CHAPTER SIX
Reforming Compensation
The military’s compensation structure creates perverse incentives that are at odds with voluntary service. Why does the Pentagon offer a retirement system that vests after twenty years? And why does the Pentagon allow that pension to be drawn at exactly the same moment instead of at age sixty-five? Is this pension structure not (1) using coercion to keep some employees in place longer than they otherwise would choose to remain and (2) creating an incentive for them to retire promptly, at the moment of peak proficiency?
Military service is motivated by selfless love of country, but not that alone. The military relies on other incentives, too, including pay, health care, retirement, training, education, camaraderie, travel, and adventure. Unfortunately, the military’s complex compensation system is overly expensive and inefficient—two factors that are closely related. Culturally, some military members like to believe that money and service are at odds, a belief that allows them to neglect how crudely and poorly the Pentagon compensation structure is designed.
The economic incentive is and has always been part of Pentagon planning for manpower. Incentives matter. As economist Eric Hanushek wrote in 1977, “Many of the largest personnel problems [in the US military] are exacerbated, if not caused by, the incentive system.” The main theme underlying the Total Volunteer Force (TVF) is that coercion is less efficient than volunteerism. And compensation reform is perhaps the easiest way for the Pentagon to improve talent incentives in place of coercion.
The stark reality is that military pay offers no monetary rewards for excellence. The armed forces spend more funds moving personnel around the globe than on performance rewards. More than 90 percent of personnel spending in 2016 went for base pay, compared to 3.9 percent on permanent change-of-station (PCS) moves and half of an percent on “incentive and special pays.” That phrase is a misnomer, as incentive pays do not reward performance; rather, they are designed to compensate career specialization and provide broadly available retention bonuses.
Three reforms to military compensation will increase efficiency and morale while reducing gross expenses. The first is to adjust the standard military pension in a way that gives more options to service members and establishes a trajectory to eliminate twenty-year cliff vesting. The second reform is an adjustment to the base pay schedule that ends the use of tenure and replaces it with role and responsibility pays. The third reform is to utilize existing incentive pays to make the new job-matching program work efficiently—a way to reward individuals who volunteer for tougher assignments. We will lump the latter two into a single section below, but first we will review the problems identified in chapter 1.
Leader/Talent
The Leader/Talent results presented in chapter 1 include four elements concerning compensation (figure 6.1). Scores range from always true (+2) to always false (−2), and all four compensation elements were scored negatively by military respondents. The average score on compensation does not mean that military paychecks are low. Rather, it means that the compensation process is evaluated poorly regarding rewards for performance, good work, efficiency, and retention. The single lowest score goes to “bonuses are used effectively to reward good work” for the simple reason that the military has no compensation mechanism to reward or penalize performance. In addition, there is a huge gap between military and civilian organizations on two of the four elements: “pay is closely aligned with performance” and “bonuses are used effectively to reward good work.”111
FIGURE 6.1. Leader/Talent Measures of Compensation Elements in the US Military
Source: Author’s Leader/Talent survey
Changing compensation is not necessary to make effective reforms to other personnel processes such as promotions and job-matching, but it will improve them. And even by itself, a modernized compensation system can enhance talent management and morale.
Rather than suggest merit-based pay, an approach that has proved controversial in government practice, the TVF instead recommends shifting base pay away from seniority and toward the specific role a service member fills. In essence, paying individuals for the exact job they perform is the best way to reward merit. The more meritorious individuals will take on harder assignments, in more remote locations, involving greater responsibility—and consequently receive higher pay to reflect those choices.
Unheeded Compensation Commissions
Every four years, the Pentagon conducts a high-level review of compensation known as the Quadrennial Review of Military Compensation, or QRMC. President Obama canceled the one due in 2016 on the grounds it was “not required” in light of the congressionally created Military Compensation and Retirement Modernization Commission, or MCRMC.112 That commission published its recommendations in early 2015, and its proposal to change the military retirement system was adopted into law that year. It was an impressive achievement on one hand, given that dozens of blue-ribbon panels had failed in similar endeavors over the decades, but the change itself was not as monumental as many observers seem to think. The retirement pension for new active-duty members will be reduced by one-fifth (starting in 2018) in order to develop a new defined-contribution savings asset, but four-fifths will remain, unaltered in vesting at twenty years and payment immediately upon separation. The incentives have not been altered in any fundamental way.
The first QRMC was convened in 1965, and as Bernard Rostker noted in his history of the All-Volunteer Force (AVF), subsequent QRMCs often splintered over varying controversies. The implicit question about the purpose of compensation was often debated, despite the seemingly obvious answer—settled thanks to research by Beth Asch, John Warner, and (separately) Curtis Gilroy—to “attract, retain, motivate, and separate personnel.”113
The single most important compensation commission assembled in the past century was the Gates Commission, created at the presidential level in March 1969 at the beginning of the Nixon administration in order to evaluate the question of whether to replace the draft with a system of manning the military entirely with volunteers. The commission was led by Thomas S. Gates, who had served as secretary of defense under Eisenhower from 1959 to 1961 and who was initially skeptical of the proposal. Reporting just under a year after its formation, the members made a shockingly unanimous recommendation to end the use of conscription. What is neglected by history is that the Gates Commission also recommended a set of comprehensive compensation reforms it said were vital for making the AVF operate effectively. Those reforms, aside from increasing base pay to attract volunteers during accessions, were “unheeded in the years that followed,” according to Todd Harrison’s 2012 study114 of what kinds of compensation are actually valued by active-duty troops: “As a result, the fundamental structure of military compensation remains largely unchanged despite the transition to an all-volunteer force.”
