A good-news story about cost control

I am working on an article for FCW about changes in the procurement system over the past 25 years, with the theme “has the performance of the procurement system improved?” Along the way, I have been looking into why cost growth in the development of new weapons systems has slowed.

Development of new military weapons and new IT systems are two of the biggest-ticket elements of government purchasing, though we spend far more on the first than the second. There are some big differences between the management approaches the government uses in those kinds of acquisitions, but weapons systems always and IT ones frequently require development of new technology beyond what has been done before.

This “never done it before” feature creates a situation with significant risk for cost growth and performance uncertainty.

Now, however, there seems to be pretty plausible evidence that the cost growth for developing new weapons systems has decelerated since 2009. That trend appears in evidence provided by the Defense Department and is confirmed by the Government Accountability Office.

I recently spoke about this with Frank Kendall, who was principal deputy undersecretary and then undersecretary of Defense for acquisition, technology and logistics (AT&L) from 2010 to 2016. Those efforts began when Ash Carter was in that job, before he went on to become secretary of Defense. (Full disclosure: Carter has returned to his old job at the Kennedy School, so we are colleagues.)

Kendall, a West Point grad with a master’s degree in engineering from Caltech, is a veteran of many senior career positions in industry and government, and he received a Presidential Rank Award while at DOD. We talked about a few areas that he believes were the most important in driving cost performance improvements in weapons system: performing affordability analysis upfront, conducting “should cost” analysis throughout the process and negotiating good prices for system production toward the end of the cycle. Some of them have potential applicability to IT acquisition.

One such change is affordability analysis, which is undertaken before a request for proposals is even issued. DOD often started programs without an affordability analysis and then cancelled them after lots of money had been spent because the original requirements busted the budget. DOD has now begun to require that programs make rough order-of-magnitude estimates of budgets for different activities a number of years out and use that information to set “affordability caps” for new programs officials are contemplating.

They then do a rough estimate of possible costs for various potential requirements during the “analysis of alternatives” process that occurs before DOD issues an RFP.

That is the responsibility of a DOD-wide Cost Assessment and Program Evaluation unit along with the program itself (which is suspected of being too optimistic in its estimates). A think tank may be brought in, and occasionally DOD will competitively hire several contractors to do a conceptual design for the requirement.

If an affordability cap is breached later on, the options include changing the requirements to reduce costs, finding another program whose affordability cap can be reduced to cover the increase or cancelling the program.

“I had to make clear,” Kendall told me, “that we would not just raise the caps; they had to be taken seriously to be effective.”

He believes that the single most important change in the past few years is the spread of should-cost analyses for programs under development. Such analyses are conducted every year and are filled with concrete suggestions for ways the programs might reduce costs -- such as introducing a second source for items being bought, redesigning features of a product or going after contractor IP upfront to be able to do competition downstream.

Acquisition executives in the three military services develop overall cost-savings targets for each program’s should-cost analysis, and program managers are supposed to track progress the same way they would with other performance metrics. In Kendall’s view, those analyses have been crucial to creating counterpressures -- a new performance metric -- among contracting or other acquisition officials. Without that new metric, they would focus only on the traditional metric of “getting money out the door,” which has created incentives against cost control. Should-cost analysis was, therefore, a key part of creating a culture that paid attention to cost control.

If one sees should-cost analysis less as a set of formulas or algorithms and more as an encouragement to think creatively about savings opportunities, the approach is fully applicable to IT systems. What is needed, as has occurred at DOD, is a requirement from senior agency leaders (CIO or undersecretary for management) to prepare such analyses and develop savings targets, which could be a performance metric for those programs.

An important barrier to engagement by DOD’s program managers was the traditional and deeply held fear that any savings a program realized would be scooped up by headquarters and not benefit the program in any way. AT&L decreed that savings would stay with each service to be used for other purposes by the program that saved the money (the approach the Army has adopted) or go to the service as a whole (which is what the Navy has done).

The final area for changes has been negotiating prices for the production of weapons systems after development has been completed. This element has the least direct relevance for IT because a number of years are spent developing weapons, which are then produced in quantities of perhaps several hundred based on the final design. Actual production is usually undertaken by the prime contractor that developed the system, and the government negotiates a sole-source fixed price with that contractor. The government has the option of injecting competition into the production process by bringing in a second source, but the knowledge the prime contractor has gained often makes this infeasible.

There is no counterpart to this in IT, where only one copy of a system is produced.

Instead, DOD has concentrated on developing information that could be used in negotiating the pricing of production contracts. Over the years, the department had developed more and more information on contractors’ costs based on the production cost data they were required to submit. AT&L concentrated on developing the ability to analyze the data to help officials during negotiations.

“We now negotiate all the elements of costs,” Kendall told me. “You go through and challenge everything that looks suspicious.”

After that comes hard bargaining, mostly about the profit rate the contractor will be allowed. “We often end up arguing about 1 to 2 percent of profit,” he said. “But for industry, that 1 to 2 percent is very important.”

On F-35 negotiations, the two sides reached an impasse. “We held the line and just said no,” Kendall added. “We were worried about precedent setting. Give them once, and they’ll want [it] again. We said, ‘If you don’t like this, take me to court.’” They didn’t.

Starting in 2009, DOD increased the number of employees in the acquisition workforce (not just contracting officials but program managers, cost estimators, engineers, etc.) after the cutbacks of the previous 15 years. That workforce is now 20 percent larger than it was in 2009. Kendall believes those new people were essential for his efforts. They provide resources for developing and managing should-cost analyses, determining negotiating positions and spending more time negotiating rather than concluding a deal too quickly because of resource constraints.

This is a good-news story about people inside the government, starting at the leadership level, taking seriously the challenge to control costs in an area that produces major benefits but also costs a fortune and has for many years been regarded as impervious to improvement. Kendall is concerned -- and he strikes me as in no way a political partisan -- that pressures are building in the Trump administration to restrict the role of the Office of the Secretary of Defense in this area and return more authority to the military services, which he believes could be devastating for cost-control efforts. (An official at GAO shared the same worry with me.)

Meanwhile, though, we should celebrate this new success in controlling costs. Some of it might be applicable to IT, but even what is not should make IT folks glad because all of us want a government that works better.