Defense Logistics Agency: Process for Selecting Defense Distribution
Center Site Contained Weaknesses (Letter Report, 04/09/98,
GAO/NSIAD-98-96).

Pursuant to a congressional request, GAO reviewed the process used by
the Defense Logistics Agency (DLA) to select the site for its new
defense distribution center, focusing on whether: (1) the process was
sound; and (2) there was any evidence that the site selection had been
predetermined.

GAO noted that: (1) DLA officials believed that the consolidation of its
eastern and western regional distribution headquarters would produce
savings; (2) DLA's establishment of a steering group and decisionmaking
criteria indicate that DLA recognized the need for a credible process to
guide its decisionmaking in selecting a site for its consolidated
distribution headquarters; (3) however, the process used by DLA to
support the site selection decision contained a number of weaknesses;
(4) among the weaknesses in the process were the absence of sufficient
information concerning personnel facilities requirements for the new
center, unrealistic cost comparisons between the competing sites, and
the use of subjective data for two noncost criteria; (5) subsequent
changes to the process, made at the request of the selecting official,
did not correct these weaknesses and created concerns about the
perception of bias; (6) the cumulative effect of these weaknesses raised
questions about the soundness of the site selection process and the
ultimate decision; (7) although various persons from the western
location raised concerns about whether the decision had been
predetermined, GAO found no evidence to validate these concerns; and (8)
likewise, GAO found no evidence that prior studies examining the
consolidation issue influenced the current site selection process or
outcome.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  NSIAD-98-96
     TITLE:  Defense Logistics Agency: Process for Selecting Defense 
             Distribution Center Site Contained Weaknesses
      DATE:  04/09/98
   SUBJECT:  Reductions in force
             Military facilities
             Base closures
             Centralization
             Site selection
             Defense cost control
             Cost analysis
             Military downsizing

             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed 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.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to Congressional Requesters

April 1998

DEFENSE LOGISTICS AGENCY - PROCESS
FOR SELECTING DEFENSE DISTRIBUTION
CENTER SITE CONTAINED WEAKNESSES

GAO/NSIAD-98-96

Defense Logistics Agency

(709310)


Abbreviations
=============================================================== ABBREV

  BRAC - base realignment and closure
  DLA - Defense Logistics Agency
  DOD - Department of Defense

Letter
=============================================================== LETTER


B-279301

April 9, 1998

The Honorable Richard W.  Pombo
The Honorable Gary A.  Condit
House of Representatives

On September 16, 1997, the Defense Logistics Agency (DLA) announced
its decision to consolidate its eastern and western regional
distribution headquarters into a new center to be located in New
Cumberland, Pennsylvania, the site of the eastern regional
headquarters.  At that time, DLA officials expected the consolidation
to occur over about a 2-year period beginning October 1, 1997. 
Through the consolidation, DLA expects to reduce its regional
headquarters and co-located support personnel by about 673 for a
savings of about $28.3 million annually. 

This report responds to your request that we review the process used
by DLA to select the site for its new center.  Specifically, this
report addresses whether (1) the process was sound and (2) there was
any evidence that the site selection had been predetermined. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

DLA officials believed that the consolidation of its eastern and
western regional distribution headquarters would produce savings. 
DLA's establishment of a steering group and decision-making criteria
indicate that DLA recognized the need for a credible process to guide
its decision-making in selecting a site for its consolidated
distribution headquarters.  However, the process used by DLA to
support the site selection decision contained a number of weaknesses. 
Among the weaknesses in the process were the absence of sufficient
information concerning personnel and facilities requirements for the
new center, unrealistic cost comparisons between the competing sites,
and the use of subjective data for two noncost criteria.  Subsequent
changes to the process, made at the request of the selecting
official, did not correct these weaknesses and created concerns about
the perception of bias.  The cumulative effect of these weaknesses
raised questions about the soundness of the site selection process
and the ultimate decision. 

Although various persons from the western location raised concerns
about whether the decision had been predetermined, we found no
evidence to validate these concerns.  Likewise, we found no evidence
that prior studies examining the consolidation issue influenced the
current site selection process or outcome. 


   BACKGROUND
------------------------------------------------------------ Letter :2

Since 1993, DLA has operated with two defense distribution regional
headquarters, an eastern headquarters in New Cumberland,
Pennsylvania, and a western headquarters in the vicinity of Stockton,
California.  These regional headquarters provided operational
oversight to over 20 geographically dispersed distribution depots.\1
See figure 1 for DLA's distribution structure prior to the decision
to consolidate the regional headquarters. 

   Figure 1:  DLA's Defense
   Distribution Structure Prior to
   Decision to Consolidate
   Regional Headquarters

   (See figure in printed
   edition.)

Source:  DLA and Defense Base Closure and Realignment Commission 1995
Report to the President. 


--------------------
\1 Distribution depots receive, store, and issue wholesale and retail
material in support of the armed forces worldwide. 


      ACTIONS RESULTING FROM THE
      BRAC PROCESS
---------------------------------------------------------- Letter :2.1

DLA reduced its number of distribution regions from three to two in
1993 and soon thereafter began exploring the potential of having just
one.  Although a study was initiated, it was not finalized and no
proposals or recommendations were approved.  During the 1995 base
realignment and closure (BRAC) process, DLA examined the military
value of the eastern and western regional headquarters and found that
they rated nearly equal.\2 At that time, DLA officials concluded that
changing the command and control structure would present significant
risks in the efficient management of day-to-day operations and the
ability to effectively support two major regional conflicts
simultaneously.  Further, they determined that span of control of
future operations and the requirement to continue to accommodate
contingency, mobilization, and future total force requirements made
two regions essential. 


