Military Attrition: DOD Needs to Follow Through on Actions Initiated to Reduce Early Separations

Published by the Government Accountability Office on 1999-02-24.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

                          United States General Accounting Office

GAO                       Testimony
                          Before the Subcommittee on Personnel, Committee on
                          Armed Services, U.S. Senate

For Release on Delivery
Expected at
2:00 p.m., EST
                          MILITARY ATTRITION
February 24, 1999

                          DOD Needs to Follow
                          Through on Actions
                          Initiated to Reduce Early
                          Statement by Mark E. Gebicke, Director, Military
                          Operations and Capabilities Issues, National Security and
                          International Affairs Division

GAO/T-NSIAD/-99-80            omcw)       1
Mr. Chairman and Members of the Subcommittee:

We are pleased to be here today to discuss our work on the recruiting and
attrition of the military services' enlisted personnel. At the request of this
Subcommittee, we began a body of work in this area in 1995. As you are aware,
enlisted attrition is a complex problem, and many processes and people are
involved in recruiting and retaining qualified enlisted personnel. Because of the
complexity of this issue, we segmented the enlistment process into its
component parts and issued separate reports on (1) recruiter selection and
incentive systems and their effects on recruiter performance,' (2)the screening of
incoming recruits for criminal history information to ensure that only qualified
persons are allowed to enlist,2 (3) reasons for attrition during the first 6 months of
an enlistee's term, 3 and (4) reasons for premature attrition after first-term
enlistees have completed 6 months of service.4 Other products related to these
issues are listed in appendix I.
Today, I would like to present a summary of our work to date and the status of
the actions initiated by the Department of Defense (DOD) to improve recruiter
and recruit quality and to reduce attrition. Keeping qualified enlisted personnel in
the military has become even more crucial as the services' struggles to meet
their recruiting goals have become even more intense in recent years.
RESULTS IN BRIEF                                              -. -

For at least the last decade, about one-third of all DOD enlistees have failed to
complete their initial terms of service. This attrition rate represents a costly
problem. We calculate that the services spent $1.3 billion to recruit and train the
72,670 enlistees who entered the services in fiscal year 1993 but were separated
before the end of their first term.5 Clearly, the services did not receive a full
return on their investment.

Our work since 1995 has shown that to decrease this attrition rate, the services
needed to (1)revise their recruiter selection and award systems to create more
incentives to recruit enlistees who are likely to complete basic training; (2)
improve the medical, physical, and criminal screening of incoming recruits to
ensure that only those fully qualified are enlisted; and (3)gather more

   Military Recruiting: DOD Could Improve Its Recruiter Selection and Incentive Systems
(GAO/NSIAD-98-58, Jan. 30, 1998).
2 Military Recruiting: New Initiatives Could Improve Criminal History Screening (GAO/NSIAD-99-

53, Feb. 23,1999).
3 Military Attrition: DOD Could Save Millions by Better Screening Enlisted Personnel

(GAO/NSIAD-97-39, Jan. 6, 1997).
4 Military Attrition: Better Data. Coupled With Policy Changes, Could Help the Services Reduce
Early Separations (GAO/NSIAD-98-213, Sept. 15, 1998).
s Fiscal year 1993 was the most recent accession year for which a full 4 years of attrition history
was available.
comprehensive data on why enlistees are being separated early to allow DOD
and the services to craft effective retention policies. Our work has resulted in 20
recommendations. However, because of the complexity of the attrition problem,
it may take several years before the effects of DOD's actions can be seen.
Consequently, we may not see a significant lowering of attrition for a number of

