Skip Navigation

Mr. Chairman and Members of the Subcommittee, I appreciate the opportunity to appear before you this morning to discuss military pay and benefits. I will briefly outline the Administration's proposals to enhance military compensation and then focus on two questions that are central to the current debate about that compensation. First, are service members leaving in large numbers because of an inadequate retirement system? Second, what does the much publicized "pay gap" between military and civilian personnel really mean?

The Congressional Budget Office (CBO) recently completed analyses of both of those issues at the request of the Senate Committee on Armed Services. In answer to the first question, CBO found that the Military Retirement Reform Act of 1986 (known as REDUX) is not having a discernable, widespread effect on the retention of midcareer personnel. In answer to the second question, CBO found that the 13 percent pay-gap figure cited in the press does not accurately measure what it purports to and that, in general, the whole notion of relying on a pay-gap estimate to set pay raises is inappropriate.

CBO did not examine overall trends in retention and can neither verify nor dismiss reports of growing problems. CBO's analysis indicates only that if the services are indeed experiencing serious retention problems, the Military Retirement Reform Act is not the cause, and repeal of that act may not solve the problems. Similarly, CBO did not estimate how large a rise in military pay is needed. Instead, the analysis indicates only that estimates of the so-called pay gap are not a useful guide for policy.
 

RECENT PROPOSALS TO IMPROVE MILITARY PAY AND BENEFITS

Reports of recruiting and retention problems in the armed forces have received increasing attention in the news media. Undoubtedly, failure to provide military personnel with adequate pay and benefits would, over the long run, threaten U.S. military capability. Members of this Subcommittee in particular are aware of the problems caused by the so-called hollow force of the 1970s, when low-quality recruits, a lack of experienced personnel, and insufficient training weakened the effectiveness and reputation of the U.S. military.

The Administration's budget for fiscal year 2000 would substantially increase military compensation. It would provide an across-the-board pay raise of 4.4 percent (0.1 percentage point above the employment cost index, or ECI) and additional targeted pay raises of up to 5.5 percent for some mid- and senior-level personnel. It would also modify the REDUX retirement system, increasing inflation protection for retirees and raising the level of benefits paid to personnel who retire with 20 years of service (from 40 percent of basic pay to 50 percent). CBO estimates that those changes would increase discretionary spending by a total of $14 billion through 2005 (compared with a baseline in which annual pay raises matched CBO's projection of the ECI). Significant increases in direct spending would not occur until after 2006, when service members who are under the post-1986 REDUX system reach 20 years of service.

Other proposals--including one offered by the Senate Armed Services Committee--would increase compensation even more. For example, S. 4, which was reported out of that committee earlier this month, provides for pay raises to significantly reduce the so-called pay gap. Those raises, together with enhanced educational benefits for families of veterans, make S. 4 more costly than the Administration's plan. CBO estimates that the bill would increase spending by a total of $23 billion through 2005--$18 billion in discretionary spending and about $5 billion in direct spending.

Before adopting such proposals, the Congress might want to consider whether they offer realistic and cost-effective solutions to the military's personnel problems. If the civilian job market continues to be as strong as it has been, the Department of Defense (DoD) may indeed have to devote considerably more resources to attracting and retaining a high-quality force. But if the department spends $14 billion or $23 billion unwisely, without identifying and focusing on the real causes of its personnel problems, it may be left both with those problems and with insufficient resources to solve them.
 

THE EFFECTS OF REDUX ON MIDCAREER RETENTION

CBO's recent analysis concluded that REDUX is not causing a large exodus of midcareer personnel. That finding comes from comparing the retention decisions of thousands of service members who began active duty shortly after REDUX was implemented but who, because of their participation in different accession or commissioning programs, were under different retirement systems. Using standard statistical methods, CBO isolated the effects of being under REDUX from the effects of other factors that might influence retention. The analysis controlled for such demographic variables as age, sex, marital status, and education as well as for occupation in the military and the possible effects of changes in the services' personnel policies during the reductions that followed the Cold War.

If the new retirement system was having a marked effect on retention, as the Joint Chiefs of Staff have reported, CBO's analysis would most likely have shown that effect. Instead, CBO found that, in general, being under REDUX had no discernable effect on the midcareer retention decisions of people who began active duty in 1987. The Navy was the only service in which REDUX had a statistically significant impact on retention of enlisted personnel between their 6th and 11th years of service. For enlisted members of the Army, Marine Corps, and Air Force, CBO found no statistically significant link between REDUX and retention in those years of service. The analysis of officer retention, which focused on years of service 8 through 10, identified a negative effect for the Air Force but not for the Navy. CBO was unable to examine Army officers, but an analysis conducted by Dean Dudley, a professor of economics at West Point, concluded that REDUX did not have a statistically significant impact on their retention.