--------------------
\2 DLA assessed the military value of its activities using four
distinct measures:  mission scope, mission suitability, operational
efficiencies, and expandability.  Out of a potential 1,000 points for
military value, the western regional headquarters scored 896, while
the eastern regional headquarters scored 882. 


      SITE SELECTION PROCESS
---------------------------------------------------------- Letter :2.2

Subsequent to BRAC 1995, DLA officials revisited the consolidation
issue and began another study.  According to DLA headquarters
officials, a preliminary assessment was made and the study was
terminated before any recommendations were made.  Meanwhile, DLA
continued to restructure its distribution organization.  Then, in
February 1997, DLA reinitiated an effort to consolidate the two
regions.  This was expected to result in the creation of a single
distribution command, known as the Defense Distribution Center. 
Under this plan, the center was to assume the regional distribution
functions and manage the two primary distribution sites and all
remaining distribution depots. 

Two steering groups were established to work concurrently on the
consolidation effort.  A "missions and functions steering group" was
established to determine the distribution center staffing
requirements and organizational design.  A "site selection steering
group" (hereafter referred to as the steering group) was established
to develop a decision process to determine a recommended site.  The
Principal Deputy Director of DLA was responsible for selecting the
site in consultation with the DLA Director and other senior officials
who made up DLA's executive leadership team. 

The site selection process included the use of a contractor, KPMG
Peat Marwick, LLP to assist the steering group in developing the
decision model and identifying the data needed.  The contractor and
an experienced DLA facilities engineer gathered, validated, and
evaluated the data used in the model. 

The steering group had 12 members, with 6 voting members from various
offices within DLA headquarters and the Office of the Secretary of
Defense; 2 voting members each from the eastern and western regions;
and 2 nonvoting members--the steering group chair (a military
colonel) and a Department of Defense (DOD) Inspector General
representative, who participated as an advisor.  The group developed
cost and other site selection criteria that were approved by the
selecting official and the executive leadership team.  (See table 1.)



                                Table 1
                
                  Site Selection Criteria and Weights

                                                Maximum
                                             points for        Maximum
                                                   each     points per
                                            criterion\a       category
----------------------------------------  -------------  -------------
Cost criteria--800 total points
----------------------------------------------------------------------
Investment costs over a 1-year period                              300
Facilities and information technology               300
Operations costs over a 5-year period                              500
Utilities, service orders, janitorial               160
Real property maintenance                           110
Personnel                                           130
Travel                                              100

Noncost criteria--200 total points
----------------------------------------------------------------------
Work                                                               100
Work environment                                     70
Commute time                                         30
Home                                                               100
Cost of living                                       40
Median house price                                   30
Crime index                                          20
Unemployment rate                                    10
======================================================================
Total point allocation                                           1,000
----------------------------------------------------------------------
\a In this model, the maximum points that could be awarded to either
site was 1,000.  Each site's data inputs were evaluated and the one
with the best outcome received 100 percent of the points allocated to
that criterion, while the other site received a fraction of the
points depending on the relative difference between the two scores. 

Source:  DLA. 

While costs played a major role in the evaluation, the preferred site
did not necessarily have to be the site with the lowest costs. 
Rather, it was the one with the highest point total. 

Three sites were initially considered for the new center:  DLA
headquarters at Fort Belvoir, Virginia, and the existing eastern and
western regional headquarters.  The Fort Belvoir site dropped out
before completion of the first data request due to a lack of
available space at the headquarters building. 


         REQUESTS FOR DATA
-------------------------------------------------------- Letter :2.2.1

The steering group sponsored two rounds of data requests from
officials at the competing sites.  According to KPMG, the responses
to the first data request were not used because some of the questions
were not clear and the respondents did not fully understand them. 
Thus, a second data request was required.  For the second data
request, the questions were redefined in an attempt to be clearer and
obtain better information.  The steering group expected that
responses to the second data request would be analyzed using the
decision model to identify the preferred site. 

Analysis of the second data request was completed on August 11, 1997. 
Although the steering group was not convened to review the results,
the steering group chair and KPMG jointly presented the results to
the selecting official (DLA's Principal Deputy Director), who told us
that he was dissatisfied with how some aspects of the approved
methodology had been implemented and saw the need for additional
data.  He requested revisions, including having a second facilities
engineer, not previously involved in the process, oversee the
collection and validation of some new data--essentially a third data
request.  The steering group was not made aware of the selecting
official's actions, including the third data request, until the group
was given both the results of its work on the second data request and
the revisions, on September 15, 1997.  On the basis of the results of
the revised study, the eastern location was selected by the Principal
Deputy Director of DLA as the site for the new center also on
September 15, 1997. 


         STATUS OF THE
         CONSOLIDATION
-------------------------------------------------------- Letter :2.2.2

As of October 1, 1997, DLA had officially established its command and
control of all distribution functions at New Cumberland,
Pennsylvania.  As of February 25, 1998, personnel performing
distribution headquarters functions at the western location were
reporting to management at the eastern location, and DLA was in the
process of implementing other aspects of the consolidated operation
at that location. 


   WEAKNESSES IN THE SITE
   SELECTION PROCESS
------------------------------------------------------------ Letter :3

The process used by DLA to support the site selection decision for
its consolidated distribution headquarters contained several
weaknesses, including insufficient data on personnel and facilities
requirements, a questionable methodology for evaluating and comparing
costs, and subjective responses used by steering group members for
two criteria.  Subsequent changes to the process, made at the request
of the selecting official, did not correct these weaknesses and
created concerns about the perception of bias.  Also, these actions
significantly altered investment and operating cost results between
the second and third data requests.\3 (See apps.  I and II.)


