Military Personnel: Financial Cost and Loss of Critical Skills	 
Due to DOD's Homosexual Conduct Policy Cannot Be Completely	 
Estimated (23-FEB-05, GAO-05-299).				 
                                                                 
From the passage of the homosexual conduct policy statute, in	 
fiscal year 1994, through fiscal year 2003 the military services 
separated about 9,500 servicemembers for homosexual conduct. This
represents about 0.40 percent of the 2.37 million members	 
separated for all reasons during this period. Questions have been
raised about the costs of separating servicemembers for 	 
homosexual conduct. Also, in the post-September 11th environment,
there has been concern about the separation of servicemembers	 
with critical occupations or important foreign language skills	 
in, for example, Arabic. GAO was asked to determine (1) the	 
military services' annual financial costs from fiscal year 1994  
through fiscal year 2003 for certain activities associated with  
administering the Department of Defense's (DOD) policy on	 
homosexual conduct--e.g., the recruitment and training of	 
servicemembers to replace those separated under the homosexual	 
conduct statute--and (2) the extent to which the policy has	 
resulted in the separation of servicemembers with critical	 
occupations and important foreign language skills. GAO provided  
DOD with a draft of this report for comment, and DOD provided	 
additional information on separations for homosexual conduct	 
compared with other unprogrammed separations.			 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-05-299 					        
    ACCNO:   A18138						        
  TITLE:     Military Personnel: Financial Cost and Loss of Critical  
Skills Due to DOD's Homosexual Conduct Policy Cannot Be 	 
Completely Estimated						 
     DATE:   02/23/2005 
  SUBJECT:   Cost analysis					 
	     Data collection					 
	     Education or training costs			 
	     Foreign languages					 
	     Homosexuality					 
	     Losses						 
	     Military discharges				 
	     Military law					 
	     Military personnel 				 
	     Military policies					 
	     Military recruiting				 
	     Military training					 
	     Occupational surveys				 
	     Policy evaluation					 
	     Policies and procedures				 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-05-299

                 United States Government Accountability Office

                     GAO Report to Congressional Requesters

February 2005

MILITARY PERSONNEL

  Financial Costs and Loss of Critical Skills Due to DOD's Homosexual Conduct
                     Policy Cannot Be Completely Estimated

                                       a

GAO-05-299

[IMG]

February 2005

MILITARY PERSONNEL

Financial Costs and Loss of Critical Skills Due to DOD's Homosexual Conduct
Policy Cannot Be Completely Estimated

                                 What GAO Found

The total costs of DOD's homosexual conduct policy cannot be estimated
because DOD does not collect relevant cost data on inquiries and
investigations, counseling and pastoral care, separation functions, and
discharge reviews. However, DOD does collect data on recruitment and
training costs for the force overall. Using these data, GAO estimated
that, over the 10-year period, it could have cost DOD about $95 million in
constant fiscal year 2004 dollars to recruit replacements for
servicemembers separated under the policy. Also, the Navy, Air Force, and
Army estimated that the cost to train replacements for separated
servicemembers by occupation was approximately $48.8 million, $16.6
million, and $29.7 million, respectively.

Approximately 757 (8 percent) of the 9,488 servicemembers separated for
homosexual conduct held critical occupations, identified by DOD as those
occupations worthy of selective reenlistment bonuses. GAO analyzed and
selected the top 10 most critical occupations for each year from fiscal
year 1994 through fiscal year 2003. About 59 percent of the servicemembers
with critical occupations who were separated for homosexual conduct were
separated within 2.5 years of service. The typical military service
contract is for 4 years of service. Also, 322 (3 percent) of separated
servicemembers had some skills in an important foreign language such as
Arabic, Farsi, or Korean. A total of 98 servicemembers had completed
training in an important language at DOD's Defense Language Institute and
received a proficiency score; 63 percent of such servicemembers had
proficiency scores that were at or below the midpoint on DOD's language
proficiency scales for listening, reading, or speaking. Students can
graduate from the basic program with proficiencies somewhat below the
midpoint of this scale.

Number of Separations of Active Duty Servicemembers for Homosexual Conduct
by Fiscal Year and Military Service

Fiscal year Army Air Force Marines Navy Totala

                    1994        136    185           36    258            615 
                    1995        184    235           69    269            757 
                    1996        199    284           60    315            858 
                    1997        197    309           78    413            997 
                    1998        310    414           76    345          1,145 
                    1999        271    352           97    313          1,033 
                    2000        574    177       104       358          1,213 
                    2001        626    190       111       290          1,217 
                    2002        432    125       105       222            884 
                    2003        378    142           62    187            769 
                   Total   3,307      2,413      798      2,970         9,488 
                 Percent         35     25            8     31             99 

Sources: Defense Manpower Data Center (data); GAO (analysis).

aPercents do not equal 100 because of rounding. United States Government
Accountability Office

Contents

  Letter

Results in Brief
Background
Costs of Certain Activities Associated with DOD's Homosexual

Conduct Policy Can Be Estimated

Servicemembers with Critical Occupations and/or Important Language Skills
Have Been Separated for Homosexual Conduct

Agency Comments and Our Evaluation 1 3 5

12

16 23

Appendixes

    Appendix I: Scope and Methodology 25 Appendix II: Financial Cost Estimate
      Tables 29 Appendix III: Critical Occupation Data Tables 31 Appendix IV:
                                   Comments from the Department of Defense 42

Tables	Table 1: Table 2:

Table 3: Table 4:

Table 5: Table 6:

Table 7:

Table 8:

Table 9:

Number of Separations of Active DutyServicemembersfor
Homosexual Conduct by Fiscal Year and Military Service 8
Number of Servicemembers Separated for Homosexual
Conduct with Some Proficiency in an "Important Foreign
Language," Fiscal Years 1994 through 2003 21
Estimated Average Annual Recruiting Cost by Military
Service and DOD, Fiscal Years 1994 through 2003 29
Total Estimated Recruiting Costs to Replace Enlisted
Personnel Separated for Homosexual Conduct, Fiscal
Years 1994 through 2003 30
Individuals Separated for Homosexual Conduct during
Selected Intervals, Fiscal Years 1994 through 2003 31
Individuals with Critical Occupations Separated for
Homosexual Conduct during Selected Intervals, Fiscal
Years 1994 through 2003 32
Individuals with Intelligence-Related Occupations
Separated for Homosexual Conduct during Selected
Intervals, Fiscal Years 1994 through 2003 33
Individuals with Training in Important Languages
Separated for Homosexual Conduct during Selected
Intervals, Fiscal Years 1994 through 2003 34
Sample of Critical Occupations 35

                                    Contents

Table 10: Sample of Intelligence-Related Occupations 37

Table 11: Languages Spoken by and Proficiency Levels for Individuals
Separated for Homosexual Conduct from Fiscal Year 1994 through Fiscal Year
2003 Who Were Trained in a Language at the Defense Language Institute 39

Table 12: Languages Spoken by and Proficiency Levels for Individuals
Separated for Homosexual Conduct from Fiscal Year 1994 through Fiscal Year
2003, as Reported through Service Personnel Files 40

Figures	Figure 1: Figure 2: Figure 3: Figure 4: Figure 5:

Figure 6:

Figure 7:

Figure 8:

Separations for Homosexual Conduct by Race,
Fiscal Years 1994 through 2003
Separations for Homosexual Conduct by Gender, Fiscal
Years 1994 through 2003
Separations under DOD's Homosexual Conduct Policy by
Reason, Fiscal Years 1994 through 2003
Average Annual Recruiting Cost Estimate by Military
Service and DOD, Fiscal Years 1994 through 2003
Estimated Recruiting Costs to Replace Enlisted
Personnel Separated for Homosexual Conduct, Fiscal
Years 1994 through 2003
Distribution of the Amount of Time Served by Individuals
with Critical Occupations prior to Separation for
Homosexual Conduct, Fiscal Years 1994 through 2003
Distribution of the Amount of Time Served by Individuals
with Intelligence-Related Occupations prior to
Separation for Homosexual Conduct, Fiscal Years 1994
through 2003
Distribution of the Amount of Time Served by Individuals
Trained in Important Languages prior to Separation for
Homosexual Conduct, Fiscal Years 1994 through 2003

9 10 11 13

14

19

20 22

Abbreviations

DOD Department of Defense
FY fiscal year
GAO Government Accountability Office

Contents

This is a work of the U.S. government and is not subject to copyright
protection in the United States. It may be reproduced and distributed in
its entirety without further permission from GAO. However, because this
work may contain copyrighted images or other material, permission from the
copyright holder may be necessary if you wish to reproduce this material
separately.

A

United States Government Accountability Office Washington, D.C. 20548

February 23, 2005

Congressional Requesters

In 1993 Congress enacted a homosexual conduct policy statute which
declared that the "presence in the armed forces of persons who demonstrate
a propensity or intent to engage in homosexual acts would create an
unacceptable risk to the high standards of morale, good order and
discipline, and unit cohesion that are the essence of military
capability."1 During the 10 years following this declaration, the military
services separated about 9,500 servicemembers for homosexual conduct under
the statute. This represents about 0.40 percent of the 2.37 million
members separated for all reasons during this period. In the
post-September 11th environment, questions have been raised about the
financial costs associated with the Department of Defense's (DOD) policy
on homosexual conduct,2 especially in light of concerns about the shortage
of personnel with skills in critical occupations and foreign language
training.

