Skip Navigation

Mr. Chairman:

I appreciate the opportunity to appear before your Subcommittee today to testify about military retirement. For more than a decade, the military retirement system has been the subject of close scrutiny. Over this period, no fewer than five major published studies--including several performed under the auspices of the Department of Defense--have reviewed the system, identified problems, and recommended changes. Yet only a few changes--none of which could be called a major restructuring--have been enacted by the Congress, which suggests some support for the current system. Today I will try to review both the strengths and weaknesses of the current military retirement system, and then turn to alternatives that could be considered by the Congress over the next few years.

In evaluating military retirement, I will focus on two major criteria: costs and the ability of the military to retain career personnel. Assessing costs and retention effects requires that one consider not only retirement pay itself, but also other parts of the military compensation system. Given the emphasis at these hearings on retirement, I will assume that other parts of the compensation system remain largely unchanged.

While my discussion today will emphasize effects of retirement on costs and retention, I realize that military retirement is an issue that some treat as a barometer of Congressional concern for the military. This suggests the importance of considering any reductions in retirement in the same legislative package as the improvements in military compensation the Congress now has before it. Such a package approach could minimize adverse effects on morale and retention, and I will return to the notion of a package approach throughout my testimony.
 

CURRENT MILITARY RETIREMENT SYSTEM

The key features of the current military retirement system are well known. Military personnel can retire after 20 or more years of service and immediately receive an annuity, which lasts for life. The annuity is adjusted twice a year for changes in the Consumer Price Index (CPI). The initial amount of the annuity equals 2.5 percent of basic pay times years of service. For a typical retiree leaving the military today after 20 years of service, the annuity amounts to about $7,000 for an enlisted person and $16,000 for an officer. Military personnel make no explicit contribution toward their retirement while on active duty.

These provisions apply to active-duty personnel who retire without disability. There are special systems for disability retirees and for retirees who are part-time reservists, and there is also a contributory program that allows military personnel to provide annuities for their survivors. I will limit my remarks today to the nondisability system for active-duty personnel, which is by far the most costly part of the military retirement program.

Effects on Retention

Retirement annuities have important effects on willingness to remain in the military, particularly for those with nearly 20 years of service. Retention rates for these personnel are well above 90 percent. Despite these high rates, there has been concern in recent years--particularly in the Navy and the Air Force--about losses of senior careerists. Any changes in military retirement annuities that lead to a further decline in senior career levels could heighten this concern.

Important as it is to retaining senior careerists, the military retirement system may also provide some undesirable incentives. There is a strong push to leave after completing 20 years of service in order to begin receiving one's annuity. Indeed, much of the recent decline in senior careerists in the Navy and Air Force has occurred because of the large number of persons who have left the military immediately upon reaching retirement age. To arrest that decline, several studies have recommended reducing annuities for those who leave after completing only 20 years of service.

Moreover, because benefits are so distant, the current retirement system provides little incentive to stay in the military at the first reenlistment point, which usually occurs after three or four years of service. Yet first-term reenlistees provide critical mid-career personnel and the base for future senior careerists. Most studies have recommended improvements in benefits so as to increase the incentive to remain in the military at the first reenlistment point. These increased benefits could be financed with the savings from reducing benefits for 20-year retirees.

The clear link between military retirement and retention, coupled with current service needs, suggests that any change in the retirement system should be designed with an eye toward improving overall retention. CBO's estimates suggest that overall numbers of career personnel are increasing in each of the services. Indeed, retention problems today are generally limited to specific skills. Nonetheless, the prospect of increases in military strengths, coupled with the need for more experienced personnel to operate highly technical equipment, suggests the importance of continued growth in career force levels, particularly in the Air Force and Navy.

Flexibility

The current retirement system may have created Inflexibility in the personnel management system. Because substantial benefits are available after 20 years of service, but none before, personnel managers may be reluctant to urge or require that some leave before completing 20 years. This is particularly true for enlisted personnel, since they are not even eligible for the severance pay that is available to officers. Yet in some services, it might be appropriate to encourage military careers of about ten years. Ten-year careers might be considered for skills such as the combat arms, where there is a need to man weapons and perform unit supervisory roles but less of a need for senior supervisors.

Costs

Costs are a concern that have led many to scrutinize military retirement. In fiscal year 1982, nearly $15.1 billion in annuities will be paid to 1,415,000 retired personnel. This compares to $1.2 billion paid to 435,000 retired personnel in 1964. Costs and numbers of retired personnel have grown sharply because of the large numbers of retirees among those who served in World War II and Korea; costs have also increased because of adjustments based on the CPI.