 In response to our recommendations, DOD and the services have already
initiated many actions to reduce attrition. The services are beginning to tie
recruiter awards more closely to recruits' successful completion of basic training
and to use better tools to select new recruiters. To improve the screening of
applicants for service, DOD has revised its medical history forms, and the
services have worked toward making their preentry physical fitness programs
more rigorous. DOD has also agreed to strengthen criminal background
screening by requiring full fingerprint checks and other technological
improvements. However, for all of these actions to result in a reduction of
enlisted attrition, DOD and the services must vigilantly follow through on these
actions and assess their effects. This will take continued monitoring and
One of DOD's major efforts will take a long time to complete: the creation of a
database on reasons for premature separations. We initially made this
recommendation in 1997. Inthe meantime, DOD has made good progress
toward creating databases on medical separations-and on enlistment waivers.
Also, DOD has just completed phase one of its effort to improve separation
codes. It has now standardized separation codes and developed common
definitions of these codes. DOD's next step is to standardize the application of
these codes among all the DOD agencies that use them. This effort is expected
to be completed in 18 months. Cireating accurate data on why enlistees are
leaving before the end of their first terms will enable the services to craft
successful policies to retain enlistees and allow DOD to set reasonable goals for
reducing attrition.

DOD faces a significant challenge in retaining the hundreds of thousands of new
recruits it enlists each year. While each new enlistee signs a contract for a term
ranging from 2 to 6 years, most first-term contracts are for 4 years. Between
fiscal years 1982 and 1993, 31.7 percent of all enlistees did not complete their
first terms of service: 11 percent of enlistees were separated during their first 6
months, and 20.7 percent between their 7t and 48 th month.
First-term attrition is costly in that IDOD estimated in 1998 that the services'
recruiting and training investment in each enlistee during the first term was an
average of $35,532. Using DOD's lower recruiting and training cost estimates for
fiscal year 1993 (ranging from $19,143 to $24,885), we calculated that the
services spent $1.3 billion on the 72,670 enlistees who entered the services in

fiscal year 1993 and departed prematurely. Because these enlistees were
separated early, the services did not get a full return on their investment.

Recruiters are the first of many military personnel who are involved in the
transformation of civilian recruits into productive servicemembers. Our work
showed, however, that recruiters did not have adequate incentives to ensure that
their recruits were qualified. Their grueling schedules and heavy workloads
exerted great pressure on them. We believe that the services should not
measure recruiters' success simply by the number of people who sign enlistment
papers stating their intention to join a military service but also by the number of
new recruits who go on to complete basic training. We also believe that the
selection of recruiters themselves is critical.

During our review of this issue, we identified practices in each service that we
believed would enhance recruiters' performance and could be expanded to other
services. Specifically, we recommended that the services (1)link recruiter
awards more closely to recruits' successful completion of basic training, (2) use
experienced field recruiters to personally interview all potential recruiters, (3)
develop or procure personality screening tests that could aid in the selection of
recruiters, and (4) provide opportunities for recruiters to interact with drill
instructors and separating recruits to allow them to-become more familiar with
what new recruits typically undergo.

Data maintained by the Defense Manpower Data Center on enlistees who
entered the services in fiscal year 1994 indicates that DOD's attrition rate was
14.4 percent at the 6-month point intheir first terms. 6 This means that in fiscal
year 1994, 25,430 enlisted personnel were separated from the services within
the first 6 months of service. Of this number 21,229, or about 83 percent, were
assigned separation codes indicating that they (1)were medically unqualified for
military service, (2) demonstrated character or behavior disorders, (3)
fraudulently or erroneously entered the military, or (4)failed to meet minimum
performance criteria. This data, along with our body of work, indicates that better
medical, physical, and criminal screening of recruits could result in a reduction of
attrition rates.

61n  past years, 6-month attrition has steadily risen, from 10.8 percent in fiscal year 1987 to 14.4
percent in fiscal year 1994. However, overall 48-month attrition has also risen, from 31.7 percent
in fiscal year 1987 to 35.8 percent in fiscal year 1993 (the latest year for which 4-year data was

Ways to Improve the Medical
Screening of Recruits

Our work on attrition during basic training indicated that the services were
enlisting persons with disqualifying medical conditions for two primary reasons:
(1) applicants conceal their medical histories, and (2) the services waive medical
conditions that, according to DOD directives, are disqualifying.