You asked us to determine (1) the military services' annual financial
costs for certain activities associated with administering DOD's policy on
homosexual conduct-the recruitment and training of servicemembers to
replace those separated under the homosexual conduct statute, inquiries
and investigations of homosexuality cases, counseling and pastoral care
for affected individuals, separation functions, and discharge reviews-and
(2) the extent to which the policy has resulted in the separation of
servicemembers with critical occupations and important foreign language
skills.

To identify various types of costs associated with the policy on
homosexual conduct, we interviewed officials from a variety of DOD and
service offices, including the Office of the Under Secretary of Defense
for Personnel and Readiness, DOD's Office of Accession Policy; and offices
in the military services responsible for budget, criminal investigation,
chaplaincy, separation, and discharge review. The Air Force, Army, and

1 10 U.S.C. S: 654(a)(15).

2 The homosexual conduct policy statute is implemented through DOD
Directives 1332.14 (enlisted administrative separations); 1332.40
(separation of regular and reserve commissioned officers); and 1304.26,
which specifies qualification standards for enlistment, appointment, and
induction.

Navy provided data on training costs by occupation. While we requested the
same training-cost data inputs, each of the services used their own
methods to calculate the reported training-cost estimates.

To address the extent to which the homosexual conduct policy statute has
resulted in the separation of enlisted servicemembers with "critical"
occupations, we adopted the military services' definition of a "critical"
occupation as an occupation that was part of the selective reenlistment
bonus program. The selective reenlistment bonus program for enlisted
military personnel is DOD's primary tool for addressing short-term
retention problems in critical occupations by providing servicemembers who
reenlisted following the expiration of their service contracts with up to
$60,000.3 We collected and analyzed this information for fiscal years 1994
through 2003. Because intelligence occupations, as a group, have enduring
importance for the military that is independent from their periodic
inclusion in the selective reenlistment bonus program, we identified
servicemembers separated under the homosexual conduct policy statute who
had such occupations. We defined the knowledge of a foreign language as
"important" if it was related to (1) an occupation included in the
selective reenlistment bonus program or (2) a language identified by
combatant commanders and the Joint Staff as a deficiency in their periodic
readiness assessments. We also analyzed separated members' occupations and
foreign language skills by their length of service. The Defense Manpower
Data Center (Data Center) provided information on occupations, foreign
language skills, and the length of service of separated servicemembers.

The principal limitation of our analysis is that, for privacy reasons, we
did not review separated servicemembers' personnel records, including
training histories, which have implications for estimating training costs.
For example, from data provided by the Data Center, we matched separated
servicemembers to specific occupations, but we cannot state whether such
individuals completed all of the training associated with their
occupations. Much of our analysis depended on the quality of information
that the services provided the Data Center with and the steps that the
Data Center took to ensure the accuracy and completeness of the data.
According to Data Center officials, since 1998, the Data Center has made a

3 We last reported on selective reenlistment bonuses in GAO, DOD Needs
More Effective Controls to Assess the Progress of the Selective
Reenlistment Bonus Program, GAO-04-86 (Washington, D.C.: Nov. 13, 2003).

special effort to ensure that the services provide accurate information
about the number of servicemembers separated for homosexual conduct.

Although we did not validate the budget/financial systems used to produce
the cost estimates used in this report, we determined that the estimates
were sufficiently reliable for the purpose of this report. We assessed
reliability by (1) reviewing existing information about the data and the
systems that produced them and (2) interviewing agency officials
knowledgeable about the data and the manner in which they were
collected. We conducted our review from August 2004 through
February 2005 in accordance with generally accepted government auditing
standards. A detailed description of our scope and methodology is
presented in appendix I.

Results in Brief	The total costs of DOD's homosexual conduct policy cannot
be estimated because DOD does not collect relevant cost data on inquiries
and investigations, counseling and pastoral care, separation functions,
and discharge reviews. DOD does collect data on recruitment and training
costs for the force overall. Using these data, we estimated that it would
have cost DOD about $95 million in constant fiscal year 2004 dollars from
fiscal year 1994 through fiscal year 2003 to recruit replacements for
enlisted servicemembers separated for homosexual conduct.4 DOD does
calculate cost estimates related to recruiting enlisted personnel, which
we applied in broad terms, for servicemembers separated under the
homosexual conduct policy statute as a replacement cost. We calculated
that the estimated average annual cost to recruit an enlisted
servicemember over the 10-year period to be about $10,500.5 Most of the
services were able to estimate total training costs-recruit (or basic)
training and occupation-specific training.

4 We are not suggesting by this cost estimate that the services
specifically recruit one-for-one replacements of servicemembers who have
been separated for homosexual conduct.

5 This figure is in constant fiscal year 2004 dollars. DOD compiles the
basis of this cost estimate pursuant to DOD Instruction 1304.8 as part of
its military personnel procurement resources report to Congress. It is
constructed by averaging the DOD estimated recruiting costs for each year
over the period. The annual DOD recruiting cost figure is calculated as a
weighted average of the services' recruiting costs.

The estimated training costs for the occupations performed by Navy members
separated for homosexual conduct from fiscal year 1994 through fiscal year
2003 was about $48.8 million ($18,000 per member).6 The comparable Air
Force cost estimate was $16.6 million ($7,400 per member).7 The Army
estimated that the training cost of the occupations performed by Army
members separated for homosexual conduct over the 10-year period was about
$29.7 million ($6,400 per member).8 The Marine Corps was not able to
estimate occupation-related training costs. However, other types of costs
such as those related to inquiries and investigations of cases, counseling
and pastoral care, separation functions, and discharge reviews are not
estimable because DOD does not collect data necessary to develop such
estimates.

The military services separated 9,488 members9 pursuant to the homosexual
conduct policy statute from fiscal year 1994 through fiscal year 2003,
some of whom were in critical occupations or had important foreign
language skills. Seven hundred fifty-seven (about 8 percent) of these
separated servicemembers held critical occupations10 ("voice interceptor,"
"data processing technician," or "interpreter/translator"), as defined by
the services. About 59 percent of the members with critical occupations
who were separated for homosexual conduct were separated during their
first 2.5 years of service, which is about 1.5 years before the expiration
of the initial service contract of most enlistees. Such contracts are
typically for 4 years. Also, 322 members (about 3 percent) had some skills
in an important foreign language such

6 The per-member cost estimates in parentheses are a weighted average of
separated servicemembers' occupations for which we have data (for the
Navy, this is 2,706 of 2,970 members). The weighted average is computed by
multiplying the occupational training costs for each occupation by the
proportion of total students and summing the products. By doing this, the
occupations with the most students are weighted the most in computing the
average.

7 We have data for 2,241 of 2,413 Air Force members.

8 We have data for 3,339 of 3,348 Army members.

9 Of the 9,488 servicemembers considered in our analysis, 136 were
officers.

10 The occupations most frequently cited for selective reenlistment
bonuses are in appendix III.

as Arabic, Farsi, and Korean.11 A total of 98 members separated under the
homosexual conduct policy statute completed language training at the
Defense Language Institute and received a proficiency rating; 62 members,
or 63 percent, were at or below the midpoint on DOD's listening, reading,
or speaking proficiency scales.12

In commenting on a draft of this report, the Under Secretary of Defense
(Personnel and Readiness) provided information on separations for
homosexual conduct compared with other unprogrammed separations from
fiscal year 1994 through fiscal year 2003.

                                   Background

    Homosexuality and the Military

The prohibition against homosexual conduct is a long-standing element of
military law.13 But in January 1993, President Clinton sought to fulfill a
campaign promise to "lift the ban" on homosexuals serving in the military.
This led to the policy familiarly known as "don't ask, don't tell." In
exchange for the military services' silence ("don't ask") about a person's
homosexuality prior to induction, gay and lesbian servicemembers, as a
condition of continued service, would have to agree to silence ("don't
tell") about this aspect of their life. Failure to maintain silence can
result in

11 Servicemembers with critical occupations and important foreign language
skills are not necessarily mutually exclusive groups because some critical
occupations such as cryptologic linguists and interrogators require a
foreign language skill. Thus a servicemember could be included in both the
critical occupations and important foreign languages groups.

12 To assess language proficiencies, DOD uses an 11-point scale. DOD
describes the midpoint on this scale as "limited working proficiency
plus." According to the Defense Language Institute, students can graduate
from the basic program with proficiencies somewhat below the midpoint of
this scale. For foreign-language-related issues in the federal government,
see GAO, Foreign Languages: Human Capital Approach Needed to Correct
Staffing and Proficiency Shortfalls, GAO-02-375 (Washington, D.C.: Jan.
31, 2002). We stated in this report that in fiscal year 2001, the Army had
a 25 percent shortfall in cryptologic linguists and a 13 percent shortfall
in human intelligence collectors in several key languages taken as a
whole.

