Skip Navigation

Mr. Chairman, Mr. Cardin, members of the Task Force, I am pleased to have this opportunity to discuss budgeting for government insurance programs. My statement makes three points:

• The current cash basis of accounting used in the federal budget provides policymakers with neither the information nor the incentives to control the cost of government insurance programs.
  
  
• Noncash accounting alternatives can contribute to improved design and operation of these programs and to a reduction in costs to taxpayers.
  
  
• Those alternatives have disadvantages and weaknesses of their own, however. Whether the gains from noncash accounting for insurance programs would outweigh the drawbacks is a matter of judgment.

SOME DISTINCTIVE FEATURES OF INSURANCE

The treatment of an activity in the budget depends in part on the substance of the transaction. In assessing current budgeting for insurance, therefore, it is useful to recall the essential character of those programs. Insurance makes people better off by giving them a chance to pay a small fixed premium to avoid the risk of a catastrophic loss. By pooling premiums and risks, insurance provides compensating payments to those who suffer a loss. Private insurance provides protection against many hazards, especially those for which a large population is exposed to risk and losses are highly predictable. In cases in which the exposed group is small or substantial uncertainty exists about the probability and magnitude of actual losses, the government may be able to expand the range of insurance beyond that offered by private insurers. The government can step in and fill gaps in available coverage because of its power to tax. That power gives government the unique ability to pay off insured losses that are far larger than anticipated when the insurance was issued.

The financial viability of a private insurance pool depends on the balance between premium income and losses. If losses are consistently greater than premium income, a private insurer may default on the contract, leaving the insured worse off than if no insurance had been available. Financial balance is also important to most government insurance programs--and all of the largest ones--because they are intended to protect the insured from losses without simply shifting those losses to others.

To be meaningful, the measure of financial balance for an insurance program must cover the entire period the insurance is in force. For short-term insurance contracts, such as fire or flood insurance that usually covers losses for one year, the appropriate comparison is between the premium fixed when the insurance is issued and losses in that year. For long- or indefinite-term insurance, such as whole life or deposit insurance, expected premiums over the life of the insurance must be compared with the corresponding long-term losses. Further, to allow for the time value of money, both expected premiums and expected losses should be expressed in present values.

Insured losses can be reduced through contract terms that give the insured incentives to avoid and limit losses. Reductions in fire insurance premiums for structures with smoke detectors, nonflammable building materials, and electric circuit breakers reduce losses by giving financial incentives to the insured to minimize risk. Thus, the terms of the insurance contract are the key to controlling insurance losses. Once the premiums, deductibles, exclusions, and conditions for the payment of benefits are set, however, the ability of the insurer to control losses is severely limited. To be effective, actions to control costs must be taken at the contract design stage, and well before the insured event has occurred. In fact, efforts to control insurance payouts after the loss are likely to be counterproductive. For example, delayed compensation for losses covered by private fire insurance can raise payouts for the cost of temporary housing while the insured structure is being replaced. Such actions can also reduce the value of insurance and diminish the size of the insurance pool.

Lower losses mean reduced social costs. For the insured, lower losses permit lower premiums and a larger number of participants in the insurance pool. For government insurance, lower losses can lower the risk that costs will be shifted to taxpayers.
 

SHORTCOMINGS OF CURRENT BUDGETING FOR INSURANCE

The cash basis of accounting that is used in the budget for most transactions records payments as they are made or received by the federal government. For insurance programs, the budget tallies premiums when they are collected and insured claims when they are paid. For long-lived federal insurance, including the largest programs that insure deposits and pensions, cash accounting records premiums as received but may not recognize the corresponding outlays for years.

That accounting is well suited to measuring the government's cash balance and borrowing needs and is comparable to a checkbook register of deposits and checks written. Current-period cash flows, however, provide little information that is useful in monitoring the financial balance of long-term government insurance programs. Indeed, cash-basis accounting provides no budgetary links between program changes enacted today and losses outside the budget window. In fact, the cash-basis accounting system encourages policymakers to achieve their budgetary targets by focusing on current-period net outlays for insurance programs. The incentive is to bring revenues up to the budget period and push outlays to the future, with little consideration of how the one affects the other.

