After a 2009 law increased centralization of defense acquisition decisions, reform began to swing in the other direction starting around 2015. The new emphasis is on agility and innovation. It seeks to bypass tedious processes in hopes of bringing advanced capabilities to the warfighter faster and at lower cost. The Army has moved in lock step by creating Futures Command which puts a cultural emphasis on experimentation and prototyping. But many observers wonder whether the overarching reforms will produce lasting change or whether they are just another swing of the reform pendulum. By examining analogous reform efforts from 1969-1971, this post finds that the problem of lining up funds is the critical weakness in acquisition reform. It argues that the 2015-2018 reforms may grind down unless they conquer new territory: the program budget.
Enter, Ostrowski
Deputy to the chief of Army acquisition Lt. General Paul Ostrowski described the governing principles of the new acquisition processes in his service. They can be summarized in four major points:
- Cross-Functional Teams (CTFs) will coordinate technology handoffs between Futures Command and the Program Executive Offices (PEOs). With his charge of the CFTs, General Ostrowski reports to two executives, the Army acquisition chief (his formal boss who controls the PEOs and reports to the civilian Secretary) and the commanding general of Futures Command (his informal boss who reports to the military Chief of Staff).[1] While dual-hatting Ostrowski may upset the ‘proverbs of administration,’ it offers a solution to the technology transition problem not yet resolved at the Office of the Secretary of Defense (OSD) level.
- Keep as much effort “left” of Milestone B as possible. Milestone B is the critical decision point that takes a promising design into full-scale development. Keeping left means more work on experiments and prototypes to reduce risk before a project is handed off to the PEOs. It implies relatively more emphasis on Futures Command functions than had existed.
- Delegate the Milestone Decision Authority of smaller programs to the PEOs. Empowering the PEOs to authorize programs into new stages of acquisition cuts a fair amount of the bureaucratic approval process. Information, however, must continue to flow along the service headquarters and OSD channels.
- Use streamlining authorities provided by Congress. Rapid prototyping and rapid fielding pathways are available for “Middle Tier” acquisition programs, exempting managers from the 5000-series regulations and a lengthy requirements process. A key element of such pathways is the use of Other Transactions Authority (OTA), which exempts certain types of work from contract regulations.
All four reforms embarked upon by the Army from 2015-2018 have strong analogs in the Packard era reforms. After a swing toward extreme centralization in the 1960s, David Packard initiated a bout of decentralization from 1969-1971. Program decisions were delegated to the services through “participatory management.” “Fly before you buy” prototypes were emphasized as a best practice. Before its expansion, the 5000-series streamlined bureaucracy. The Army even recognized the technology transition problem at the time. It created “Task Forces” to build a bridge between the technology centers and program offices. The handoff phase is called the “valley of death” because so many attempts have failed.
The Acquisition Process
Just five years after Packard initiated his reforms in 1969, the Army released a critical review of the acquisition system.[2] In it, two clashing viewpoints from either side of the “valley of death” are immediately recognizable. On the left-hand side, the panels from requirements & concepts and science & technology reported on unproductive processes. They concluded that “The professional military officer has been thoroughly trained to try to bring order out of chaos. The R&D process is by nature disorderly and attempts to regularize and institutionalize it will usually stifle it.” As for cost-effectiveness analyses, the most important consideration in progressing to full-scale development, they “are of spotty quality and… can also be distorted in presenting results.”
On right-hand side, the production and costing panels had quite different views. They found that “There is a need within the Army for realistic acquisition planning early… The Army should take the necessary steps to emphasize the necessity of establishing a highly visible baseline cost estimate which will remain with the project throughout the review cycle and the subsequent acquisition process.” Those on the right of the “valley of death” stressed process and analysis represented by the “requirements pull” approach. Those on the left stressed agility and creativity found in the “technology push” approach. There is little doubt which side would be victorious when the competing cultures clash at full-scale development.
Before clashes occurred, the reform pendulum swung back towards tight control of early acquisition planning in 1976 with OMB Circular A-109 and in 1977 with the institution of Milestone 0. Not even subsystem prototyping was permitted without authorization from the Secretary of Defense. The bureaucratic process of consensus building quickly reestablished itself, even though layering of decisions was a unanimous problem not only for the Army, but Navy and Air Force. As the Army colorfully illustrated:
“OSD is now hydra-headed. Questions pour out of these many heads. The questions can overlap, or deal with the same issues. They appear not to be coordinated at OSD level. The result is tri-service organizational entropy gain.”
