An illustrative example of a federal deposit insurance system without reserves
Ron J. Feldman
- Executive Vice President and Senior Policy Adviser
Published September 1, 1998 | September 1998 issue
The article, "When Should the FDIC Act Like a Private Insurance Company?" argues that the current reserve policy for federal deposit insurance is fatally flawed. The article suggests, in general terms, alternative funding principles focusing on the setting of fair premiums, increased borrowing authority for the FDIC, and a deposit insurance system that no longer maintains a fund that holds reserves. This example provides a little more detail on each of those three facets of a federal deposit insurance system without reserves. It also reviews how Congress could choose to manage a reserveless system. The primary goal of this example is to increase discussion about the issues raised by moving from the current system rather than to argue that there is a single model for the new regime.
Setting fair premiums. Under this construction, Congress would eliminate the current links between premium setting and reserves. In the place of reserves, Congress could instruct the FDIC to set actuarially fair premiums in a cost-effective manner over a demarcated time horizon. Setting a fair premium for deposit insurance poses extremely difficult technical challenges, the scope of which goes far beyond this article. Suffice it to say that a number of forecasts of bank failures by government and private analysts have been widely off the mark over the last 15 years. However, if the government hopes to manage its risk and limit subsidies to insured banks, it has little choice but to try to price insurance correctly, and it is clear the current regime cannot achieve that goal.
The cost-effective caveat means that the expense of improving the premium should not exceed the benefits that the improvement provides. It would not be cost effective, for example, to send out an army of highly skilled and well-paid premium setters to reside at each insured bank and estimate the likelihood of the bank's failure on a daily basis. As a result, the FDIC must investigate how to acquire robust yet cheap information. The assessment of a bank's riskiness by market participants represents one likely source of such data. The Minneapolis Fed's 1997 Annual Report describes a plan for creating a source of market information and incorporating it into premiums. A host of analysts have proposed alternative methods that rely on market data to address the difficulty in setting fair premiums. This data could include risk premiums paid on uninsured bank liabilities or the prices set in reinsurance markets where firms or investors take on some of the risk of insuring deposits. Others have created pricing models based primarily on nonmarket information such as regulatory bank ratings or capital levels. The FDIC's research staff estimated the fair premium using such nonmarket signals and found that recent deposit insurance assessments were not actuarially fair.1
Determining the time horizon over which premiums should pay for claims also raises a host of options. There are reasons why the FDIC should not try to have one year's premiums pay for that year's losses. One of the great difficulties in offering deposit insurance is the fact that bank failures tend to occur in bunches. Bank failures are often "lumpy" because events that cause one bank to fail, such as a crisis in agriculture or energy sector or a sudden movement in interest rates, can cause many banks to fail. As a result, the claims made against the FDIC can increase or decrease suddenly.
Indeed, this failure pattern is one reason why the federal government can offer deposit insurance and private firms may not. Larry Summers, undersecretary of Treasury, recently noted to Congress, "the Federal government is uniquely capable of spreading risk over time. A private insurer ... can only lose so much money in any given period without being declared bankrupt. By contrast, the capacity of the Federal government to borrow for the purpose of meeting short-term contingencies dwarfs that of any private sector entity."2
Setting deposit insurance premiums to pay for costs over fairly short periods of time, therefore, can lead to extreme fluctuations in assessments. The FDIC has found that such variability can cause high volatility in banks' net income.3 In contrast, by taking advantage of its risk-spreading ability and setting insurance premiums to pay for costs over a period longer than a year, the FDIC could adjust premiums to new forecasts of losses without large increases or decreases in the assessment.
Yet, estimating future losses could become more difficult over longer time horizons. Moreover, if the premiums are set to pay for long-run costs, the temptation may exist to charge rates based on long-term averages that may not reflect the risks that banks are taking in the current period. These factors must be balanced when determining the time horizon for premium setting. Whatever the time period, setting fair premiums without reserves requires the FDIC to borrow funds, as Summers indicated.
Borrowing authority. Providing the FDIC with permanent and indefinite borrowing authority ensures that the insurer has access to needed funds. It also smoothes out assessment rates. If a given year's premiums do not pay for that year's outflows, the FDIC could borrow the necessary funds to resolve the failing bank. By definition, the FDIC should set fair premiums such that banks will pay back the Treasury over some prespecified period of time.
Eliminate the fund. Congress, in this example, would eliminate the fund as an accounting vehicle in order to reduce tendencies to focus on reserves. The premiums paid by banks would flow into the general fund of the government. The FDIC, the Treasury, and the budgetary agencies of the president and Congress would keep regular tallies of premium inflow and outflows related to expenditures of the deposit insurance system. The FDIC would also continue to acquire the information necessary to produce financial statements and set premiums. Thus, policymakers and the insurer could continue to find out, if so desired, if bank premiums have paid for insurance costs over the long run.