--------------------
\3 The data and results are presented for information purposes only. 
We cannot validate the results of either data collection effort
because of problems identified with the process. 


      INSUFFICIENT DATA ON
      PERSONNEL AND FACILITIES
      REQUIREMENTS
---------------------------------------------------------- Letter :3.1

Because information on staff size and functions was being determined
concurrently with the site selection process, the steering group was
not given complete information on the staffing requirements,
organizational design, and facility requirements for the new
headquarters.  Because these requirements have a substantial impact
on space utilization and costs,\4 it is important that they be
properly defined in advance of facility space planning. 

The steering group was initially told by senior DLA management that
an estimated 400 persons would be needed to staff the consolidated
distribution headquarters.  Therefore, the first data request asked
officials at the competing sites for facility requirements and costs
based on the requirements of 400 staff.  Subsequently, the missions
and functions steering group provided an estimated personnel strength
of 347 persons, which was used in the second data request.  In both
instances, the competing locations were not given a more detailed
breakdown of the operational functions or the number of persons
associated with them. 

Respondents from the regional locations stated that they could only
estimate floor plan requirements and then compute associated costs. 
According to a KPMG official, the floor plans developed by each site
contained certain unrealistic aspects and did not present a clear
picture of the investment costs that would be required. 

Although the structure and functions of the new headquarters were
determined prior to the third data request by the mission and
functions steering group, DLA officials considered the information
too sensitive to release to the competing locations because it could
lead to speculation about layoffs.  For the third data collection
effort, requested by the selecting official, a DLA facilities
engineer provided the regional respondents with the number of
functions and the number of staff per function, but not the identity
of the functions.  As a result, the third data request also resulted
in hypothetical floor plans and associated costs.  Officials at the
eastern location told us that, after becoming aware of the final
plans for staffing, they made some good guesses in identifying some
of the functions and could have implemented their floor plan if
required to do so.  However, they did not believe the floor plan
presented the optimal solution.  For example, although they knew they
needed a law library, they did not know which function it would be
associated with; therefore, their floor plan resulted in locating
their legal staff offices at one end of one building and the law
library at the opposite end of a second building.  Although some
changes were to be expected, according to regional officials, the
lack of definitive information meant that neither site's floor plan
would have been fully implemented if selected. 

DLA officials are still in the process of finalizing the floor plan
for their headquarters location in New Cumberland, but the officials
said they do not expect the costs to exceed the estimates provided in
the site selection competition.  Nevertheless, questions still exist
regarding what differences might have existed between the plans and
costs initially provided by the two competing locations if they had
had a clearer picture of the performing functions and space
requirements. 

Both steering group members and regional respondents said the lack of
information on personnel and facilities hampered their ability to
perform their tasks.  A KPMG official agreed, noting that although he
had been involved in numerous site selections, this was the first
time he had participated in a selection process in which the
functions and related staffing had not been determined before the
site selection process began. 


--------------------
\4 These costs are reflected in the facilities and information
technology and personnel categories in the site selection decision
model. 


      QUESTIONABLE COST COMPARISON
      METHODOLOGY
---------------------------------------------------------- Letter :3.2

Analysts often assign varying weights to evaluation criteria in this
type of analysis to distinguish the relative importance of individual
criteria.  This approach is also used to assign different weights
between cost and noncost variables to distinguish their relative
importance.  However, DLA's site selection steering group assigned
different weights to individual cost criteria, which produced a
distorted picture of the comparative costs of the two competing
locations.  For example, a dollar spent on a service order became
more significant or of more value than a dollar spent on real
property maintenance.  The distortion caused by this weighting was so
significant for the second data request that, even though the eastern
location was $3.8 million more expensive than the western location
based on a straight comparison of costs, the assignment of points
made it appear that the eastern location had come out ahead in the
cost categories.  (See app.  I for the results of the second data
request.)

The DOD Inspector General representative who participated in the
steering group's proceedings told us that he had questioned the
weighting of individual cost elements and recommended to the group
that costs be evaluated on a straight comparison basis between
locations.  Further, KPMG officials told us that they had also told
the group repeatedly that this methodology was not comparing dollars
equally.  Individual steering group members we spoke with could not
recall their rationale for a disproportionate ranking of dollars and
did not understand the potential impact of their actions until they
saw the results of the second data request. 


      SUBJECTIVE EVALUATION OF TWO
      NONCOST CRITERIA
---------------------------------------------------------- Letter :3.3

Specific criteria for evaluating work environment and commute time
were not established because the steering group could not determine
data sources for these two noncost criteria.  Thus, steering group
members subjectively determined point values for these criteria at
each location.  For work environment, members told us they considered
everything from distance between the parking lot and the office to
perceived lifestyles to ongoing working relationships.  For commute
time, because some members had been to each location only twice,
their experience with commute time consisted of traveling between
their hotel and the site.  Some members argued that commute time was
not really a valid criterion, because it was a matter of personal
choice.  The subjectivity of these responses and the inconsistency in
what members considered made the determinations of questionable
value.  These two criteria represented 100 points, or half the points
awarded in the noncost category and 10 percent of the total points. 


      SELECTING OFFICIAL REQUESTED
      CHANGES AFFECTING THE
      ANALYSIS
---------------------------------------------------------- Letter :3.4

DLA's site selection decision support model was reviewed and approved
by its Principal Deputy Director who served as the selecting
official, in consultation with DLA's executive leadership team,
before data collection efforts were initiated to ensure the
objectivity of the process.  However, after being briefed on the
results of the second data request analysis, the selecting official
requested changes in the analysis and asked for additional data.\5 He
stated that he did this to better ensure the comparability of data
between the sites.  This produced the requirement for a third data
request.  The selecting official decided that equal points should be
awarded to each site for personnel costs, that real property
maintenance costs be reassessed, and that changes should be made to
requirements for furniture and space.  The selecting official's
actions to negate the impact of one criterion and to have data
reassessed after receiving the results of the analysis and without
consulting the steering group created concerns among various steering
group members about the perception of bias. 


