBank	69.7
11 Wells Fargo	23.8	38.2		11 U.S. Bank	67.5
				12 BankBoston	64.9
				13 Fleet	63.8
				14 NationsBank (TX)	60.0
				15 First Chicago	58.4
				16 Bank of New York	56.1
				17 Wachovia	51.6
				18 Republic	50.2
				19 CoreStates	45.4
				20 Barnett	42.8
				21 Mellon	38.8

Source: Consolidated Reports of Condition and Income

Putting Uninsured Depositors and
Other Creditors of Large Banks at Risk

We propose that Congress amend FDICIA so that uninsured depositors cannot be protected fully when a troubled bank is rescued under the TBTF policy. Congress can implement this reform in several ways described below, but the basic idea is to require some form of coinsurance—the practice of leaving some risk of loss with the protected party. Regardless of the formulation, we think all uninsured depositors should be subject to the same coinsurance plan, including banks with uninsured deposits at their correspondent. In addition, we propose treating unsecured creditors holding bank liabilities that have depositlike features the same as uninsured depositors. Also, we propose that the coinsurance plan under a TBTF bailout be phased in over several years to avoid any abrupt change in policy.

Our other key recommendation is for regulators to incorporate a market assessment of risk in the pricing of deposit insurance. In particular, after implementation of this proposal, we recommend that regulators include the risk premium depositors and other creditors receive on uninsured funds in the assessments on TBTF banks.

Lastly, to help uninsured depositors and other creditors monitor and assess bank risk, we propose that regulators disclose additional data on banks' financial condition. We think that banks will have an incentive under our proposal to disclose additional information that the market requires. Regulators should consider mandated, cost-effective disclosures only if voluntary release does not occur.

Coinsurance under TBTF
To address the potential problems created by 100 percent coverage under the TBTF exception, we turn to the conventional technique that private insurance companies have used to mitigate moral hazard. Analysis identified coinsurance about 30 years ago as a primary means for insurance underwriters to combat moral hazard.⁸ In that vein, we propose that uninsured depositors and other creditors face a small but meaningful loss when regulators exercise the TBTF policy.

Congress could formulate a coinsurance policy in several ways. A simple approach would set a "maximum loss rate" that uninsured depositors could face, say at 20 percent, and phase it in over time. Congress could set an initial, maximum 5 percent loss and have it rise 5 percentage points a year for the following three years. Under the maximum loss rate approach, uninsured depositors would receive no protection if they would suffer losses of less than 20 percent of their uninsured deposits, for example, after they receive proceeds from the liquidation of failed bank assets. This approach may be objectionable if the rate is set so high as to effectively eliminate TBTF coverage. To address this, the maximum loss rate could be lowered or Congress could apply the coinsurance rate to the loss that uninsured depositors would have faced if the FDIC did not protect uninsured depositors. This "loss reduction rate" would ensure that uninsured depositors always receive some coverage above what they would have received from the liquidation of the failed bank's assets.** In 1990, the American Banker's Association proposed a coinsurance formulation somewhat similar to the loss reduction approach for eliminating 100 percent coverage at TBTF banks.

Congress must address a critical trade-off when choosing a coinsurance policy. The potential loss should be set high enough so that uninsured depositors will be induced to monitor the financial condition of their bank, but not so high as to eliminate the stabilizing benefits of government protection. Certainly, the details of a co-insurance policy under this maxim will always be somewhat arbitrary. However, policy-makers do not escape the difficult decision of determining how much market discipline to impose on depositors by maintaining the status quo.

The TBTF exemption in FDICIA should also not allow full coverage of all unsecured creditors of failed banks. Indeed, Congress recently indicated a preference for imposing post-resolution losses on such creditors by passing a national depositor preference statute in 1993 (which put unsecured creditors further back in the queue for receiving revenue realized from the resolution of a failed institution). In scaling back potential coverage for unsecured creditors, we think Congress must address the same critical trade-off it faces when putting uninsured depositors at risk. Thus, we recommend that unsecured creditors holding depositlike bank liabilities should face the same coinsurance policy as uninsured depositors. All other unsecured creditors should receive no special coverage. This formulation would protect, for example, holders of very short maturity bank liabilities like the overnight loans banks make to each other, certain commercial paper and foreign deposits but not holders of longer-term liabilities or claims arising from the provision of services or court judgments.

Incorporating Risk into Insurance Assessments
Prior to 1993, the FDIC charged all banks the same assessment for deposit insurance regardless of their risk of making a claim against the insurance fund. Banks did not face the cost of excessive risk taking because the insurer—like the depositor with 100 percent insurance coverage—did not charge risk penalties when the probability of bank failure increased.

FDICIA required the FDIC to create a system of risk-based insurance assessments because a flat-rate assessment encourages excessive risk taking. The initial risk-based assessments turned out to be too high for low-risk banks and too low for the highest-risk banks.⁹ Rather than correcting this failure, the FDIC was, in essence, forced back to flat-rate assessments (due to restrictions on the insurance reserves it can hold). Since 1996 over 90 percent of banks have paid the statutory minimum assessment, which has fallen from $2,000 per year to zero.