13 10 U.S.C. S: 654(a)(13).

separation from the military.14 In November 1993, Congress passed the
homosexual conduct policy statute and stated that the military's
suspension of questioning should remain in effect unless the Secretary of
Defense considers reinstatement of questioning necessary to effectuate the
policy set out in the statute.15 The statute also sets out the findings of
Congress in addition to the homosexual conduct policy. Included in the
findings section is a description of the differences between military and
civilian life, which forms a rationale for the institution of the policy.

Military life is fundamentally different from civilian life in that the
extraordinary responsibilities of the armed forces, the unique conditions
of military service, and critical role of unit cohesion, require that the
military community, while subject to civilian control, exist as a
specialized society [which] is characterized by its own laws, rules,
customs, and traditions, including numerous restrictions on personal
behavior, that would not be acceptable in civilian society.16

In short, Congress indicated that because of the unique nature of military
life, the military services may need to treat individuals who engage in
homosexual acts, as defined by the statute, differently than they would be
treated in civilian society.

    Separations for Homosexual Conduct during 1994-2003 Period

According to our analysis of the information provided by the Defense
Manpower Data Center, 9,488 servicemembers were separated for homosexual
conduct from fiscal year 1994 through fiscal year 2003.17 This figure
represents servicemembers who were on active duty at the time of their
separation, including members of the Reserves who were on active duty for
31 or more consecutive days. According to a Data Center official,

14 10 U.S.C. S: 654(b) and DOD Directive 1304.26, Qualification Standards
for Enlistment, Appointment and Induction (Mar. 4, 1994). For a discussion
of issues associated with the "don't ask, don't tell" policy, see
Congressional Research Service, Homosexuals and U.S. Military Policy:
Current Issues (Mar. 17, 1999).

15 Pub. L. No. 103-160, S: 571(b)-(d), (10 U.S.C S: 654, notes).

16 10 U.S.C. S: 654(a)(8).

17 In commenting on a draft of this report, the Under Secretary of Defense
(Personnel and Readiness) stated that 9,501 servicemembers were separated
for homosexual conduct from fiscal year 1994 through fiscal year 2003.
(See appendix IV.) According to the Servicemembers Legal Defense Network,
9,682 servicemembers were separated for homosexual conduct during the same
period. The Network reports information on these separations at
www.sldn.org.

118 reservists (other than those who served on active duty) were separated
for homosexual conduct from fiscal year 1993 through fiscal year 2003.
Because these separated reservists represent a small number of total
separations under the homosexual conduct policy statute, we did not
include them in our analysis. This exclusion is consistent with DOD's
reporting practice in this area, which reports only active duty personnel
separated for homosexual conduct. The figure also does not include
servicemembers who were in the Army National Guard, the Air National
Guard, or the Coast Guard. According to a Data Center official, the
official tracking of separations for homosexual conduct began in 1997 at
which time it was decided to include only the members of the Air Force,
Army, Marines, and Navy on active duty. The data also do not include
servicemembers who, for example, were separated for a "pattern of
misconduct," which could include several reasons for separation, including
homosexual conduct.

The Data Center also provided data on the characterization of service at
separation for service members separated for homosexual conduct from
fiscal year 1994 through fiscal year 2003. For "characterized" separations
(5,763 servicemembers), DOD granted "honorable" separations to 4,710
servicemembers (82 percent); "general (under honorable conditions)"
separations to 766 (13 percent); and "under other than honorable
conditions" separations to 287 servicemembers (5 percent). DOD also
granted "uncharacterized," or entry-level separations to 3,304
servicemembers who were separated for homosexual conduct during this
10-year period. The Data Center also classified as "bad conduct," the
separation of four servicemembers, which is a type of punitive separation
applicable to enlisted personnel only. (See Manual for Courts Martial,
Rule 1003(b)(8).) The Data Center did not have characterization-of-service
data for 417 servicemembers who were separated for homosexual conduct
during this 10-year period.

Table 1 and figures 1 and 2 show the number of separations by military
service, race, and gender, respectively, from fiscal year 1994 through
fiscal year 2003.

  Table 1: Number of Separations of Active Duty Servicemembers for Homosexual
                  Conduct by Fiscal Year and Military Service

     Fiscal year          Army   Air Force     Marines     Navy        Totala 
         1994              136      185               36    258           615 
         1995              184      235               69    269           757 
         1996              199      284               60    315           858 
         1997              197      309               78    413           997 
                           310      414               76    345         1,145 
                           271      352               97    313         1,033 
                           574      177              104    358         1,213 
                           626      190              111    290         1,217 
         2002              432      125              105    222           884 
         2003              378      142               62    187           769 
        Total            3,307     2,413             798   2,970        9,488 
       Percent              35            25           8    31     

           Sources: Defense Manpower Data Center (data); GAO (data).

                aPercents do not equal 100 because of rounding.

Figure 1: Separations for Homosexual Conduct by Race, Fiscal Years 1994
through 2003

1%

Other (133) Black (1,129)

Unknown (1,550)

White (6,676)

Sources: Defense Manpower Data Center (data); GAO (analysis).

Figure 2: Separations for Homosexual Conduct by Gender, Fiscal Years 1994
through 2003

                                 Female (2,586)

                                  Male (6,887)

         Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: Gender information was not available for 15 of the 9,488
servicemembers separated for homosexual conduct during this period.

The homosexual conduct policy statute states three reasons for separation,
namely, that a servicemember has (1) "engaged in, attempted to engage in,
or solicited another to engage in a homosexual act or acts...;" (2)
"stated that he or she is a homosexual or bisexual, or words to that
effect...;" or (3) "married or attempted to marry a person known to be of
the same biological sex." In addition, the statute provides mitigating
factors that may prevent separation in cases arising under the first two
categories.18 Figure 3 shows the distribution of separations by these
three reasons from fiscal year 1994 through fiscal year 2003.

18 10 U.S.C. S: 654(b).

Figure 3: Separations under DOD's Homosexual Conduct Policy by Reason,
Fiscal Years 1994 through 2003

In 1992 GAO reviewed DOD's policy on homosexuality, including the costs
associated with replacing personnel separated under the policy and the
cost of investigating allegations of homosexuality.19 We concluded that
"DOD does not maintain records of the costs associated with administering
its policy [on homosexuality]; nor does it record the costs of
investigating alleged cases of homosexuality. Accordingly, our analysis
was limited to estimates of the costs of recruiting and training
individuals to replace personnel discharged for homosexuality."

We also noted that the total cost of replacing personnel discharged for
homosexuality would need to include other factors such as out-processing
and court costs.

1%

Married or attempted to marry a person known to be of the same biological
sex (57)

Engaged in, attempted to engage in, or solicited another to engage in a
homosexual act or acts (1,520)

Stated that he or she is a homosexual, bisexual, or words to that effect
(7,900)

         Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: The figure displays information on 9,477-rather than all 9,488
servicemembers separated for homosexual conduct during the 10-year
period-because the statutory reason for separation was missing for 11
former servicemembers.

    Previous GAO Report on Costs Associated with DOD's Homosexual Conduct Policy

19 See GAO, Defense Force Management: DOD's Policy on Homosexuality,
GAO/NSIAD-92-98 (Washington, D.C.: June 12, 1992).

The cost data in this report and the 1992 report are not comparable
because, at the time of the 1992 review, we did not include the estimated
training costs for the occupations of servicemembers who were separated
for homosexual conduct.

  Costs of Certain Activities Associated with DOD's Homosexual Conduct Policy
  Can Be Estimated

Though the total costs associated with DOD's homosexual conduct policy
cannot be determined because neither DOD nor the services collect relevant
cost data, some costs can be estimated. For example, DOD does collect
estimates of the costs to recruit enlisted servicemembers, a portion of
which can be associated with DOD's homosexual conduct policy. In addition,
upon our request, the services were able to calculate the estimated costs
associated with the training of personnel by occupation. However, DOD was
unable to estimate the costs associated with other activities related to
DOD's homosexual conduct policy, namely, those related to investigations
and commanders' inquiries, counseling and pastoral care, and the
processing and review of separations.

    DOD Collects Data Related to Recruitment Costs

While not specific to individuals discharged for homosexual conduct or
other reasons, DOD does collect data related to the cost to recruit
servicemembers. Collected data related to DOD's annual average recruiting
cost estimate for enlisted servicemembers are shown in figure 4. Taken
together, available data show that the average annual recruiting cost
estimate for enlisted personnel from fiscal year 1994 through fiscal year
2003 was about $10,500 per member in constant fiscal year 2004 dollars.20

20 This figure is an average of DOD's reported cost per recruit. Each of
the services annually reports recruiting costs to DOD that are weighted by
the size of the force to determine an average cost per recruit. DOD's
reports on recruiting do not include the cost per recruit for officers and
medical personnel.