A summary of the compensation recommendations made by the Gates Commission in 1970 included:
There have been major negative consequences of ignoring the Gates Commission. The first is galloping inflation in the net cost of military personnel since 1973, even as the size of the force on active duty has steadily declined. The second is a hugely inefficient bubble of active-duty manpower around the twentieth year of service (YOS), coinciding exactly with the twenty-year defined benefit (DB) cliff pension, unchanged for a century, which establishes perverse and coercive work incentives that are at odds with voluntary service.
FIGURE 6.2. Continuation Rates of Officers by Years of Service
Source: Military Compensation and Retirement Modernization Commission, Final Report (2015)
Consider figure 6.2 from the MCRMC that shows the continuation rate of officers in the armed forces by years of service. There is a distinct twenty-year notch. Very few individuals leave the service after the twelfth YOS, and an alarming number quit promptly at the twentieth YOS. This is rational behavior given the incentives, but the incentives are wasteful.
Pentagon data on retirements by years of service also show a bubble at the twenty-year mark (table 6.1). There were 451,522 total non-disability military Army retirees who received retired pay in September 2011. More than half of the retirees left during the twentieth year of service (234,462), and another 20,000 left through early retirement. In sum, 56 percent of Army retirees left active-duty Army service at the first retirement opportunity. The same pattern holds in the other services.
|
TABLE 6.1. |
||||||
|
Military Retirement Bubble at 20 Years of Service115 |
||||||
|
Service |
|
Retirees |
(20 YOS) |
Early Retirees |
||
|
Army |
451,522 |
51.9% |
4.4% |
|||
|
Navy |
392,816 |
51.2% |
3.6% |
|||
|
USAF |
541,417 |
47.9% |
3.0% |
|||
|
USMC |
85,464 |
47.4% |
5.4% |
|||
|
Source: Author’s calculations from DOD actuary data |
||||||
The retirement bubble at twenty years is followed by a smaller bubble at the thirty-year mark, which is when the multiplier of years in service is maximized. Serving beyond thirty years does not increase the multiplier. Across the Department of Defense (DOD), there were 38,600 retirees with thirty YOS compared to 6,034 retirees with thirty-one YOS and roughly 10,000 retirees for all years beyond that combined. These two bubbles are clear evidence that the vesting cliff and multiplier cliff create strong incentives to stop working on active duty.
As for costs, the twenty-year defined benefit pension is growing exponentially more expensive. The Defense Business Board reported in 2011 that annual outlays exceeded $50 billion in 2010 and will more than double before today’s lieutenants become generals. Liabilities of the program are $1.3 trillion (roughly 10 percent of US GDP) and will rise to $2.6 trillion in 2035. This is “unsustainable,” say Pentagon actuaries and countless scholarly studies.
There are other cost drivers. The tenth QRMC, published in early 2008, reported that net costs of pay and benefits were respectively $10,600 and $17,800 higher for enlisted members and officers than for their civilian peers. Mark Cancian described the updated cost pressures and the widespread concern that they are harmful to national security in a 2015 article in Joint Forces Quarterly:
Since 2001, pay per Active-duty Servicemember has grown over 80 percent in 2001 dollars (about 50 percent in constant dollars). Military pay has increased 40 percent more than civilian pay since 2000, and enlisted Servicemembers are now paid more than 90 percent of their civilian counterparts with comparable education and experience (officers earn more than 83 percent of their civilian counterparts).116
So? Isn’t higher pay the market premium for serving in uniform, given the danger and other unique burdens involved? Yes, but as Cancian noted, when active-duty military personnel get more expensive, DOD adapts by using more reservists, contractors, and capital.
Recall the problem identified by Hanushek, cited in chapter 3, of implied contracts with absolute career control by the military, which must compensate accordingly. Hanushek also decried the undifferentiated nature of these implied contracts—paying soldiers the same whether or not they are based in the least desirable locales, paying pilots a bonus whether or not they actually fly, and so on. We should extend the Hanushek problem further by analyzing the additional cost necessitated by a closed labor pool. The nature of a centrally managed labor market involves trying to perfectly match the number (and specialty) of requirements with people, i.e., spaces with faces. The challenge to precisely match numbers puts the personnel commands on a knife-edge with costs on each side: overmanning and shortfalls. In fact, the eleventh QRMC in 2012 included a chapter that examined four critical fields—remotely piloted vehicles, special ops, mental health, and linguists—and even calculated the “requirements ratio,” which was significantly below 1.0 for these four specialties.
The military’s closed labor pool is a profound cost driver. Hanushek observed in his 1977 paper just four years after the AVF’s adoption that the closed nature of the military labor pool and other unreformed compensation policies would cause higher cost burdens than expected. Economically, the armed forces have engineered a highly inelastic labor supply. If the labor supply curve shifts inward (the incessant threat of the one individual most likely to quit), the military can only maintain equilibrium quantity of labor by escalating compensation. The standard economic toolkit offers only one solution to this problem: thicken the labor supply in order to make it more elastic. This is the effect of opening the labor pool to later entry, shown in table 6.2, leading to pay at, rather than above, equilibrium.
As an analogy, imagine trying to hire a plumber with five years of experience or more. If conscription were an option, this uniformed plumber could be paid below-market wages. But in a closed market where plumbers need to be accessed untrained five years prior and retained in a system of widely equal pay scales based on retaining the marginal individual, he would necessarily be paid above-market wages. To alleviate that pressure, the market must be opened to lateral entry at the mid-rank specialties. The closed system in place now is a one-way dynamic in which individuals have the ability to move out of but never into the ranks, albeit with rare exceptions. A closed labor market puts upward pressure on compensation.