Two types of modifications to insurance programs reduce the short-term deficit but raise long-term costs. One is to delay payments for losses that have already been incurred. Experience with the cash-conservation strategy for dealing with insolvent, insured thrift by the Federal Savings and Loan Insurance Corporation (FSLIC) in the 1980s and early 1990s showed that approach increased losses by more than $50 billion.⁽¹⁾ Prompt resolution of insured, failed financial institutions is a lower-cost strategy, even though it requires a higher cash-basis deficit in the current period. Another approach to deficit reduction is to raise insurance premiums. In order to increase premium income while retaining existing participants, it may be necessary to expand promised benefits or otherwise enhance the other terms of the insurance. If so, that strategy will produce a short-term cash inflow and deficit reduction but a long-term increase in costs and in the deficit.

As an example, consider the case of the Pension Benefit Guaranty Corporation (PBGC), the federal program that insures private defined benefit pensions. Every year since it came on budget in 1981, PBGC has collected more in premiums and other income than it has paid in pension benefits and administrative expenses. In 1997, the excess of net cash inflows totaled $1.2 billion. Consequently, the cash-basis federal budget deficit was $1.2 billion lower in 1997 as a result of that outstanding commitment by the government to insure over $1 trillion in earned private pension benefits.

The PBGC may in fact be a moneymaker for the U.S. government. But PBGC's annual cash flows are not reliable evidence for that supposition. A more relevant indicator would be the present value of expected future outlays and premium receipts. Unfortunately, no systematic estimate of that measure exists. Yet, in its 1997 annual report, PBGC disclosed assets in excess of liabilities of nearly $4 billion. PBGC also reported, however, that "reasonably possible" future losses, which are not yet classified as liabilities, are in the range of $21 billion to $23 billion. Thus, the cash-basis indication of a gain to the government from its pension insurance program may be seriously misleading.

Legislated changes in federal pension insurance indicate that while cash-basis budgeting influences policy, the Congress has not ignored the long term. In 1986, 1987, and 1994, policymakers improved the financial condition of PBGC by restricting the ability of a sponsor of an underfunded pension system to terminate the plan, requiring faster funding of plan benefits, and authorizing PBGC to change premiums that increase with underfunding. Yet each legislated change also raised premiums from $2.60 to $8.50 per participant in 1986, to $16 in 1987, and to $19 in 1994. While higher premiums may improve financial balance in the program and raise income in the short term, they can weaken the financial condition of the insurance fund, if they are not closely related to plan risk. Specifically, if premiums are too high for low-risk plans and too low for high-risk plans, the high-risk plans will stay in the insurance pool, but the low-risk plans will have an incentive to switch to uninsured, defined contribution plans. With the available information, it is impossible to know if those legislated changes improved PBGC's financial condition.
 

NONCASH-BASIS ACCOUNTING FOR INSURANCE PROGRAMS

The most straightforward fix for the shortcomings of the current treatment is to drop net cash inflows as the program's contribution to budget outlays and replace that cash measure with the annual change in the insurance fund's long-term position. The new measure would be based on the present value of all expected future outlays for insurance in excess of expected future premium receipts. The annual budget cost of (or gain from) the insurance program would be the increase (or decrease) in the program's financial shortfall. Use of this new measure would shift the focus of efforts to control costs from current-period premiums and outlays to the net position of the insurance fund.

The budgetary mechanics of using noncash measures have already been developed for credit reform. The essential element of the new accounting is that increases (or decreases) in the insurance fund deficit must be covered by a payment from (or to) the on-budget program account to (or from) the off-budget financing account. The financing account would also collect all premium and investment income and pay all pension claims and administrative expenses. PBGC, as an example, already has an off-budget "trust fund" to account for assets and liabilities of terminated pension funds taken over by the federal government as well as an on-budget revolving fund. The revolving fund is a natural candidate for the role of on-budget program cost account and the off-budget trust fund could serve conveniently as the financing account for PBGC.⁽²⁾
 

DISADVANTAGES OF NONCASH ACCOUNTING FOR INSURANCE

Three disadvantages stand against the advantages of noncash accounting for government insurance programs: intrinsic difficulties in forecasting future insured losses, increased volatility in the federal budget estimates, and greater demands on the budget process. The most significant is probably the first. Analysis, however informed and technically capable, cannot foresee with accuracy the timing and magnitude of insured claims for deposits in failed financial institutions or for guaranteed pensions in terminated pension plans sponsored by bankrupt employers. Indeed, one of the reasons the government offers such insurance is that private firms may be unwilling to do so in the face of pervasive uncertainty about the probability and severity of future losses.

