The Costly Alternative to Controlling Cost

June 1, 1983

The original plan was to build the F-15 fighter at a rate of 144 aircraft a year, but it didn’t work out that way. Instead, procurement of the initial inventory was stretched out from six years to nine, which added $2 billion — the price of an additional wing of F-15s — to the program cost.

This is not an isolated example. According to a study by Air Force Systems Command, this sort of thing happened all too often during the past decade. It now takes longer and costs more to build a weapon system than it used to.

Real cost growth — the increase beyond that caused by inflation — averaged more than five percent a year on major Air Force systems in the 1970s. And systems are getting more expensive in other ways, too. In constant dollars, the price tag on an F-15 has advanced capabilities that could only be dreamed of when the F-100 was built. It has, however, become progressively more expensive to add increments of capability to weapon systems.

Development time for current systems is typically 11.5 years, almost double the time for systems in the 1950s and 1960s.

The Systems Command study finds that, over time, the leading reasons for cost and schedule growth have changed. Before 1970, technical problems were the most frequent cause of programs getting into trouble. That changed in the 1970s, though, when program instability became the biggest reason why procurements went astray.

Several factors contributed to the acquisition turbulence. Air Force funding expectations were too optimistic. In fact, procurement budgets actually declined in constant dollars below the levels of the 1950s and 1960s. This, along with unanticipated double-digit inflation, crated a funding gap. The response frequently decided upon was to cut back on quantities and stretch out programs over more time. That wrecked the budget as well as the schedule, since delays, additional overhead, and inefficient production rates drove up costs.

There were other destabilizing influences. Frequent requirements changes were made to systems already in advanced stages of development. Increasingly active participation in program management by the Pentagon and Congress has an impact, too.

The A3 Study

The AFSC analysis, called the Affordable Acquisition Approach (A3), is the best of several recent studies on the ravaging effects of cost growth on system acquisition. These studies have aroused the attention of Congress and have inspired commentary, much of it irresponsible, in the public media.

He real question is whether the historical cost growth pattern can be broken. Top-level Air Force and Defense Department officials say it can be. They cite new acquisition initiatives, such as multiyear procurement, as reasons why history need not repeat itself.

The Air Force, for example, claims savings of nearly a quarter billion dollars so far on multiyear procurement of the F-16 fighter. The F-16 was one of the first programs approved for the multiyear approach, which eliminates the stops and starts and inefficient production rates that are common in year-by-year contracting.

The A3 conclusions are similar to those reached in previous reviews of the acquisition process, but this study is different for several reasons. The most important is that tit is based on empirical data from analyzing 109 past system acquisitions. A more dramatic reason, though, is that is projects hypothetical but specific consequences that might occur if history did repeat itself.

Even if the procurement budget sails untouched through Congress, the Air Force will experience a shortfall in its weapons buy unless it either kills some programs or has succeeded in totally wiping out cost growth.

Should cost growth continue at the 1970s’ annual rate of five percent for the next five years, the buying power of the Air Force budge could be cut by up to twenty-three percent. And that assumes the Air Force gets every penny for procurement called for in the President’s program and the Five-Year Defense Plan. This is by no means certain.

The FY ’84 Program Objective Memorandum (POM) forecasts an average annual AFSC procurement authority between now and FY ’88 of $19.2 billion, or about 2. Times the average of the past six years. The increase is to fund continued modernization of the strategic and tactical forces.

The chart on the facing page shows the impact of various cost growth and funding level combinations. The worst case hypothesized in the A3 study is that the procurement budget is held at the FY ’82 level over the entire period, and that the cost growth climbs to seven percent a year. The sixty-three percent a year. The sixty-three percent shortfall resulting from this combination means that the Air Force could carry out only thirty-seven percent of its planned acquisitions.

The Problem Factors

The A3 team analyzed three decades’ worth of system procurements and came up with a list of fourteen factors that seemed to contribute to program cost and schedule growth. The chart above shows the frequency with which each factor occurred over the thirty-year period. Some factors have since declined in importance, and others have become more significant. For example, funding instability — found in fifty-six percent of all the programs examined — occurred forty-eight percent of the time in the years before 1970, but sixty-four percent of the time since then.

The study notes that the factors are interdependent. “For example,” it says, “technical problems may require additional time and money to solve. This may cause funding problems and lead to a requirements change if it is perceived that correcting he technical problem is too costly. At the same time, the presence of technical problems and cost growth may weaken the support that the program has within the Air Force, the DoD, or the Congress, resulting in a program with funding problems, requirements changes, and external management involvement.”

Inflation is a perpetual problem. It is bad enough by itself, but it feeds on cost growth like a loan shark’s interest rates. Selected Acquisition Reports (SARs) on major Air Force R&D programs showed that inflation caused by real cost growth added up to nearly twice as much as the amount of the original overruns.