At the time of our review, the services asked applicants to provide medical
 records only when they divulged past medical problems. Also, the services did
 not ask all applicants to provide the names of their medical insurers or medical
providers. Further, questions on the forms used to collect applicants' medical
histories were nonspecific, ambiguous, and not tied to medical conditions that
most often resulted in recruits' separations. Finally, DOD did not have data on
which disqualifying medical conditions represented acceptable attrition risks.
That is, because DOD's physical enlistment standards were not empirically linked
to performance in the military, but rather were based on military experience and
expert judgment, personnel throughout the services found it difficult to determine
which medical conditions should be waived and which ones should not.

Our recommendations for improving the services' medical screening included
requiring all applicants to provide the names of their medical insurers and
providers, revising the forms used to collect applicants' medical histories, and
using a newly proposed database of medical diagnostic codes to determine
whether adding medical screening tests to the entrance examinations and/or
providing more thorough medical examinations to selected groups of applicants
could cost-effectively reduce attrition at basic training.
Ways to Improve the Physical
Fitness of Recruits
While recruiters use standard criteria in screening applicants for military service,
physical fitness is not among these criteria. Thus, the services have no
assurance that recruits will be able to pass their physical fitness tests in basic
training. Infact, of the 176,000 recruits entering the services in fiscal year 1994,
approximately 7,200 (or 4 percent) were discharged in the first 6 months of
service because they failed to meet minimum performance critera such as
physical fitness requirements. According to basic training personnel throughout
the services, recruits who are in good physical shape have a greater chance of
meeting overall military performance standards. Those struggling to meet
physical requirements are often correspondingly demotivated to meet other
military requirements.
As we began our work, the services were beginning to invigorate their Delayed
Entry Programs, the structured programs supervised by recruiters before their
recruits begin active duty. For example, the services were encouraging recruits

to maintain or improve their physical fitness so that they would be able to meet
the initial physical conditioning requirements of basic training. However, at the
time of our review, only the Marine Corps conducted regular physical fitness
training for its recruits and required them to take a physical fitness test while in
the Delayed Entry Program. The Marine Corps reported that attrition was lower
among recruits who passed the test.

To maintain recruit quality and increase a recruit's chances of graduating from
basic training, we recommended that the Army, the Navy, and the Air Force
implement the Marine Corps' practice of administering a physical fitness test to
recruits before they report to basic training. We also recommended that the
services incorporate more physical fitness training into their Delayed Entry
Ways to Improve the Screening
of Recruits for Criminal Histories

Another criterion for screening applicants for military service is moral character.
A criminal history does not automatically eliminate applicants from consideration.
Rather, the services may grant "moral character waivers" to enlisted persons with
minor or serious criminal histories based on the information they receive. For
example, between fiscal year 1990 and 1997, 13 percent of all enlistees had
received moral character waivers for offenses ranging from minor traffic
violations to serious crimes, and moral character waivers represented 62 percent
of all waivers granted (including, for example, waivers for physical and medical
standards and number of dependents). The services strive to gather the most
complete information possible on their applicants' criminal histories in order to
make informed decisions on whom they wish to enlist.
In response to concerns about the moral character qualifications of enlisted
personnel, we undertook a review to (1) determine the extent to which relevant
criminal history information on potential enlistees is available to the military
services and (2) identify any federal government initiatives that could improve the
process of obtaining criminal history information. During our review, we found
that the services have extensive procedures for encouraging applicants to self-
report criminal history information. Among other things, the services repeatedly
query each applicant, providing as many as 14 opportunities to disclose any
criminal offenses to as many as seven different service and Military Entrance
Processing Station (MEPS) officials. The services also conduct periodic
inspections and investigations to ensure the integrity of the entire recruiting
process, which includes disclosure of disqualifying information.
The services cannot always obtain or substantiate all available criminal history
information because of certain service policies and federal, state, and local laws
and policies that sometimes preclude access. First, the services do not use
fingerprints to substantiate the majority of enlistees' criminal histories. Without

full fingerprint searches, the services cannot detect undisclosed aliases and
ensure that they have all available criminal history information. Second, federal
law and state and local laws and policies, which generally limit or prohibit
disclosure of criminal history information, impede the recruiting community's
access to certain criminal history information. Also, state and local governments
sometimes charge fees or require fingerprints to release the information. Third,
available criminal history databases (not controlled by DOD) are incomplete. For
example, only 46 percent of the Federal Bureau of Investigation's records are

Of further concern is the services' practice of sending enlistees to basic and
follow-on training and in some cases to first-duty stations before the results of
criminal record checks are received. This practice results in training costs that
could be avoided.

Several DOD and Department of Justice initiatives have begun that could
improve the process of obtaining criminal history information. These initiatives
have the potential of making available to DOD and the services more complete
information upon which to make moral waiver decisions and expedite the process
for obtaining record checks. However, DOD and the services have not
formulated a coordinated approach for using these initiatives to better ensure that
the military does not enlist and train individuals with undesirable backgrounds.
To enable DOD to benefit from the most complete-criminal history information
possible on all applicants for military service, we recommended that it (1) develop
and monitor a DOD-wide plan to use the recent initiatives that have been
undertaken by the Defense Security Service and the Federal Bureau of
Investigation to improve background investigations; (2) require all national
agency checks for enlistment into the military services to be based on a full
fingerprint search; and (3) direct the services, after the new initiatives are in use,
to stop sending enlistees to training and to first-duty stations without having all
available criminal history information.

After enlistees have been recruited, screened, and trained, many still fail to
complete their contractual terms of service. In our work on attrition, we found
that the separation codes assigned to enlistees, which form the primary basis for
DOD's tracking of attrition rates, captured only general categories of discharge.
That is, these codes are not specific enough to capture exactly why separations
are occurring and are used inconsistently by the services. The assignment of
separation codes requires a degree of subjectivity, which may mask true reasons
for separation and make it more difficult to analyze why attrition is occurring and
to determine what can be done to decrease it.

Over 70 percent of male enlistees who entered the services in fiscal year 1993
and were separated between their 7th and 48'h month of service had misconduct,
medical/physical conditions, performance problems, or drug use recorded as the
reason for separation. Over 71 percent of the women in this group were
separated for pregnancy, medical/physical problems, misconduct, performance
problems, or parenthood. While this data, which is based on separation codes,
gives general categories of discharge, it is not specific enough for the services to
determine which of these enlistees might have been rehabilitated or otherwise
encouraged to complete their first terms of enlistment.
Similarly, data now available on quality-of-life issues that underlie attrition--such
as enlistees' perceptions that military benefits are eroding, that career
opportunities are limited, and that deployment rates are too high--are not at
present tied to service efforts to prevent the attrition of first-term personnel. If the
services could use quality-of-life surveys to collect data from first-term enlistees
who separate early, they could focus their attention on improvements that would
have the most impact on reducing the attrition of these personnel.

All the services are concerned about attrition, and service leaders are conveying
this concern to their local commanders. Two services, the Army and the Air
Force, have even set numerical goals for reducing first-term attrition. Research
supports the positive effect of command emphasis on finding ways to get more
high-quality personnel to complete their first terms. .However, setting numerical
goals for reducing attrition without complete information on its underlying causes
or guidance on what specific actions should be taken to reduce it may turn these
goals into arbitrary ceilings.
While command emphasis on attrition is critical to the services' efforts to reduce
it, this emphasis must be linked to clear policy changes that target specific
groups of enlistees the services wish to retain without lowering standards. Better
guidance to commanders on what actions should be taken to deal with identified
problems or what accommodations could be made to retain certain categories of
enlistees is also needed. An example of a successful policy change to reduce
attrition is the Army's recent decision to retain enlistees who in past years would
have been separated early because they had been deemed ineligible for
reenlistment. In this case, Army commanders targeted a group of enlistees they
wished to retain-primarily those with performance problems-and made formal
policy changes to do so. The Army will now be able to measure the effects of its
policy change in terms of how many more enlistees it was able to retain and why.
Our interviews with first-term enlistees, supervisors, and service officials indicate
that other types of enlistees could be targeted for remedial action if specific

mitigating actions are taken. For example, enlistees who commit minor
disciplinary infractions, who fail physical fitness or career development tests, who
are one-time drug users, or who become pregnant may simply need to be
provided further counseling, optional testing, other job choices within the service,
or remedial training by their commissioned or noncommissioned officers.
 Finally, granting honorable discharges to enlistees who deliberately seek ways
 out of fulfilling their service commitments simply encourages others to do
 likewise. One Army unit we visited had already begun to attempt to close these
"escape routes" and impose more punitive measures against certain enlistees,
particularly those found to use drugs. Some enlistees could be motivated to
remain in the service if they knew that there were no easy ways out and that
serious negative consequences were associated with behavior or performance
that warranted discharge.

Among our many recommendations to help reduce enlisted attrition, we
suggested that DOD issue guidance for implementing the assignment of
separation codes that will provide it with a reliable database for managing
attrition and for the services to set appropriate targets for reducing it. We also
recommended that the services (1) use existing quality-of-life surveys or create
new ones to identify quality-of-life initiatives aimed at reducing attrition; (2) collect
more complete data on specific groups of enlistees whom the services wish to
target for remedial action and issue guidance and formal policy changes to local
commanders indicating what specific actions-such as more counseling, optional
testing, further job choices, or remedial training-can be taken to prevent the
early discharge of enlistees in these groups; and (3) reassess the
appropriateness of providing favorable types of discharge to enlistees whose
behavior or performance led to their early separation to ensure that proper
incentives exist to encourage enlistees to complete their first terms.

Recruiting and retaining well-qualified personnel are among the strategic goals
included in DOD's performance plan for fiscal year 2000 required under the
Government Performance and Results Act of 1993 (P.L. 103-62, Aug. 3, 1993).
In line with these goals and in response to recommendations contained in our
reports on recruiting and attrition, DOD and the services have initiated many
actions. While we see these actions as positive steps, it is likely to be some time
before a drop in enlisted attrition rates will be seen. The effect of each particular
action may not be precisely measured, but it is important that DOD and the
services follow through on their actions and monitor their effects. These changes
represent an opportunity for DOE) and the services to continually improve their
recruiting and training processes and retain more qualified enlisted personnel.

As of February 1999, DOD reported the following changes, among others, in
response to our work in the recruiting area:

   -   The services were beginning efforts to link recruiters' awards more closely
       to enlistees' successful completion of basic training. For example, the
       Army Recruiting Command has proposed a new recruiter incentive
       program called the "Basic Training Pride in Ownership Incentive Program
       for Fiscal Year 1999," which will link a new soldier's successful completion
       of basic training to selection criteria for quarterly, semiannual, and annual
       awards to recruiters. The Navy's recruiting teams receive 10 percent of
       their points when applicants enter the Delayed Entry Program, 40 percent
       when they are sent to basic training, and 50 percent when they graduate
       from basic training.

   -   All services but the Army now use experienced recruiters to personally
       interview all potential recruiters. The Army uses a team of recruiters to
       interview volunteer recruiters, but for nonvolunteers, the Army uses a two-
       person team, one of whom is a recruiter, to assist in the nomination and
       screening of potential recruiters.

   -   The Army is testing an instrument to predict success as a recruiter. It will
       complete testing and present the results on the use of this "Recruiter
       Selection Profile" by March 1999.

   -   The Air Force is testing an instrument called the "Emotional Quotient
       Inventory" to screen potential recruiters and plans to report on its success
       at the end of the testing period.

   -   All services were working toward closer associations between recruiter
       schools and their basic training organizations.

   -   All services were working to develop more rigorous and structured
       physical fitness programs for enlistees while they are inthe Delayed Entry

   -   In response to our reports on attrition, DOD has initiated the following
       actions, among others:

   -   The Military Entrance Processing Command is formulating procedures to
       comply with the requirement for all applicants to provide the names of their
       medical providers.

   -   A new form to collect applicants' medical histories has just been
       completed. DOD states that this new form will allow medical practitioners
       to obtain precise medical histories that will be reviewed by Military
       Entrance Processing Station physicians in determining applicants' medical

        eligibility. This process will also allow MEPS physicians to request
        additional medical records or information on applicants prior to their
        scheduled physical examinations.

