Evolutionary Acquisitions and Spiral Development in DOD Programs: Policy Issues for Congress
Evolutionary Acquisition and
Spiral Development in DOD Programs:
Policy Issues for Congress
Gary J. Pagliano and Ronald O’Rourke
Specialists in National Defense
Foreign Affairs, Defense, and Trade Division
Summary
The Department of Defense (DOD) in 2001 adopted a new approach for developing
major weapon systems, called evolutionary acquisition with spiral development
(EA/SD), as its preferred standard. EA/SD is intended to make DOD’s acquisition
system more responsive to rapid changes in military needs. EA/SD poses potentially
important challenges for Congress in carrying out its legislative functions, particularly
committing to and effectively overseeing DOD weapon acquisition programs. This
report will be updated as events warrant.
Background
Origin and Aims of EA/SD. DOD in 2001 adopted a new approach for
developing new weapon systems, called evolutionary acquisition with spiral development
(EA/SD), as its preferred standard. EA/SD, which is referred to informally (though not
entirely accurately) as spiral development, is an outgrowth of the defense acquisition
reform movement of the 1990s, and is part of DOD’s effort to make its acquisition system
more responsive to rapid changes in threats, technology, and warfighter needs. It is also
intended to increase DOD’s control over program costs, DOD program-manager
accountability, and participation of high-tech firms in DOD weapon acquisition programs.
DOD’s goals in using EA/SD are to:
!get useful increments of new capability into the hands of U.S. personnel
more quickly;
!take better advantage of user feedback in refining system requirements
and developing subsequent increments of capability;
!mitigate technical development risk in weapon programs that are to
employ new or emerging technologies; and
!facilitate the periodic injection of new technology into weapons over
their life cycles, so as to better keep pace with technological changes.
Description of EA/SD. Under DOD’s previous weapon acquisition method, now
known as single step to full capability (SSFC), DOD would first define a specific
performance requirement to be met, and then work, usually for a period of more than 10
years in the case of a complex weapon system like an aircraft or ship, to develop and build
a design that, upon first deployment, was intended to meet 100% of that requirement. The
core idea of EA/SD is to set aside the quest for 100% fulfillment of the requirement in the
initial version of the weapon and instead rapidly develop an initial version that meets
some fraction (for example, 50% to 60%) of the requirement. Field experience with this
initial version is then be used to develop later versions, or blocks, of the weapon that meet
an increasing fraction of the requirement, until a version is eventually developed that
meets the 100% standard.
Figure 1 below details the process for each block. Each block includes four phases
for conceiving, developing, producing, and sustaining (i.e., supporting) a weapon system.
Each phase is governed by certain acquisition rules and regulations, including entrance
and exit criteria, and is subject to the requirements process, including the Initial
Capabilities Document (ICD) and Capability Development Document (CDD). Each block
includes its own acquisition contracts and fully funded budgets for a defined time period.
Figure 1. The 5000 Acquisition Model Utilizing Evolutionary Acquisition
and Spiral Development
As shown in Figure 1, spiral development occurs as the second phase within a block.
Spiral development is an iterative process for developing a weapon system’s capabilities
in which the developer, tester, and user to interact with one another so as refine (i.e.,
spiral down to a specific understanding of) the system’s operational requirements. Spiral
interaction can change the course of a system’s technology development.
Although EA/SD differs from SSFC in its use of block development from the outset
of a program, from a program-management perspective, EA/SD is similar in some areas
to SSFC, including the development milestones and reviews that are used at each
development stage. EA/SD, however, is intended to be more flexible than SSFC in terms
of permitting changes in a program’s requirements or development path resulting from
changes in threats, technology, or warfighter needs. EA/SD is also intended to be more
flexible than SSFC regarding entry points into the acquisition process. Under SSFC, the
dominant entry point was the beginning. Under EA/SD, in contrast, programs can enter
various phases of any block (A, B, or C in Figure 1), depending on the maturity of the
program.
Under EA/SD, the final desired capability of the system can be determined in two
ways — at the beginning of the program, with the content of each deployable block
determined by well-understood (i.e., mature) key technologies, or along the way, with the
content of each block determined by success or failure in developing less-well-understood
(i.e., emerging) technologies or the evolving needs of the military user. Applying EA/SD
at the outset of large weapon acquisition programs, such as the ballistic missile defense
program, can create significant initial uncertainty regarding the design and ultimate cost
of the systems that will eventually be procured under the program, the number of systems
to be procured, and the schedule for procuring them. Applying EA/SD to other programs,
particularly those intended to develop more up-to-date subsystems for improving existing
weapons such as the F-16 fighter or M-1 tank, can produce much less uncertainty
regarding the program’s ultimate outcome.