One of the ways A3 studied the data was to group the intertwined factors into four clusters — technical factors, risk management factors, government management factors, and funding instability factors — to look at their relative importance over time.

  • Technical Factors. Far fewer programs encountered technical problems in recent years, and trouble arose less frequently from going into full-scale development before the bugs had been worked out of a system. Engineering instability — meaning design changes initiated by the program office or the contractor — declined sharply as a source of trouble. (Requirements changes, which originate somewhere above the program office level, are quite another matter, about which more will be said.) Technical complexity, defined for A3 purposes as the existence of large numbers of interfaces, subsystems, and components, rose in importance. Overall, technical factors caused less cost and schedule growth in the 1970s than they once did.
  • Risk Management Factors. In the early 1970s, the Pentagon adopted a conservative “fly-before-buy” acquisition philosophy. There was less reliance on concurrency, in which a production decision is made before development is completed. Systems can be fielded quickly with concurrency, but glitches are more likely, too. The reduced range of Minuteman I is an example of concurrency leading to less capability than was desired. The A3 findings for the risk management cluster came as no surprise. Problems arising from concurrency and shortened acquisition cycles are less frequent. Test requirements have increased. Program managers have to spend more time on the “ilities” — the reliability, supportability, and design-to-cost work generated by a cautious acquisition philosophy — but these are comparatively minor factors in the over all cost and schedule problem.
  • Government Management Factors. More and more, program directors have had the Air Staff, the Defense Department, and Congress looking over their shoulders. The number of programs experiencing such “external management impact” has grown from less than half prior to 1970 to more than two-thirds in recent times. Changes to system requirements generated somewhere above the program manager’s head occur somewhat more often than they once did. Emphasis on joint service developments has given the program office more interfaces to worry about. Half of the 1070s programs — compared with a fourth in the 1950s and 1960s — ran into cost or schedule growth as a result of weak high-level support. Overall, government management factors figure increasingly in program troubles.
  • Funding Instability Factors.Fully two-thirds of the acquisition programs in the 1970s suffered from seesaw funding or other budget decisions at the Air Staff, Defense Department, or congressional levels. The damage was offset to a degree because Air Force cost estimators during this period were more accurate than their predecessors had been. The number of programs in which cost estimates based on initial program definition turned out to be low declined from forty-eight percent to thirty-six percent. “This is not to say that over one-third of the programs with cost-estimating problems is a suitable situation, but only to point out that the number of programs with such problems have decreased,” the A3 study comments.

The chart above shows the five factors most often affecting cost and schedule growth. Technical problems have declined in importance. So have technical advance impacts — an awkward term that means going into full-scale development with immature technology. Technical complexity is slightly up as a driver of cost and schedule growth. The big surges since the days when systems cost less and were fielded sooner have come in external management impact and in funding instability.

What Can Be Done

The Air Force’s acquisition woes are not worse than those of the other services, but, thanks to the A3 study, they are better documented and analyzed. Unfortunately, the study has been seized upon by some as evidence that the Air Force needs a tighter collar and a shorter chain in fiscal matters. A more reasonable conclusion is that the Air Force has responsibly faced up to its procurement problems. A3 confirms that recent Air Force and Defense Department efforts on cost and schedule growth are pointed in the right direction, and it provides a basis for further efforts. AFSC has combined all of its cost-control actions into a major push called “Project Cost.”

The Air Force had a lot of help getting into its present predicament, and will need a lot of help getting out of it. Some of the bigger variables in cost and schedule growth depend on action above the System Command level. But smaller factors — as we as factors that have declined in relative importance — still cause trouble. Thus, the A3 study calls for both a top-down and a bottom-up attack on cost.

Key elements in that attack are already in progress. Multiyear procurements, which the Air Force had been advocating for years, were part of the Acquisition Improvement package that Deputy Defense Secretary Frank C. Carlucci introduced in 1981. As noted earlier, multiyear contracting can smooth out instability and lead to efficient production rates if the technique is used properly.

Another Carlucci initiative was Preplanned Program Improvement (P2I), which provides for upgrade modifications to systems after they are deployed. It decreases the temptation to design high-risk features into a developing system, since those features can be added later after the technology matures. It may even cut down on the number of requirements changes inserted into ongoing programs for whatever reason. According to one Pentagon analyst, there were more than 500 design changes to the F-15 in its second year of production. (Systems Command sources say, however, that such a figure would have to count extremely minor items.)

To protect programs further from unnecessary change, AFSC is putting tighter control on “baselines,” which define a system in terms of cost, performance, support, and schedule. Requests for baseline changes are screened more carefully nowadays. Earlier this year, Systems Command began requiring new program directors to sign “contracts” committing themselves to the baselines of their programs within forty-five days of taking charge. Future baselines will reflect firm agreement among AFSC, the Air Staff, operating commands, and supporting commands.