The challenge of projecting insurance costs is greater than doing so for direct loans and guarantees under credit reform.⁽³⁾ Direct federal loans and federal guarantees of private loans have the simplifying advantage of a fixed term to maturity and a limited dollar amount. When the government originates or guarantees a $1 million loan that must be repaid in five years, the government's liability is limited both in total exposure and time. For deposit or pension insurance, neither limit is in force. When the government insures a depository institution or a pension fund, it is taking on an open-ended commitment to indemnify all present and future insured depositors and pensioners against losses (up to a specified maximum per account) for the life of the institution.

Limited foresight means that the best estimates of the insurance funds' deficits will be highly uncertain. The extent of that uncertainty is not likely to be conveyed by a single number appearing in the budget. Indeed, the solidity of any such number will be illusory. An estimated change in the PBGC deficit of $10 billion, for example, may be no more than the midpoint of two equally probable estimates with different signs, minus $5 billion and plus $25 billion. Yet policymakers may be persuaded by the $10 billion estimate to take action that moves the insurance program farther from financial balance.

Moreover, uncertainty leaves room for unjustified optimism about the beneficial future effects of current policy adjustments. Faced with an urgent need to pay for current spending, policymakers may not be able to resist the temptation to claim large, if speculative, future savings from policy proposals. Noncash accounting would increase the opportunities for painless fiscal solutions.

Estimates of the long-term financial condition of government insurance programs are not only uncertain, they are also volatile. Year-to-year changes in interest rates, the level and composition of economic activity, and regional swings in real estate or commodity prices can all have major effects on the estimated future cost of deposit and pension insurance. Annual reestimates on the order of tens of billions of dollars should be expected.

The volatility of estimates, however, need not be an insurmountable obstacle to noncash budgeting for insurance. Insurance cost estimates vary over time because of uncontrollable changes in the economy and natural phenomena and because of controllable changes in policy. By distinguishing controllable and noncontrollable causes, the effects of each may be treated differently in the budget. For example, during the late 1980s and early 1990s, when outlays for deposit insurance threatened to distort the size of the budget deficit, CBO reported the deficit with and without outlays for deposit insurance. That display makes sense because the economic effects of deposit insurance occur when losses are being incurred, which is much earlier than outlays for resolving failed institutions.

The Budget Enforcement Act of 1990 also distinguished the budgetary effects of policy changes from those caused by forces outside the control of the government in its pay-as-you-go provisions governing entitlement programs. In that case, the Congress put in place a procedure to force offsetting policy actions for expansions initiated by legislation but exempted the effects of uncontrolled events from such discipline. Following these precedents, the budget might distinguish year-to-year changes in the deficit that result from policy and those from nonpolicy causes.

Finally, changing the budgetary treatment of government insurance programs will create new demands on the budget process. Both the analytical requirements and the complexity of the proceedings would increase if cash-basis accounting is replaced with changes in the insurance fund deficit. Although it is difficult to know when more is too much, there are limits to the capabilities of the budget process that cannot be ignored.
 

CONCLUSION

Like virtually all other budget issues, this one calls for judgment. On the one hand, the current budgetary treatment of government insurance programs is far from ideal. On the other, the alternative that would most directly address the shortcomings of the current system is not without significant downside risks. Once again, policymakers are confronted with a choice between imperfect alternatives.

1. General Accounting Office, Budgeting for Federal Insurance Programs, GAO/AIMD-97-16 (September 1997); and Ron Feldman, "How Weak Recognition and Measurement in the Federal Budget Encouraged Costly Policy: The Case of 'Supervisory Goodwill'," Public Budgeting and Finance, vol. 16, no. 4 (Winter 1996), pp. 31-44.

2. For details, see Congressional Budget Office, Controlling Losses of the Pension Benefit Guaranty Corporation (January 1993), pp. 21, 37.

3. Congressional Budget Office, Budgetary Treatment of Deposit Insurance: A Framework for Reform (May 1991).