Those results are not surprising. In both cases in which a negative effect could be identified clearly, it was fairly small--a reduction of 3 or 4 percentage points in retention. Such relatively modest effects are generally consistent with what many analysts predicted when REDUX was enacted. Given the range of factors outside CBO's model that could influence retention and the statistical uncertainty surrounding the estimates, it is possible that REDUX is having a small negative effect on retention in each of the services. For the most part, CBO's results neither demonstrate such an effect nor rule it out.

Nevertheless, CBO's analysis indicates that REDUX has not caused a large midcareer exodus among the first officers and enlisted personnel to enter the service under that system. Any large declines in midcareer retention that have been observed in the armed forces probably result from other factors. Frequent, unscheduled deployments and attractive civilian job opportunities for people in particular occupations are among those other possible explanations. In CBO's equation for Air Force officers, for example, being a fighter pilot had a much larger negative effect on retention than being under REDUX did. It is unlikely that repealing REDUX would compensate for those other factors.

The effects of REDUX on retention may become more pronounced in the future, however. Some officers and enlisted personnel who began active duty in 1987 may not be fully aware of which system they are under. Many others--at least before the recent spate of media attention that REDUX has received--were not fully aware of the financial implications of being under that system. Moreover, the leaders of the military services have now made repeal of REDUX their top readiness priority and a symbol of the nation's commitment to support its military personnel. In that climate, failure to make some change could itself have a negative effect on morale and retention in the near term.

Most analyses of retention--both today and when REDUX was enacted--indicate that dollars spent on deferred compensation, such as retirement pay, have less impact on retention than dollars spent on the pay and benefits that service members receive while still on active duty. Thus, one option to consider is whether changing other benefits--improving access to military health care, for instance, or eliminating the out-of-pocket costs paid for housing--might provide a strong symbol of commitment without increasing the military's reliance on deferred compensation. Changes in the structure of the military retirement system, which offers service members no incentive to save and which provides no vesting until 20 years of service, may indeed be warranted. But any proposed change in that system deserves careful analysis and consideration.
 

IS THE "PAY GAP" A USEFUL GUIDE FOR PAY POLICY?

CBO has not analyzed overall trends in retention or tried to identify what specific policies--increases in bonuses, special pay, or across-the-board pay raises--would allow DoD to meet its personnel goals most efficiently. But CBO has looked carefully at the so-called pay gap that is often cited to support demands for large increases in basic pay.

CBO's analysis of the pay gap concludes that the figure commonly cited as a measure of that gap (13 percent) has no value in determining the appropriate level of military pay. It does not indicate whether the nation is treating military personnel fairly in the sense of offering them a standard of living comparable with that of other U.S. citizens. Nor does it indicate what DoD must pay to meet its requirements for skilled, experienced, high-quality personnel.

Although it is called a measure of the pay gap, the 13 percent figure is not actually based on a comparison of civilian and military pay levels. Instead, it is a comparison of changes in pay over time. It is calculated by comparing increases in military basic pay since 1982 with increases in the ECI, an index of wages in private industry. The choice of 1982 as a starting point for the comparison is essentially arbitrary, and at most, the resulting number is a measure of changes in relative pay since that date. It indicates that overall military pay has increased more slowly than civilian pay but not whether military pay is higher or lower than civilian pay.

In addition, the pay-gap calculation is faulty even as a measure of the changes in relative pay. On the military side, it looks only at basic pay without taking into account other factors that affect regular military compensation, such as food and housing allowances, the speed of promotion, and the financial benefits from the tax-free status of many allowances. On the civilian side, it looks at average wage growth for workers who are typically older than members of the military and more likely to hold a college degree. Those demographic differences distort the comparison because wages of young high school graduates--a group more representative of the military--have risen less rapidly than wages of other civilians. Moreover, the most commonly quoted measure of the pay gap does not compare military and civilian wages over the same time interval.

After adjusting for those factors, CBO found that since 1982, the wages of enlisted personnel actually rose by between 3 percent and 10 percent relative to those of high school graduates with a similar age distribution. The wages of officers, in contrast, fell by between 6 percent and 12 percent over that period relative to the wages of college-educated civilians with a similar age distribution. Thus, the so-called pay gap is actually negative for enlisted personnel and positive only for officers.