Congressional management. A system without reserves could give the insurer significant discretion and removes a familiar tool that policymakers use to set insurance goals. To both ensure that the FDIC acts in the public interest and to give Congress a new management focus, policymakers should focus on the premium setting policy and the closure of sick banks. Congress could, for example, require the FDIC to come before the appropriate committee on a regular basis and review the premiums they plan to charge for the year and discuss the long-run health of the insurance system.
To further isolate premium setting and legislative reporting from political pressure, Congress could also establish an independent board of experts whose sole directive is to set fair premiums. Versions of such a system currently exist for government pensions. Technical experts appointed by the president for staggered 15-year terms set the assumptions for military pensions. An expert panel picked from the private sector also helps set assumptions for nonmilitary pensions. This organizational structure was designed to shield pension assumptions, which determine the amount of funding the government must pay into retirement funds among other potentially costly outcomes, from political pressure.
Analysts have also suggested a range of options for creating the proper incentives for the FDIC to shut down failing banks, including systems that require the shuttering of banks where the market value of the banks' financial cushion, or equity, falls below a given amount. Others have called for contracts that hold the FDIC's board of directors financially responsible for forbearance. Policymakers may want to look at such options anew if the FDIC is given additional discretion in running the insurance program.
Congress would also have to look at its own incentive system and, in this light, should consider adjusting its current budgetary procedures for the FDIC to move to a premium focused regime. The annual difference between the cash inflows from premiums and other sources and the cash outflows to resolve failed banks currently affects the bottom line of the budget. When inflows exceed outflows, for example, deposit insurance reduces the deficit/increases the surplus. Like reserves, these cash flows can deceive Congress as to the financial status of deposit insurance. Deposit insurance reduced the deficit in almost every year from 1977 to 1987 while expected losses for the FDIC rose. In 1988, bank deposit insurance suddenly went into the red, increasing the deficit by billions of dollars. The losses in the bank insurance fund occurred when the number of bank failures shot up and the cash flows from the fund--even those that would be repaid by the sale of assets at failed banksshot up.
The Congressional Budget Office, among others, has noted the weaknesses of the cash-based system and in a 1991 report offered a number of alternative budgetary systems.4 In one option, the premium setting authority would announce their estimates of fair premiums in synch with the budgeting process. Congress would then have three options: take no action, appropriate funds to "buy down" the premium if it wants to relieve banks of the fair cost, or enact reforms to deposit insurance that alter the future loss forecast of the premium setting board. Only in the case where Congress appropriates funds would federal deposit insurance affect the budgetary bottom line. The Treasury would track the other cash flows into and out of federal accounts but they would not affect the deficit/surplus under this plan.
This budgeting reform has several ancillary benefits in addition to improving the premium setting process. The ability of Congress to reduce the fair rate should encourage policymakers to regularly weigh the costs and benefits of offering deposit insurance. Estimates of bank failures in the future are likely to have a wide range of possible outcomes. Under this regime, the insurer would have to charge a premium that reflects the uncertainty they face in offering deposit guarantees. This may lead to premiums that Congress considers too high. Requiring an appropriation to reduce that charge would force Congress to face the costs of offering a program whose future expenses are not known with a high degree of certainty. In addition, Congress would have less incentive to require regulators to forbear. Waiting to close a bank would only raise premiums, as expected losses increase, and would not affect the budget's bottom line as it had in the past.
Additional questions. This simple sketch does not answer all the questions that it raises. It does not, for example, inform Congress what should happen to the money that has built up in the BIF after the transition to a new system without reserves. Controversy may also continue to exist in this system over the fate of funds generated when the FDIC, in retrospect, sets premiums higher than necessary to cover long-run losses. Currently, banks demand rebates when they feel they paid premiums that were too high.
On the one hand, the FDIC should reset premiums on a regular basis in the new system. Thus, the FDIC would lower its premium if experience suggests that existing estimates of long-run costs were too high. This ongoing process should keep the premium as fair as possible. Banks should have no more access to rebates for fairly set premiums than the average consumer has access to her previously paid home insurance premiums when she does not make a claim over a period of time. On the other hand, a consumer has access to a variety of home insurers if they believe their current insurer charges too high a rate. The FDIC has a monopoly, thus preventing banks from shopping around. The budgetary regime described above would provide a compromise of sorts by allowing Congress to provide rebates but requiring that such outflows result from appropriations that increase the budget deficit/reduce the surplus. Setting premiums according to new types of market measures, including the prices that firms or investors charge to bear some of the risk of deposit insurance, would indirectly provide a check on the FDIC's premium setting ability. Clearly, a thorough vetting of rebate policy would be part of any effort to reform the current reserve policy.
3 Conference material from Confidence for the Future: An FDIC Symposium, January 29, 1998, Panel 3: Striking a Balance within the Current Framework p. 6, and Federal Deposit Insurance Corporation. 1995. Assessments; Final Rule, Federal Register, August 15, 1995, p. 42724-42737. [1053k PDF]