Earlier this month at the American Enterprise Institute (AEI), Secretary of the Navy Ray Mabus argued that those looking for savings in the defense budget should look to Department of Defense agencies he considers “pure overhead.” According to a report by Sydney Freedberg of Breaking Defense, Mabus claimed, “If you want to look at real money, 20 percent of the Pentagon budget – 20 percent, one dollar out of every five – is spent on the ‘fourth estate’[:] the Office of the Secretary Defense, the defense agencies, the organizations run by the undersecretaries.” While the Navy secretary was arguing in favor of cutting spending for department-wide agencies instead of the Navy budget—as though the two were mutually exclusive—he is onto something.

Mabus cited three agencies within the Pentagon that he believes are driving up costs:

Cutting DFAS seems like a no-brainer. “They write our checks,” Mabus complained at the AEI event, “We tell them who to write the checks to, we tell them how much to write them for, and they write the checks. Last year they charged us $300 million.” According to the agency’s website, it was established in 1991 to save money by streamlining accounting functions. But adding a bureaucratic layer between the original payer and ultimate payee seems like a counterproductive way to do so.

The DLA and DOT&E are more complicated. Concerning the latter, Mabus argued that the agency’s tests do not provide useful information. He specifically cited DOT&E criticism of the Navy’s troubled Littoral Combat Ship (LCS) program. Pentagon testing agency claims the LCS lacks survivability, but the Navy secretary countered that the testers do not take account of the ship’s ability to maneuver away from hostile fire. Moreover, he argued, the alternative that would meet DOT&E’s survivability standard was a far more expensive destroyer.

The Navy secretary’s criticism of the Pentagon’s testing agency was obviously self-serving, but he does have a point. There will always be a knowledge gap between those doing the testing and those who will eventually use any weapon. DOT&E has five different offices within the agency to deal with different types of weapon systems. But each office must have expertise on the concept of operations the military services will employ in all those areas if it wants to provide realistic testing. Moreover, there is always a level of technological uncertainty in the development of any new system.

On the other hand, the Pentagon’s testing office was a response to real problems in weapons acquisition. The creation of DOT&E was one of the few lasting achievements of the military reform movement of the 1980s. As political scientist Daniel Wirls explained his book Buildup,

Troubled and troubling weapons programs, including the M-1 tank and the Maverick missile, imparted renewed relevance to a long-standing aphorism of reformers: Fly Before You Buy. That is, rigorous testing of any weapon, preferably under battlefield conditions, should precede any procurement decision. The Office of Operations Testing and Evaluation law was to assure that weapons would undergo thorough and realistic tests prior to procurement decisions.

It is unclear what is the best path forward. Mabus raises a good point about DOT&E, but the services did not do any better when left to their own devices.

DOT&E is not the only place where gaps in information persist, but the path seems much clearer with regard to the Defense Logistics Agency. In his remarks at AEI, Mabus complained that the DLA is not producing the savings on fuel purchases that it should:

The theory is that the Defense Logistics Agency can buy fuel cheaper if they buy it for everybody at once. Well, the thing of it is we use different aircraft fuel than the Air Force because we operate in the maritime environment and they don’t we use different fuel for our ships than anybody else. So it’s really not these big bulk fuel purchases, it’s service-specific fuel purchases [done by an outside agency].

This problem is common with centralization at the Pentagon. Just as joint development of weapon systems fails to achieve economies of scale because the different services have different needs, economies of scale on bulk fuel purchases prove illusory for the same reason.

There is another problem with DLA’s fuel sales to the individual services though. The agency purchases oil at market prices and then sells it to the services at a fixed rate. When the rate is higher than the market price—as it was for much of the past year or so—the additional money goes into a “rainy day fund” meant to shield the military from fluctuations in the oil market. Again, in theory, this idea is understandable. If the price of oil spiked during a security crisis, the services could end up paying far more for fuel than they would have otherwise—or be left unprepared for operations. However, a fixed price gives the services no indication when they should conserve fuel for that hypothetical “rainy day” or when to allocate funds to other priorities.

Mabus might have had self-serving reasons for doing so, but he hit on something that costs taxpayers a great deal of money: the trend toward greater centralization of functions at the Pentagon since the end of World War II. Centralization was supposed to lead to savings. Instead, attempts to cut down on redundant functions shared by the military services only spawn new forms of redundancy. Knowledge gaps in the inherently uncertain process of develop technologically sophisticated weapons get exacerbated. And price signals that could lead to better allocation of funds disappear in attempts to find economies of scale that ultimately prove illusory.