 Figure 4: Average Annual Recruiting Cost Estimate by Military Service and DOD,
                         Fiscal Years 1994 through 2003

Dollars

18,000

16,000

14,000

12,000

10,000

8,000

6,000

4,000

2,000

         1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Fiscal year

Air Force

Marine Corps
Navy
Army

                                  Source: DOD.

Note: All figures are in constant fiscal year 2004 dollars. Tabular data
related to cost in this and other figures are in appendix II.

The total estimated cost to recruit potential replacements for the 9,352
enlisted servicemembers separated under DOD's homosexual conduct policy
during the 10-year period21 was about $95 million in constant fiscal year
2004 dollars. (See table 4 in appendix II.) Estimated recruiting costs by
military service are shown in figure 5.

21 Of the 9,488 servicemembers considered in our analysis, 136 were
officers, and recruitment costs per officer were not available.

Figure 5: Estimated Recruiting Costs to Replace Enlisted Personnel
Separated for Homosexual Conduct, Fiscal Years 1994 through 2003

                          Dollars in thousands 16,000

                                     14,000

                                     12,000

                                     10,000

                                     8,000

                                     6,000

                                     4,000

                                    2,000 0

With the exception of the Marine Corps, the services were able to compute
cost estimates to train members, by occupation, upon our request. We asked
the military services to provide total and per-capita training-cost
estimates of the occupations performed by servicemembers who were
separated under the homosexual conduct policy statute for fiscal years
1994 through 2003. These figures include estimates of all training costs
related to selected occupations, including recruit training. The Navy
estimated that the total training cost for the 10-year period was $48.8
million and the estimated per-capita cost was about $18,000. The
comparable total estimated cost for the Air Force was $16.6 million, and

         1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Fiscal year

Air Force

Marine Corps
Navy
Army

                      Sources: DOD (data); GAO (analysis).

          Note: All figures are in constant fiscal year 2004 dollars.

    Most Military Services Can Compute Estimates of Costs to Train Personnel

the per-capita cost estimate was $7,400. The Army estimated that the
training cost for selected Army occupations for the 10-year period was
about $29.7 million. The estimated average training cost of these
occupations was about $6,400 per member.

    Other Types of Costs Associated with the Homosexual Conduct Policy Cannot Be
    Estimated

Investigations and Commanders' Inquiries

Counseling and Pastoral Care

Processing Separations from Military Service

We also examined the availability of other cost-estimate data associated
with homosexual conduct, including investigations and inquiries,
counseling and pastoral care, processing separations from military
service, and the review of such separations by service boards. For these
cost categories, we found that relevant data (for example, a system that
records the time spent on specific tasks for specific reasons) are not
collected, and, as a result, these types of costs cannot be estimated.

Investigative cost estimates were not available for our inquiry because
DOD law enforcement organizations do not generally investigate adult
private consensual sexual misconduct as a matter of investigative priority
and because of resource limitations. As the Navy notes in a policy
statement on this subject, "if there is no victim, there is virtually no
circumstance where the [criminal investigative service] will investigate
sexual misconduct." Sexual misconduct cases under these circumstances are
referred to commanders for appropriate disposition. And because commanders
do not record the time they spend on sexual misconduct inquiries, it is
not possible to estimate the cost of conducting them.

The estimated cost of counseling services, including pastoral care
provided through the chaplains corps, is also not determinable.
Servicemembers separated for homosexual conduct are not required to seek
counseling. Army and Navy chaplains, for example, record the types of
tasks they perform-religious ministry, outreach, or pastoral care-but they
are not required to compute the time they spend performing these
activities. Consequently, it is not possible to estimate the cost of
conducting such tasks. Furthermore, chaplains are not required to
differentiate "pastoral care" in their task reports by topics covered such
as homosexual conduct or sexual harassment.

The estimated cost of separating servicemembers also cannot be determined.
Separation procedures are handled by salaried employees who work in the
personnel offices of various military installations and who have multiple
responsibilities other than coordinating a servicemember's

separation from the military. They too do not compute their time spent on
the various activities they perform.

Review of Separations by Service Servicemembers who have been separated
for homosexual conduct have

Boards	occasionally requested service discharge review boards to review
whether their separations were properly granted. The estimated costs
associated with this activity also cannot be determined. Officials
associated with such boards told us that they are not required to compute
the estimated cost of reviewing servicemembers' requests and that they do
not record the number of reviews associated with DOD's homosexual conduct
policy. But service discharge review board officials were able to identify
for us at least 119 reviews associated with homosexual conduct (the Army,
72 reviews, fiscal years 1993-2003; Navy, 24 reviews, and Marines, 11
reviews, fiscal years 2000-2003; and Air Force, 12 reviews, fiscal years
2001-3). The service discharge boards conducted about 33,200 reviews
during these same time periods.

  Servicemembers with Critical Occupations and/or Important Language Skills Have
  Been Separated for Homosexual Conduct

From fiscal year 1994 through fiscal year 2003, the military services
separated members who had some training in critical occupations and/or
important foreign languages pursuant to the homosexual conduct policy
statute. Most servicemembers who had such occupations were separated
during their first 2.5 years of service. Also, DOD separated
servicemembers who had some language skills in Arabic, Chinese, Farsi, and
Korean. Relatively few of these separated servicemembers had proficiency
scores in listening to, reading, or speaking these four languages that
were above the midpoint on DOD's language proficiency scales, although
students can graduate from the basic program with proficiencies somewhat
below the midpoint of this scale.

    Most Separated Servicemembers Who Had Critical Occupations Were Separated
    during Their First 2.5 Years of Service

Servicemembers with critical occupations were separated for homosexual
conduct from fiscal year 1994 through fiscal year 2003. Examples of
critical occupations, as defined by the military services, include "voice
interceptor," "data processing technician," and "interpreter/translator."
The occupations most frequently cited as "critical," that is, eligible for
selective reenlistment bonuses are listed in appendix III. (See table 9.)
We found that 757 (about 8 percent) of the 9,488 servicemembers discharged
for homosexual conduct during this time period held critical occupations.

We determined the separation rate for these individuals at four time
intervals: recruit training, advanced individual training, and two 1-year
periods thereafter. The length of recruit training varies between the
services:

o  84 days in the Marine Corps,

o  63 days in the Army,

o  56 days in the Navy, and

o  42 days in the Air Force.

Overall, 1,747 (about 19 percent) of the 9,239 servicemembers separated
under the homosexual conduct policy statute were separated during recruit
training.22 An additional 1,037 servicemembers (about 11 percent) were
separated during advanced individual or occupation-related training.
Advanced individual training occurs after recruit training, and the length
of training varies widely by occupation. For the purpose of our analysis,
we considered advanced individual training as 100 days following recruit
training, which is about the average number of days for this type of
training. For example, for the Marines, this would mean between the 85th
and 185th day of service. Generally, 5,446 servicemembers (about 59
percent) were separated by the end of the 365-day period following
advanced training, or within about 1.5 years of service.

22 The Data Center has length-of-service data for 9,239 of the 9,488
servicemembers who were separated for homosexual conduct during the
10-year period.

Before new recruits are sent to recruit training, they are required to
take an enlistment oath and sign a contract to serve one of the military
services for a specified period of time, generally from 2 to 6 years and
typically for 4 years. Consequently, a separation within 1.5 years is well
before the end of a typical service contract for enlisted personnel. By
comparison, we reported in 1998 that for fiscal years 1982 through 1993,
about 32 percent of all enlistees were separated during their first term
of service: 11 percent of enlistees were separated during their first 6
months (versus about 30 percent of servicemembers who were separated for
homosexual conduct during their first 6 months) and about 21 percent of
all enlistees from their 7th through 48th month.23

Next, we analyzed the length of service for 755 servicemembers separated
for homosexual conduct who had critical occupations.24 The separation rate
for this group was lower than for the total population separated for
homosexual conduct. Generally, 267 servicemembers (about 35 percent) were
separated within about 1.5 years of service, and 443 servicemembers (about
59 percent) were separated within about 2.5 years of service. Figure 6
shows the separation rate of servicemembers who had critical occupations
by various time periods.

23 GAO, Military Attrition: Better Data, Coupled With Policy Changes,
Could Help the Services Reduce Early Separations, GAO/NSIAD-98-213
(Washington, D.C., Sept. 15, 1998).

24 The Data Center has length-of-service data for 755 of the 757 separated
servicemembers who held critical occupations.

Figure 6: Distribution of the Amount of Time Served by Individuals with
Critical Occupations prior to Separation for Homosexual Conduct, Fiscal
Years 1994 through 2003

3%

4%

Period 2: Separated within 3 to 6 months of military service (advanced
individual training)

59%

Period 3: Separated within 6 months to 1.5 years of military service

Period 4: Separated within 1.5 to 2.5 years of military service

Period 5: Separated after more than 2.5 years of service

Sources: Defense Manpower Data Center (data); GAO (analysis).