The problem here is that, even if uninsured depositors and other creditors have the potential for bearing losses, TBTF banks may still take on too much risk if insurance assessments are not risk-based. This occurs because taxpayers would continue to bear some of the banks' risk. Thus, we recommend that any FDICIA reform to eliminate the potential for 100 percent coverage at large banks be accompanied by efforts to make their assessments more sensitive to risk.

To this end, we recommend market-based, risk-adjusted insurance assessments for TBTF banks. We suggest that the FDIC include the risk premium implicit in rates paid on uninsured deposits (as well as other market measures of risk) into insurance assessments for the largest banks. The FDIC, for example, could charge the largest banks with a risk premium above policy-makers' comfort level, say that on the A-rated corporate bond, much higher assessments in order to halt exploitation of insured status. Alternatively, the FDIC could incorporate the risk premium more gradually so insurance assessments rise or fall incrementally with changes in the risk premium.

Proposals to incorporate risk premiums on uninsured deposits into FDIC assessments date at least as far back as 1972. Critics of this reform usually argue that depositors and other market participants will not price risk correctly, either because they believe they will receive a bailout or because they do not have adequate information. As a result, assessments based on market risk premiums will not accurately price risk. Our proposal makes clear that uninsured depositors cannot receive full protection. Recent empirical reviews find that suppliers of funds to banks charge risk premiums if they will, in fact, face losses when their bank defaults.¹⁰ As for the lack of information about banks' behavior toward risk, we add one additional reform to our proposal.

More Disclosure
The more information depositors and other creditors have on the financial and operational condition of their bank, the more the risk premium on uninsured deposits and other bank funding will reflect true underlying risk. Thus, we recommend ensuring that market participants have data that make the financial condition of banks less opaque.

Regulators, for example, receive nonpublic information on loans with late repayments that could be made public along with other information or assessments that regulators gather or produce. Some of this data may contain new information for market participants, and disclosure of such information may improve the ability of funders to evaluate their banks.¹¹

Banks will have an incentive under our proposal to reveal more information about their business. The more germane data banks provide, the less opaque they are and the lower the interest rate they will have to pay uninsured depositors. Regulators should consider requiring cost-effective disclosure only if banks do not voluntarily disclose the information that markets need to properly assess bank financial condition.

Of course every uninsured depositor does not have the same skills or desire to analyze disclosures provided by banks and regulators. But existing firms, and firms that will develop in the future, will surely provide analysis of banks' condition for uninsured depositors just as specialized firms provide evaluations of mutual funds and the claims-paying ability of insurance firms. The lack of interest in bank conditions currently manifested by uninsured depositors and the absence of firms providing evaluations of banks for these depositors reflect current incentives, not innate features of the human or business condition.

Addressing the Fundamental Trade-off

We began this essay noting that deposit insurance is a two-edged sword. On the one hand, it protects the small saver and limits the probability of banking panics. On the other hand, too much deposit insurance leads banks to take on more risk than they would otherwise. All deposit insurance systems must face this fundamental trade-off.

Our plan clearly adds market discipline. It also addresses the concern of increased instability in four ways.

First, we put only a small percentage of uninsured depositors and other creditors' funds at risk. Limiting the amount of the potential loss should help contain spillover effects from any one large bank failure to other banks and the rest of the economy. A bank that has uninsured deposits at a large failing bank, for example, is not likely to suffer losses so great as to bankrupt it. Thus, depositors do not have cause to run all banks with deposits at large failing institutions for fear of interbank exposure.¹² Moreover, the small potential loss makes it less likely that the benefit of pulling deposits out of a bank will outweigh the costs of making alternative investments.

Second, our reform should increase the information that markets have on banks' financial condition. Depositors are less likely to run sound banks if they are more transparent.

Third, by increasing market discipline, we make it more likely that banks will not take excessive risk.

Finally, we call for a gradual increase in the coinsurance rate. Policy-makers and analysts will have time to review the response of depositors and other creditors and adjust the rate. Likewise, incorporating a market-based risk premium into the insurance assessment need not occur at once.

Can we guarantee that our plan achieves the right balance? No reform plan can make such a claim. We may have too much or too little of uninsured depositors' funds at risk. Under our plan, uninsured depositors at risk of loss might pull their funds from all large banks regardless of their solvency. But one cannot reasonably judge the plan's potential costs in the abstract. Rather, interested parties must compare this proposal to the current system and alternative proposals.

