Overcompensating?

A Tool to Refine Policies for Retaining Military Personnel

By John Ausink

John Ausink is a policy analyst at RAND.

As U.S. military operations expand overseas from Afghanistan to Iraq, they place greater strains on U.S. personnel. Defense policymakers, who wish to retain qualified personnel, would benefit from a better understanding of the likely effects of new personnel compensation policies. Changes in policies relating to basic pay, retirement packages, targeted bonuses, reenlistment incentives, and voluntary termination incentives can produce unintended as well as intended results.

To estimate the likely effects of new compensation policies, we at the RAND Corporation have developed a computer software tool called the Compensation, Accessions, and Personnel Management (CAPM) system. Based in part on a version of the well-known Annualized Cost of Leaving econometric model, the CAPM system enables decisionmakers both to study the effects of policy changes on retention rates and personnel levels and to compare the outcomes of various policy packages.

Analysts have always used econometric models to help policymakers study the recruitment, supply, and retention of military personnel. The introduction of the All Volunteer Force in 1973 made the use of these models even more important. But the personnel data used in the models have often been separated from other quantitative tools used by policymakers to project the effects of changes in compensation policies.

To close this gap, the CAPM system integrates the personnel data with the policy projections. This combination allows policymakers to alter the personnel data, to make alternative policy choices, and to quickly compare the results of a wide variety of options.

Two Examples

There have been times in the past when the military services have expressed concern about meeting recruiting goals. One possible response to this problem would be to relax educational standards for recruits.

We used the CAPM system to explore the potential ramifications of such a policy. We began by taking U.S. Air Force data on recruits from fiscal year 2000 and projecting the likely future inventories of these enlisted personnel. We assumed that about 50 percent of the recruits in fiscal year 2000 were high-aptitude white males (as measured by their performance on the Armed Forces Qualification Test) and that 15 percent were low-aptitude white males. Leaving end-strength goals the same, we then projected alternate future inventories, assuming that the recruiting goal for high-aptitude white males had been reduced to 40 percent of total recruits and that the goal for low-aptitude white males had been raised to 25 percent.

The CAPM system showed the likely — and expected — result. Given that retention rates are generally lower for low-aptitude personnel than for high-aptitude personnel, the total number of recruits required in the short run to maintain the same end-strength in the long run had increased by the fourth year of the projection. Thus, an initial effort to make recruiting easier appeared to have the long-term effect of increasing the need for new recruits.

Another example further illustrates the utility of the CAPM system. In January 2001, the U.S. Department of Defense increased basic military pay across the board — partly to increase overall retention, but also partly to tie subsequent increases in pay more closely to promotions rather than to the length of time spent in any one pay grade. We used the CAPM system to compare the likely retention rates associated with the change, again limiting our analysis to air force personnel.

As expected, and as shown in the figure, retention rates generally increased as a result of the pay increase. What came as a surprise was that retention rates for those eligible for retirement after the 20th year of service decreased for some cohorts in the projection. In particular, the model showed that there would be a fairly significant decrease in retention for those who completed 22 years of service in fiscal year 2001.

Closer examination of the assumptions of the model provided an explanation. For those who entered military service before September 8, 1980, retirement pay is a multiple of the basic pay received in the final year of service. In contrast, for those who entered military service between September 8, 1980, and August 1, 1986, retirement pay is based on the average basic pay for the highest-paying 36 months of the person’s career (usually the last three years).

The model indicated that the increase in basic pay raised retirement pay enough to encourage more people who were still under the “old” retirement plan to leave the air force rather than to reenlist. The compensation changes generally had the opposite effect for those under the newer retirement system. We found this exercise to be a good example of how an unexpected model prediction could lead to a closer examination of the underlying assumptions, which could lead in turn to a better understanding of the motivation for changes in retention behavior.

A Coordinated Effort

When areas of policymaking authority such as recruiting, reenlistment, and retirement are closely related to each other, policies established in one area may have profound and sometimes unintended consequences in other areas. This is particularly true in the military, which has long planning horizons, vast amounts of data that can drive the decisionmaking process, and customarily short tours of duty for decisionmaking personnel.

Although the underlying econometric model of CAPM breaks no new analytical ground, the ease with which CAPM allows analysts and decisionmakers to study the probable effects of policy changes can help everyone involved better understand the potential consequences. CAPM is still a work in progress, but the model is an example of the capabilities a user should expect in such analytical tools, and its structure provides an excellent foundation upon which to build further improvements.

Related Reading

Background and Theory Behind the Compensation, Accessions, and Personnel Management (CAPM) Model, John Ausink, Jonathan Cave, Manuel Carrillo, RAND/MR-1667-AF/OSD, 2003, 58 pp., ISBN 0-8330-3428-6.

A New Tool for Compensation, Accessions, and Personnel Management in the U.S. Military, RAND/RB-100-AF, 2003, 1 p.

A Tutorial and Exercises for the Compensation, Accessions, and Personnel Management (CAPM) Model, John Ausink, Albert A. Robbert, RAND/MR-1669-AF/OSD, 2003, 62 pp., ISBN 0-8330-3430-8.

Users’ Guide for the Compensation, Accessions, and Personnel Management (CAPM) Model, John Ausink, Jonathan Cave, Thomas Manacapilli, Manuel Carrillo, RAND/MR-1668-AF/OSD, 2003, 84 pp., ISBN 0-8330-3429-4.