In sum, there are two institutional distortions in the Pentagon’s compensation system that artificially drive up costs. Worse, as Harrison has shown, a great deal of military compensation is valued far less by young troops than it costs the services. The Total Volunteer Force would remedy both with two general reforms. First, transform base pay by shifting from tenure to role and responsibility pays. In addition, add assignment incentive flexibility to help fill difficult jobs. Second, transform the retirement system. The MCRMC-inspired reforms of 2015 did not go nearly far enough. A detailed case for each major reform is presented next.
Transform Base Pay from Tenure to Role and Responsibility
Concerns about military compensation often focus on the escalation in pay and benefits. Total compensation per service member has nearly doubled since 2001, while the size of the active force is actually lower. The Government Accountability Office noted that service member earnings are equivalent to the 80th percentile of comparable civilian compensation as of 2008, up from the 70th percentile just four years earlier.117 While escalating costs may be a problem, the more vexing issue is that the service chiefs have essentially no authority to shape and thereby control pay escalation. Rather, Congress determines annually what pay raise shall be applied to the one-size-fits-all pay table. Intelligent critics can complain that Congress should exercise more restraint, but there is an unavoidable incentive problem: military pay raises should not be a single lever managed by a legislative vote. In other words, why is there a universal pay table in the first place, instead of granting each service chief the authority to establish pay and allowances for their services within a fixed personnel budget? Flexibility is the solution to both the political economy problem that drives inflation and the readiness problem rooted in a rank structure constrained by the industrial-era hierarchy.
|
TABLE 6.3. |
||||
|
DOD Personnel Spending for FY 2016 (in $ millions) |
||||
|
|
|
Total ($m) |
Total (%) |
|
|
Pay, Officers |
33,374 |
29.5% |
||
|
… Incentive Pays |
488 |
0.4% |
||
|
… Special Pays |
1,189 |
1.1% |
||
|
… Allowances |
555 |
0.5% |
||
|
Pay, Enlisted |
68,655 |
60.7% |
||
|
… Incentive Pays |
245 |
0.2% |
||
|
… Special Pays |
1,730 |
1.5% |
||
|
… Allowances |
2,356 |
2.1% |
||
|
PCS moves |
4,460 |
3.9% |
||
|
Total |
113,054 |
|||
|
Source: Department of Defense FY 2017 President’s Budget (M-1). |
||||
In this section, I will showcase the current pay table—the essential nature of which has been unchanged since the mid-1940s—and an alternative pay table that could improve flexibility, morale, and readiness, and even allow cost savings.
Active-duty pay and benefits total $33 billion for officers and $74 billion for enlistees, about half of each being base pay and the rest including retirement accruals, housing, subsistence (meals), social security taxes, and roughly 5 percent for special pay and allowances. Table 6.3 details a few of these items for each service.
Military base pay has been calculated on two dimensions since at least 1949: rank (pay grade) and cumulative time in service. The official pay table, which applies to all four armed forces, is organized in columns for time in service every two years (up to forty) and rows for pay grades (E-1 to O-10). Table 6.4 is an abridged version, but the reality is that most careers follow a narrow path along the diagonal. For example, there are no O-1s in uniform at the ten-year point, except for the rare individual with prior enlisted time, and there are certainly no O-5 and higher officers with less than two years of cumulative time in service.
Note that monetary rewards in the base pay table are greater for tenure than rank. For example, an O-1 typically is promoted to O-2 at the two-year point. Base pay rises $453 per month due to the higher rank, whereas the tenure step is worth $475. For O-2 to O-3, rank is worth $593, while time in service is worth nearly $800. The next promotion to O-4 at year twelve pays roughly $600, whereas an additional decade of service yields $1,200 more in monthly salary.
The use of tenure-base pay is intended as a crude proxy for level of responsibility—crude in the sense that a ten-year captain or sergeant is assumed to be 20 percent more responsible than a four-year captain or sergeant, and 10 percent less responsible than a sixteen-year person of equivalent rank. The use of tenure-base pay is an anachronism of a day when personalized algorithms that factor in occupation, location, merit, and detail-difficulty pays were impossible to imagine, let alone implement. Those days are gone. Why use crude proxies when the technology exists to fairly and impersonally designate pays for each role, while also fairly and impersonally adding bonus pays for critical roles that are difficult to fill?
Pay has always been supplemented by untaxed allowances for housing and subsistence, which are also unrelated to merit. These in-kind payments have roots in military tradition long before the nation was founded, but we know that the Navy and Marine Ration Chart of 1794 prescribed daily rations. For example, Monday rations were “1 pound of bread; 1 pound of pork; ½ pint of peas or beans; 4 ounces of cheese; and ½ pint of distilled spirits.”118 Most sailors were allotted one ration per day, specialists such as carpenters and boatswains received two, lieutenants three, and captains six. Today, common allowances include $530–$2,000 monthly for housing, more for overseas housing, money for uniforms, and a $250 monthly payment for jobs that require family separation.
There are over sixty special and incentive pays, according to the Defense Finance and Accounting Service,119 including long-standing pays for sea duty, flying duty, hazardous duty in combat, reenlistment, and other retention bonuses used for force-shaping. Special pays in 1949 were narrowly available, including sea and foreign duty pay for enlisted members of different ranks (ranging from $8 to $22.50) and aviation or submarine crew pay.