We identified servicemembers separated under the homosexual conduct policy
statute who had intelligence-related occupations (a partial list of these
occupations is in appendix III, table 10); not all of these occupations
were related to the selective reenlistment bonus program. We identified
730 separated servicemembers who held intelligence-related occupations
from fiscal year 1994 through fiscal year 2003. The separation rate is
similar to the separation rate of servicemembers who held occupations that
were related to a selective reenlistment bonus: 274 of these
servicemembers (about 38 percent) were separated within about 1.5 years of
service, and 450 servicemembers (about 62 percent) were separated within
about 2.5 years of service. Figure 7 shows the separation rate of
servicemembers with intelligence-related occupations by various time
periods.

Figure 7: Distribution of the Amount of Time Served by Individuals with
Intelligence-Related Occupations prior to Separation for Homosexual
Conduct, Fiscal Years 1994 through 2003

4%

Period 1: Separated within 3 months of military service (recruit training)

4%

Period 2: Separated within 3 to 6 months of military service (advanced
individual training)

Period 3: Separated within 6 months to 1.5 years of military service

                                      62%

Period 4: Separated within 1.5 to 2.5 years of military service Period 5:
Separated after more than 2.5 years of service

Source: Defense Manpower Data Center (data); GAO (analysis).

Note: Parts may not sum to equal cumulative percents because of rounding.
(See appendix III for frequency counts.)

    Some Servicemembers with Training in Important Languages Were Separated for
    Homosexual Conduct

DOD separated several hundred members with training in important foreign
languages. During fiscal years 1994 through 2003, DOD separated 322
servicemembers for homosexual conduct who had some skills in a foreign
language that DOD had considered to be especially important. A total of
209 separated servicemembers attended the Defense Language Institute for
training in one of these important languages. Ninety-eight of these 209
completed training and received a proficiency rating, and 62 members (63
percent of the 98) had proficiency scores at or below the midpoint on
DOD's language proficiency scales for listening, reading, or speaking. To
assess listening, reading, and speaking proficiencies, DOD uses an
11-point scale. DOD describes the midpoint as "limited working
proficiency, plus." According to the Defense Language Institute, in order
to graduate from the basic language program, students are expected to
achieve at least a "limited working proficiency" in listening and reading
and

an "elementary proficiency, plus" in speaking a foreign language. Both of
these levels are below the midpoint on DOD's proficiency scale. Table 2
shows the number of servicemembers separated for homosexual conduct who
had some skill in an important foreign language.

Table 2: Number of Servicemembers Separated for Homosexual Conduct with
Some Proficiency in an "Important Foreign Language," Fiscal Years 1994
through 2003

 Number of students with listening proficiencya Number of students with reading
           proficiencya Number of students with speaking proficiencya

Number of separated servicemembers

     Who attended Defense Language Institute Language Institute students with
                                                           proficiency scores

Below midpoint

Above midpoint

Below midpoint

Above midpoint

Below midpoint

Above midpointLanguage

      Arabic     54  20  10 (50)  5 (25)   8 (40)   7 (35)  20 (100)    0 (0) 
     Chinese     20    6  1 (17)   0 (0)   0 (0)    5 (83)   4 (67)    1 (17) 
      Farsi      9     2 2 (100)   0 (0)   1 (50)   1 (50)   2 (100)    0 (0) 
      Korean     50  25  21 (84)   2 (8)  17 (68)     1 (4)  24 (96)    0 (0) 
     Russian     42  25  11 (44)  8 (32)   5 (20)   9 (36)   19 (76)   4 (16) 
      Serbo-                                                          
     Croatian    8     4  2 (50)  1 (25)   1 (25)     0 (0)  3 (75)    1 (25) 
     Spanish     24  15   5 (33)  5 (33)   1 (7)    5 (33)   12 (80)    1 (7) 
    Vietnamese   2     1 1 (100)   0 (0)   0 (0)   1 (100)   1 (100)    0 (0) 
      Total                                                           
      number    209  98     53      21          33       29        85 
     Percent    100  47     54      24          34       30        87 

Sources: Defense Manpower Data Center (data); GAO (analysis).

Notes:

1. "Important" foreign languages are those for which servicemembers are
eligible to receive selective reenlistment bonuses or those identified as
"deficiencies" by combatant commanders and the Joint Staff in their
periodic readiness assessments.

2. The table does not include the number and percentage of students with
scores at the midpoint but includes such information only for students
below or above the midpoint.

aPercentages in parentheses. The Data Center has length-of-service data
for 205 of the separated servicemembers who received training in an
important foreign language.

We analyzed the length of service for the 205 separated servicemembers who
had received training in an important foreign language at the Defense
Language Institute. Figure 8 shows the separation rate for these

servicemembers. About 131 (64 percent) were separated within about 2.5
years of service.

Figure 8: Distribution of the Amount of Time Served by Individuals Trained
in Important Languages prior to Separation for Homosexual Conduct, Fiscal
Years 1994 through 2003

1%

Period 2: Separated within 3 to 6 months of military service (advanced
individual training)

64%

Period 4: Separated within 1.5 to 2.5 years of military service

Period 5: Separated after more than 2.5 years of service

Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: No servicemember with training in critical languages was separated
for homosexual conduct in Period 1, the first 3 months of military
service, which generally corresponds to recruit training.

We further analyzed the occupations of the 54 separated servicemembers who
received training in Arabic at the Defense Language Institute. We were
able to match 42 (about 78 percent) with an occupation that utilizes a
foreign language, many in intelligence-related occupations such as
"cryptologic linguist" or "communications interceptor." However, these 42
members might have had limited experience in their occupation because 36
servicemembers (about 86 percent of the 42) were listed as "helpers" or
"apprentices," or had the lowest skill level associated with the
occupation.

Agency Comments and 	In commenting on a draft of this report, the Under
Secretary of Defense (Personnel and Readiness) provided information on
separations for

Our Evaluation	homosexual conduct compared with other unprogrammed
separations from fiscal year 1994 through fiscal year 2003. DOD also
provided technical changes, which we made where appropriate. The
department's written comments are incorporated in their entirety in
appendix IV.

Unless you publicly announce its contents earlier, we plan no further
distribution of this report until 3 days from its issue date. At the time,
we will send copies of this report to the Secretary of Defense; the
Secretaries of the Army, the Air Force, and the Navy; the Commandant of
the Marine Corps, the Director of the Office of Management and Budget; and
interested congressional committees. We will also make copies available to
others upon request. In addition, the report will be available at no
charge on the GAO Web site at http://www.gao.gov.

Please contact me on (202) 512-5559 ([email protected]) or George
Poindexter, Assistant Director, on (202) 512-7213 ([email protected]),
if you or your staff have any questions concerning this report. Major
contributors to this report were Lisa Brown, Alissa Czyz, Joe Faley,
Nicole Gore, Catherine Humphries, Tom Mills, Charles Perdue, and Jen
Popovic.

Derek B. Stewart, Director Defense Capabilities and Management

List of Congressional Requesters

The Honorable Martin T. Meehan
Ranking Minority Member

Subcommittee on Terrorism,
Unconventional Threats and Capabilities
Committee on Armed Services
House of Representatives

The Honorable Neil Abercrombie
Ranking Minority Member
Subcommittee on Tactical Air and Land Forces
Committee on Armed Services
House of Representatives

The Honorable Tom Allen
The Honorable Robert Andrews
The Honorable Tammy Baldwin
The Honorable Danny Davis
The Honorable Susan A. Davis
The Honorable Diana DeGette
The Honorable William Delahunt
The Honorable Eliot Engel
The Honorable Barney Frank
The Honorable Sheila Jackson-Lee
The Honorable James R. Langevin
The Honorable Carolyn Maloney
The Honorable George Miller
The Honorable Jim Moran
The Honorable Jerrold Nadler
The Honorable Eleanor Holmes Norton
The Honorable Christopher Shays
The Honorable Adam Smith
The Honorable Pete Stark
The Honorable Lynn Woolsey
House of Representatives

Appendix I

Scope and Methodology

To conduct our work, we interviewed individuals at a variety of Department
of Defense (DOD) and service offices, including the office of the Under
Secretary of Defense for Personnel and Readiness; DOD's Office of
Accession Policy; DOD's Defense Manpower Data Center; and offices in the
military services responsible for budget, investigation, chaplaincy,
separation, and discharge review.

To determine the estimated financial costs associated with DOD's
homosexual conduct policy, we obtained information on the estimated costs
to recruit enlisted personnel from fiscal year 1994 through fiscal year
2003 from DOD's Office of Accession Policy. DOD includes this information
in the Military Personnel Procurement Resources Report. DOD calculates
recruiting cost per enlisted member by dividing a military service's total
expenditures for recruiting enlisted personnel by the service's total
number of accessions. Recruiting expenditures include, but are not limited
to, the costs associated with recruiting personnel, enlistment bonuses,
advertising, communications, recruiting support, and recruiting command
resources. We computed an average of the reported figures for fiscal years
1994 through 2003. DOD does not include per-capita recruiting costs
associated with commissioned officers in its procurement resources report.