--------------------
\5 The results of the second data request analysis showed that the
eastern location had won on points while the western location was the
less costly site.  (See app.  I.)


         PERSONNEL COSTS
-------------------------------------------------------- Letter :3.4.1

The selecting official disagreed with the steering group's use of
personnel cost as a criterion.  As a result, following a briefing on
the second data request analysis, the selecting official, acting
independently of the steering group, decided to eliminate personnel
costs as a consideration.  The selecting official reasoned that the
grade structure of the new headquarters should be independent of the
chosen site, making personnel costs irrelevant, so he gave equal
points to both locations.  (See app.  II for the results of the third
data request.) The steering group, however, had considered personnel
cost to be an important criterion and had based it on average grade
levels of the current structure at each location.  The steering group
reasoned that, although the employees of the new site would be
downsized, the grade structure at the new headquarters would be
similar to that of the region where it was located.  DLA headquarters
officials told us that in setting up the new distribution
headquarters at New Cumberland, they expect to restructure and
downgrade positions at that location to meet the requirements for the
new organization.  To what extent this would lessen the higher
personnel costs for the eastern location is unclear given employee
bumping rights and save pay provisions that would likely be
associated with such restructuring.\6 Also, the importance of
personnel costs as a decision factor should not be minimized since
savings in this area can mean the potential for significant recurring
savings in the long term.\7 The results of the second data request
showed that the eastern location had higher average grade levels,
resulting in a $3.1 million difference in personnel costs between the
two locations over a 5-year period.  (See app.  I.)


--------------------
\6 Employee retention through the exercise of reduction-in-force
bumping rights can add significantly to personnel costs because civil
service procedures allow employees moving into lower graded jobs to
retain the pay of their former graded positions.  See Defense Force
Management:  Challenges Facing DOD as It Continues to Downsize Its
Civilian Work Force (GAO/NSIAD-93-123, Feb.  12, 1993). 

\7 Savings in personnel and base operating costs have provided a
major component of total savings from the BRAC process and have often
helped to offset any investment costs required to implement BRAC
decisions. 


         REAL PROPERTY MAINTENANCE
         COSTS
-------------------------------------------------------- Letter :3.4.2

The selecting official had the facilities engineer responsible for
the third data request reassess real property maintenance costs.  The
selecting official told us that he did this in the interest of
obtaining more realistic data.  For the second data request, the
steering group used the Navy Public Works Center real property
maintenance estimates used in the 1995 BRAC process.  These estimates
had been reviewed in 1995 by the DOD Inspector General, who found the
procedures used to be reasonable and the cost estimates to be
consistently generated, generally supported, and reasonably
accurate.\8

According to the facilities engineer responsible for the second data
request, the data had also been used in DLA's real property
maintenance project development, budgeting, and execution processes. 
The DLA facilities engineer made minor changes to the Navy Public
Works Center data as part of his data validation efforts before
approving their use for the second data request analysis. 

The selecting official told us he believed that the data used in the
second data request were not realistic, based on his personal
knowledge of the conditions of the two sites, past experience in the
distribution area, and knowledge of flaws in repair and maintenance
data used for BRAC.  He reasoned that the BRAC data were not
comparable because the Navy Public Works Center had different people
with different perceptions and evaluation criteria assess the
individual sites.  For example, he said the eastern location's
database included about $47,000 for the cost of painting a building
with a dryvit exterior, which, according to the facilities engineer,
does not need cyclic painting.  However, the facilities engineer had
already removed this item from the eastern location's database during
his validation of the second data request. 

A DLA facilities engineer, not previously involved in the process,
reassessed real property maintenance costs for the third data
request, producing a significant change in costs between the two
locations.  The results of the second data request had shown a
difference of about $643,000 in the real property maintenance
category in favor of the western location.  The results of the third
data request produced a difference of about $182,000 in favor of the
eastern location. 

For the third data request, the facilities engineer went one step
further than the previous engineer and had the sites submit
justification for removing additional projects from the Navy Public
Works Center database.  The projects submitted for removal included
some cyclical projects that the respondents did not believe were
needed during the 5-year time frame covered by the analysis.  The
eastern location submitted and received approval for removing about
$791,000 worth of projects.  According to western location officials,
they had one such project, valued at $95,000, but they did not submit
it because their efforts to have it removed during the second data
request were unsuccessful.  Both facilities engineers said that if
the western location had resubmitted this project for the third data
request it would have been considered and may have been removed.  The
removal of this item alone would not have significantly impacted the
cost or point spread in the final analysis of the third data request. 

However, the facilities engineers later identified an error of about
$210,000 in the western location's database for costs that KPMG
agreed should have been excluded.  The correction reducing the
western location's real property maintenance costs was not made in
the final analysis because, according to KPMG, it was identified
after the third data request analysis was completed.  Although these
reductions--the possible removal of the $95,000 project and the
$210,000 correction--in the western location's costs would have
changed the dollar and point spread advantage for real property
maintenance costs from the eastern site to the western site, they
would not have been enough alone to effect the overall outcome of the
study.  (See app.  II.)


--------------------
\8 Department of Defense Inspector General Report No.  95-241,
Summary Report on the Audit of Defense Logistics Agency 1995 Defense
Base Realignment and Closure Data Collection Process (June 19, 1995). 