It must be understood that incentive pays are not designed to reward performance; rather, they are designed to compensate career specialization and skills retention. And this is what makes the history so instructive. Differential pay for occupational specialties is relatively rare today, with important exceptions: doctors, pilots, and especially Navy nuclear propulsion specialists (Hanushek said it “only occurred in extreme cases” in the 1970s), but it was common a century ago in the armed forces. Today, most troops of the same rank get roughly the same monthly paycheck, only because the US military devolved away from differential pay step by step during the twentieth century. In 1893, for example, monthly pay in the US Navy was highly differential across eighty enlisted ratings, from $9 (for apprentices, third class) to $70 (for machinists), whereas most sailors earned $24 to $35.120
In contrast, the use of tenure in pay calculations emerged in the late nineteenth century. The Navy pay chart for 1860 has five columns for “years on sea service” for officers, with the lowest, “Under 7” increasing by two-year increments up to the highest, “Over 13.” Either an officer was promoted to the higher (and higher paid) rank or stayed in the senior lieutenant pay bracket. In 1870, this had been replaced by a binary tenure column for each rank, with a bonus after the five-year term in that rank.
Today, differential pay is utilized for some specialties, but not to the degree necessary for optimal retention. Instead, large sums are spent on retention bonuses. One Rand study121 noted DOD’s reenlistment bonus spending rose from $625 million in 2002 to $1.4 billion six years later.
Hardship pays are common, to include extra pays for locations overseas with below-standard quality of life. These range from $50 to $150 per month. However, these pays are not set dynamically (reflecting supply and demand). Though they rarely utilize it, the services have authority for a special pay that could radically reshape job-matching. It is known as the assignment incentive pay (AIP). According to the DOD, it is designed to encourage members to volunteer for difficult-to-fill jobs or assignments in less-desirable locations. The monthly statutory maximum payable is $3,000.
The TVF would shift monthly compensation away from cumulative years of service and eschew tenure pays altogether. The alternative pay structure envisioned for the TVF is presented in table 6.5. Baseline pay is presented in a single column. The next column adds a monthly pay for the role: maintenance officer, cyber NCO, cryptographic/linguist, and so on for every occupational code across all the services. The role pay can also include different levels of responsibility, such as higher pay for command positions or other critical roles. The role pay could range from $0 to $10,000 per month. Included in these role pays is the level of responsibility, which can be much more nuanced than the tenure assumptions in the status quo. For example, there may be two USAF majors in the human intelligence (HUMINT) specialty, but one is a staff role at the Pentagon while the other is at a remote facility in the Middle East. The latter role would have a higher pay supplement, coded into the requirement. Moreover, the role pay would be publicly transparent, particularly visible on the online talent marketplace when the job is open.
The sum of these two pays—baseline and role—would serve as the base pay equivalent in current retirement formulas. TVF compensation should not be misconstrued as a way to stealthily slash Pentagon pension burdens. In addition, unit commanders would receive a 50 percent pay supplement during their months actively in command.
The role bonus would be composed of increments for skills and occupation (rather than skills alone). This authority would allow services to compensate the individuals who take on tougher jobs (including command) that involve higher career risk, longer hours, and greater stress. In principle, there is no reason to pay a senior O-3 in an easy job more than a junior O-3 in a demanding job. The same principle applies to E-5s, O-5s, E-6s, and so on.
The column “+AIP” stands for the assignment incentive pay, which already exists in current regulations, as mentioned earlier. The armed forces currently use bonus pays as flexible compensation for certain hazardous, remote, and otherwise challenging jobs. This flexibility is effective and should be expanded to enable decentralized job-matching in an evenhanded and objective manner to prohibit local favoritism. Jobs that remain unfilled after a given time should be paid more using a standardized incentive pay. As noted in chapter 3, wage flexibility is a core principle in labor markets.
The TVF recommendation is that the AIP should be used in conjunction with job-matching. The AIP should only be used objectively, not subjectively. That is, the AIP could not be granted at the discretion of a commander to award members of the unit. Rather, it would be applied according to an algorithm for unfilled requirements.
A posted requirement that receives no applicants for thirty days would get an automatic, publicly listed AIP for some percentage (say 5 percent) of base pay. The AIP would increase incrementally every fifteen days up to a dollar amount set by the service personnel command. Importantly, the AIP algorithm would reset to zero for each job posting. Patterns of unfilled jobs would show up in AIP data for roles throughout the workforce. Branch managers would be able to analyze the data and adjust the role supplements accordingly. The use of data—trending high and low labor volunteers over time—will allow compensation to be fine-tuned for roles throughout the military, generating untold cost savings and nimble force-shaping that are beyond reach today.
What if jobs remain unfilled even when the AIP has maxed out? After 180 days, qualifications should be relaxed to allow individuals one rank lower to apply.
There is one area where subjectivity should be allowed, and that is to lower or nix AIP for up to one-third of manpower requirements. The personnel center or local commander could use his judgment to designate non-AIP positions, giving a powerful signal regarding which ones are in fact high priority.
The consequence of shifting to TVF compensation would be a sea change in how force-shaping is done. Role pays could be adjusted nimbly to align hundreds of occupational labor supply-and-demand equilibria, from the half dozen Army helicopter repair specialties to the score of Air Force acquisition specialties. In the near term, monthly pay would not change much. In the long term, compensation would align with civilian peers in equilibrium, which will likely lower net costs to the military.
One last recommendation to change direct compensation is to reform the way travel for training and education programs is managed. A severe constraint on service flexibility is the Joint Federal Travel Regulation (JFTR), because it requires any program that lasts fewer than six months to be compensated with expensive “temporary duty” travel pay. Consequently, the services design most training programs to be longer than six months, which requires service members to incessantly move their permanent base locations. This regulation should be amended so that a greater variety of broadening programs can be offered. For example, active-duty troops could have the option to participate in congressional and business internships, or other brief training programs, without per diem reimbursement. To be clear, this would not mandate the end of temporary duty compensation, nor would it mandate any service branch to change its education and training programs; it would simply allow services more options to design and redesign training and education opportunities for members.