We also requested that each of the four military services provide
estimated training cost information for occupations performed by enlisted
servicemembers who were separated for homosexual conduct from fiscal year
1994 through fiscal year 2003. In order to provide total estimated
training costs, we asked the services to provide estimates of both fixed
and variable costs1 associated with each occupation. Estimated
occupationrelated training costs include, but are not limited to, military
and civilian pay for instructors, operations and maintenance, student
transportation, ammunition, supplies, and flying costs (if any). We
reviewed the services' general methodology for developing training-cost
estimates and found

1 Total costs are the total costs of producing any given level of output.
Total cost can be divided in two parts: fixed costs and variable costs.
Fixed costs are those that do not vary with output. All costs that vary
directly with output are variable costs.

Appendix I Scope and Methodology

them acceptable. We used weighted averages2 to estimate the average
per-member occupational training costs for the Air Force, Army, and Navy.
The Marine Corps was unable to provide this information. Additionally, we
excluded from our analysis the training costs associated with medical and
health-care-related occupations because the services could not reasonably
estimate them. Service officials told us that the length of training and
other factors necessary to achieve a health-care-related proficiency
varies widely, as do the costs associated with them.

To assess the extent to which DOD separated members with critical
occupations or important foreign language skills, we obtained
occupation-and foreign-language-related data (for fiscal years 1994-2003)
on servicemembers separated for homosexual conduct from the Defense
Manpower Data Center's Active Duty Personnel Transaction File, which is a
compilation of data provided by each of the military services. Our
analysis was limited to active duty personnel and did not include 118
reservists who were separated for homosexual conduct because they
represent a small number of total separations under the homosexual conduct
policy statute. This is consistent with DOD's reporting practice in this
area. The department reports only active duty personnel separated for
homosexual conduct. The Data Center provided information on an
individual's branch of service, occupation, rank, length of time in
service, and language skills.

With respect to the occupational data, we adopted the military services'
definition of a "critical" occupation as an occupation that was part of
the selective reenlistment bonus program. The selective reenlistment bonus
program for enlisted military personnel is DOD's primary tool for
addressing short-term retention problems in critical occupations by
providing servicemembers who reenlist following the expiration of their
service contracts with up to $60,000. The Army, Marines, and Navy list
their 10 most critical occupations in their annual budget justifications.
The Air Force, however, does not prioritize its critical occupations in
its budget justification. The services determine reenlistment bonus
amounts by multiplying (1) a servicemember's current monthly basic pay by

2 In calculating a weighted average, each value is multiplied by its
"weight," and this product is summed for all values. The "weight" is
derived as a proportion of the total. With respect to a service's
occupational training costs, the costs of training for an occupation (the
value) would be multiplied by that occupation's weight (that occupation's
proportion of total servicemembers for all occupations). This product
would be summed for all occupations to calculate a service's weighted
average of occupational training costs.

Appendix I Scope and Methodology

(2) the member's number of additional years of obligated service by (3) a
bonus multiple that can range from 0.5 to 15. For the Air Force, we used
this bonus multiple to determine a list of the 10 most critical
occupations for each year from fiscal year 1994 through fiscal year 2003;
the Air Force occupations with the 10 largest bonus multipliers in a
specific year were deemed by us to be the most critical. For example, in 1
year we included Air Traffic Control in the list of the top 10 Air Force
occupations because it had a bonus multiplier of 7, which is the largest
multiplier that the Air Force used from fiscal year 1994 through fiscal
year 2003. In contrast, Pararescue, and all other occupations that had a
bonus multiplier of 5 for that year, were not included on our list of most
critical Air Force occupations. This is because there were at least 10 Air
Force occupations whose bonus multipliers were 5.5, 6, or 7. Note that, in
other years, depending on the bonus multipliers for all jobs, Pararescue
could be included as an occupation on the "top ten" list.

To assess the extent to which DOD separated individuals for homosexual
conduct in intelligence-related occupations, we compiled a list of
servicelevel occupation titles that could be categorized as
"intelligence-related" by their relationship to DOD's occupational codes.
DOD occupation codes are a way of organizing service-level occupations
into general categories. Each separated servicemember whose occupation
matched an intelligence-related DOD occupational code was considered to
have an intelligence-related occupation.

Finally, with respect to separations for homosexual conduct of individuals
with important language skills, we identified separated servicemembers
with foreign language skills using language data drawn from the Defense
Manpower Data Center. The Data Center provided two types of language data.
The first type addresses the language skills of servicemembers who
attended the Defense Language Institute's Foreign Language Center.
Language proficiency data for these students are based on the Defense
Language Proficiency Test score they received when tested at the
completion of their course of study. The other type of language data in
the active duty file is information reported to the Data Center by the
services. The language proficiency data in this file are based on multiple
sources- from servicemembers themselves or from the official Defense
Language Institute proficiency test.

Although we did not validate the budget/financial systems and processes
used to calculate the cost estimates used in this report, we determined
that the estimates were sufficiently reliable for the purpose of this
report. As

Appendix I Scope and Methodology

previously discussed, we assessed the reliability of these data by (1)
reviewing existing information about the data and the systems that
produced them and (2) interviewing agency officials knowledgeable about
the data to determine the steps taken to ensure the accuracy and
completeness of the data.

We assessed the reliability of the Defense Manpower Data Center's Active
Duty Military Personnel Transaction file by (1) performing electronic
testing of the required data elements, (2) reviewing existing information
about the data and the system that produced them, and (3) interviewing
agency officials knowledgeable about the data. We determined that the data
were sufficiently reliable for the purpose of this report. We conducted
our review from August 2004 through February 2005 in accordance with
generally accepted government auditing standards.

Appendix II

                         Financial Cost Estimate Tables

  Estimated Cost of Recruiting Servicemembers Separated for Homosexual Conduct

Table 4 shows that the total estimated cost to recruit potential
replacements for enlisted servicemembers separated for homosexual conduct
from fiscal year 1994 through fiscal year 2003 was about $95 million. To
compute this cost, we multiplied the number of servicemembers as shown in
table 1 (less the number of officers) by the data in table 3 for each
service and each year. For example, we multiplied the number of Army
members who were separated for homosexual conduct in fiscal year
1994-136-from table 1 by the Army's average annual recruiting cost for
fiscal year 1994 ($9,597) from table 3 in order to compute $1.305 million
in table 4. The sum of these calculations for the 10-year period is about
$95 million in constant fiscal year 2004 dollars.

 Table 3: Estimated Average Annual Recruiting Cost by Military Service and DOD,
                         Fiscal Years 1994 through 2003

                            Constant FY 2004 dollars

                                  Fiscal year

Service   1994    1995    1996    1997    1998    1999    2000    2001    2002    2003 
 Army   $9,597 $11,053 $10,460 $11,547 $13,059 $14,278 $14,078 $15,509 $16,200 $16,536 
 Navy    6,937   8,214   8,573   8,466   8,803 10,124  10,162  11,221  13,121   13,394 
Marine                                                                         
 Corps   7,362   5,732   6,595   6,313   6,560   8,208   8,353   8,831   8,453   9,356 
  Air    4,832   4,805   4,873   5,306   5,126   6,636   8,244   9,928   9,934   9,376 
 Force                                                                         
  DOD    8,315   8,953   7,606   9,519   8,928 10,134  10,913  12,906  13,715   14,206 

                                  Source: DOD.

                   Appendix II Financial Cost Estimate Tables

Table 4: Total Estimated Recruiting Costs to Replace Enlisted Personnel
Separated for Homosexual Conduct, Fiscal Years 1994 through 2003

                              Dollars in thousands

     Fiscal year          Army   Air Force    Marines      Navy         Total 
                        $1,305     $879         $265      $1,755       $4,204 
                         2,023     1,086        395       2,152         5,656 
                         2,040     1,345        389       2,632         6,406 
                         2,263     1,613        492       3,446         7,814 
                         4,035     2,097        499       2,958         9,589 
                         3,855     2,289        788       3,159        10,091 
                         8,110     1,443        860       3,587        14,000 
                         9,585     1,807        980       3,221        15,593 
                         6,638     1,192        879       2,860        11,569 
                         6,091     1,322        580       2,478        10,471 
        Total        $45,945      $15,073      $6,127    $28,248      $95,393 
       Percent              48            16     6          30     

Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: All figures are in constant fiscal year 2004 dollars.

Appendix III

                        Critical Occupation Data Tables

Length of Service of Most servicemembers separated for homosexual conduct
were separated

within 1.5 years of entering military service (approximately periods 1-3
inServicemembers Who table 5). The first and second periods on the table
correspond to different Were Separated for phases of enlisted personnel
training: recruit training (Period 1) and Homosexual Conduct advanced
individual training (Period 2), when a servicemember is initially

trained in an occupation. The exact number of days in each period varies

by service.1

Table 5: Individuals Separated for Homosexual Conduct during Selected
Intervals, Fiscal Years 1994 through 2003

                                Number Percenta

              Period 1: Period 2:  Period   Period       Period 5:      
                recruit advanced   3: next  4: next      subsequent     
    Service    training individual 365 days 365 days            periods Total 
                        training                                        
     Marine         153         76      289      123                139   780 
     Corps                                                              
      Army          583        407      918      522                811 3,241 
      Navy           47        260    1,154      568                886 2,915 
Air Force        964        294      301      245                499 2,303 
     Total        1,747      1,037    2,662    1,458              2,335 9,239 
     number                                                             
    Percent          19         11       29       16                 25   100 

                 Period 1: Period 2:      Period   Period       Period 5: 
                   recruit advanced       3: next  4: next                
    Service       training individual     365 days 365 days  subsequent   
                           training                            periods    
     Marine             20             10       37       16            18 101 
     Corps                                                                
      Army              18             13       28       16            25 100 
      Navy               2              9       40       19            30 100 
Air Force            42             13       13       12            22 102 

Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: The Data Center has length-of-service data for 9,239 of the 9,488
servicemembers who were separated for homosexual conduct during the
10-year period.

aPercents may not add to 100 because of rounding.