         NEW FACILITY REQUIREMENTS
-------------------------------------------------------- Letter :3.4.3

The selecting official requested revisions to the requirements for
furniture and space.  These changes significantly affected the
relative position of the competing locations within the facilities
and information technology cost category.  The eastern location's
cost advantage in the investment cost category, which includes
facilities and information technology costs, went from only about
$19,000 in the second data request analysis to about $1.7 million in
the third data request analysis.  The facilities engineer who
developed the third data request told us that the requirements were
instituted to ensure a level playing field.  However, some steering
group members disagreed with the changes and told us that the
requirements gave an advantage to the eastern location, which already
had modular systems furniture required by the new data request. 

In developing the second data request, the steering group had voted
to disregard the selecting official's direction for the new
headquarters to include modular systems furniture.  While DLA
officials told us that such furniture was used at two other newly
renovated sites, various steering committee members told us they did
not believe that an official standard requiring such furniture
currently existed within DLA.  Nonetheless, this became a requirement
under the third data request, consistent with the selecting
official's earlier guidance.  The facilities engineer who developed
the data request required that the competing regions resubmit floor
plans to include the systems furniture and stipulated that its life
expectancy not exceed 10 years within the 5-year time frame covered
by the analysis.  As a result of these new requirements, the western
location submitted a cost of about $901,000 for purchasing new
systems furniture in the third data request because it could not
verify the age of the systems furniture stored in its warehouse. 

Moreover, in addition to requiring systems furniture, the third data
request included other new requirements such as a minimum of 22
meeting rooms and floor-to-ceiling walls for conference rooms and
offices to meet what the facilities engineer described as an
idealistic view of what DLA offices should look like.  The engineer
said that he used these requirements to ensure that both proposals
would be based on comparable work space.  Neither the steering group
nor officials from the competing locations agreed with all of these
requirements.  For example, they protested the need for 22 meeting
rooms, calling it excessive and wasteful.  Officials at both
locations told us that they currently had more people with fewer
meeting rooms and had encountered no difficulties in doing their
work. 


         IMPACT OF CHANGES
-------------------------------------------------------- Letter :3.4.4

The changes made for the third data request analysis had the effect
of improving the position of the eastern location, including shifting
the cost advantage (on an absolute dollar basis) to the eastern
location.  However, it should be noted that these costs were largely
one-time costs that could easily be offset over time should there be
significant recurring savings in another cost area, such as
personnel.  Steering group members told us they had no role in the
third data collection effort.  They received the results of the
second and third data requests on the same day, September 15, 1997. 
According to the steering group's minutes, the group accepted the
results of the third data request because the request was at the
discretion and authority of the selecting official.  However, they
cautioned the selecting official that the process of the third data
request would appear biased to outside parties, considering they had
not been consulted regarding this phase.  The results of the third
data request analysis showed that the eastern location scored much
better in both cost and point totals than the western location--the
eastern location was the least costly by about $2.1 million. 
However, the results of the second data request showed the western
location was the least costly by about $3.8 million.  Again, because
of problems identified in the process, we could not validate either
set of data. 


   NO EVIDENCE TO VALIDATE CLAIMS
   OF PREDETERMINATION
------------------------------------------------------------ Letter :4

Allegations had been made that DLA officials had selected the eastern
site for the Defense Distribution Center before the site selection
study took place.  We found no evidence to validate concerns that the
site selection decision was predetermined.  Previous studies examined
the consolidation issue but left the two regions intact.  We found no
evidence that the prior studies influenced the current site selection
process or outcome. 

DLA officials told us they had considered consolidating their
regional distribution headquarters for a number of years and had
eliminated one of three regional headquarters in 1993.  Subsequently,
they had studied options for consolidating the two remaining regional
headquarters; however, the study was not finalized and no proposals
or recommendations were approved.  The issue of consolidating the two
regions had also been separately addressed as part of DLA's BRAC
deliberations in 1995.  Even though DLA's BRAC 1995 assessment
emphasized the importance of retaining two regions, we learned that
following BRAC 1995, DLA officials once again began revisiting the
issue and began another study.  A DLA official told us that the
post-BRAC study was justified because the 1995 BRAC process had
produced decisions to close six depots.  (See
fig.  1.) However, a DLA headquarters official told us that, although
a preliminary assessment was made, this study was not completed and
no report was issued. 

We were provided documents that various officials from the western
location said raised concerns about whether the decision had been
predetermined.  However, we found no evidence to support that the
information provided in these documents reflected the official
position of DLA or influenced the current site selection process. 
For example, a 1995 briefing document from a previous study indicated
a planned future staffing level of 387 at the eastern location and
the phaseout of staff at the western location.  According to DLA
headquarters officials, the briefing document was preliminary and
this study was not finalized. 

The selecting official and DLA officials associated with the most
recent consolidation study told us that all the previous studies were
outdated, given changes in DLA's structure.  Thus, the selecting
official said that he did not consider them in the most recent study
effort. 

Additionally, claims were made by DLA officials at the western
location that actions were taken to better position the eastern
location in the competition for the consolidated distribution center. 
These actions included DLA's moving a general flag officer, its
Defense Distribution Systems Center, and the DLA Operations Support
Office to the eastern location in 1996.  While this may have given
the appearance that the eastern site was being preselected, we found
no support indicating that this was considered in the site selection
process.  Moreover, DLA officials told us the flag officer would have
moved to the western location if it had been the selected site. 


   CONCLUSIONS
------------------------------------------------------------ Letter :5

DLA's efforts to establish a steering group and formulate
decision-making criteria indicate that DLA recognized the need for a
credible process to guide its decision-making.  However, the process
used by DLA to support the site selection for its consolidated
distribution headquarters contained a number of weaknesses, and
raised questions about the soundness of the decision-making process. 
The evaluation was completed without adequate information concerning
facility requirements, which forced an assessment based on
hypothetical costs; technical weaknesses further skewed the results. 
Subsequent changes to the process, requested by the selecting
official, did not correct these weaknesses and created concerns about
the perception of bias.  Additionally, an incomplete assessment of
personnel costs minimized opportunities to fully assess the potential
for long-term recurring savings. 