The point of CBO's analysis is not that the pay-gap calculation is faulty and could be improved. The message is more fundamental: that shifts in the relative wages of military and civilian personnel, cumulated over many years, are not a useful guide for policy. No matter how carefully calculated, they say nothing significant about either the fairness of military pay or the adequacy of that pay to support DoD's personnel needs.

Would a comparison based on the absolute level of military and civilian pay rather than increases in pay prove helpful? The answer would depend on how the results of the comparison were used. A comparison of pay levels might indicate whether the nation was treating military personnel fairly--in the sense of offering them either a lifestyle comparable with that of other U.S. citizens or wages comparable with what they might expect in civilian jobs. Some people would argue that asking military personnel to work for less than they could earn in private-sector jobs is unfair, even though a strong sense of duty and patriotism might still motivate military personnel to serve.

The results of such a pay-level comparison might be surprising. Based on current military pay, CBO finds that enlisted service members earn more than about 75 percent of U.S. male high school graduates of the same age. Similarly, officers earn more than about 75 percent of U.S. college graduates of the same age. Those findings suggest that the financial rewards of military service are fair in the sense that they are more than comparable with what private-sector jobs offer most U.S. citizens. One could still argue, however, that service members deserve even more because of their exceptional quality and devotion and the sometimes arduous and dangerous conditions in which they work.

Even though careful comparisons of pay levels might say something about fairness, they would not answer the most pressing question that DoD's pay policy must address: whether current pay levels allow the department to meet its requirements for personnel. Being a member of the military is a unique profession. Sometimes that uniqueness entails advantages. For example, flying an F-16 as an Air Force pilot in Europe may be more rewarding and challenging than flying for a private airline. As a result, although DoD may need to pay its pilots more than it does today, it probably does not need to match the salaries that civilian airlines offer. In other cases, the special character of the military can be a drawback. DoD may have to offer high school graduates assigned to combat specialties more than they could earn in the private sector to induce them to accept military discipline and other sacrifices associated with serving in the armed forces. Because of the differences between military and civilian life, there is no reason to assume that matching civilian salaries--either in specific occupations or overall--would allow DoD to meet its personnel needs.

But if measures of the pay gap are not a helpful guide for policy, what approach can DoD and the Congress use in setting pay? Over the long run, the best approach may be to ask whether retention patterns allow the military to achieve its goals for experienced personnel. If DoD is consistently failing to attract and retain personnel across a wide range of occupations, then increases in military pay relative to civilian pay may be warranted. If retention or recruiting is a problem only for some services, some occupations, or a few years of service, then other solutions--such as bonuses or changes in the pay tables--may be more cost-effective than across-the-board increases. In the case of recruiting problems, for example, studies have repeatedly shown that additional recruiters, more advertising, or higher enlistment bonuses are more cost-effective than across-the-board pay raises.

In the short run, it can be difficult to determine whether retention is slipping or whether observed changes reflect normal year-to-year variations. If no clear trend is apparent, it may be appropriate to set that year's military pay raise equal to the growth in civilian earnings--as indicated by the ECI or by a similar measure tailored to reflect the demographic characteristics of military personnel. Matching civilian wage growth is reasonable in the short run because if nothing else changed (such as DoD's needs, the attractiveness of the military lifestyle to young people, or civilian unemployment rates), a policy that held relative wages constant might hold the supply of manpower to DoD constant as well.

But the pay gap approach--which adds up growth in civilian wages over a decade or more and compares that with growth in basic pay over the same period--is something quite different. Over a long period, many factors other than wages will have changed. If DoD is consistently unable to meet its requirements for experienced enlisted personnel, then a pay raise may be needed. But it is arguably of little interest, and even less relevance, to know that the earnings of enlisted personnel have risen relative to civilian earnings since 1982.

Although the notion of a 13 percent pay gap has little to support it, that figure has been widely publicized--and just as widely misunderstood. Because of the diverse forms that military compensation takes (housing and subsistence allowances, bonuses, the tax advantage), service members may have difficulty evaluating how they fare compared with civilians. Depending on how senior military leaders treat this issue, eliminating the so-called pay gap, like repealing REDUX, could easily become an important symbolic goal. That could force DoD either to disappoint the expectations of service members or to devote more resources to compensation--and fewer resources to other needs--than are truly warranted. Some increase in military pay relative to civilian pay may be appropriate, especially if civilian unemployment continues at its current low level and declines in retention become widespread. But the best way to determine the amount of that increase may be to examine DoD's ability to fulfill its requirements for personnel.