Specifically, the services should have the flexibility to waive TDY (temporary duty assignment) payments for optional training programs. Many graduate programs and fellowships are offered to senior officers as one-year assignments. Yet the same benefits—exposing officers to civilian colleges and private-sector firms—could be achieved with three- or four-month programs instead of nine or twelve months, which is the status quo. The roadblock is JFTR.
Transform the Military Pension
In January 2015, the Military Compensation and Retirement Modernization Commission issued its final report with fifteen recommendations for reform. The first recommendation was to lower the standard defined benefit pension by one-fifth and to add a 401(k)-style program. Unlike dozens of earlier reform commissions and boards, MCRMC achieved its goal. Congress was persuaded to develop legislation implementing the pension proposal, which it adopted later that year. Congress enacted a new Blended Retirement System—mandatory for new service members starting on January 1, 2018—which reduces the monthly DB payment from 50 percent of base pay to 40 percent. Technically, the base pay multiplier was reduced from 2.5 percent to 2 percent of base pay per year of service at the time of retirement. The blended plan also adds a supplemental defined contribution (DC) program.
MCRMC’s proposals were a small step in the right direction, but policymakers should not consider its report as the last word or even a comprehensive analysis of alternatives. The commission’s overall conclusion was that “several key features of the compensation system continue to meet the needs of the All-Volunteer Force.”122 I believe that is an incorrect assessment. The blended retirement is a step in the right direction, but it is a small step and is likely to be negligible in managing talent.
The standard military twenty-year cliff vesting creates a perverse incentive for both labor supply and labor demand, as the Gates commission noted decades ago. There are two main problems: the vesting date (at twenty) and the initial payment date (immediately upon retirement). The latter effectively pushes active-duty troops to immediately leave at the twenty-year point in order to begin drawing their pensions. And numerous studies have documented the reluctance of the military services to lay off senior officers in the years before the twenty-year point. This is the single greatest challenge for talent management, and it remains knotted.
A smarter approach would offer pension payments starting at the age of fifty-five or some length of time after the vesting date. Reservists get their pensions at age sixty irrespective of the age they retire. With the savings this extension will generate, monthly payments could be made larger. Another alternative would give entering service members the option to take higher blends of DC versus DB mixes, even allowing them to opt for full DC-only plans. MCRMC did not consider this, but it is a logical next step.
The latter half of this chapter highlights the inefficiency of the military DB structure using an analysis of the value of work. During the first year of service, annual base pay for a typical officer in any service is $35,212. By the twentieth year, base pay is $99,374. However, that last year of work will also qualify the individual for a lifetime stream of retirement income worth half of base pay, and the discounted value of that pension over a typical lifetime is $869,772. In other words, the twentieth year of service is worth a combined sum just under one million dollars.
What’s Wrong with the Defined Benefit
The existing DB retirement is utilized by just one in six enlisted troops from all service branches. Roughly half of enlistees serve on active duty for ten years or less, and only 13 percent serve beyond twenty years of service. Those who serve for 20 years earn a generous retirement, but most do not stay in that long. In contrast, nearly half of all officers serve for twenty years or more, qualifying for the lifetime DB payments. That disparity alone is a hangover from the draft era, when enlistees were poorly paid, poorly valued labor. To be sure, young men are more naturally fit as warriors than older men, a reality that has not changed under the AVF and the force structure of the enlisted corps and is particularly true for ground forces. Even today, the notion of a full career for an enlistee is encouraged only for the best who are promoted to the middle and upper enlisted ranks as NCOs.
Another challenge is cost. In the coming years, the Pentagon is likely to experience large pension pressures. Annual outlays for military pensions exceed $50 billion and will double during the next two decades, while liabilities of the program are predicted to grow from $1.3 trillion—roughly one-tenth the size of the US GDP—today to $2.6 trillion in 2035, according to the Defense Business Board in 2011. At best, the MCRMC reforms will shave one-fifth of the additional pension burden, but it will take two decades for the next generation of retirees to depart the service and the impacts of the reforms to be clear.
Long before a retirement benefit was the norm in the private sector, the military offered a simple deal: work for twenty years or longer on active duty and in return receive half of your base salary after you retire, for as long as you live. That’s what came to be called a defined benefit, and it was the norm in private-sector firms in the 1950s. However, the private sector shifted away from unsustainable DB plans starting in the 1980s, partly because of cost and partly because they lock in mediocre employees while nudging out entrepreneurial employees. More common today is a DC asset like the 401(k) that is owned by individual workers.
The MCRMC final report acknowledged a widespread consensus: “Services, as well as service members, would benefit from additional flexibility in the compensation system.” However, its blended proposal shifts the pension from being 100 percent DB to a blended plan that has a DB/DC ratio of 80/20. And the underlying DB structure remains the same, with no change at all to the timing of the vesting cliff. There are four structural elements that could be adjusted: the multiplier, vesting year, benefits initiation, and asset ownership. The MCRMC alternative leaves three core structures unchanged, and simply tweaks the multiplier (from 2.5 to 2 percent).
Importantly, MCRMC does not consider discontinuing the use of the DB pension. Instead, it asserts that the “current DB plan should be maintained because of its strong retention-pull effect on the Services’ force profiles.” This claim is not substantiated with references to research, scholarly or otherwise, merely a passing mention that it is the Pentagon’s preference. Retention-pull of the current compensation structure is a core problem. To be sure, a DB pension is a magnet for some personnel, but which personnel? For undifferentiated labor, it doesn’t matter, but for talent it is critical. Logic suggests that risk-averse officers are more attracted to a pension, and innovative officers are more attracted to a flexible, permeable force structure.
