Is there really $125B in Pentagon Waste?

Imagine you're the Deputy Secretary of Defense, and in walks a group of consultants who tell you that you can save $125 billion over five years if you act on their recommendations. They say you need a clear science and technology strategy, to establish rigorous management processes, and so forth. They say savings will start immediately. You flip through their PowerPoint brief stacked with charts and buzz words like optimize, modernize, and agile enterprise. You get weary when a slide shows, in pictographs, how you can re-program the efficiency savings to 50 brigades, 10 carrier strike groups, or 83 F-35 fighter wings over the same five year time frame. Slow down, you think. You don't want Congress getting a hold of the idea that there is $125 billion in waste can disappear overnight because you know there are no easy answers of the kind peddled by consultants.

That is, unless you're Deputy Secretary of Defense Bob Work. As Craig Whitlock and Bob Woodward reported in the Washington Post, Mr. Work did in fact get worried when he saw the chart depicting the savings. He got worried, however, that the DoD could easily save those billions and that Congress would want it back! The journalists wrote that the report by private consultants unambiguously laid out "a clear path" to $125 billion over five years. Former chairman of the Defense Business Board Robert L. Stein called the study's data "indisputable" because it came from the Pentagon itself. Because the DoD feared that Congress would "use the findings as an excuse to slash the defense budget," officials decided to suppress the "internal study that exposed $125 billion in administrative waste."

Frank Kendall, Undersecretary of Defense for Acquisition, Technology, and Logistics (AT&L), fired back that the consultants' study was "shallow" and failed to grasp basic obstacles to restructuring the public sector. The Washington Post journalists shrugged at the response and suggested that Pentagon officials were just trying to "wait things out" rather than tackle the hard issues at hand. It makes one wonder whether the journalists actually read the report. Perhaps a brief foray into the 125 billion-dollar report will give the reader a feel for the argument at hand.

The report, little more than a deck of PowerPoint slides, starts out by showing potential savings based on projected productivity gains. The consultants assumed 7% annual productivity gains to get to the $125 billion, whereas 3% annual gains only get you $60 billion in savings. Why 7% productivity? Because private sector industries "commonly show similar gains." Never mind the fact that economists generally agree that productivity has been flat for years. The industry benchmarks used in the report are not representative of defense support activities. Productivity in "office supplies, stationary, and gift store" is not representative of productivity in defense, even if they are DoD inputs. Moreover, there is no generally acceptable way of measuring productivity in the DoD, even for support functions. This is because the defense "marketplace" is only a loose facsimile of the kind you read about in economics text books.

Supposing that the DoD could achieve 7% productivity gains, how would the consultants recommend the DoD gets there? Productivity gains amounting to $46-89 billion can be achieved through "contract spend optimization" which requires more rigorous vendor negotiations, aggregating spending to gain economies of scale, increasing productivity, and eliminating "gold plating." This is basically all they had to say on regarding half of the total savings. The supposedly specific and actionable recommendations merely restate the objectives of existing job functions. For example, to anyone familiar with defense vendor negotiations, it is hard to believe that regulations on contracting officers could get any more rigorous. Lawyers are brought into the discussions at ever earlier stages in preparation for award protests. In short, the recommendations are empty.

In almost all cases, productivity gains were implied by consolidating information technology (IT) infrastructure and leveraging its output, big data. But even smaller firms find implementing IT solutions difficult, fraught with unforeseen challenges and cost overruns. Striving for bigger and better data is often a fool's errand, even for relatively simple operations.

An early example of the DoD implementing streamlined IT efficiencies came during the Korean War. The Munitions Board attempted to centralize the inventory data of machine tools using IBM punch cards to optimize allocations. The attributes of each machine were recorded and contractors could read the cards to locate their tools instead of doing on-site inspections. But machine tools are more complicated than the Munitions Board had assumed. The IBM cards completely failed to convey important nuances in machine tools leading to expensive errors and lost time. The initiative and countless other more like it sought to strip away complexity and replace it with the pristine order that consolidation imposes.

The out-sized share of the PowerPoint went to specific discussions on IT consolidations, whereas the consultants hand-waved issues of contracting and civilian fringe costs. This is concerning because the Task Group Chair, Mr. Phil Odeen, also has positions at Booz Allen Hamilton and Globant, both of which are IT services providers. According to the meeting minutes, he did most of the talking. Was the report a well-intended set of recommendations, a business pitch, or both?

The easiest way to see that the consultants were blowing smoke comes at the end. They write that "Technical solutions... will not, by themselves, achieve the savings." The technical solutions are a tool that must be supported by a culture change. For example, there needs to be "strong, consistent top leadership," "clear decision-making authority" and "defined accountability at all levels." Here, the consultants pay lip service to what really matters, and reference their earlier work on culture change. The only problem is that their conclusions are antithetical to all present discussions on defense reform!

The consultants want more centralized planning and more rigorous regulations. Congressional leaders, such as Rep. Mac Thornberry and Sen. John McCain, are calling for decentralization and fewer regulations. Legislation is being discussed would devolve some of the planning and programming decisions from AT&L down to the services. Further, AT&L itself may decentralize, separating its R&D functions from acquisition and logistics. The Pentagon is in the awkward position of being pulled in two directions.