Although various officials from the western location raised concerns
about whether the decision had been predetermined, we found no
evidence to validate that the information they provided to us
reflected official DLA positions.  Also, we found no evidence that
prior studies examining the consolidation issue influenced the
current site selection process or outcome. 



   RECOMMENDATION
------------------------------------------------------------ Letter :6

Because of the weaknesses in the process supporting DLA's site
selection decision and subsequent questions raised about the
soundness of the decision-making process, we recommend that the
Secretary of Defense independently and expeditiously reassess DLA's
site selection decision, taking into consideration issues and
questions raised in this report. 


   AGENCY COMMENTS AND OUR
   EVALUATION
------------------------------------------------------------ Letter :7

DOD provided written comments on a draft of this report, and they are
included in their entirety in appendix III along with our evaluation
of them.  DOD nonconcurred with the report's findings pertaining to
(1) insufficient data on personnel and facilities requirements, (2)
questionable cost comparison methodology, (3) subjective evaluation
of two noncost criteria, and (4) selecting official's requested
changes affecting the analysis. 

DOD noted that DLA could have made the site selection decision
unilaterally but chose to put a process in place that solicited input
from the Office of the Secretary of Defense, the DOD Inspector
General, DLA headquarters, and DLA regional experts.  It further
stated that DLA structured the evaluation process based on other
successful models (including BRAC) and adjusted it to accommodate the
special considerations felt to be important by representatives of the
sites most impacted.  We agree with DOD that sound and supportable
decision-making processes are needed in making consolidation
decisions.  Our concern is that the process DLA decided to use was
not well implemented.  In particular, it contained weaknesses in
methodology.  The cumulative effect of these weaknesses raised
questions about the soundness of the site selection process and the
ultimate decision.  We believe that the majority of issues raised in
DOD's response were already adequately addressed in our report and,
accordingly, we made only minor modifications to the report regarding
the requirement for systems furniture. 

DOD partially concurred with our recommendation.  While disagreeing
with the report's findings, DOD nonetheless agreed with our
recommendation that an expeditious review of the site selection
decision should be done, taking into account issues and questions
raised in this report.  However, DOD did not set a time frame for
doing so.  Also, DOD did not specifically address that portion of our
recommendation that stated that the Secretary of Defense should
independently conduct the reassessment.  We continue to believe it is
important that an independent and expeditious assessment be made by
the Secretary. 


   SCOPE AND METHODOLOGY
------------------------------------------------------------ Letter :8

To assess the soundness of the process DLA used to recommend and
select a site for the Defense Distribution Center, we reviewed
supporting documentation for the criteria, weights, and analysis used
in the selection process.  We interviewed all participants in the
process.  Participants included the

  -- Site Selection Steering Group--the steering group chair, an Air
     Force colonel in the DLA Logistics area; a DOD Inspector General
     representative; four representatives from DLA headquarters
     offices, four regional headquarters representatives, including
     two from each region; and two officials from the Office of the
     Secretary of Defense, including one from the Comptroller and one
     from the Logistics offices;

  -- KPMG Peat Marwick, LLP contractor personnel;

  -- DLA facility and installation officials involved in this
     process;

  -- the DLA selecting official, the Principal Deputy Director;

  -- the DLA Executive Director, Logistics Management;

  -- the Commander and Deputy Commander at Defense Distribution
     Region West and Defense Distribution Region East, respectively;
     and

  -- DLA officials from both Defense Distribution Region East and
     Defense Distribution Region West that responded to data
     requests. 

We visited both sites evaluated in the analysis and reviewed proposed
floor plans.  We traced and verified selected data inputs used to
support DLA's analysis to verify the reliability of selected DLA and
KPMG data validation.  We also reviewed documents from BRAC and other
DLA consolidation studies, as available, to compare methodologies
used.  Documentation associated with studies other than the BRAC
process was limited. 

To address the question of site selection predetermination, we
interviewed DLA officials who had participated in or had knowledge of
BRAC studies and DLA consolidation studies and reviewed documents
relevant to these studies.  We also interviewed participants in the
Defense Distribution Center site selection process to determine
whether they had prior knowledge of these studies.  Additionally, to
follow up on allegations of predetermination, we spoke to
representatives of the DLA Council of American Federation of
Government Employees union locals from both the eastern and western
regions. 

Given the sensitive nature of this assignment, we met with senior DLA
officials on two separate occasions to brief them on the results of
our work and to solicit their comments on preliminary drafts of this
report.  We incorporated their comments, as appropriate, to enhance
the technical accuracy and completeness of our report. 

We conducted our work from October 1997 to February 1998 in
accordance with generally accepted government auditing standards. 


---------------------------------------------------------- Letter :8.1

We are providing copies of this report to the Chairmen and Ranking
Minority Members of the Senate Committees on Armed Services and on
Appropriations; the House Committees on National Security and on
Appropriations; Members of Congress of the affected congressional
districts; the Director, Office of Management and Budget; the
Secretary of Defense; and the Director, Defense Logistics Agency.  We
will also make copies available to others upon request. 

Please contact me at (202) 512-8412 if you or your staff have any
questions concerning this report.  Major contributors to this report
were Barry W.  Holman, James R.  Reifsnyder, Kathleen M.  Monahan,
Jacqueline E.  Snead, and Gary W.  Ulrich. 