Consider the case of a highly valuable mid-career officer with expertise in cyber warfare. She is at the ten-year mark of her uniformed military service with a computer science degree, a passion for hacking, extensive military training, and valuable experience. According to the MCRMC’s calculus, this officer’s current DB package will be worth $712,000 if she stays an additional decade, which is about $70,000 per year. That amount is not a large incentive for a cyber-skilled officer. The point is not that such highly talented officers need more dollars to incentivize their service, because money is not the object. Career control and flexibility are the key incentives that talented individuals want, but will never have, with the existing DB superstructure. A larger point is that efficient force-shaping cannot be achieved in the modern era using twenty-year cliff vesting.
Rather than offer a set of alternatives—with analysis of trade-offs among them—the MCRMC offered one. To understand a fuller set of options and why they are necessary, it is helpful to consider some of the useful background published in the MCRMC report, as well as its Interim Report published in June 2014:
The Value of Work
How much is the military’s DB pension worth at the twenty-year mark? For the typical NCO, the discounted net present value is $201,282, according to MCRMC estimates.123 The actual cost to the United States government would be significantly more, probably twice as much, but the perceived value for an individual who faces an uncertain future discounts the future dollars; hence this estimate is what the typical retiree would trade for if it were a lump sum. For a typical officer (lieutenant colonel/O-5) at the twenty-year point, MCRMC says the discounted net present value of the DB is closer to $712,000.
As the Gates Commission noted in its 1970 report, the twenty-year cliff pension distorts workplace incentives. A pension cliff of twenty years is illegal in a private-sector pension—three times longer than what is allowed—because it is so coercive as to be deemed abusive to the employee. Simple analysis reveals how strong and perverse the incentive is.
First, consider the value of work just before and after the retirement point. An NCO who enlisted at age eighteen can retire at the age of thirty-eight. More to the point, the newly retired NCO begins to draw a pension immediately. At the twenty-year mark, exactly, he qualifies for the DB pension, which is 50 percent of base monthly pay for the remainder of his life, whether he lives to thirty-nine or ninety-nine. Base monthly pay for an enlistee at the rank/pay grade E-8 at nineteen years of service is $4,878, which steps up to $5,009 at the twenty-year mark. At first blush, the value of work—that is, staying in uniform instead of retiring—is $5,009 per month. But two other factors must be considered.
Because the monthly pension payments would begin immediately upon retirement, his decision to not retire means forgone pension income. Working after twenty years on active duty is suddenly half as valuable as before: the net income benefit of working is half the gross amount in a full paycheck. The second, and more sizable, marginal impact of preretirement work is the increasing lifetime value of the DB pension for each additional month of work up to the twenty-year threshold. If an enlisted individual leaves the service after nineteen years and eleven months, he abandons a future stream of pension payments that could easily total a million dollars.124 Because individuals discount the value of future payments, the perceived value of that pension income stream to the individual recipient is far less—I calculate it to be worth $533,000 to the typical veteran using a discount rate of 5 percent per year. It is worth even less to the enlistee on the first day of basic training due to an additional two decades of discounting, even if he plans on serving twenty years.
Individuals have what is called a present-bias for income for at least three reasons, and this is why individuals can be analyzed with a different discount rate than organizations or large groups. First, individuals face a risky future, especially a mortality risk. Suppose a veteran dies in an accident just three years after retiring: there is no future income stream. A second reason is that individuals are impatient, which is not a criticism and is the reason people use loans to purchase big-ticket items like automobiles, appliances, and houses. A third reason is uncertainty regarding inflation and other unforeseeable events. Large organizations can hedge against uncertainty in a way that individuals cannot.
Consider the perspective of the value of work for a master sergeant in the Air Force near the end of his twentieth year. His monthly base pay is $5,009, but the value of working during that last month rather than quitting is much more. The expected value of the last month of work in the US military before the twenty-year cliff is $538,284—base pay plus the entire expected value of the retirement package. Then the very next month’s paycheck has a net marginal value of just over $2,504: his base pay minus the forgone retirement payment, which is half that amount. The incentive effects of these income differentials must be large.
Looking at the retirement cliff, the mid-career NCO calculates the average value of each month’s work before and after retirement with the pension as the primary decision factor. With ten months until retirement, each month is worth about one-tenth of its retirement value, or $50,000. With five months to go, each month is worth twice that. The average monthly value escalates until the twenty-year vesting threshold is reached.
At the threshold of twenty years in uniform, for everyone on active duty, the economic value of work collapses. This probably explains why over half of the retirees from every branch of service choose that moment to retire.
In my analysis here, I use some strong simplifying assumptions to model the cost of the DB pension versus the value to the retiree. I assume an average lifespan of eighty years. I assume inflation is zero and cost-of-living adjustments are zero. I assume that the government earns zero interest when it sets aside dollars to pay future pensions. Most importantly, I assume that individuals have a discount rate of 5 percent more than the government’s because the government faces that much less risk and uncertainty. Most scholars treat individual discount rates as equivalent to institutional discount rates, so my assumption of a discount gap is driving a large difference in the perceived value that veterans see in their retirement pay and how the Pentagon sees it.
This means that future income streams are discounted at a rate of 5 percent each year, with projected income streams in future decades worth less than the stream in the present. For example, Charlie would value 95 cents today equally with a promise to pay Charlie 100 cents a year from today. By using a different discount, I am suggesting that the cost of the government offering an annuity is much more than the benefit an individual derives from getting one.
Monthly retirement pay for the current retiring officer at twenty years is $49,874 (this is half of the average of the highest three years of pay). Assuming a life expectancy of eighty-four years and a retirement age of forty-one years, I calculated the discounted value of that income stream at the moment of retirement to be $925,163. The cost to the government is, I assume, not discounted and therefore stands at $2.14 million.
To understand the powerful incentive of the military’s lifetime defined benefit, I calculated the value of a year’s work for the typical officer over twenty-four years of a career. Figure 6.3 shows how this value grows over time, in which the value of serving each year is that year’s base pay plus the total value of a DB pension divided by the years remaining until vesting. For example, two years before retirement, the total value of working that year is $520,000 ($100,668 base pay plus half of the discounted lifetime value of the pension, which at that moment is $839,150). One year before retirement, a year’s work is worth just under $1 million to the typical American officer.
FIGURE 6.3. Officer’s Value of Work by YOS (pay plus pension)
Source: Author’s calculations
There is a collapse of work value at year twenty-one. Monthly base pay is suddenly half as valuable as before—the net income of working is half the gross amount in a full paycheck. The incentive effects almost certainly explain the twenty-year retirement bubble. It is also notable how quickly the value of each additional year of work rises, assuming promotion to O-6 at year twenty-two, and this is because pay rises dramatically (and consequently so does the pension value) with years of tenure between twenty and twenty-eight years of service, regardless of role, responsibility, or rank.
This raises the question of whether the new blended system will smooth out the twenty-year value of work, and the answer is no (see figures 6.4 and 6.5, based on calculations I made for this book). The new blended retirement program does not change the shape of the work value ramp at all, only its peak, which will be 20 percent lower because of the smaller DB and only slightly higher because of the new asset component (adding 5 percent of matched base pay).
In the final analysis, there are two critiques of the current military pension. One is that the perceived value of the pension to individual veterans is far lower than its cost to the government. I made a number of assumptions to make that critique, but the one that really matters is the gap in discount rates. Even if you do not buy that argument, the second critique is that the twenty-year cliff severely distorts work incentives. There are no assumptions needed, because the shape of the value of work over time is the same, regardless of discount rates and cost-of-living adjustments.
TVF Alternatives
A well-designed pension would aim to optimize management of talent by flattening the annual value of work, primarily to avoid exponential rising threshold spikes in pensions or other forms of delayed compensation. This section introduces two alternative designs for a military pension that would better optimize permeability while lowering overall costs to the military.
FIGURE 6.4. Officer’s Value of Work by YOS (pay plus pension), Comparing Pension Plans
Source: Author’s calculations
FIGURE 6.5. NCO’s Value of Work by YOS (pay plus pension), Comparing Pension Plans
Source: Author’s calculations
The first alternative retains the DB structure, but shifts vesting to year ten and initial payment out to age fifty-five. The second alternative ends the use of DB payments in lieu of a full DC savings asset. Legislators and military leaders could avoid concerns about overly radical reforms by retaining the status quo as an option. Entering service members should be given a menu of the existing blended DB or one of the two TVF alternatives. In addition, service chiefs should be empowered to select one of the alternatives for their new accessions. There is no reason that the Navy should be required to offer the same retirement package as the Air Force. Each service has unique talent needs across multiple dimensions (age distribution, physical fitness, rank structure, and occupational skills, to name a few) that call for tailored compensation structures rather than a one-size-fits-all plan. Table 6.6 summarizes the key design features of each plan, as compared to the two plans in effect now.
The “TVF DB” is a modernized defined benefit, with a far less coercive vesting cliff of ten years, whereas the payout of benefits begins over a decade later at the age of fifty-five (whether the member is retired at that point or not). It pushes out the date pension payments begin for two reasons: to end coercive incentives and to save money. However, an earlier vesting YOS will incur major new costs for the younger service members who will qualify for it. In order to keep a lid on those costs, the services should change incentives only gradually. The multiplier need not be the same for every year served. TVF DB is modeled with an initial multiplier of 1 percent for years one to ten, 2 percent for years eleven to fifteen, and 3 percent for every year thereafter. The discounted value of the TVF DB pension for a twenty-year retiree is three-quarters of the status quo, even though the full cost to the government is slightly more than half of the status quo.
I also modeled the value and cost of a completely new kind of military pension along the lines called for by the Defense Business Board. This “TVF Save” savings plan is a DC-only plan. It would discontinue the DB plan completely for new officers and enlistees. Any full DC plan allows for greater flexibility in managing the force, allowing individuals to take off-ramps and on-ramps between active duty, reserve status, and the private sector.
The plan I modeled begins contributions during the third year and scales up employer matching from 1 percent to 20 percent during the first five years. A generous direct employer contribution that equals 25 percent of base pay per year is added to the personal savings account beginning in year six, plus a match of funds saved by the individual up to a maximum of 25 percent of base pay. If members save the maximum amount, the Pentagon would contribute 50 percent of base pay to the savings account monthly—far in excess of the private-sector norm. Yet it would cost the military far less than the status quo. Under this plan, service members create an asset that is valued by the twenty-year retiree at the same level as the status quo DB pension, meanwhile saving the government $1.3 million per retiree.
The TVF Save plan would incur additional costs as well, due to the number of service members who leave between six and twenty years and will receive a generous savings asset. It is impossible to model the net savings or cost military-wide, but overall costs of this plan must be considered in balance with increased permeability and therefore overall compensation that will fall to a lower equilibrium due to increased labor supply.
Value of work under the two TVF plans is radically different than the two status quo plans. Members will be likely to leave the service more frequently in the absence of twenty-year vesting. However, these new pension structures also enable a wide-scale continuum of service careers with frequent reentry of reservists and other veterans to active-duty status. The military stands to enhance its talent pool with continuum careers, and it will also generate a much thicker labor supply than it currently has—one that it can be selective in accessing.
Additional Background
The January 2015 final report from the MCRMC is not the first major government assessment to recommend changing the US military’s compensation structure. A few years after the AVF was in place, on March 13, 1978, the US Government Accountability Office (GAO) was calling for an end to the twenty-year cliff in a clearly titled report, Retirement Security: The 20-Year Military Retirement System Needs Reform:
Twenty-year retirement, in conjunction with present personnel management policies, is an inefficient means of attracting new members, causes the services to retain more members than are needed up to the 20-year point, provides too strong an incentive for experienced personnel to leave after serving 20 years, and makes it impossible for the vast majority of members to serve full careers.
This echoed the findings of the President’s Commission on Military Compensation (PCMC) in 1978, which, in the words of a recent Rand summary, “criticized the military retirement system on three counts: that it was inequitable because there were no retirement benefits for those not reaching 20 YOS; that it inhibited flexible force management because managers were reluctant to separate low performers as they neared 20 YOS (emphasis added); and that it was ineffective because it had little effect on recruiting and early retention, but an extremely strong effect on retention after 10 or 12 YOS.”
On November 15, 1996, the GAO issued a report titled Military Retirement: Possible Changes Merit Further Evaluation. From the summary:
The military retirement system strongly influences the broad shape of the force, since it provides an increasing incentive for service members to stay in the military as they approach 20 years of service and encourages them to leave thereafter, helping DOD to retain midcareer personnel and yielding a relatively young force; (6) however, the system can also impede effective force management because military personnel are not entitled to any retirement benefits unless they have served 20 years, and the services have been reluctant to involuntarily separate personnel with less than 20 years of service beyond a certain point due to the financial consequences for service members and the impact on morale; (7) some analysts, including several of GAO’s roundtable participants, believe the military retirement system is an obstacle to achieving the right force size and composition. (emphasis added)
In 2006, the Defense Advisory Committee on Military Compensation (DACMC) issued its report, which took aim at the status quo pension system as inefficient, inflexible, and inequitable. It proposed a hybrid plan, what many scholars call a “blended” plan, with DB and DC components. Its DB component would retain the 2.5 multiplier, vest earlier (year ten), but pay out at age sixty. Once again, the underlying motivation cited by DACMC was the need for flexibility in force management. In 2011, the Defense Business Board issued a report that offered the latest harsh critique and proposal for pension reform.
Indeed, before the AVF was implemented in 1973, the Gates Commission recommended a series of personnel reforms that it considered necessary. Its proposal called for an introduction of “vesting” over time in contrast to the binary all-or-nothing cliff at twenty years. The incentive effect of the pension is carefully discussed in the following passage, something lacking altogether in the hundreds of pages of the MCRMC’s final report. Here is the Gates Commission:
Because military retirement benefits are budgeted each year out of current funds, servicemen have never acquired vested retirement rights except those which arise after one has served long enough (19½ years) to be eligible to retire. This policy has a number of undesirable effects.…
Retirement benefits have the additional disadvantage of being worth too much to the individual who is beyond his tenth or eleventh year of service. He cannot afford to separate from the service because of the benefits for which he will qualify after another 9 or 10 years. Because of this potential loss, the military rarely discharges individuals who have served more than say, 10 years.
Also, the substantial retirement income available after 20 years of service (when some enlisted men are only 37 years old) induces many individuals to retire as soon as they are eligible. The combination of retirement income and civilian earnings is very attractive. Men who retire early are often those with superior civilian earning opportunities and they are precisely the individuals the services would like to retain longer. The Department of Defense group organized to study compensation recognized the shortcomings of the present retirement system.
Conclusion
The Total Volunteer Force would reform military compensation by transforming base pay and retirement. Reforms to the military pension would not affect the retirement payments and promises to veterans or troops currently on active duty. Instead, their impact would be limited to future service members with the simple exception being that service chiefs should have the authority to allow their service members the option to transition to one of the new options. As for transforming base pay, this chapter has shown that the long-standing pay tables are outdated. The roots of the tables date back to the beginning of the republic, but US military pay tables in the mid-nineteenth century had more occupational flexibility and less tenure rigidity than in 2016. A move to role-based pay is long overdue, and the ability to use assignment-pay incentives already exists in law.
This book makes the case that the US military offers many lessons, positive and negative, about organizational design. The final three chapters focus on three areas of talent management where the armed forces are self-graded as deficient: job-matching, performance evaluations, and compensation. Yet it would be an error to look to the private sector for best practices. Indeed, it was exactly this kind of hubris in the 1960s when Robert McNamara centralized HR policies at the Pentagon that created the problems that persist today. Unfortunately, the centralized rules and regulations have become cultural norms, and reforms to them are seen as taboos violating an ancient trust. Fortunately, the organizational assessment presented in chapter 1 revealed that the US armed forces are world-class in terms of most leadership metrics, particularly the values and sense of purpose that are unique as well as sacred.
Other firms would do well to study how the US military builds those bonds of trust, self-sacrifice, and voluntary service despite the red tape. It is the spirit of volunteerism inherent in this new generation of senior officers that is already breaking through the bureaucratic concrete. We can look to the world-class performance evaluations pioneered by the Marines or the peer ratings done by the Army Rangers to find hope. After a decade at war, American troops are coming home with little tolerance for regulatory barriers to excellence.