Yet being pulled in two directions is par for the course in the Pentagon. For example, the landmark 1986 Goldwater-Nichols Act attempted to decentralize operations, but, as Rep. Thornberry said, it "unintentionally fueled a runaway, outmoded bureaucracy." The public should take little solace, however, that any prospective legislation would turn out differently. As scholar Fred Thompson found, decentralization requires giving operating managers the "maximum authority feasible, or, in the alternative, subjecting them to a minimum of constraints." But Congress doesn’t trust the lower echelons and wants to increase constraints! For example, the legislators want binding unit cost and schedule targets earlier in the acquisition process, at Milestone A, which is at initiation of Technology Development or prototyping. (Note that the troublesome F-35 went through a Milestone A equivalent in 1996. Certainly that is too early to smooth over programmatic risks with extensive before-the-fact controls.)

Ultimately, one must acknowledge that the fundamental principles of complex operations are decentralization, redundancy, and self-organization. While policymakers admit the inherent complexity in defense and claim they desire to decentralize operations, they shy away from actual implementation because of its superficial illogicalities. Hence they attempt to centralize, eliminate duplication, and control operations through extensive regulations. While in many ways it mirrors the de facto way of doing business in the 1940s and 1950s, a return to direct budget appropriations to decentralized operating units who are controlled using after-the-fact evaluations would go a long way to effecting real defense reform.

But such decentralization means a loss of coherence at the top. It means you should be not expect a reasonable answer from seemingly simple questions. Sen. McCain, for example, fumed that "The DoD cannot tell us how many civilian contractors they employ, no one knows how many people work for the Department of Defense." While that line was trumpeted around as a clear indication that the DoD didn't have its management in order, one has to sympathize with the poor analyst in some cubicle in the Pentagon who knows that the data are flat out not there, nor should they be. Just for starters, are we talking about direct charging workers from prime contractors, or down to the second tier, or third, or fourth, and so forth? Do we include indirect workers? Or which types of indirect? Are we talking physical headcounts or full-time equivalents based on hours charged? The analyst would think, “For there to be any chance of having a clean answer at my fingertips, it would require far more comprehensive cost controls.” The Senator incentivized defense officials to cover themselves with more centralization and new regulations at the same time he cried that the DoD is too centralized and bloated with regulations.

The answer turned out to be about 740 thousand. But the better question is whether that matters. Is there any basis for judging 740 thousand contractors as too high or too low, even if the number were not arbitrary in some sense? The Sen. McCain’s statement makes clear that he views the defense enterprise as a vast administrative problem, where decentralized operations can be neatly aggregated to inform policy. In this way, the Senator champions the utopian view of consolidation and big data at the same time he champions decentralization and competition. At least the consultants who would save $125 billion on baseless assumptions were consistent, if shallow, in their views.

Defense Reform 2016?

Congress again returns to a beloved topic, defense reform, and its policy lineup lacks any thought or imagination.

For now, let's examine the appeal of decentralization. Rep. Mac Thornberry believes that this is done by allowing the services to have more say in planning and programming through an increased role on the requirements generation and acquisition milestone processes. He also would like to break up his staff in OUSD(AT&L) by bringing back the Director of Defense Research and Engineering (DDR&E).

Note there are two "decentralizations" going on.

First, there is a decentralization of power from the OSD staff to the operating services. More decisions will be made by service executives than staff. The focal point of decision making on the services will resemble a weakened form of what had existed between 1949-58 and 1969-86.

Second, there is decentralization in the OSD staff itself. OUSD(AT&L), which is in the line of command but not an operating unit, was itself the product of a merger between DDR&E and manpower and installations functions from Goldwater-Nichols in 1986. Where decision making focuses on the OSD staff, they inevitably will come down to the detail of operations, requiring greater scale and specialization. This process grows staff into a hierarchy within a hierarchy.

The old myth of the staff had it that it that staff wasn't an operating unit, nor in the line of command. It collecting information, advised, and evaluated on behalf of the commander, and had no authority of its own.

However, with the 1958 Reorganization Act, DDR&E was created in order to bypass the long chain where lower echelons in the staff have to run up to the Secretary of Defense and then back down the appropriate chain in the services. The staff to the Secretary gained "legal" authority to exercise power over directing the services.

While the services are mission oriented with their own straight-line hierarchies, the staff is organized functionally. The functions, such as R&D, logistics, etc., are shared by all missions. Numerous lines of command in the staff have authority affecting operating decisions in the service hierarchy but no responsibility for the resulting outcomes.

This is how you get, as Admiral Rickover once lamented, a six month delay "just because one staff person with no responsibility but with authority had on his own decided that the policy was wrong."

What is ironic is that the second decentralization, fragmenting OUSD(AT&L), doesn't really appear to be reducing its overall power. While it reduces its authority in some aspects, which are devolved to the services, it appears that the strong mandate pursued by DDR&E will provide it new areas of authority which, by being out front of the acquisition process, might grow those areas.