David R.  Warren, Director
Defense Management Issues


SITE SELECTION STEERING GROUP'S
DECISION MODEL RESULTS
=========================================================== Appendix I

                       East                   West               Difference
              ----------------------  --------------------  --------------------
                  Points                Points                Points
              received\a        Cost  received        Cost  received        Cost
------------  ----------  ----------  --------  ----------  --------  ----------
Cost criteria
--------------------------------------------------------------------------------

Investment costs over a 1-year period
--------------------------------------------------------------------------------
Facilities           300     $33,721       193     $52,494       107     $18,773
 and
 information
 technology

Operations costs over a 5-year period (discounted)
--------------------------------------------------------------------------------
Utilities,           148     971,159       160     897,935        12      73,224
 service
 orders,
 janitorial
Real                  34     931,700       110     288,204        76     643,496
 property
 maintenance
Personnel            123  55,111,403       130  51,968,016         7   3,143,387
Travel               100      24,062        95      25,371         5       1,309
================================================================================
Total cost           705  $57,072,04       688  $53,232,02        17  $3,840,025
                                   5                     0

Noncost criteria
--------------------------------------------------------------------------------
Work                  54                    70                    16
 environment
Commute time          26                    30                     4
Cost of               40                    40                     0
 living\b
Median house          30                    20                    10
 price
Crime index           20                     6                    14
Unemployment          10                     4                     6
 rate
================================================================================
Total points         885                   858                    27
--------------------------------------------------------------------------------
\a In this model, the maximum points that could be awarded to either
site was 1,000.  Each site's data inputs were evaluated.  The site
with the best outcome received 100 percent of the points allocated to
that criterion, while the other site received a fraction of the
points depending on the relative difference between the two scores. 
The lower scoring site's points equaled n * (a/b) where n equals
total points possible, a equals the lowest data value input for the
subject criterion, and b equals the highest data value input for the
same specific criterion.  For example, the eastern location received
100 points for travel costs, the maximum number points allocated for
this criterion, because it had the lower cost.  The western location
received 95 points, which was determined by dividing $24,062 by
$25,371 and multiplying the result by 100 points. 

\b Both sites were given equal points for the cost of living because
comparable data were not available, according to the steering group. 

Source:  DLA. 


SELECTING OFFICIAL'S REVISED
ANALYSIS RESULTS
========================================================== Appendix II

                             East                West             Difference
                      ------------------  ------------------  ------------------
                        Points
                      received              Points              Points
                            \a      Cost  received      Cost  received      Cost
--------------------  --------  --------  --------  --------  --------  --------
Cost criteria
--------------------------------------------------------------------------------

Investment costs over a 1-year period
--------------------------------------------------------------------------------
Facilities and             300   $79,923        13  $1,795,6       287  $1,715,6
 information                                              22                  99
 technology

Operations costs over a 5-year period (discounted)
--------------------------------------------------------------------------------
Utilities, service         160   986,610       136  1,159,43        24   172,825
 orders,                                                   5
 janitorial\b
Real property              110   141,089        48  323,356\        62    182,26
 maintenance                                               c
Personnel\d                130                 130                   0
Travel                     100    24,062        95    25,371         5     1,309
================================================================================
Total cost                 800  $1,231,6       422  $3,303,7       378  $2,072,1
                                      84                  84                  00

Noncost criteria
--------------------------------------------------------------------------------
Work environment            54                  70                  16
Commute time                26                  30                   4
Cost of living\e            40                  40                   0
Median house price          30                  20                  10
Crime index                 20                   6                  14
Unemployment rate           10                   4                   6
================================================================================
Total points               980                 592                 388
--------------------------------------------------------------------------------
\a In this model, the maximum points that could be awarded to either
site was 1,000.  Each site's data inputs were evaluated.  The site
with the best outcome received 100 percent of the points allocated to
that criterion, while the other site received a fraction of the
points depending on the relative difference between the two scores. 
The lower scoring site's points equaled n * (a/b) where n equals
total points possible, a equals the lowest data value input for the
subject criterion, and b equals the highest data value input for the
same specific criterion.  For example, the eastern location received
100 points for travel costs, the maximum number points allocated for
this criterion, because it had the lower cost.  The western location
received 95 points, which was determined by dividing $24,062 by
$25,371 and multiplying the result by 100 points. 

\b For the third data request the facilities engineer changed the
threshold for service orders to clarify definition problems.  He
requested that the data request respondents capture the costs only
for the maintenance and repair projects greater than $2,000. 

\c This figure is overstated by about $210,000, due to an error made
in a Defense Distribution Region West database.  Defense Distribution
Region West officials asked for a correction, which would have
resulted in the maximum points going to the west.  According to KPMG,
the change was not made because it was identified after the third
data request analysis was completed, and given the large point spread
and dollar difference in the overall results, the correction would
not have had a significant impact on the outcome. 

\d The selecting official decided to give both sites equal points for
personnel costs, because he reasoned that the costs would be
identical after formation of the Defense Distribution Center,
regardless of the site chosen. 

\e Both sites were given equal points for cost of living because
comparable data were not available, according to the steering group. 

Source:  DLA. 




(See figure in printed edition.)Appendix III
COMMENTS FROM THE DEPARTMENT OF
DEFENSE
========================================================== Appendix II



(See figure in printed edition.)

Now on p.  6. 

See comment 1. 

Now on p.  7. 

See comment 2. 



(See figure in printed edition.)

Now on p.  8. 

See comment 3. 

Now on p.  8. 

See comment 4. 

See p.  9.
See comment 5. 



(See figure in printed edition.)

See p.  10.
See comment 6. 

See p.  11.
See comment 7. 

Now on p.  13. 

See p.  15. 