Baselines can still be changed for legitimate reasons, such as reassessment of the evolving threat over the long period a system is in development. Advocates of a change, however, will have to demonstrate the absolute necessity of their proposal, and must show that disruption to the program will be minimal. In addition, a baseline change proposal must now come in with a “zero cost alternative,” meaning at least one idea of how to do it without any more money.

In support of realistic budgeting, the Air Force and the Defense Department are making greater use of independent cost estimates. The independent estimates is compared with the one the program manager turns in, and the higher of the two becomes the budget figure unless there is a strong rationale for doing otherwise.

The Air Force is also dusting off the “Should Cost” technique, which sends teams to contractor plants in search of wasteful practices that make production of a system coast more than it ought to. Teams are empowered to look into everything from wage settlements to the amount of scrap material thrown away.

A3 proposes that full-scale development be viewed as a commitment and that production be a “tough gate” in the system acquisition process. Early development is relatively cheap, accounting for only three percent of the life-cycle cost of a typical system. Full-scale development adds another twelve percent. The big expenses begin with production. Since nearly all programs that reach full-scale development go on to production., the Air Force must be sure — very sure — of programs it allows to proceed into full-scale development, the study says. Beyond that point, however commitment must be unswerving. “Once in production at an economical rate, stay there,” the study advises. “It’s cheaper than stretching out a program.”

The study says further that a strong Planning, Programming, and Budgeting System (PPBS) effort will be needed. It does not describe the difficulties in achieving that.

In theory, the Defense Department’s PPBS is supposed to match up military requirements with the resources to procure them. In practice, PPBS limped along badly all through the 1970s, and the defense acquisition process grew apart from it. Repairing PPBS was one of the Carlucci initiatives.

But even if the fiscal machinery in the Pentagon has been fixed, that may not be enough. As a recent study by the American Enterprise Institute points out, Pentagon budgets are not arrived at in reality by establishing military requirements and then funding to the level indicated. Budgets are sized in terms of percentage increases or decreases from previous budgets.

During hearings on FY ’84 funding, Defense Secretary Caspar Weinberger ask repeatedly that Congress think in terms of requirements instead of budget percentages. On the Hill and in the media, a tendency persists to contrast proposed increases in the defense budget with cuts elsewhere in federal spending — the so-called “fairness” issue. Argument continues about whether the proper increase for defense would be ten percent, five prevent, or some other figure.

Should the Air Force somehow succeed in its seemingly impossible task of reaching zero cost growth, the weapons buy in the outyears could still fall short as a result of budget cuts.

Consequences of Failure

The fate of future procurement budgets lies with Congress and the President, and to lesser extent with the priority that the Defense Department puts on systems procurement vs. other needs. The Air Force will have its hands full with the cost growth part of the problem.

Instances of zero cost growth are rare. The Air Force brought it off in acquisition of the Cobra Judy missile-tracking radar, but that was a one-of-a-kind system that basically took mature Cobra Dane technology and put it aboard a ship. The original C-141 buy came in on cost. Too.

There is evidence, however, that recent initiatives are having a positive effect. Both the B-1B and F-16 programs are tracking very near their baselines.

The consequences of failure to achieve zero cost growth, combined with lower-than-projected budgets, are illustrated by two examples in the A3 study.

The situation hypothesized is that the Air Force has succeeded in holding real cost growth to two percent a year, and that procurement budgets average $15 billion a year — midway between the current budget and the planned level.

As the chart on p. 88 shows, this would mean a thirty percent drop in buying power.

To illustrate the depth of such a cut, the first A3 example lops a sample thirty percent out of the Air Force procurement budget from the FY ’84 POM. The following programs are lost: WASP, MRASM, Infrared Maverick, Navstar, the derivative fighter, HARM, glcm, lantirn, the F-16 buy, and part of the F-15 buy.

This is not to say that the Air Force would react to a budget cut of that size by cutting these specific programs, but rather to show how much a reduction of such scope would hurt. Other cutback options would be similarly painful.

In the second example, the drop in purchasing power is the same, but this time the response is to stretch out programs and slow production rates. Doing this would increase costs by somewhere between $6 billion and $14 billion. Should the higher-figure turn out to be correct, the penalty would be equivalent to the cost of more than 400 F-15s, or more than fifty B-1Bs, or about eighty C-5Bs.

No matter how difficult a target zero cost growth is, the A3 data should supply plenty of motivation for the Air Force to go after it with a vengeance.

When programs get into trouble or if funding expectations are not met, the intelligent decision will be a hard one. It means canceling weak programs rather than weakening healthy programs by stretching them out and perpetuating the old cycle of cost escalation. — End