Programs Using EA/SD. Although DOD used EA/SD for years on a somewhat
limited basis, DOD decided in 2001 that EA/SD would henceforth be the “preferred” (i.e.,
standard or default) acquisition strategy for all types of weapon acquisition programs —
newly initiated programs, existing programs for developing new weapons, and programs
for upgrading weapons already in existence. EA/SD was elevated in prominence that year
when DOD announced that it was applying EA/SD to its ballistic missile defense
program1 and that the Navy’s program for a new family of surface combatants would be
an EA/SD program.2
Several defense programs are now using EA/SD. The ballistic missile defense
program is a more complex case than others because it includes multiple weapon systems,
some existing and some in initial development, in different phases and blocks of
development. In addition, although the ballistic missile defense program has embraced
most of the EA/SD model (notably, the possible absence of ultimate cost and timeline
projections), it differs from other programs being pursued under EA/SD because it
operates under different oversight rules instituted in January 2002 by Defense Secretary
Rumsfeld.
GAO Report. A November 2003 General Accounting Office (GAO) report on
EA/SD prepared at the direction of the Senate Armed Services Committee (see
Legislative Activity section below) concluded the following:
1 Statement of Lt. Gen. Ronald T. Kadish, USAF, Director, Ballistic Missile Defense
Organization, on The Ballistic Missile Defense Program, Amended FY2002 Budget, Before the
Senate Armed Services Committee, July 12, 2001, pages 2-3, 6-8, 14. See also CRS Report
RL31111, Missile Defense: The Current Debate, coordinated by Steven A. Hildreth.
2 For more on these new surface combatants, see CRS Report RL32109, Navy DDG-1000
(DD(X)) and CG(X) Ship Acquisition Programs: Oversight Issues and Options for Congress, by
Ronald O’Rourke, and CRS Report RL33741, Navy Littoral Combat Ship (LCS) Program:
Oversight Issues and Options for Congress, by Ronald O’Rourke.
DOD has made major improvements to its acquisition policy by adopting
knowledge-based, evolutionary practices used by successful commercial companies.
If properly applied, these best practices can put DOD’s decision makers in a better
position to deliver high-quality products on time and within budget....
The next step is for DOD to provide the necessary controls to ensure a knowledge-
based, evolutionary approach is followed. For example, the policy does not establish
measures to gauge design and manufacturing knowledge at critical junctures in the
product development process. Without specific requirements to demonstrate
knowledge at key points, the policy allows significant unknowns to be judged as
acceptable risks, leaving an opening for decision makers to make uninformed
decisions about continuing product development.
DOD was responsive to the requirements in the Defense Authorization Act for
Fiscal Year 2003 [see Legislative Activity section below]....
This [GAO] report makes recommendations that the Secretary of Defense
strengthen DOD’s acquisition policy by requiring additional controls to ensure
decision makers will follow a knowledge-based, evolutionary approach. DOD
partially concurred with our recommendations. DOD believes the current acquisition
framework includes the controls necessary to achieve effective results, but department
officials will continue to monitor the process to determine whether other controls are
needed to achieve the best possible outcomes. DOD agreed it should record and
justify program decisions for moving from one stage of development to next but did3
not agree with the need to issue a report outside of the department.
Issues for Congress
EA/SD poses potential issues for Congress regarding DOD and congressional
oversight of weapon acquisition programs. Some of these issues appear to arise out of
uncertainty over how EA/SD differs from the SSFC approach; others appear to arise out
of the features of EA/SD itself.
DOD Acquisition Policy and Oversight. One issue for Congress, addressed
in the GAO report, is whether DOD has established adequate rules and regulations for
conducting internal oversight of EA/SD programs. Some observers have expressed
concern about this issue, particularly with regard to the spiral development phases of
programs. In support of this concern, they have cited budget justification documents for
the ballistic missile program, which have included some references to block development
but have provided incomplete information on how much funding is spent for specific
blocks, over what period of time, and on what progress has been made to date in each
block. Supporters of EA/SD argue that DOD is fully aware of the need for adequate
oversight and will take steps to ensure that it is provided. Potential questions for
Congress include:
!How does DOD oversight for EA/SD programs compare to DOD
oversight of SSFC weapon acquisition programs in terms of frequency
3 U.S. General Accounting Office. Defense Acquisitions[:] DOD’s Revised Policy Emphasizes
Best Practices, but More Controls Are Needed. Washington, 2003. (November 2003, GAO-04-
and nature of reviews, information required to be submitted to reviewing
authorities, and evaluation and reporting by reviewing authorities?
!Will DOD oversight procedures, and review bodies be the same for all
EA/SD programs, or will they vary from program to program?