The following are GAO's comments on the Department of Defense's (DOD)
letter dated April 6, 1998. 

GAO COMMENTS

1.  We did not assume that each data request result would be
implemented without any changes.  Our point is that, while we would
not expect the selected site to implement the floor plan exactly as
submitted, we do believe the requirements should have been more fully
defined and shared with the data request respondents.  This was of
particular importance in this study because investment costs were a
major factor in the site selection criteria.  Likewise, better
clarity of personnel requirements by function could have led to
better estimates of space requirements and cost. 

2.  We agree that the site selection steering group was given the
responsibility to develop the criteria and weights for the decision
support model and followed its established process to do so. 
Although one can assign different weights to costs as compared to a
straight cost comparison, it is not a methodology that we have
typically seen in such analyses, and as noted in our report, steering
group members we spoke with could not recall their rationale for
using this approach.  Furthermore, both the Defense Logistics
Agency's (DLA) contractor and the DOD Inspector General advised
against such a methodology.  Varying weights can be assigned to
evaluation criteria in this type of analysis to distinguish the
relative importance of individual criteria, particularly when
distinguishing between cost and noncost variables.  However,
assigning different weights to individual cost criteria, reduced
DLA's ability to perform the most meaningful comparison of cost.  We
saw no reason that costs should not have been evaluated on a straight
comparison basis to provide a more accurate picture of costs. 

3.  On the basis of our discussions with steering group members and
our review of the site selection backup documentation, we do not
believe that the steering group had and used a valid basis for
evaluating work environment and commute time.  Each steering group
member ranked each of the sites between 1 and 10 for quality of work
environment and then commute time, with 10 being the most favorable
score.  Average scores were calculated and the highest average scores
received the maximum points in the site selection analysis.  The
weakness in this method was that the basis for the ranking was not
clearly established.  Work environment and commute time were not
clearly defined; had they been, the steering group might have
identified objective measures for assessing these criteria.  No data
were used, and group members' knowledge of commute times and working
environment was limited.  As a result some group members used
commutes from nearby hotels and based working environment on personal
working relationships.  Alternatively, we note that another DLA site
selection study, pertaining to the issue of consolidating cataloging
functions, used quantitative factors to assess quality of life at
work, including factors such as individual office space per person;
average commute time measured in average number of miles traveled;
availability of public transportation; types and numbers of amenities
such as day care, gym, and credit union; parking fees; and distance
to the airport.  Clearly, the method used in the current study raised
questions about the soundness of the method used to evaluate these
two criteria.  According to steering group meeting minutes, senior
DLA leadership expressed concern about the subjective evaluation of
these criteria.  Additionally, while it is true that these criteria
constituted only 10 percent of the total points assigned overall,
they constituted 50 percent of the noncost criteria. 

4.  To what extent time requirements did not allow for consultation
with the site selection steering group about the changes requested by
the selecting official is unclear.  The time constraint--making the
site selection decision by the end of the fiscal year--appeared to
have been self-imposed.  Additionally, we agree that comparable data
and information should have been used in the site selection process. 
However, to avoid questions about the objectivity of the evaluation,
standards need to be clearly stated and agreed to up front.  The
selecting official had approved the site selection decision support
model before data collection efforts were initiated to ensure the
objectivity of the process.  While it is within his authority and
discretion to make changes, he did not do so until after he saw the
results of the second data request analysis and did so without
consulting the steering group.  His actions did not correct the
weaknesses we identified but resulted in negating personnel costs,
reassessing real property, and establishing new facilities
standards--the impact of which dramatically altered the resulting
costs and point values in the site selection analysis.  These actions
make it difficult for us to be certain that DLA had the best
comparable data it needed for its analysis. 

5.  We agree that locality pay was not relevant after Fort Belvoir
was removed from the site selection process.  As DOD stated, the
locality pay was the same for both of the remaining sites.  However,
the steering group was correct in initially identifying personnel
costs as an important criterion.  The importance of personnel costs
should not be minimized since savings in this area can mean the
potential for significant recurring savings in the long term.  As our
report notes, the results of the second data request showed that the
eastern location had higher average grade levels, resulting in a $3.1
million difference in personnel costs between the two locations over
a 5-year period.  While it may be difficult to project bumping rights
along with voluntary early retirement and separation incentive pay,
it can be done.  For example, DLA officials planned to conduct a mock
reduction-in-force to determine the effects on personnel, but had not
yet done so.  It should also be noted that, absent definitive data,
DLA and other DOD components previously used standard factors in
prior base closure rounds to project some personnel impacts and
costs. 

6.  While we agree that the real property maintenance data should
have been comparable, DLA's site selection backup documentation
indicated that the data were reviewed and some modifications were
made to it to ensure comparability between the two competing
locations before it was used in the second data request analysis. 
Our concern relates to the decision-making process.  The DLA Chief of
the Real Property Maintenance Team was approved by the steering group
as the facilities engineer responsible for validating the data.  He
told us that he validated the data as a routine matter of prudent
facilities engineering management.  During his data validation of the
responses to the second data request, he removed the requirement for
the eastern site to repaint a nonpaintable exterior surface. 
Subsequently, the selecting official told us that he based his
decision to reassess real property maintenance on his personal
knowledge and experience.  Having the data reassessed after they had
already been validated raised concerns among various steering group
members about the perception of bias. 

7.  DLA officials suggested that not requiring systems furniture and
other facilities requirements would result in a substandard work
environment and indicated that these requirements were used at two
other DLA locations not part of this site selection process.  We have
modified our report to reflect DLA's point about these other
locations.  However, as noted in our report, members of the steering
group told us they did not perceive this as an official DLA standard
for furniture and workspace. 

*** End of document. ***