Admiral Hyman Rickover from the Navy immediately recognized Packard’s flaw. He wrote a memo to Packard on May 22, 1970, arguing that layers of bureaucracy still expected to evaluate and approve most decisions. After all, they controlled the money. A tremendous flow of information was necessary to justify funds even though the information was of no use to the “doers.” For example, the Army Bushmaster program underwent 17 separate cost-effectiveness analyses before final approval. Rickover found that attempts to streamline the acquisition process merely create new systems and documents to “prove” compliance with the directives so that funding would be made available.[3]
The Program Budget
The Packard era reforms sought to purge the Defense Department of centralization and “whiz kid” systems analysts yet failed to address the most important aspect of that system: the program budget. When Robert McNamara entered the DoD, systems analysis and program budgeting were two sides of the same coin. One could not be performed without the other. They joined under the Planning-Programming-Budgeting System (PPBS).[4] The program budget was devised as a means for unified control of plans and activities. Before then, budgets merely controlled funding to organization and object, or the means through which programs would be accomplished. This meant that the “program” emerged as an historical artifact. By contrast, the PPBS is a future directive of program policy.
The two major problems with the PPBS are classification and time. First, classification means that the budget must be prepared by program which is at cross-purposes with organization. The Army once had strong in-house technical services that had organic appropriations in the budget. The program budget quickly led to the abolition of their statutory role in 1962. By 1974, the remnants of the Army technical services complained that their funding was never assured but came from bits and pieces of program funding decided by outsiders. Their demise left the Army with little technical knowledge with which to buy smartly. Second, the program budget requires two years lead time for programming to accommodate the one year allocated for budgeting procedures. When program expenditures occur several years after funds are appropriated, plans become locked-in and redirection is difficult. The program budget depends entirely on the accuracy of long-range forecasts, which, as time has shown, are woefully inadequate for directing future technologies.
The disconnect between the budget and acquisition processes is a longstanding problem. Milestone decisions do not authorize program funding, they must be anticipated by two years to make money available. This is often impossible simply because the information doesn’t exist so far in advance. For smaller programs, waivers are available. However, as General Ostrowski commented, no one applies. “It’s easier to do the documents than to ask for the permission not to. Because that’s the culture that you’ve been raised in.”[5] How long efforts can go unprogrammed is opaque. Their funding sources are limited and ad hoc. For fiscal year 2017, Congress made nearly $1.6 billion available compared to $500 million the year before.[6] On inflation-adjusted terms, the sums are comparable to the prototyping funds made available during the Packard era. It is unclear whether these sources will disappear as they had in the past. For all practical purposes, efforts then and now must be programmed into the budget before full-scale development. Layers of bureaucracy can quickly reinject themselves through control of funds.
Research, Development, Test & Evaluation (RDT&E) requires flexibility and agility. However, the program budget requires plans be set in stone and justified to layers of bureaucracy. Epistemologically, the program budget embodies antiquated notions of “scientific management” that stand in stark contrast to new and more powerful concepts of complexity and adaptation. The Army in 1974 recommended that “Line item attention to R&D projects [should] not begin until start of engineering development.” But critical uncertainties are still unresolved by that time. Roland McKean, one of the two founders of the PPBS, later admitted that he wanted the entire RDT&E appropriation exempted from programming.[7] Reforming the program budget, particularly for RDT&E, is the first and most important aspect to meaningful reform intended to change the culture of process.
Conclusion
National policy objectives and the warfighter’s life depend on a qualitative superiority of weapon systems. Yet centrally planned technology has proven an inadequate means for tapping the nation’s talent. Another paradigm finds that all RDT&E efforts are a search of the unknown, not just the science and technology activities. Failure is a precondition for success and redundancy insures viable alternatives. Accountability is best accomplished at the production decision when fully tested systems can be compared. Therefore, the entire RDT&E budget appropriation should be organizationally based with programmatic accounting. This does not solve the “valley of death” problem which also plagues the private sector. Yet there is no singular procedure for solving technology transitions. Diversity in RDT&E and selection in production acts as a filter, exposing which organizations performed the complex tasks well and which did not. Selection prior to full-scale development requires nothing short of an oracle.
Endnotes:
[1] Freedberg Jr. (2018, September 23). “Futures Command Won’t Hurt Oversight, Army Tells Congress.” Breaking Defense.
[2] U.S. Army. (1974, April 1). “Report of the Army Materiel Acquisition Review Committee: Volume II.”
[3] “Weapon Systems Acquisition Process.” (1971). Hearings, Committee on Armed Services U.S. Senate.
[4] Hitch & McKean. (1960). The Economics of Defense in the Nuclear Age. RAND Corp.
[5] Bold. (2018, January 25). “Ostrowski Outlines Army’s Plan to Streamline Acquisition.”
[6] Judson. (2018, April 9). “New Army Acquisition Chief Takes on Rapid Buying.” Defense News.
[7] McKean & Anshen. (1967). “Limitations, Risks, and Problems.” Program Budgeting. Second Edition.
Leave a Reply