Congressional Program Commitment and Oversight. Another issue for
Congress is how to carry out its responsibility to allocate defense spending. EA/SD poses
potentially significant issues for congressional oversight, particularly for newly initiated
weapon acquisition programs, in three areas:
!Ambiguous initial program description. Programs initiated under
EA/SD may not be well defined at the outset in terms of system design,
quantities to be procured, development and procurement costs, and
program schedule. These are key program characteristics that Congress
in the past has wanted to understand in some detail before deciding
whether to approve the start of a new weapon acquisition program.
EA/SD can thus put Congress in the position of deciding whether to
approve the start of a new a program with less information than it has had
in the past.
!Lack of well-defined benchmarks. A corollary to the above is that
Congress may not, years later, have well-defined initial program
benchmarks against which to measure the performance of the military
service managing the program or the contractor.
!Funding projections potentially more volatile. Although projections
of future funding requirements for weapons acquisition programs are
subject to change for various reasons, funding projections for EA/SD
programs may be subject to even greater volatility due to each program’s
inherent potential for repeated refinements in performance requirements
or technical approaches. As a result, any long-range projections of future
funding requirements for EA/SD programs may be even less reliable than
projections for systems pursued under the SSFC approach.
Supporters of EA/SD argue that it can improve congressional oversight of DOD
weapon acquisition programs because the information that DOD provides for a given
program will focus on the specific block that is proposed for development over the next
few years. This information, they argue, will be more reliable — and thus better for
Congress to use in conducting its oversight role — than the kind of long-range
information that used to be provided under the SSFC approach.
Under SSFC, DOD provided information about the entire projected program,
stretching many years into the future. Such information, supporters of EA/SD argue, may
appear more complete, but is not very reliable because it requires projecting program-
related events well into the future. DOD’s history in accurately projecting such events,
they argue, is far from perfect. As a result, they argue, information provided in
connection with an SSFC weapon acquisition program can give Congress the illusion —
but not the reality — of understanding the outlines of the entire program. On the other
hand, critics of EA/SD contend that it has the potential for drawing Congress into
programs to a point where extrication becomes difficult if not impossible, and without a
clear idea of a program’s ultimate objectives.
Potential questions for Congress and DOD regarding congressional oversight of
EA/SD programs include the following:
!What might be the impact, on congressional approval of new weapon
acquisition programs and subsequent congressional oversight of those
programs, of having limited initial detail in terms of system design,
quantities to be procured, procurement schedules, and total costs?
!How might congressional oversight of weapon development programs be
affected if program information with longer time horizons but potentially
less reliability is exchanged for program information with potentially
greater short-term reliability — but, without previously available, if
imperfect, estimates of full program costs?
!To what extent might DOD’s new preference for EA/SD be influenced,
as some critics contend, by the knowledge that it might relieve DOD of
the responsibility for providing specific answers to congressional
questions regarding system architecture, effectiveness, time lines, long-
term strategic implications and cost?
Legislative Activity
FY2007 Defense Authorization Act (H.R. 5122/P.L. 109-364). Section 231
of H.R. 5122/P.L. 109-364 (conference report H.Rept. 109-702 of September 29, 2006)
would, among other things, require DOD to review and revise policies and practices on
weapon test and evaluation in light of new acquisition approaches, including programs
conducted pursuant to authority for spiral development granted in Section 803 of P.L.
FY2004 Defense Authorization Bill (H.R. 1588/S. 1050). In its report
(S.Rept. 108-46 of May 13, 2003, page 346) on S. 1050, the Senate Armed Services
Committee expressed support for incremental acquisition and directed GAO “to assess
current acquisition policies and regulations and to determine whether: (1) the policies
support knowledge-based, evolutionary acquisitions; (2) the regulations enforcing these
policies provide the necessary controls to ensure the Department’s intent is followed; and
(3) the policies are responsive to concerns expressed by the committee in [P.L. 107-314].”
As discussed above, GAO submitted the required report in November 2003.
FY2003 Defense Authorization Act (H.R. 4546/P.L. 107-314). Section 802
of the conference report (H.Rept. 107-772 of November 12, 2002) on the FY2003 defense
authorization act (H.R. 4546/P.L. 107-314 of December 2, 2002) required DOD to report
on how it planned to apply to EA/SD programs certain statutory and regulatory
requirements for major DOD acquisition programs. Section 803 set forth conditions to
be met before a DOD acquisition program can be pursued as an EA/SD effort, and
required DOD provide annual status reports for the next five years on each research and
development program being pursued under EA/SD. Section 132 required the Air Force
to submit to Congress a list of programs that it had designated as acquisition reform
“pathfinder programs,” set forth conditions under which those programs can proceed, and
applied to them the requirement for filing status reports established under Section 803.
These provisions are also discussed on pages 455-456 and 667-668 of the report. The
Senate Armed Services Committee, in its report (S.Rept. 107-151 of May 15 [legislative
day, May 9], 2002) on the FY2003 defense authorization bill (S. 2514), included similar
provisions and commented extensively on the EA/SD process (see pages 94 and 333-335).