    -   DOD has revised DOD Directive 6130.3, "Physical Standards for
        Appointment, Enlistment and Induction" to include International
        Classification of Disease codes to note any medically disqualifying
        conditions. DOD has asked that the services use these codes to identify
        all medical waivers and separations. Once a database of these codes has
        been created on all enlistees, DOD will be able to determine whether
        attrition risks are higher for various types of medical conditions and
        whether it would be cost-effective to add more medical screening tests to
        preentrance physical examinations.

While DOD has not had time to implement recommendations contained in our
just-released report on the screening of recruits for criminal histories, it generally
concurred with the draft report's findings and recommendations. DOD plans to
form a working group to develop and monitor a DOD-wide plan to use recent
initiatives to improve background investigations. DOD also agreed to require full
fingerprint searches for all potential enlistees, but implementation will depend on
the availability of needed technology. In addition, DOD agreed to stop sending
enlistees to training and to first-duty stations before it has received all available
criminal history information. However, DOD said that before directing such a
change, it would need to develop a system to ensUre-the prompt availability of
information on criminal histories and to allow flexibility for exceptions when
delays occur.
In responding to our report on enlisted attrition between 7 and 48 months of
service, DOD agreed to (1) prepare a report on quality-of-life issues that could be
addressed to reduce attrition, (2) provide local commanders with guidance and
formal policy changes related to specific types of attrition the services target for
remedial action, (3) reassess the! appropriateness of providing favorable types of
discharges to enlistees whose behavior or performance led to their early
separation to ensure that proper incentives exist to encourage enlistees to
complete their first terms, and (4) prepare a report by October 1999 documenting
service initiatives related to our recommendations.


Though DOD and the services have taken many positive steps toward improving
their selection of recruiters and their screening of recruits, the creation of a
database of specific, consistent, and complete reasons for enlistees' premature
separations has not yet been completed. We first made this recommendation in
January 1997. Our follow-on work, in 1998, confirmed the need for better

information on why enlistees were not completing their first terms, and our work
on criminal background checks reemphasizes the need for better data on
enlistment waivers.

 DOD has begun to lay the foundations for creating a database on why enlistees
 are being separated medically, and it has begun efforts to solve the problem with
 its waiver data. In April 1998, it began its data collection effort into nonmedical
 separations. DOD has formed a joint working group consisting of representatives
 from the Office of the Secretary of Defense, the services, the Defense Finance
 and Accounting Services, and the Defense Manpower Data Center to develop
 common definitions and applications to the separation codes. The working group
 has just completed phase one of a two-phased project. DOD reports that the
working group has standardized the current separation codes and developed
common definitions across the services that are in line with current DOD policy
and guidance. Inthe next phase, the working group will standardize the
application of the codes among all DOD agencies that use them. The efforts are
expected to be completed inthe next 18 months. We believe that collecting such
data is key to the services' ability to craft policies that increase the proportion of
first-term personnel who complete their obligations.

Mr. Chairman, this concludes my prepared statement. We would be happy to
respond to any questions that you or other Members-of the Subcommittee may
APPENDIX I                                                            APPENDIX I

                          RELATED GAO PRODUCTS

Military Attrition: DOD Needs to Better Analyze Reasons for Separation and
Improve Recruiting Systems (GAO/T-NSIAD-98-117, Mar. 12, 1998).

Military Attrition: DOD Needs to Better Understand Reasons for Separation and
Improve Recruiting Systems (GAO/T-NSIAD-98-109, Mar. 4, 1998).

Military Attrition: Better Screening of Enlisted Personnel Could Save Millions of
Dollars (GAO/T-NSIAD-97-120, Mar. 13, 1997).

Military Attrition: Better Screening of Enlisted Personnel Could Save DOD
Millions of Dollars (GAO/T-NSIAD-97-102, Mar. 5, 1997).

Military Recruiting: More Innovative Approaches Needed (GAO/NSIAD-95-22,
Dec. 22,1994).