Costs over the next two decades should not grow so rapidly, however. Our forecasts suggest only about a 10 percent increase in active-duty retirees over the next two decades and--apart from CPI adjustments and wage growth--similar increases in costs. CPI adjustments and wage growth could, of course, cause much larger increases.

Indeed, the most important question concerning retirement costs should not be their size today or their growth over the next few decades. Rather, concern should focus on whether alternatives to the current system--when evaluated with other military compensation reforms--would result in acceptable retention levels at less cost. Let me turn now to a discussion of some possible alternatives.
 

ALTERNATIVES TO THE CURRENT SYSTEM

Studies of the military retirement system have produced dozens of alternative schemes. In order to provide a flavor for the range of proposals, I have grouped them into several categories, emphasizing those proposals that seem most widely discussed today.

Early Vesting

Almost all studies of military retirement have proposed that those leaving the military with fewer than 20 years of service receive some benefits; currently they receive none. This "early vesting" of retirement benefits should improve the flexibility of the system, since military managers might be willing to separate those with fewer than 20 years of service if they have some pension rights. By moving benefits nearer to the first-term reenlistment point, early vesting might also strengthen the incentive to reenlist and so increase numbers of careerists and improve the pattern of retention.

Early vesting could add to retirement costs. Long-run costs for retirees from active duty might eventually increase by roughly $1 billion a year in today's dollars. This estimate assumes that those with between 10 and 19 years of service receive a deferred annuity, beginning at age 60, based on the same formula as the one used for today's retirees (2.5 percent times basic pay times years of service). While costs would eventually increase, the rise would not begin for many years--until those leaving reached age 60. Thus it might be possible to garner the advantages of early vesting while offsetting any added costs by other changes. One such offsetting change could involve reform of the provisions for cost-of-living adjustments.

Cost-of-Living Provisions

Currently, military annuities are adjusted twice a year, in September and March, based on increases in the Consumer Price Index. The Administration has recommended switching to a once-a-year adjustment of retirement pay. This change has been approved by the Senate in its current authorization bill, contingent on the same change being made for civil service retirees. Similar provisions were contained in last year's authorization bill. If once-a-year increases were implemented by skipping next September's CPI increase, and then providing annual increases starting in March 1982, savings in fiscal year 1982 (based on CBO economic assumptions) would amount to $440 million for the military retirement system alone.

Much more far-reaching proposals have also been made. Senator Stevens, for example, introduced a bill (S. 677) that would apparently include the military. For federal retirees who are younger than age 60, this bill would provide an annual raise equal to only half the amount of the CPI increase. Those between the ages of 60 and 64 would still only receive annual raises, but these would equal the full CPI increase, while retirees age 65 or over would receive twice-a-year adjustments based on the full CPI increase. CBO has not formally estimated the costs of this legislation. But savings in the first year of implementation could amount to about $400 million for the military alone, with much larger savings in future years as military annuities were sharply reduced.

Indeed, with one important change, S. 677 would provide a variant of the "two-tiered" annuity that most studies have recommended. A two-tiered annuity would provide less military retirement pay in the years immediately after a person left the military, but more in old-age years. If S. 677 were modified to provide a catch-up increase at age 60, making up for the half-CPI increases up to that age, then it would provide such a two-tiered annuity.

Changes in the cost-of-living provisions--especially large changes like those in S. 677--could affect retention of career personnel. Rates of retention for those with more than 20 years of service might actually improve because the annuity gained by leaving the military would be reduced. But, facing the prospect of lower annuities, those at all years of service up to 20 years--and particularly those with more than about 10 years of service--might leave in larger numbers. Thus it would be important to couple any major change in cost-of-living provisions with improvements in retirement--such as early vesting--or other changes in pay that would offset these declines in retention.

Social Security Offsets

Military personnel contribute to social security and receive benefits based on their military service. As with all covered workers, military members' contributions to social security are matched by their employer, the Department of Defense. Several studies have recommended reflecting this matching contribution by reducing military annuities at the time social security benefits begin. This reduction, often called a social security offset, would equal part or all of the social security benefits attributable to military service.

Such an offset could cut costs. If the social security offset for military retirees was patterned after the one now in effect for those receiving military survivor benefits, then savings could eventually reach about $2.5 billion a year in today's dollars. Depending on whether the Congress applied such changes to current retirees or to those on active duty, these savings might not begin for as many as 40 years, however.

A social security offset could result in lower rates of retention for many in the military because of the reduced size of their annuities. Thus, as with the cost-of-living changes just discussed, it might be appropriate to couple such changes with other increases designed to improve retention.