1 Period 1, recruit training, includes the following intervals for each of
the services: Marines, 0 to 84 days; Army, 0 to 63 days; Navy, 0 to 56
days; and Air Force, 0 to 42 days. Period 2, the average time for advanced
individual training (100 days), includes the following intervals for each
of the services: Marines, 85 to 185 days; Army, 64 to 164 days; Navy, 57
to 157 days; and Air Force, 43 to 143 days. Period 3 spans 1 year from the
end of the advanced individual training period, and period 4 spans 1 year
from the end of period 3. Period 5 includes all subsequent time periods.

                  Appendix III Critical Occupation Data Tables

Length of Service of Most servicemembers separated for homosexual conduct
who had critical Separated Servicemembers occupations were separated
within 2.5 years of entering the military. Two Who Had Critical and a half
years corresponds approximately to the end of the 4th period in

  table 6.Occupations

Table 6: Individuals with Critical Occupations Separated for Homosexual
Conduct during Selected Intervals, Fiscal Years 1994 through 2003

                                Number Percenta

              Period 1: Period 2: advanced Period     Period  Period 5: 
                recruit                    3: next   4: next            
    Service    training individual         365 days 365 days subsequent Total 
                        training                              periods   
     Marine           0                  1        0        0          3 
     Corps                                                              
      Army           21                 19       47       38         39 
      Navy            0                  1      135      102        207 
Air Force          0                  9       34       36         63 
     Total           21                 30      216      176        312 
     number                                                             
    Percent           3                  4       29       23         41 

                  Period 1: Period 2:      Period 3: Period 4:   Period 5:    
                    recruit advanced       next           next 
     Service       training individual      365 days  365 days   subsequent   
                            training                              periods     
     Marine               0             25         0         0       75       
      Corps                                                    
      Army               13             12        29        23       24       
      Navy                0             <1        30        23       47       
    Air Force             0              6        24        25       44       

         Sources: Defense Manpower Data Center (data); GAO (analysis).

               aPercents may not add to 100 because of rounding.

                  Appendix III Critical Occupation Data Tables

Length of Service of Most servicemembers who had intelligence-related
occupations were Separated Servicemembers separated for homosexual conduct
within approximately 2.5 years of Who Had Intelligence-entering military
service. Two and a half years corresponds approximately Related
Occupations to the end of the 4th period as shown in table 7.

Table 7: Individuals with Intelligence-Related Occupations Separated for
Homosexual Conduct during Selected Intervals, Fiscal Years 1994 through
2003

                                Number Percenta

              Period 1: Period 2: advanced Period   Period    Period 5: 
                recruit                    3: next  4: next             
    Service    training individual         365 days 365 days subsequent Total 
                        training                              periods   
     Marine           0                  0       14       14         20 
     Corps                                                              
      Army           32                 23       84       49         62 
      Navy            0                  1       84       74        129 
Air Force          0                  3       33       39         69 
     Total           32                 27      215      176        280 
     number                                                             
    Percent           4                  4       29       24         38  100a 

Period 1: recruit Period 2: advanced Period 3: next Period 4: next Period
5: Service training individual training 365 days 365 days subsequent
periods

Marine Corps 0 0 29 29 42

Army 13 9 3420 25

Navy 0 <1 2926 45

Air Force 0 2 23 27 48

         Sources: Defense Manpower Data Center (data); GAO (analysis).

                 aPercents may not add to 100 due to rounding.

                  Appendix III Critical Occupation Data Tables

Length of Service of The same pattern is true for servicemembers separated
for homosexual Separated Servicemembers conduct who were trained in an
important language. Most servicemembers Who Had Important Foreign were
separated by the end of the 4th period-or approximately 2.5 years Language
Skills after entering military service-as shown in table 8.

Table 8: Individuals with Training in Important Languages Separated for
Homosexual Conduct during Selected Intervals, Fiscal Years 1994 through
2003

                                     Number

Period 1: recruit Period 2: advanced Period 3: next Period 4: next Period
5: subsequent Service training individual training 365 days 365 days
periods Total

Marine Corps 0 0 1 3 2

Army 0 023 28 28

Navy 0 0149 12

Air Force 0 2 24 27 32

Total number 0 2 62 67 74

Percent 0 1 30 33 36

                                    Percenta

Period 1: recruit Period 2: advanced Period 3: next Period 4: next Period
5: subsequent Service training individual training 365 days 365 days
periods

Marine Corps 0 0 17 50 33

Army 0 029 35 35

Navy 0 040 26 34

Air Force 0 2 28 32 38

         Sources: Defense Manpower Data Center (data); GAO (analysis).

               aPercents may not add to 100 because of rounding.

                  Appendix III Critical Occupation Data Tables

Occupations Most A sample of occupations eligible to receive a selective
reenlistment bonus Frequently Cited for is shown in table 9. Because each
service's designation of critical Selective Reenlistment occupations
changes annually, the column on the far right of the table

shows the number of times from fiscal year 1994 through fiscal year
2003Bonuses that an occupation appeared on the military services' "top
ten" list of critical occupations.

                    Table 9: Sample of Critical Occupations

Most frequently cited occupations receiving selective Total number of
years in which the occupation Service reenlistment bonuses, FY 1994-2003
received a selective reenlistment bonus

Army Automatic Test Equipment Operator

Engineer Tracked Vehicle Crewman

Noncommunications Interceptor/Analyst

Special Forces Communications Sergeant

Voice Interceptor (Persian/Vietnamese)

Aircraft Pneudraulics Repairer

Broadcast Journalist

Diver

Explosive Ordinance Disposal Specialist

Interrogator (Chinese/Korean)

OH-58D Helicopter Repairer

Petroleum Supply Specialist

Psychological Operations Specialist

Radar Repair

Satellite Communications Systems Operator-Maintainer

                    Signal Intelligence Analyst (Chinese/Korean)            3 
       Navy           Aviation Structural Mechanic (Equipment)              4 
                     Aviation Structural Mechanic (Structural)              4 
                         Cryptologic Technician (Technical)                 4 
                             Data Processing Technician                     4 
                         Electrician's Mate (Nuclear Field)                 4 
                              Fire Control Technician                       4 
                          Machinist's Mate (Nuclear Field)                  4 
                                      Mineman                               4 
                                 Missile Technician                         4 
                               Operations Specialist                        4 

                  Appendix III Critical Occupation Data Tables

                         (Continued From Previous Page)

             Most frequently cited occupations receiving selective

Total number of years in which the occupation received a selective
reenlistment bonusService

                       reenlistment bonuses, FY 1994-2003

Air Force Combat Controller

Air Traffic Control

Communication Computer System Programmer

Far East Crypto Linguist

Mid East Crypto Linguist

Pararescue

Slavic Crypto Linguist

Communication Computer System Control

Electronics Signals Intelligence Exploitation

Interpreter/Translator

Marines Aircraft Flight Engineer, KC-130

Electronic Switching Equipment Technician

Ground Mobile Forces Satellite Communications Technician

Air Command and Control Electronics Operator

Computer Technician

Consolidated Automatic Support System Technician

Cryptologic Linguist, Arabic

Surface Air Defense Systems Acquisition Technician 4

Technical Controller 4

Aircraft Navigation Systems Technician Identification Friend or
Foe/Radar/Tactical Air Navigation 3

Computer System Technician, Honeywell Data Processing
System 6 3

Counterintelligence Marine 3

Cryptologic Linguist, Korean 3

Cryptologic Linguist, Spanish 3

Field Artillery Radar Operator 3

Interrogation-Translation Specialist 3

Marine Air Ground Task Force Plans/Operations Specialist 3

Nonappropriated Funds Audit Technician 3

Radio Technician 3

Test Measurement and Diagnostic Equipment Technician 3

Weather Forecaster 3

    Sources: Service-submitted budget justification (data); GAO (analysis).

                  Appendix III Critical Occupation Data Tables

Examples of Intelligence-Related The following is a sample of the type of
intelligence-related occupations

Occupations	that we included in our analysis. All the occupations in table
10 have an intelligence-related DOD occupation code, which was the
criterion used to identify intelligence-related occupations.