Alternative Proposals
Some analysts have argued that FDICIA's regulatory reforms—in conjunction with existing market discipline provided by stockholders, bondholders and bank managers fearing for their jobs—adequately address the potential problems that 100 percent coverage at TBTF banks can create. Yet, as we noted, stockholders are the beneficiaries of excessive risk taking and are unlikely to act as a bulwark against it. Moreover, experience over the last 15 years suggests to us that these other sources of market discipline are not sufficient to control the deleterious effects of moral hazard. And while the FDICIA regulatory reforms were a step in the right direction, reviews of these changes do not suggest that they adequately curtail risk taking.¹³ Moreover, we are skeptical that additional regulations can effectively substitute for putting uninsured depositors and other creditors at risk. Certainly, promulgating Draconian regulations could prevent the problems arising from 100 percent coverage, but such a strategy would impair the ability of the banking system to efficiently allocate financial resources.

Eliminating federal deposit insurance altogether offers another alternative for addressing the moral hazard of 100 percent coverage at the largest banks. We do not find this reform credible. Banking panics were not eliminated in the United States under private deposit insurance systems. Consequently, it is likely that banking regulators, in order to address the threat of such panics, would still want to fully protect depositors at large failing banks after privatization. Thus, the public would still assume that their deposits are fully protected at banks considered TBTF and moral hazard would still remain a problem.

In contrast to privatization options, we view reforms requiring TBTF banks to hold subordinated debt as differing from our own in implementation detail rather than intent or substance.¹⁴ Both plans seek to increase market discipline by putting those who supply funds to TBTF banks at risk, and both allow the FDIC to incorporate additional market data into their assessment-setting process. And, like our plan, credible subordinated debt reforms must address the trade-off between depositor safety and bank risk.

Further Reform
Unfortunately, our proposed reform will almost certainly not be the final action that Congress must take to eliminate 100 percent coverage. For example, uninsured depositors may divide their funds into multiple insured accounts in order to avoid the potential for loss. In response, Congress may have to curb the ability of depositors to receive more than $100,000 in insurance coverage through multiple accounts, or apply a co-insurance policy to all deposits.

A Unique Opportunity

Economic theory and experience both suggest that providing 100 percent government protection for bank depositors and other creditors can be extremely costly. One hundred percent coverage quashes market discipline, encourages banks to take on too much risk and can massively misallocate society's limited resources. It can lead to the very type of bank failures and costs to society that deposit insurance aims to prevent. These costs are so high that Congress should prohibit bank regulators from providing 100 percent coverage as is currently allowed at the nation's largest banks. We think that a credible reform is to amend FDICIA so that uninsured depositors and other creditors cannot be fully protected under the TBTF exception but do not suffer losses so large as to eliminate the stabilizing benefits of government protection. In another step necessary to address moral hazard, we recommend that the FDIC use market signals, from at-risk depositors and other available sources, to make insurance assessments of TBTF banks risk-related.

Our reform is unlikely to be the last that Congress will need to pass. However, the potential for future reform should not dissuade policy-makers from using the current window of opportunity to avoid the mistakes of the past. The FDIC sought to increase depositor discipline in the early 1980s through a resolution method, called modified payoff, in which uninsured depositors would receive no extraordinary coverage, but rather a proportion of their money based on the liquidation value of the bank's assets. The FDIC had "experimented" with the method and "hoped to expand the modified payoff to all banks regardless of size."¹⁵ But the failure of Continental soon after the pilot program began led to its demise. Now, Congress can establish a commitment in law to impose losses on uninsured depositors and other creditors at banks deemed TBTF. Without such a credible commitment, Congress may find itself in the uncomfortable position of conducting the post-mortem of another financial debacle.

Fixing FDICIA The Minneapolis Fed Proposal To guarantee that uninsured depositors cannot be protected fully under the too-big-to-fail (TBTF) policy, and to thereby minimize the impact of the moral hazard problem, Congress should amend the Federal Deposit Insurance Corp. Improvement Act of 1991 (FDICIA) as follows: Prohibit the full protection of uninsured depositors and other creditors when TBTF is invoked. Incorporate the risk premium that depositors and other creditors receive on uninsured funds into the assessments on TBTF banks. Require the disclosure of additional data on banks' financial condition.

Notes

* In addition to potentially more TBTF banks, recent research suggests that large banks have not reduced their exposure to a significant risk since FDICIA. Specifically, the largest 10 percent of publicly traded bank holding companies had significant interest-rate risk both before and after FDICIA. If FDICIA had reduced the interest-rate risk that large banks are taking, the relationship between interest-rate movements and bank stock prices should not be as strong after FDICIA as before it. Instead, the research finds no decline in the level of interest-rate risk at large banks since FDICIA. See David E. Runkle, "Did FDICIA Reduce Bank Interest-Rate Risk?" Working Paper, Federal Reserve Bank of Minneapolis, December 1997. Return

** An example where an uninsured depositor suffers a $10,000 loss clarifies the difference in the two approaches. A 20 percent rate under the maximum loss formulation would provide no insurance if the $10,000 loss is less than 20 percent of the depositor's total uninsured deposits and complete coverage for any losses above 20 percent. In contrast, a 20 percent coinsurance rate under the loss reduction formulation would reduce a $10,000 post-liquidation loss to $2,000.

Notes