More Minor Provisions

The changes discussed above represent major revisions to the retirement system that are likely to have substantial effects on costs and retention. There are also more minor provisions that would be likely to affect costs and retention more modestly, but could be considered on equity or other grounds.

For example, current law offers military retirees a "look-back", provision. This provision allows new retirees to base their retirement pay either on recent military pay raises or on recent increases in the CPI, whichever are more favorable. Such a provision was repealed for civil service retirees last year. Savings from repealing the provision for the military would probably be minimal unless CPI increases again exceed increases in military basic pay, as was the case several times in the 1970s. If that situation recurs, then savings could be substantial.

Military personnel who can time their retirement precisely can also take advantage of the "rounding" feature. Under that provision, those who retire with more than six months' service receive a full year's credit for retirement purposes; those retiring with less than six months receive no credit at all for that year. The Defense Manpower Commission, in its 1976 review of the military retirement system, recommended prorating retirement credit according to the portion of the final year spent on active duty. Such a change might be viewed as equitable and could result in modest cost savings.

Last year the Congress also decided to base military retirement annuities on the average pay over the three years when pay was highest, rather than basing the calculation on retirement pay at the day of retirement. This "high-3" provision, however, applies only to new recruits and so will have little effect for 20 years. The Congress could phase in high-3 more quickly, perhaps over the three years recommended by one earlier study. Phasing in high-3 over three years might reduce costs by several hundred million dollars over the next five years. But it might also be viewed as unfair by those near retirement age and could adversely affect retention, which suggests coupling such a change with other pay increases designed to improve retention.

Finally, sharp increases in the CPI, coupled with much smaller increases in military pay in the 1970s, have left some retirees receiving substantially more in retirement pay than those retiring today with the same rank and years of service. If the Congress perceives this as inequitable, it could limit cost-of-living raises for such retirees--or otherwise gradually reduce their retirement pay--until it matched the pay of those retiring today. Such a revision might be coupled with a provision basing future cost-of-living increases on CPI or wage growth, whichever is less, in order to avoid similar problems in the future.
 

ACCRUAL ACCOUNTING FOR MILITARY RETIREMENT

My testimony so far has focused on changes in the retirement system that would affect benefits. But one key revision to the system--implementing accrual accounting--would not change benefits at all. Accrual accounting would, however, change the military budget so as to reflect the liabilities that are building up today because of military personnel now on active duty. Currently, the federal budget reflects only the costs of those already retired.

Accrual accounting could greatly improve management of military manpower by making retirement costs more obvious. For example, the 5.3 percent across-the-board pay raise recommended by the Administration will increase the retirement pay for all those retiring in 1982 and beyond. But retirement costs in the 1982 budget are hardly affected at all by this special raise. Under an accrual accounting system, the future liabilities caused by the 5.3 percent raise would be recognized immediately because the amount that would have to be set aside in 1982 to fund future retirement costs would increase by about $650 million. Similarly, under accrual accounting, any savings from changes in retirement benefits--such as those discussed above--would be immediately visible in the federal budget. This should improve the debate over changes in retirement.

Implementation of accrual accounting could cause some onetime shifts in the defense budget. Pending receipt of the details of the Administration's accrual accounting proposals, CBO has not estimated the size of these changes. But it is likely that budget authority and outlays in the defense budget would increase, since the budget would reflect not only the liabilities that are building up because of today's military employees but also the costs of paying off the unfunded liability for past employees. Such a one-time shift in the defense budget might be a small price to pay for the management improvements that could stem from making the full costs of military retirement visible in the budget.
 

SUMMARY

I have tried to emphasize today that retirement is an important part of the military compensation system, and plays a key role in retaining career personnel. At the same time, my testimony has suggested that there are changes in the retirement system that could improve the pattern of military retention while also reducing long-run costs. Such changes might include a social security offset or shifts in cost-of-living provisions.

Of course, any major change carries some risk of adverse effects on career retention at a time when we want to preserve the trend toward increases in numbers of careerists. Thus, it is probably desirable to couple any reduction in retirement benefits with improvements in military compensation designed to ensure adequate retention. Such improvements could include early vesting of retirement benefits as well as other revisions.

Retirement is also a highly emotional issue, and any discussion of changes may trouble today's military personnel at a time when the importance of their service is being stressed. This problem underscores the need to make changes in the retirement system gradually in order to minimize the effects on those now serving. Gradual change means, of course, that savings are not likely to occur quickly. But, if this Subcommittee works to implement some form of accrual accounting for military retirement, then at least future costs and savings will be clearly reflected in today's federal budget.