    Table 10: Sample of Intelligence-Related Occupations Service Occupation

  Air Force Airborne Far East Crypto Linguist Airborne Romance Crypto Linguist

Airborne Slavic Crypto Linguist

                  Airborne Warning Command and Control System

Electronic System Security Assessment

Far East Crypto Linguist (Chinese)

Far East Crypto Linguist (Korean)

Far East Crypto Linguist (Vietnamese)

Imagery Interpreter

                           Intelligence Applications

Interpreter/Translator

                            Mid East Crypto Linguist

Mid East Crypto Linguist (Arabic)

Mid East Crypto Linguist (Hebrew)

Mid East Crypto Linguist (Persian)

                         Signals Intelligence Analysis

Army	Counterintelligence Agent Imagery Ground Station Operator
Intelligence Analyst Interrogator (Chinese/Korean) Psychological
Operations Specialist Signal Intelligence Analyst (Chinese/Korean) Voice
Interceptor (Persian/Vietnamese)

                  Appendix III Critical Occupation Data Tables

                         (Continued From Previous Page)

Service Occupation

           Marine Corps Air Command and Control Electronics Operator

Airborne Radio Operator/Loadmaster

                           Counterintelligence Marine

                          Cryptologic Linguist, Arabic

                          Cryptologic Linguist, Korean

Cryptologic Linguist, Persian, Semitic

 Cryptologic Linguist, Spanish Fleet Satellite Communications Terminal Operator

High Frequency Communication Central Operator Imagery Interpretation Specialist

Intelligence Specialist

Interrogation-Translation Specialist

Non-Morse Intercept Operator/Analyst Navy Air Traffic Controller Aviation
                         Antisubmarine Warfare Operator

Cryptologic Technician (Collection)

Cryptologic Technician (Interpretative)

Cryptologic Technician (Technical)

Electronic Warfare Technician

Operations Specialist

Radioman, Surface Warfare

Sources: DOD (data); GAO (analysis).

Tables 11 and 12 describe characteristics of the language speakers in the
population of those separated for homosexual conduct from fiscal year 1994
through fiscal year 2003, as reported by the Data Center. The table lists
the median proficiency level for all speakers of each language. DOD's
language proficiency scale includes 11 possible values, ranging from 00 to
as high as 50.2 In tables 11 and 12, the median proficiency is the middle
value if all proficiency scores for students in that language are placed
in numerical order.

2 DOD's language proficiency scale is as follows: 00-no proficiency;
06-memorized proficiency; 10-elementary proficiency; 16-elementary
proficiency, plus; 20-limited working proficiency; 26-limited working
proficiency, plus; 30-general professional proficiency, plus; 36-general
professional proficiency plus; 40-advanced professional proficiency;
46-advanced professional proficiency, plus; and 50-functionally native
proficiency.

                  Appendix III Critical Occupation Data Tables

Two tables are provided rather than one because the service-provided data
set contains an unknown mixture of self-assessed and Defense Language
Proficiency Test data. For the language institute-trained population of
language speakers, however, all proficiency data resulted from tests. Note
the high percentages of service members in both groups without a reported
proficiency score; individuals with no data available are included as
those without any recorded proficiency in speaking, listening, or reading.
This means that the Data Center did not have any information from any
source on the servicemembers' ability to use their reported language.

Table 11: Languages Spoken by and Proficiency Levels for Individuals
Separated for Homosexual Conduct from Fiscal Year 1994 through Fiscal Year
2003 Who Were Trained in a Language at the Defense Language Institute

Language

Total number of servicemembers

Median proficiency

     Number (and percent) of reported servicemembers with no proficiency data
                                                                    available

           Arabic, Modern Standard            54        20            34 (63) 
              Chinese, Mandarin               20        26            14 (70) 
                   French                      3        26     
                   German                      1        20     
                   Hebrew                      2        N/A            1 (50) 
                   Korean                     50        20            25 (50) 
      Persian, Iranian (includes Farsi)        9        20             7 (78) 
                   Russian                    42        26            17 (40) 
               Serbo-Croatian                  8        26             4 (50) 
                   Spanish                    24        26             8 (35) 
                   Tagalog                     1        26     
              Vietnamese, Hanoi                2        20             1 (50) 
                    Total                     216                    111 (51) 

Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: N/A = not available.

                  Appendix III Critical Occupation Data Tables

Table 12: Languages Spoken by and Proficiency Levels for Individuals
Separated for Homosexual Conduct from Fiscal Year 1994 through Fiscal Year
2003, as Reported through Service Personnel Files

                                               Number (and    
                                               percent) of    
                                                     reported     Number (and 
                                                                  percent) of 
                                               servicemembers        reported 
                   Total number of      Median with no         servicemembers 
                                                                         with 
      Language     servicemembers  proficiency   proficiencya  no proficiency 
                                                              data availablea 
      Achinese                   2           0        2 (100) 
       Amashi                    1           0        1 (100) 
Arabic, Modern                5          20         1 (20)          2 (40) 
      Standard                                                
      Chinese,                   2          30         1 (50)          1 (50) 
      Cantonese                                               
      Chinese,                   2         N/A         1 (50)         2 (100) 
      Mandarin                                                
       Danish                    1         N/A              0         1 (100) 
       French                   13          26         5 (38)          4 (31) 
       German                   10          16         5 (50)          2 (20) 
       German,                   1           0        1 (100) 
      Bavarian                                                
Haitian, Creole               1          50              0 
      Hungarian                  2          26              0         2 (100) 
     Indonesian                  1          30              0 
       Italian                   5          50         1 (20)          1 (20) 
      Japanese                   1          10              0 
       Korean                    5          20              0          2 (40) 
Old High German               1          10              0 
      Persian,                                                
       Iranian                                                
      (includes                                               
       Farsi)                    1          20         1(100)         1 (100) 
       Polish                    1         N/A        1 (100)               0 
     Portuguese,                 1         N/A              0         1 (100) 
      Brazilian                                               
       Russian                   9         N/A         1 (11)          3 (33) 
Serbo-Croatian                3          20              0               0 
       Spanish                  50          20        18 (36)         30 (60) 
      Spanish,                  59          30         6 (10)           4 (7) 
      American                                                
      Spanish,                   2          20              0               0 
      Castilian                                               
Spanish, Creole               1          30              0               0 

                  Appendix III Critical Occupation Data Tables

                         (Continued From Previous Page)

                                           Number (and    
                                           percent) of    
                                                 reported         Number (and 
                                                                  percent) of 
               Total number of      Median servicemembers            reported 
                                           with no        servicemembers with 
    Language   servicemembers  proficiency   proficiencya no proficiency data 
                                                                   availablea 
     Tagalog                 8          50         1 (12)              1 (12) 
      Urdu                   1          50              0 
Vietnamese,               1          50              0 
     Central                                              
      Total                190                    46 (35)             57 (30) 

Sources: Defense Manpower Data Center (data); GAO (analysis).

Note: N/A = not available

aIndividuals received three separate proficiency scores: one in reading,
one in listening, and one in speaking. If any one of these three scores
indicated that the individual was tested but had no proficiency, the
individual is counted in the "no proficiency" column. Likewise, if one of
the three scores was not available, the individual is listed in the "no
data available" column.

                                  Appendix IV

                    Comments from the Department of Defense

Appendix IV
Comments from the Department of Defense

GAO's Mission	The Government Accountability Office, the audit, evaluation
and investigative arm of Congress, exists to support Congress in meeting
its constitutional responsibilities and to help improve the performance
and accountability of the federal government for the American people. GAO
examines the use of public funds; evaluates federal programs and policies;
and provides analyses, recommendations, and other assistance to help
Congress make informed oversight, policy, and funding decisions. GAO's
commitment to good government is reflected in its core values of
accountability, integrity, and reliability.

Obtaining Copies of The fastest and easiest way to obtain copies of GAO
documents at no cost

is through GAO's Web site (www.gao.gov). Each weekday, GAO postsGAO
Reports and newly released reports, testimony, and correspondence on its
Web site. To Testimony have GAO e-mail you a list of newly posted products
every afternoon, go to

www.gao.gov and select "Subscribe to Updates."

Order by Mail or Phone	The first copy of each printed report is free.
Additional copies are $2 each. A check or money order should be made out
to the Superintendent of Documents. GAO also accepts VISA and Mastercard.
Orders for 100 or more copies mailed to a single address are discounted 25
percent. Orders should be sent to:

U.S. Government Accountability Office 441 G Street NW, Room LM Washington,
D.C. 20548

To order by Phone:	Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202)
512-6061

  To Report Fraud, Contact:
  Waste, and Abuse in Web site: www.gao.gov/fraudnet/fraudnet.htm

E-mail: [email protected] Programs Automated answering system: (800)
424-5454 or (202) 512-7470

Congressional	Gloria Jarmon, Managing Director, [email protected] (202)
512-4400 U.S. Government Accountability Office, 441 G Street NW, Room 7125

Relations Washington, D.C. 20548

Public Affairs	Paul Anderson, Managing Director, [email protected] (202)
512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

                               Presorted Standard
                              Postage & Fees Paid
                                      GAO
                                Permit No. GI00

United States
Government Accountability Office
Washington, D.C. 20548-0001

Official Business
Penalty for Private Use $300

Address Service Requested
*** End of document. ***