Podcast: Defense management reform with Peter Levine

Peter Levine joined me on the Acquisition Talk podcast about his new book, Defense Management Reform: How to Make the Pentagon Work Better and Cost Less. He was a career professional staff member for Congress and former Deputy Chief Management Officer. During our wide-ranging conversation, we discuss:

Why the DoD is more like an economy than a business
The balance between experimentation and discipline
Views on Middle-Tier and OTAs
How budgets can be cherry-picked to meet a strategy
The assertion of civilian control

Peter argues that the 2009 Weapon Systems Acquisition Reform Act (WSARA) should be viewed as a huge success on its own terms. The 1990s emphasis on deregulation and commercial item contracting was extremely important for less complex procurements, but created problems for major programs. Too often major programs were initiated without buying down sufficient technical risk through experimentation and analysis. This led to a great deal of cost growth in the 2000s.

By adding discipline, such as raising the status of independent cost estimates, programs started after WSARA have shown far greater stability and less cost growth. But there are no silver bullets, Peter reminds us. The valid criticism of WSARA is that it brought stability at the expense of innovation. This led to the 2015-present reforms re-emphasizing rapid acquisition, iterative development, and commercial procedures. These concepts are not new, and while they may apply well to software efforts, they does not obviate the need for cost, schedule, and technical baselines ahead of Milestone B.

Podcast Annotations

There’s a ton to learn from Peter throughout. Here he is on the Pentagon audit:

The thing you need to think about is, why are you trying to get to auditability? In the private sector, a financial statement is used to place a value on a business. That’s important to investors. The Department of Defense doesn’t have investors. Whether the Department of Defense is worth $1.5 or $2 trillion dollars isn’t important. What is important is the question of the quality of the defense it can provide.

Another thing a financial statement does is help you measure profit and loss. You assess profit and loss by comparing purchase prices to sales prices, basically. Well we’re not actually selling defense assets so we don’t have a sales price to compare things to. What we have is a military value.

That doesn’t mean it’s not important for the Department to keep good books, but the focus of those books shouldn’t necessarily be an auditable financial statement.

Peter suggests that some of the resources devoted to the audit could be more usefully employed in other information systems, including acquisition and human resources. One of the things the DoD IG hammered the Pentagon on recently was how the service acquisition executives for not knowing the number or cost of ACAT II and III programs.

The Defense Acquisition Visibility Environment (DAVE) is the primary IT system for acquisition data, including DAES and SARs. But it is only complete for MDAP programs. Lower level programs have often escaped submitted program cost, schedule, and other details. New 5000-series documents seem to be placing emphasis on registering programs through DAVE. Of course, renewed emphasis on reporting and review may go against the intention of ACAT II and III program delegation to the PEO/PMO level.

This runs into the concept of centralized policy and decentralized execution. Since the Melvin Laird years, the way programs have been devised is that the Joint Staff creates military policy from the top-down, the services create programs plans from the bottom-up, and OSD then reviews and approves program plans. Peter identifies a couple problems with the model. First, on the incentives for the services to be over-optimistic in order to get a “foot in the door” on a new program:

That’s one of the things we did in WSARA and in Goldwater-Nichols was look at why we have these cost overruns and what is the driver. It seems to be that the problem is in the inception of the program where we make these over-optimistic assumptions and then there’s major incentives in the system to be optimistic, so you have to address that upfront. That’s why we created the independent cost estimates a few decades ago and made that a major emphasis in WSARA.

And second, how program line items can be “cherry picked” to match military policy rather than devising programs from the policy:

The PPBE is a top-down approach, taking a strategy and figuring out what do I need to meet that strategy… The problem I see is that we’ve stuck with a bottom-up approach to building the budget and the POM. We have this National Security Strategy and we have this National Defense Strategy we tell the services to go build a POM.

The services go down to all their stovepipes and say, ‘tell me what you need,’ and then that comes up to OSD which tinkers at the edges. We look at what’s in a budget, and we match it to defense objectives, and we say ‘I can find these things that match this, these that match this.’ Well, I can do that with any budget… cherry-pick it to match the National Defense Strategy.

Peter says there will never be a “blank slate” to redesign the Pentagon from the top down, but “it’s important to have some, at least more, of that top-down approach so you can shape the defense program, rather than let the defense program shape you.” He points to a great article from CNAS’s Susanna Blume and Molly Parrish, Make Good Choices, DoD. There, they recommend a Bishop’s Fund, where OSD would retain some percentage of funding to program joint priorities. Peter thinks that this concept has merit for navigating a middle-ground between strictly top-down or bottom-up initiative.

Here are a couple other good lines:

We have to walk and chew gum at the same time. We have to be able to be evolutionary and revolutionary. We have to be able to experiment and have the discipline to build things that are complex. We have to be able to to multiple different things that take multiple approaches.

And: