SP-4102 Managing NASA in the Apollo Era

 

Chapter 4

The NASA Acquisition Process: Contracting for Research and Development

A SUMMARY OF NASA CONTRACTING PHILOSOPHY

 

 

[65] From its establishment to the present, NASA has contracted with the private sector for most of the products and services it uses. Compared with the organizational changes discussed in chapter 3, changes in this area have been slight Between 1958 and 1965 NASA developed procedures at once internally consistent and politically acceptable, and which would spread NASA funds widely through the economy. The greater part of the acquisition process is traditionally called procurement, but that term is not used in this chapter title for three reasons: it sets too sharp a division between the preaward phase of the contract and its postaward administration; it implies that responsibility for contracts was concentrated in a particular headquarters office, when in fact there was scarcely a key official who was not involved; and something is needed to link procurement and contracting with the planning of NASA projects. Project planning (including listing mission objectives and their rationale, preliminary estimates of schedules and costs, and risk assessment) and procurement were the elements of a double-stranded process. The more general term acquisition process covers what is meant by procurement and has been used within NASA to signify "an orderly progression of Agency programs from early concept through the development and operation of hardware . . . and [provision of] the best sources . . . in the execution of projects and the performance of supporting services." 1 A basic feature of the process is that the program sets the contracting philosophy, not the reverse.

What sort of programs are we talking about? The official NASA source evaluation board manual has enumerated those special features that bear on the process, including "technological complexity, tight time schedules, unusual reliability requirements, a general absence of quantity, and little follow-on [66] production."2 As of March 1969 NASA had bought twenty Mercury, thirteen Gemini, and thirty-eight Apollo (command service module) spacecraft, including "boilerplate" (test) models and spacecraft modified for changed mission objectives.3 Furthermore, NASA often had to contract for products whose main features could not be precisely defined in advance, so there was no clear-cut basis on which the bidder could estimate costs. In such cases NASA could not use formal advertising and turned to negotiated procurement.

With important exceptions, NASA scientists and engineers have not built flight hardware. Rather, they have planned the program, drafted the guidelines, and established the parameters within which the product is to be developed. Viewed in this light, the rationale for an in-house staff has largely been to enable NASA to perform those functions that no government agency has the right to contract out, functions enumerated by a former Director of the Bureau of the Budget as "the decisions on what work is to be done, what objectives are to be set for the work, what time period and what costs are to be associated with the work, what the results expected are to be . . . the evaluation and the responsibilities for knowing whether the work has gone as it was supposed to go, and if it has not, what went wrong, and how it can be corrected on subsequent occasions." 4 Without the experience of actually building a spacecraft or performing experiments, center personnel could not effectively select industrial contractors or supervise flight projects running to many millions of dollars. The in-house work in advanced and supporting research and technology at Lewis, Langley, or Goddard was done less for its own sake than because, without it, Government personnel would have had no prospect of keeping contractors at arm's length. Nor could a center attract and keep the best people unless they had the opportunity, as Wernher von Braun put it, "to keep in touch with the hardware and its problems."5

By separating evaluation and production, NASA acquisition philosophy has had three especially important long-range consequences: the delegation of technical direction and monitoring to the centers, the refusal to set up a production capacity already existing in the private sector, and the refusal to create operating divisions intermediate between public and private sectors.

NASA has generally preferred to stand in relation to private organizations as buyer to seller and not to establish captive research centers like the RAND Corporation, the Aerospace Corporation, and the Institute for Defense Analyses. Exceptions such as Bellcomm and the Jet Propulsion Laboratory (JPL) are more apparent than real, although, like genuine Federal contract research centers, they worked for a single sponsor who guaranteed stable annual funding.6 Bellcomm, NASA's one experiment in sponsoring an adjunct trouble-shooting organization, differed in two respects from most captive groups: it was established as a profit-making corporation, and it was dissolved when its ostensible reason for being ended. As for JPL, it performed a variety of functions, and work was assigned as though it were a NASA center. The reasons for NASA's aversion to captive organizations were complex, but they stemmed mainly from top management's desire to keep NASA's suppliers at a healthy distance, rather than (in Webb's [67] words) "to create some new kind of organism which we would then be responsible for monitoring and which would in some way become dependent on us for support . . . and where we would have to evaluate their work not only on the projects assigned, but as to continued support."7 When NASA contracted for analytical or advanced mission studies it did so ad hoc, and it was not averse to using those same Federal contract research centers it was loath to establish in its own organization.8

NASA has generally preferred not to manufacture items available from private firms, even where it had the capacity to do so. Consider the choices posed by the problem of how best to develop automatic data processing (ADP) equipment, which was essential for mission control, data reduction, flight simulation, information retrieval, financial management, and image processing of data returned by remote sensors. NASA management chose not to build an ADP capability similar to what it was developing in space electronics. The computer was a tool, not an end product, and the research and development work of IBM, Control Data, and Minneapolis-Honeywell obviated the need to match the private sector strength for strength.9 Instead, the Office of Tracking and Data Acquisition compiled an annual ADP plan based on individual plans submitted by the centers. The plan served as the basis for reports required by the Bureau of the Budget and the General Services Administration, as a point of departure for negotiations with commercial suppliers, and as a yardstick for future planning.10 This approach served NASA well. By 1967 NASA had gone from first-generation equipment based on vacuum tubes to second-generation (transistors), and third-generation (integrated circuitry) equipment in less than ten years.11 And by letting contracts for the installation, programming, and maintenance of computers-generally for one or three years with an option to renew-NASA escaped heavy fixed costs for manpower and equipment.

NASA's ADP planning also demonstrates the agency's policy of assigning the acquisition process to the field installations where engineers and procurement specialists had the competence to decide the feasibility of negotiated competition in any given instance, prepared the initial procurement plan that served as the basis for a request for proposal (RFP), and made awards up to a dollar level that was raised from time to time by headquarters.12 Even in cases where headquarters made the award-where the bidding was noncompetitive or the action involved a competitively negotiated procurement of an amount more than the centers were authorized to handle-most of the preliminary steps were taken by center employees who wrote the procurement plan or served on the source evaluation boards reporting to the Administrator. Since the technical expertise to evaluate and the management skills to select were heavily concentrated at the centers, any other course would have been less efficient.

However, a word of caution seems in order. Drawing the line between evaluation as an internal function and design and production carried out by firms under contract was not easy. The private sector has been called on to define the mayor systems of most large programs; to integrate the interrelated parts of a [68] system, such as the stages of a launch vehicle or the vehicle and the payload it was intended to place in orbit; or to do checkout analysis that required the inspection of other companies' performance. In addition, the selection of the most qualified bidder depended on the qualifications sought. Agency officials had to decide, for example, whether to negotiate only with the most technically competent source or to encourage limited competition to obviate NASA's becoming captive to any single contractor; whether to select one firm to integrate all the systems in a complex flight project like Apollo, to have the work done by several companies under NASA technical direction, or to have it done by NASA employees; and to what extent one or more contractors could share in the early planning stages of approved programs. None of these problems was purely technical. In choosing a source, NASA had to consider not only the contractor's skills but also how the selection would affect NASA's relations with Congress, other Federal agencies, and the aerospace community.

So much by way of summary. To analyze the contracting cycle without being swamped in complexities, the remainder of this chapter deals with the questions such a process raises:

 

1. How did the procurement (contracting) process develop to subserve the ends for which NASA was constituted?
 
2. To what extent did NASA staff retain full control of program planning and management? Alternatively, what was perceived within NASA to be the dividing line between functions that could and could not be delegated to outsiders, whether firms, not-for-profit research centers, or universities?
 
3. To what extent did NASA's acquisition polices achieve such goals as "maximum competition consistent with the nature of the procurement," providing "the best sources to participate in the execution of projects and the performance of supporting services" and other formal expressions of policy?13 In short, how well did the acquisition process work on its own terms?

 

ORIGINS OF GOVERNMENT BY CONTRACT

 

As one student of Federal contracting has observed, the shift from work done in-house to contracting out was the result "not so much of explicit congressional direction as of agency decisions, in which Congress acquiesced, to seek the important advantages believed to be gained from close contacts with the capabilities of the private sector in R&D." 14 One could trace the American bias in favor of private enterprise to at least the late 18th century; but for our purposes, we need not go as far back as Hamilton's Report on Manufactures. After World War II, three circumstances conspired to increase enormously the volume and dollar value of work done for the Federal service under contract and to change the character of that work: the virtues of contracting out to the private sector, the limitations of [69] formal advertising, and the demand for special skills in the management and integration of complex weapons systems. The advantages of contracting, especially for management and consulting services, were readily apparent. The Government could tap experience and capabilities already available in the private sector; it was not bound by civil service regulations pertaining to hiring, dismissal, salaries, or work termination; persons who would not work for the Government because of red tape could be hired under these conditions; and the use of contractors probably allowed a much more rapid buildup of a large work force where such manpower was needed.

At the same time, many Federal agencies, especially the Department of Defense (DOD), found the limitations of formal advertising irksome. The 1947 Armed Services Procurement Act reaffirmed that advertising would remain the norm, and other modes "exceptions," for Government contracting.15 Yet advertising had only limited application for defense and, later, for space research and development. To advertise for bids, one must usually be specific about the item required, the design must remain fixed over a long period, and the lowest bid must normally be accepted by the contracting officer. In practice, formal advertising did not lead to more competition; a 1966 RAND study showed that of some 2300 procurements studied, 45 percent resulted in three bids or less, 32 percent in two bids or less, and 8 percent in only one bid.16 Increasingly, agencies with long-leadtime programs began to waive formal advertising-as the Armed Services Procurement Act allowed them to-in favor of negotiation because it allowed them much greater scope in exploring their bidders' proposals and cost data.

The only programs since the Manhattan Project comparable to the NASA mission were the weapon systems programs managed by the Air Force since 1954. Problems that became characteristic of civilian space research and development- the difficulty of estimating costs, the general uncertainties of the process, the long lead times-had already been faced by the Air Force in the management of the Atlas and Titan intercontinental ballistic missiles. In 1954 the Air Force established a Western Development Division of the Air Research and Development Command under Gen. Bernard Schriever to direct and coordinate its ballistic missile programs. The similarity between Western Development and NASA need not be pushed too far. The former was a blank-check outfit set up to run a crash program to close the missile gap. Under Schriever's philosophy of concurrency, production and operations were telescoped together, even while research and development were proceeding. Concurrency meant "simultaneous work on basic and applied research, vehicle design, component design, test facility design and Construction, component and system testing, the creation of production facilities, and the design, proof, and test of launch site facilities without which the missile would be impotent." 17

Schriever's approach to program management was equally radical. The case for hiring systems contractors to manufacture and integrate the components was not open and shut. The Army, at its Huntsville Arsenal, was as capable of [70] developing weapon systems (e.g., Jupiter) as complex as the Air Force's Thor intermediate-range ballistic missile. In essence, the Air Force turned to private contractors because it had neither the depth of competence found in Army laboratories nor the time to recruit engineers. For various reasons, including politics, the Air Force preferred to foster a civilian aerospace industry in peacetime. The research and development capabilities were there, ready to be exploited. Moreover, many Air Force weapons managers were convinced that, in contrast to the Army's arsenal system, "intimate relations with industry did promote a significant shortening of the long period of weapons gestation."18 While retaining ultimate responsibility for its programs, the Air Force delegated to civilians every aspect of the research and development cycle. In several cases the Air Force selected a prime contractor for technical integration, testing, assembly, subcontracting, and the like; this was how the Bomarc missile and the B-58 bomber were developed.19 In its ballistic missile programs, the Air Force worked through several associated contractors for components and subsystems and hired a separate organization, the Ramo-Wooldridge Corporation, to serve as technical director of the program. Ramo-Wooldridge, excluded from production of hardware, was both "line" and "staff"; the former insofar as it did systems engineering and provided technical direction for the Western Development Division, and the latter inasmuch as it also did long-range planning studies for the Air Force. 20

The Army's arsenal system and the Air Force's use of independent contractors for systems engineering might be considered two extremes in the Government's management of its R&D programs. Yet there is a point at which extremes meet. At the Huntsville Arsenal "as much or as little in-house production was possible as was compatible with the objective of keeping the technical team on the fringes of research and development work, or assuring the most rapid development of a specific item."21 On the other hand, the Air Force, in hiring RamoWooldridge and, later, the Aerospace Corporation to provide technical direction for its missile programs, was creating its parallel to the arsenal system. NASA could use features of both approaches without duplicating either. From the beginning it was NASA policy to contract out for items that could not be procured off the shelf (table 4-1).

How far NASA was prepared to go in hiring a contractor to do integration work is less clear. Some centers, like Goddard, occasionally integrated their own flight projects. In the case of the largest projects, especially Apollo, it is not even certain that there was a single contractor or center to integrate all the systems. Rather, each center delegated its responsibility to an industrial contractor, for example, Marshall's Saturn IB and Saturn V integration to Chrysler and Boeing, respectively. With NACA laboratories, the Vanguard Division of the Naval Research Laboratory, and the von Braun team, NASA had the competence to perform those functions in-house that the Air Force had delegated to RamoWooldridge and the Aerospace Corporation. The following section examines how NASA developed and refined its policies for acquiring and launching complex space systems.

 


[71] Table 4-1.-Distribution of direct NASA procurements, FY 1960-1968.

Type

FY 1959 1

FY 1960

FY 1961

FY 1962

FY 1963

FY 1964

FY 1965

FY 1966

FY 1967

FY 1968

Net Value of Awards (Millions)

Total

$212.5

$336.7

$755.5

$1 550.6

$3 230.5

$4 593.9

$5 187.4

$5 031.6

$4 650.9

$4 132.7

Business firms

87.8

174.0

423.3

1 030.1

2 2617

3 521.1

4 141.4

4 087 7

3 864.1

3 446.7

Educational

5.8

17.0

24 5

50.2

86.9

112.9

139.5

150.0

132.9

131.5

Nonprofit

-

-

-

-

15.3

29.1

25 3

27.7

39 6

33.6

JPL

22 4

38.3

86.0

148.5

230.2

226.2

247.2

230.3

222.2

207.2

Government

96.5

107.4

221.7

321.8

628 5

692.6

622 8

512.5

366.9

287 0

Outside U.S.

(2)

(2)

(2)

(2)

7.9

12.0

11.2

23.4

25.2

26.7

Percent of Total

Total

100

100

100

100

100

100

100

100

100

100

Business firms

41

52

56

66

-

70

77

79

81

8383

Educational

3

5

3

3

-

3

2

3

3

33

Nonprofit

-

-

-

-

(3)

1

1

1

1

1

JPL

11

11

12

10

7

5

5

5

5

5

Government

45

32

29

21

20

15

12

20

8

7

Outside U.S.

(2)

(2)

(2)

(2)

(3)

(3)

(3)

(3)

(3)

1

1 Nine months operation.
2 Included in Government.
3 Less than 0.5 percent.

Source: NASA Procurement Office, Annual Procurement Report, Fiscal Year 1968, p. 70.

 

[72] HOW NASA LEARNED TO DEAL WITH ITS CONTRACTORS, 1958-1962

 

The management of the acquisition process may be considered as a set of subsidiary, interrelated problems: providing a legal framework, creating a source selection procedure, safeguarding the privileged nature of source selection documents, and finding the right organizational location for the procurement function. With modifications, NASA adopted the Armed Services Procurement Regulations of 1947, which had been extended to civilian agencies in 1949 by delegation to the General Services Administration (GSA). The procurement regulations, which listed seventeen categories exempted from the rule of awarding to the lowest responsible bidder, including "services for which competition was impracticable" and "services for experimental or developmental work," enabled NASA to negotiate R&D contracts and even go to a company on a sole-source basis.22 In an exchange of letters in 1959, GSA permitted NASA to follow the armed services regulations in those cases where they differed from GSA's Federal procurement regulations, otherwise binding on all civilian agencies.

Under Glennan, officials adopted and modified the DOD source evaluation procedure for proposals on large prime contracts-those of $1 million or more. When NASA negotiated in such cases, the usual procedure was to set up a source evaluation board (SEB). By 1961 the procedure had been elaborated to comprise the following stages: preparing a procurement request by the responsible division or center; drafting a procurement plan by the appropriate contracting officer, in which was outlined the proposed procurement, the funding, the sources to be solicited, the type of contract to be used, the schedule for completing the procurement, and the negotiation determination and findings; preparing a request for proposal if the contract was to be negotiated; and awarding the contract by the Administrator whenever the contract was for 55 million or more. Each stage of the process had its own difficulties-the precision with which procurement plans could be drawn, the decision to advertise or to negotiate, and fixing the particular stage of the R&D cycle at which to request proposals. The development of a standard selection procedure took several years, but Glennan took two key steps. The first, announced in August 1959, was the promulgation of guidelines for awarding very large contracts; this was to be done by the Administrator, assisted by ad hoc boards responsible for establishing the selection criteria for each contract. The second, determined in late 1960, clarified the role of the boards in source selection. "Instead of the [source evaluation board] selecting contractor sources, or making recommendations, it became the primary SEB function only to evaluate potential sources and order rank their findings. The selection decision rested solely with the Administrator."23

Yet it would be misleading to assert that NASA had a uniform selection procedure, even in 1961. There was no single evaluation process. "There had been almost as many approaches ... as there had been NASA programs.... NASA Headquarters had provided only the broadest guidelines for source evaluation practices . . . thereby permitting the Centers wide latitude and flexibility . . . in [73] this area."24 Moreover, there was a real danger that the formal selection process might become so cumbersome as to defeat its own ends by increasing rather than lowering the cost of R&D.

In addition, Glennan and later Webb had to come to terms with the supposedly privileged nature of source selection documents. They would both insist that certain records of contract negotiations, such as the SEB report, should not be available to congressional committees; that their disclosure could only damage relations between NASA and its contractors, jeopardizing the mode by which both sides transacted business; and that, in any case, nothing essential was omitted in those documents that NASA chose to make public. This issue, which first surfaced in 1959 when the U.S. Comptroller General and later the House Science and Astronautics Committee demanded the records of NASA's negotiations with the Rocketdyne Corporation for the F-1 engine, was to become the basis of serious criticism in 1967 when Webb refused to produce certain key documents pertaining to the North American Aviation contract for the Apollo command and service modules.25 In a very real sense, the "production of documents" controversy was part of the larger question of how far NASA was willing to go in riding herd on its prime contractors: whether, for example, it was prepared to terminate a contract in midcourse because of the contractor's incompetence or go to a second source when a contractor was unable to fulfill its responsibilities; whether NASA was willing to disallow costs and penalize contractors for overruns; finally, whether NASA had a sufficient depth of in-house skills to prevent the agency from becoming captive to its contractors.

The question facing Webb, Dryden, and Seamans in mid-1961 was how well such a contracting system would serve to organize the manned lunar landing program. In general, they accepted, while improving on, the procedures instituted by Glennan. The principle of contracting out for R&D was reaffirmed; the role of in-house staff in technical direction was stressed; and headquarters officials took it upon themselves to make procurement policy more uniform yet flexible enough for NASA to obtain space hardware whose main features could not be specified in advance. The key organizational change was the establishment of an Office of Industry Affairs in March 1963, with NASA's Procurement Division placed directly under it.* " . . . instead of being one of several divisions in the Office of Administration' the Procurement Division became the all-important division under a Deputy Associate Administrator. " 26 Procurement Director Ernest Brackett and his deputy, George J. Vecchietti, who succeeded him in February 1964, laid down the general rules of the process: to advise the head of the Office of Industry Affairs on procurement matters, to publish and coordinate agency-wide procurement policies, to make determinations respecting procurement matters, and to serve as liaison with other agencies, especially DOD's office of Installations and Logistics.27 Brackett's powers did not extend directly to the substantive nature [74] of NASA contracting. It was the center employees' responsibility to draw up the procurement plan, draft technical specifications, and develop realistic cost and budget estimates.

The other important changes of 1961-1963 focused on two closely related problems: how to structure the contract to the work desired and how to draw the line between functions that could or could not be delegated to contractors. For its largest R&D contracts NASA had to go to negotiation; and while negotiation did not automatically preclude competition, attaining competition became more difficult. There were also two situations in which competition was not feasible, and the contracting officer recommended that NASA negotiate with a single firm. The first case is represented by the contracts for the Gemini spacecraft and for ten improved Delta launch vehicles, both of which were awarded to the McDonnell Douglas Corporation. McDonnell-Douglas had been prime contractor for the Mercury spacecraft. The Mercury and Gemini designs were similar, the second program was intended to follow closely on the first, and the technical experience gained in Mercury gave McDonnell-Douglas a decided advantage over other potential suppliers. The justification for negotiating with McDonnell-Douglas for the Delta vehicles was similar. The firm had the experience and proven capability to do the work, going to a new source would cost an additional $10-20 million and delay launch schedules by eighteen to thirty months, and without McDonnell Douglas experience, it would be difficult even to prepare definitive specifications outlining the scope of the work.28 The second case was one in which, as Brackett explained,

 

the nature and scope of the work is such that very special technical, management, and organizational capabilities are required. In such situations, while it cannot be said that there is only one company capable of performing the work, a particular company nevertheless stands out among all others as possessing a superior combination of the requisite, and sometimes unique, skills.29

 

In the same category were the contracts involving the creation of Bellcomm and commissioning General Electric to perform test and checkout services in support of Apollo.

But the difficulty in working such a system was that it tended to weaken NASA's bargaining position vis-a-vis its suppliers. NASA officials wanted as many alternatives as were available, and they wanted to bring about a state of affairs where the contractor also stood to lose. In 1961 the majority of NASA R&D contracts provided for paying a contractor for all acceptable costs plus a fixed amount of fee as profit, or simply cost-plus-fixed-fee (CPFF).** Whatever their value in attracting bidders, CPFF contracts could scarcely be considered efficient even by the most liberal definition. In a "cost-plus" contract, the contractor was not penalized sufficiently for underbidding or for inadequate performance. Nor did the contractor have any real incentive to economize. Quite the [75] contrary; while the fee was fixed, cost overruns would be profitable to the contractor.

In effect, the problems plaguing NASA contract policy ran the gamut from heavy cost overruns to the expense of preparing proposals (an expense borne almost entirely by the Government), to the excessive time spent by NASA employees in judging proposals, to firms' reliance on "brochuresmanship" to win NASA contracts For each contract, several questions had to be resolved. For example, should the contractor use Government-furnished equipment or provide its own facilities, and how much should be allowed for research and development undertaken at the contractor's discretion, so-called independent research and development? Such problems were closely related and had to be attacked from many angles, from program definition to postaward administration. The principles that NASA officials enunciated were the ones behind the major reforms of 1962-1965: developing realistic specifications before proposals were solicited, including in the RFP as many of the definitive contract terms as possible, strengthening in-house capabilities for technical direction, and screening proposals to eliminate companies with no reasonable chance of receiving a contract.

 

THE BELL REPORT AND ITS AFTERMATH, 1962-1963

 

The NASA acquisition system was not solely a technical response to technical problems of managing the space program. NASA, DOD, and the Atomic Energy Commission, the three Federal agencies that spent 90 percent of the Government's R&D dollars, had to come to terms with a blurring of the public and private sectors in the late 1950s.30 When the facilities of the Atomic Energy Commission were operated under contract, when many of the leading aerospace firms did essentially all their business with the Federal Government, when a very large share of university research was funded with Federal grants-the line between private and public was no longer clear. Of all the studies devoted to the problem, one of the most searching and influential was the report to the President submitted by Budget Director David Bell on 30 April 1962.31 The task force that Bell chaired included Webb, Defense Secretary Robert McNamara, Presidential Science Advisor Jerome Wiesner, the chairmen of the Atomic Energy Commission and Civil Service Commission, and the director of the National Science Foundation. The Bell report did not set guidelines applicable to every circumstance or draft rules broad enough to cover all contingencies, but it did lay down a general rule:

 

There are certain functions which should under no circumstances be contracted out. The management and control of the Federal research and development effort must be firmly in the hands of full-time Government officials clearly responsible to the President and Congress. . . . decisions concerning the types of work to be undertaken, when, by whom, and what cost . . . must be made by full-time Government officials.32
 

[76] The report also concluded that the effects of contracting out Federal R&D work "on the Government's own ability to execute research and development work had been deleterious"; and the report's principal recommendation-that Federal agencies use their laboratories to maintain knowledge of the most advanced science and technology-was put forward as a means of making Government a more sophisticated buyer.33 The Bell report accepted as settled the need for science-based agencies to go outside Government for scarce and valuable skills; it defended the profit motive as often the most effective way to get the job done; and it strongly dismissed any notion that the Government provide hardware or services that were available from the general economy.

What effects did the Bell report have on the Federal structure in general and on NASA in particular? One of its recommendations, that pay scales for Federal scientific personnel should be made "comparable" to those in the private sector, was partially met by the 1962 Federal Salary Reform Act, which also abolished the quota on the number of supergrade positions allotted for scientists and engineers. The report had also warned the Government that its contractors or employees were too often placing themselves in situations where conflicts of interest became unavoidable; here, tentative guidelines were set forth in a Presidential memorandum of 2 May 1963.34 Most important, the report recommended broader use both of fixed-price contracts and contracts with incentive provisions. In this case, some of the pressure for change came from those agencies, especially DOD, that were represented on the Bell task force. Prior to the report's publication, DOD officials had addressed some of the problems that led to the task force's creation by (1) publishing jointly with NASA a guide to stiffen the reporting requirements imposed on their prime contractors; (2) introducing the concept of program budgeting, whereby the Defense budget, force requirements, and alternative methods of meeting them were combined on a multiyear basis; (3) strengthening the "hardware ban" so that no firm could act as general systems engineer and produce components for the same project; (4) tightening contract administration; and (5) revising in March 1962 the armed services procurement regulations to make the fixed-price contract the preferred type or (where this was impracticable) to include incentive provisions with fees up to 15 percent.35

Not all DOD management procedures could be transferred wholly to NASA, if only because NASA contracted for some products that were much more specialized than anything the military needed. In fact, the contracting styles of the two agencies differed. Unlike DOD, NASA took title to all inventions resulting from the performance of R&D contracts, as required by section 305 of the Space Act. As of 1962, NASA had no "excess profits" clause in its contracts and was not bound by DOD's statutory requirement to terminate a contract if the contractor offered or gave a gratuity to secure favored treatment.36 Further, program budgeting, which came to be almost synonymous with Defense Secretary McNamara's method of working, was only reluctantly accepted at NASA when it was extended to civilian agencies in 1965. All the same, the innovations at DOD were bound to have repercussions within NASA. The Space Act required both agencies to [77] coordinate their programs, they shared many of the same prime contractors, and since its establishment NASA had used contract administration services supplied by DOD. The 1963 changes in NASA procurement were due essentially to the example set by DOD, combined with technical problems intrinsic to the space program.37 But whatever limited steps NASA took before the summer of 1962 to reduce the cost of doing business with industry, the Bell report made it politic to intensify such changes. The first, middle, and last steps involved improving the definition of specific programs, which NASA did by revising its contracting structure: using fixed-price contracts where possible, eliminating letter contracts, and rewarding contractors for detailed, accurate estimates and for selecting their subcontractors competitively.

The sequence of changes in NASA's contracting system can be stated briefly. On 29 May 1962 Webb appointed a study group chaired by Walter Sohier, the NASA General Counsel, to review the source evaluation procedure and various methods of improving contractor performance.38 Between the establishment of the Special Procurement Study and its final report in February 1963, NASA moved toward wider use of incentive contracts. In September a NASA circular announced that the agency would favor "procurements that lend themselves to the use of contract incentive provisions" while avoiding their use where they were unsuitable. In the same month Webb named Robert H. Charles, a McDonnell executive, as his special assistant with responsibility for procurement. Before he left NASA in November 1963 to become Assistant Air Force Secretary (Installations and Logistics), Charles wrote several reports whose main conclusion-that "significant improvement in product quality . . . timeliness and cost can be achieved if the procurement process is saturated with competition before contract execution, and with performance and cost reduction incentives thereafter"-became the cornerstone of NASA contracting policy.39 The events of the following year represented the unfolding of this principle: the final report of the study group, recommending that NASA improve its handling of incentive contracts and compel its contractors to prepare better work statements; the January 1963 budget request, which provided for the establishment of an Electronics Research Center; Seamans' memorandums of 25 February and 22 March to all center directors, suggesting that the RFP, including the incentive clause, "contain the precise language of the definitive contract terms"; and the directive of November 1963, ordering that CPFF contracts be reduced substantially and that incentives be considered for all.40

The Bell report was less the effective cause of these changes than their catalyst. The report, signed by Webb, did not go beyond NASA practice or the 1960 report of the Kimpton committee. NASA had always accepted in principle the division between in-house and contracted work. Years later, Webb asserted that he had "made sure that the things I considered important either could or would be incorporated in the final report.... It may be that on certain of the most important matters, I played one of the more decisive roles."41 By endorsing the existing system the Bell report made it easier for NASA and DOD to make some [78] changes that would go down well with Congress. Any policy intended to save tax money by rewarding a contractor for good performance and penalizing it for a bad one would be welcome there. To act in the spirit of the Bell report was good public relations, and it was cited explicitly before congressional committees as justification for establishing an Electronics Research Center or writing incentive provisions into new contracts.42

To repeat, the real basis for the 1962-1963 changes must be sought in the risk-bearing features of NASA programs. It is easy enough to assume that the risks were almost entirely on NASA's side. But from the standpoint of those companies that NASA invited to bid, the risk might not seem worth taking. Why, for instance, should a business organization invest enormous capital in building a plant to produce liquid hydrogen with no guarantee that it would be needed beyond a few missions? Contracts also tended to change as the systems did. For R&D of this sort, NASA used CPFF rather than competitive bidding or fixed-price contracts. But the great, the fatal, defects of CPFF were that profit was not tied to performance and that for all intents and purposes the contractor could not lose. The firm was not penalized for overruns nor rewarded for administrative efficiency. The firm had more reason to concentrate on getting the contract-often by bidding low-than on managing it.

The difficulties with incentives were of a different order, depending on whether a new contract or conversion from CPFF was involved. Incentives were tricky: Often little hard data were available to determine the relative importance of performing on time, within cost, and to specifications. These difficulties were compounded when converting an existing contract. The problem had a five-fold aspect:

 

1. Seeking to make the buyer more sophisticated in defining the mission design and the standards for successful performance.
2. Determining the proper weight and scoring for such performance.
3. Avoiding unrealistic cost estimates.
4. Persuading the contractor to adjust risk and convert voluntarily.
5. Reducing the number of cases of sole-source procurement without returning to competitive bidding and its disadvantages.

 

These goals were not easily attained. To convert the North American Aviation and McDonnell contracts for the Apollo and Gemini spacecraft took from 1962 through 1965. Few systems were as simple as the Early Bird communications satellite, which was "well-defined, requiring little or no development, operating with little or no technical direction, and procuring an essentially off-the-shelf item."43 Others, like the J-2 engine, were difficult to manage, let alone convert, because the programs themselves were changing. Originally designed for a two stage vehicle, this engine had to be modified for the S-II stage of the Saturn V, and its firing time extended from 250 to 350 to 400 and eventually to more than ;00 seconds. When the contract had been negotiated, no firm requirement existed for such an engine, and it involved "a minimum of facilities, a minimum of tests, [79] and an engine which is considerably short of the present performance."44 In such cases, the variables affecting NASA's ability to convert a contract from CPFF to some form of incentive were the technical complexity of the project, the contractor's attitude toward conversion, the technical competence of the NASA contracting officer, and the objective desired-whether cost control, prompt delivery, or improvement in performance. All these criteria were tied to the source evaluation procedure by which the contractor was selected. As we shall see, there was often less competition than the system promised because some companies were so clearly superior for a given system that NASA had to resort to sole-source procurement. Moreover, while the SEBs only "evaluated," rarely did headquarters select someone not in effect recommended by the evaluation.45 Indeed, one official complained that in certain contracts for support services Webb and Seamans concluded "that they could do nothing else than to accept the judgment of the source evaluation board."46

 

NASA Contract Administration

By the end of 1964 NASA had revised its contracting structure within the management system established by the 1963 reorganization. In general, NASA had brought its contracting policies into line with those of DOD and had delegated to the Defense Supply Agency and the Defense Contract Audit Agency much of the responsibility for administering its contracts, a development foreshadowed by DOD's Project 60 study, begun in May 1962. In August 1963 the Project 60 policy committee, which included NASA representatives, recommended consolidating NASA and DOD field contract administration in a single agency under the Office of the Secretary of Defense, a recommendation first carried out as a pilot project in the five-State Philadelphia region beginning in April 1964.47 That June DOD consolidated its field administration service into eleven Contract Administration Service regions, which handled the vast bulk of NASA contracts, and the transition period was largely over by 1966.

In substance, NASA delegated many contract functions to Defense agencies, that is, the authority to handle property administration or to consent to smaller subcontracts. NASA retained the authority to change the terms of the contract. At the beginning of FY 1967 DOD was administering approximately 1700 NASA contracts totaling $11.7 billion; 1500 of these, worth $4.2 billion, were handled by the Defense Contract Administration Service (DCAS), set up as a result of Project 60, while another 200, worth $7.5 billion, were managed by that military department having "cognizance" of all contracts in one plant.*** The Air Force, for [80] example, had cognizance of Boeing (Seattle), Rocketdyne, and Douglas, while the Navy handled Grumman. Where NASA let a contract that would not be performed at a cognizant assigned plant, it was sent to the DCAS regional office where the contractor was located. The great advantages of this arrangement were that it allowed NASA to work through a single Government agency, DCAS, for field administration; it avoided a wasteful duplication of services; and it enabled NASA to make maximum effective use of DOD, while reserving to itself those functions that it could not delegate.

 

The Hilburn Task Force and the Origins of Phased Project Planning (1964-1965)

If 1964 was the year in which NASA attained organizational maturity, it also marked the stage at which some of its largest projects ran into serious trouble. By September the delays, particularly in Apollo, had become so grave that Seamans warned Mueller "if present trends continue we will not achieve a lunar landing in this decade and the cost of the program will be in excess of twenty billion dollars."48 Why was the situation worse than it had been a year earlier? The answer lies partly in the peculiar technical difficulties connected with Apollo, requiring as it did the simultaneous development of a launch vehicle, three spacecraft modules, and ground support equipment at the Cape. Many components operated satisfactorily when tested alone but failed when incorporated into a system, for example, the automatic checkout equipment for the Apollo spacecraft.49

The difficulty that became increasingly evident to Seamans and his staff was that the contractor's organization was not structurally adequate to the demands made upon it. North American Aviation, the prime contractor for the command and service modules, was also responsible (through its Rocketdyne Division) for the S-II stage of the Saturn V and for the F-1 and J-2 engines that would power all three stages. Top management was entirely aware that "when you let the contract, all you've done is started a process that with the greatest of care, and ability, and drive will produce a bird. All you've done is put in motion forces that have the capability but which could fail at any point along the line. " 50 There were no easy solutions to the problems generated by Apollo or, in lesser degree, of Surveyor, Ranger, Nimbus, and the Orbiting Observatories. In the case of Apollo, management could go to a second source for the propulsion system and actually considered doing this.51 NASA could stick with North American Aviation and tighten the reins by more careful definition of the work to be done or by more frequent visits by the Apollo program manager. Finally, another corporation could be hired to integrate the three stages of the Saturn V with each other and with the spacecraft. Clearly, turning a program around in midcourse is more difficult than precisely defining a program at the beginning.

However, new methods like phased project planning were not introduced simply because of drastic schedule slippages. The problem facing NASA was more [81] simple and more complex than one of slippages: simpler because the causes of the delays could probably be dealt with by improved management techniques that already existed-incentive contracts, projectized management, and closer control by headquarters; more complex because both the Bureau of the Budget and the President could be expected to "go to a greater level of detail [than before] and examine all deviations in program content, delays in launch schedules, and other factors which affect total program costs."52 Failures that might have been excusable in a new agency would be of more concern now that NASA had the experience, the manpower, and the funds to get things done. In NASA's first three years, more than half its launches failed, but it received little criticism. However, after the failure of Ranger 6 in February 1964, Congress made it clear to NASA that no more would be tolerated. Hence the tone of Webb's letter to Seamans quoted above, a letter remarkable for its frankness and realism. Management would have to "learn to predict future costs more accurately" and to reduce them "to an absolute minimum," a task that would involve "timely reexaminations of some of our basic operating practices. "

Then Webb explained what he had in mind:

 

We must, for example, find an effective means to take all actions necessary to terminate developmental costs when developmental tasks are completed before the time when we can incorporate them in a test system, and not have to continue underwriting a . . . team to protect ourselves against the eventuality that the production articles intended for later consumptions may encounter difficulties.... [Otherwise] we will find ourselves continually on the defensive with respect to the management and budgeting of the program, even though the flight missions succeed. Those who are opposed to this program will take advantage of every opportunity in these areas to discredit the agency and the program.

 

Phased project planning was only one method among several that NASA management used to bring programs under greater control. Others included the revision of NASA's contract with JPL, pressure on the centers to tighten their monitoring of flight projects, and a series of task force studies by the Office of Programming on NASA cost and schedule estimating that were carried out in the summer of 1963. The most important of these studies was conducted at Marshall and led to the conclusions that "nobody knew anything beyond his specific area of responsibility,'' and that more advanced planning was essential-a policy that the task force sold to Seamans and, through him, to Webb.53 That NASA officials were considering the phasing of projects long before they issued guidelines is evident from a July 1964 memorandum by Seamans, recommending that

 

any new projects should be planned on a phased basis with successive contracts for advanced studies, program definition, prototype design, and flight hardware and operations. This will permit the work statements, including development engineering, to evolve in an orderly manner with maximum realism.54

 

At the same time Seamans assigned Deputy Associate Administrator Earl Hilburn to study current methods of scheduling and estimating the cost of projects.

 


[82] Table 4-2.-Average schedule slippage factor for various classes of projects.

 

Project

Launched

.

Successful

Unsuccessful

All Flights

Canceled

Future Flight Articles

Overall

.

No.

Average

No.

Average

No.

Average

No.

Average

No.

Average

No.

Average

.

Total

75

1.74

28

158

103

1.70

38

2.25

134

2.25

312

2.02

Communications

10

1.65

2

1.46

12

1.61

5

1.51

-

-

17

1.58

Lunar

1

1.79

6

1.24

7

t.32

-

-

15

1.86

22

1.69

Planetary

1

1.05

2

1.00

3

1.01

2

1.00

1

1.00

6

1.01

OART

5

1.96

5

1.75

10

1.85

1

2.44

5

1.83

16

1.88

Meteorology

9

1.69

0

-

9

1.69

2

1.84

3

3.18

14

2.18

Observatories

t

2.07

1

1.67

2

1.87

-

-

7

3.97

9

3.50

Simple scientific satellites

16

1.70

7

1.63

25

1.68

4

1.58

10

2.76

39

1.95

Manned spacecraft (S/C) and all systems

9

2.27

1

1.93

10

2.24

17

2.44

66

2.23

93

2.27

Launch vehicle development

11

1.60

1

1.64

12

1.60

2

5.25

12

1.93

26

2.03

Engines

-

-

-

-

-

-

2

2.20

-

-

8

1.98

Boilerplate S/C

10

1.60

3

1.72

13

1.72

3

2.38

9

1.95

56

1.88

Biosatellite

-

-

-

-

-

-

-

-

6

1.44

6

1.44

Source: Second Interim Report to the Associate Administrator on Studies Relating to Management Effectiveness in Scheduling and Cost Estimating (15 Dec. 1964), p. 2.

 


[83] Table 4-3.-Average schedule slippage factor by center.

 

Center

Launched

.

.

Successful

Unsuccessful

All Flights

Canceled

Flight Future Articles

Overall

.

No.

Average

No.

Average

No.

Average

No.

Average

No.

Average

No.

Average

.

Total

75

1.74

28

1.38

103

1.70

38

2.25

134

2.25

311

2.02

Manned Spacecraft Center 1

19

1.92

5

1.94

24

1.92

14

2.71

61

2.303

130

2.17

Marshall

9

1.34

3

1.54

12

1.39

9

1.86

24

1.65

49

1.61

Goddard

33

1.72

4

1.63

37

1.71

6

1.67

14

3.34

57

2.11

Jet Propulsion Laboratory

2

1.42

8

1.18

10

1.23

4

3.12

11

1.98

25

1.86

Langley

9

1.79

7

1.74

16

1.77

4

1.48

9

1.67

29

1.70

Lewis 2

3

2.28

1

1.04

4

2.08

1

2.44

5

2.52

11

2.46

Ames

-

-

-

-

-

-

-

-

10

2.56

10

2.56

1 Includes all Mercury flights.
2 Includes all Centaur vehicles.
3 Lunar module and command service module for the same flight are listed separately, as are Agena Target Vehicles and Gemini launches.

 

Source: Second Interim Report to the Associate Administrator on Studies Relating to Management Effectiveness in Scheduling and Cost Estimating (15 Dec. 1964), p. 3.

 

[84] In two reports submitted in September and December 1964, Hilburn laid the basis for many subsequent changes. He concluded that slippages were greatest in propulsion systems and in launch vehicles requiring new engine development; that, on the whole, slippages were greatest in the newer centers; that the largest cost increases were due to incomplete definition of the work at the start; and that delays in one mission impinged on all future missions in that series (tables 4-2 and 4-3).55 Most significantly, Hilburn concluded "that NASA's project control can be improved; that slippages and overruns are not inevitable . . . it is felt that sweeping changes in organization or procedures are neither necessary nor desirable."56 NASA already had the remedies in hand; all that was needed was to intensify their use. Align all projects on a "vertical" basis and see that all contractors do the same; conduct project reviews to identify all areas of dubious ("soft") planning; wherever possible, put contracts on a multiple incentive basis, with special weighting for costs and schedules.57

By January 1965, agency policy was to define all programs as explicitly as possible. Phased project planning (PPP) was only one method of attack and perhaps not even the most important. Reserving to chapter 6 a detailed account of NASA program planning, several features of PPP nevertheless deserve emphasis here. First, the guidelines of October 1965 applied specifically to new projects at a time when very few new projects were starting and several that had reached the advanced study phase had to be canceled. Second, the concept, if not the name, was certainly not new. To cite one example, the construction of the 64-meter radio antenna at Goldstone, California, as part of the Deep Space Network was mentioned by Seamans as "almost a textbook case of phased project planning." 58 What he did single out as new was "the agency's ability at this point to assure that all new projects can be undertaken in this manner."59 Third, as a major benefit, NASA would be able to keep all options open as long as possible. Here, the language of the relevant policy directive is significant. PPP was "not an end in itself"; it was introduced only to ensure that research and development would now be conducted in "an appropriate number of sequential phases" with maximum competition characterizing the "phase-by-phase increments of project execution," and each phase allowing for "the fundamental concept of agency top management participation at all major decision points."60 Fourth, the exceedingly gradual way in which the details of PPP took shape is also noteworthy. Only after more than a year-from the appointment of the Hilburn task force-of studies and consultations with the directors of the program offices did senior management take the first tentative steps toward a goal that would require another three years to attain. Some reasons for the delay appear in chapter 6, in which it also becomes apparent that the goal of fostering maximum competition at every stage of the R&D cycle was inherently unworkable.

 

THE NASA ACQUISITION PROCESS EVALUATED

 

Phased project planning was predicated on the assumption that NASA employees would be responsible primarily for defining programs and providing [85] technical direction to agency contractors. The separation of evaluation and production was the key to the NASA acquisition process, and that process was central to the way NASA worked. Any assessment of the agency must therefore try to measure the success of the acquisition process.

Before proceeding with this argument, however, some preliminary remarks are in order. First, "success" is an ambiguous term, both in NASA and DOD parlance. Was the Air Force's C-5A transport plane a failure because of cost overruns or a success because it resulted in "a substantial reduction in the cost of Airlift capability"?61 In NASA, was the Syncom I communications satellite a success because it was placed in synchronous orbit or a failure because it returned no data? In such cases, even cost-effectiveness is not a reliable criterion. The C-5A experienced overruns of 60 percent, yet this was low compared with the 200-300 percent overruns of earlier weapon systems uncovered by Peck and Scherer in the early 1960s. Moreover, "the early ballistic missile programs incurred large cost overruns and their initial performance was deficient. Yet, these early programs were not counted as failures but, rather, successes, because the goals perceived during the 'missile gap' era put overriding emphasis on early deployment of some kind of deterrent capability."62

Second, NASA and DOD acquisition procedures were directed to the production of very different systems. Compared with DOD, remarkably few NASA programs have been canceled because they were beyond the state of the art or because their costs clearly outweighed any possible banefits.**** 63 The difference is not to be sought in the quality of management or the techniques used for reporting and control. Program evaluation and review techniques (PERT) could be used for weapon systems and spacecraft alike, and some of NASA's most important projects were managed by military detailees who were trained in PERT. The question, rather, is what such canceled programs as the Skybolt missile, the nuclear-powered airplane, the Cheyenne armed helicopter, and the B-70 supersonic bomber had in common that Apollo, the Orbiting Observatories, and Viking did not. The former were all planned as extensions of capabilities already existing; no one could build a "better" supersonic manned bomber unless there was already a bomber to serve as a yardstick for technological change and hence for cost effectiveness and cost-benefit studies. With Apollo and other very large NASA programs, there was nothing to serve as a measure of each program's relative success. No one had landed on the Moon before the crew of Apollo 11 did, nor had anyone successfully soft-landed on Mars prior to Viking.

To put the matter differently, the question here is less one of effectiveness than it is of efficiency. Granted that NASA did accomplish its lunar landing mission within the time specified, the Apollo program cannot be compared to [86] alternate approaches that were never adopted. But the acquisition process can be valuated in terms of what NASA management claimed on its behalf. Thus, if the role of in-house staff was to do advance planning and to provide technical direction, how well was this accomplished? How well did the source evaluation procedure succeed in promoting the avowed aim of fostering maximum competition? What benefits resulted from the extensive use of contracts with incentive provisions? The following sections are intended to provide tentative answers to these questions.

 

THE INTEGRITY OF THE NASA CIVIL SERVICE STAFF

 

What distinguished NASA employees from contractor employees was the complex of responsibilities subsumed under technical direction and supervision. Jut between the fabrication of flight hardware and the in-house capacity to plan programs was the broad grey area of contracting for nonpersonal services. Such contracts, by reason of the issues they raised, serve as a test case of NASA's ability o make explicit its criteria for contracting out.64 No one expected NASA personnel to fabricate either the Saturn launch vehicles or the Apollo spacecraft. Less clear was the course to be followed in staffing the three main tracking networks, writing computer programs at NASA installations, and operating test facilities for simulating the space environment. A case could be made for doing such work eternally or for hiring a contractor.

Within NASA, the Bureau of the Budget, the U.S. General Accounting Office (GAO), and interested congressional committees, there was no consensus as o what "support services" were, the extent of the practice, or the guidelines for contracting to the private sector. There was agreement that support services were ancillary to the agency's functions, could be performed in-house if staff were available, and were more temporary and intermittent than those usually performed by civil service employees. Beyond this, agreement broke down over costing formulas, the questions of whether and to what extent support service contracts were being used to evade agency personnel ceilings, and whether an agency was delegating functions that should be retained to exercise technical Erection. Furthermore, drawing a fine distinction between one kind of service contract and another was difficult, if not impossible. As Finger remarked in congressional testimony:

 

Some of these services require little contractor capital investment; others require substantial investment. Some of them require little skilled labor; others require special training and highly specialized skills. In some cases the function is separate and distinct; in others it is intimately related to the whole process of successful system development and flight mission operations. In some cases it is a service to the Government organization; in others it is a direct support of a development contractor's responsibility. Some of these functions are performed at a Government installation; others ... at a contractor's plant.... In some cases there are many customers for the services; in others the Government may be the only customer. In some cases the operation is a continuing and stable one; in others it is so closely related to program or project requirements that it requires continuing adjustment in numbers and skills of people assigned.65
 

[87] Finger stressed the special and urgent nature of the task that he and other officials were expected to manage. Given the target date for Apollo and the extraordinary buildup in required manpower and facilities, NASA had perforce to turn outside for many kinds of services.

NASA's position was complicated by the tendency of officials to argue as if service contracts were only occasional departures from the general norm that Government provide such services in-house. Prior to 1967, NASA guidelines specified that the agency was to contract for support services only if the capability for doing the work already existed, industry itself normally contracted for similar work, the nature of the work made full-time employment of Government employees impracticable, or skills not readily available within the Government were required.66 On the face of it, this position seemed opposed to that of the Bureau of the Budget, which since 1959 had asserted the general policy that Government should rely on private enterprise to supply its needs;67 but the difference was superficial. NASA officials argued that in a period during which the space budget had doubled and then tripled over what it had been in 1960, the agency had to go outside for certain rare and valuable skills-whether these involved performing analytical studies in support of the Apollo landing, operating the mission control center at Houston, or providing technical services at the Kennedy Space Center. Over and over they argued that it was impracticable to graft these skills onto a Government operation when abundant capacity for such work existed in the private sector. NASA, so their argument ran, did not contract for end products when it awarded a service contract; simply by virtue of being services, items such as those listed were purely means to obtain an end product designed by NASA employees. The reverse of this argument was that once the space program began to shrink, it might be necessary for NASA to build up in-house those skills for which it had contracted in the past. It may be significant that the strongest criticism of these practices-beginning with a June 1967 GAO report concerning service contracts at Goddard and Marshall-occurred just as NASA budgets and manpower were beginning to decline. 68

In the area of launch operations, NASA deliberately assigned responsibility to the flight contractor from the design phase through checkout, launch, and performance in flight. Whatever the benefits of a purely abstract cost analysis, NASA deemed it impossible either to employ civil service personnel for the job or to impose a moratorium on flight operations until the contractor work force could be converted to civil service positions. In any case, NASA did not make the mistake of trying to justify its contractor policies on the basis of cost alone-a position that would have left the agency vulnerable indeed. One official at Goddard said that "cost is not the prime determinant in decisions to contract. To accept such a premise would have . . . deleterious impact on current and future contracting operations of this Center and all of NASA."69 Where even the Bureau of the Budget could not adduce uniform cost criteria to justify contracting out, NASA officials could regard their own policies as being well within Bureau standards that were unclear to begin with. Cost was a consideration, but not the only one.

[88] Sometimes, as Finger explained, "there was a strong feeling that the function ought to be performed in-house and, therefore, even though for some functions the cost differences would be within the 10 percent [stipulated by the Bureau as the differential for contracting out] or less than 10 percent, the function could still be moved in-house if it was felt that the function ought to be performed that way." 70

All these arguments assumed that support contracts did not impair NASA's control of its own operations. Time and again, in congressional testimony and in correspondence with GAO, Finger and Webb emphasized the safeguards for maintaining the integrity of NASA programs: the "clear separation" of Government and contractor responsibilities, maximum practical competition among qualified firms, the physical separation of Government and contractor employees to avoid the supervision of the latter by the former, the imposition of a hardware ban on major support contractors like Bellcomm and General Electric, and the "recompeting" of support contracts-seeking new competitive bids instead of automatic extensions-so that industrial firms could never become too complaisant.71

But how effective, really, were these safeguards? The issue was not whether this or that center might hire firms to remove the garbage, mow the lawns, or provide guard service, even if Government unions were inclined to argue that a multiplicity of such contracts might subtly erode NASA's ability to distance itself from its contractors. The real test cases were those contracts for engineering support services that seemed directly to threaten NASA's ability to control and monitor its own programs. To examine the issues at stake, three such contracts will be described: the Bellcomm, General Electric (GE), and Boeing Technical Integration and Evaluation (TIE) contracts.+ Other than their large size, these contracts had little in common with other support contracts or even with each other. The Bellcomm and GE contracts were negotiated early in the Apollo program to anticipate or forestall problems; the TIE contract was let in the aftermath of the Apollo 204 fire. Bellcomm was a profit-making subsidiary of American Telephone and Telegraph that worked solely for NASA, while the GE and Boeing contracts were carried out by existing divisions within the parent companies. Bellcomm went out of business in May 1972 when its last contract extension expired; the Boeing TIE contract was phased out in 1970.72 In the same year, NASA gave each OMSF center the option of extending its portion of the GE contract separately. Yet each contract met one of the criteria of a support contract, in that the work "could have been done by our hardware prime contractors or in-house if sufficient . . . capacity were available."73 The questions are, why did NASA decide to contract in these cases, and how did each decision affect the role of in-house staff.

[89] Bellcomm was established in March 1962 at NASA's request to do analytical studies in support of the Apollo landing. Unlike the GE and Boeing contracts, it was strictly a Washington, D.C., operation on behalf of OMSF. Bellcomm manufactured nothing. Among the specified tasks were preparing specifications for the Apollo Program Office, indicating Apollo requirements for data from Surveyor and Lunar Orbiter, evaluating proposed manned spaceflight experiments, and defining scientific objectives for lunar missions after the first landing.74

As an Apollo Program Office report stated, the situation was unusual "because NASA was attempting to use a contractor organization as its line engineering directorate."75 The justification for creating Bellcomm proceeded at several levels: the oft-repeated "unique capability" argument,++ NASA's general practice of contracting to private industry, and the impossibility of NASA's assembling civil service manpower in a brief period, only to disperse it within six to eight years. Moreover, Apollo differed radically from the Mercury and Gemini programs that preceded it. The earlier programs used the Atlas and Titan launch vehicles developed by the Air Force, with essentially only one center (the Manned Spacecraft Center and the Space Task Group that was its nucleus) involved. Apollo, however, involved all three OMSF centers and required totally new launch vehicles, a new spacecraft, and new ground test and launch facilities-all of which had to be integrated into a functional system reliable enough for manned flight. "Management of the program require[d] a strong program office at Headquarters with a strong systems engineering capability not then in existence."76 This was NASA's principal justification for turning to the Bell System; indeed, it was argued that the creation of Bellcomm would also enable NASA to tap the capabilities of Bell Laboratories and Western Electric.

Although linked with Bellcomm in congressional testimony, the GE contract was otherwise distinct.77 Bellcomm was involved in systems engineering; GE, in "the implementation of standards to assure proper integration, reliability, and checkout of hardware" at Houston, Marshall, and the Cape.78 GE would eventually design and manage fourteen sets of Apollo Spacecraft Automatic Checkout Equipment rather than the three originally specified, support the investigation of the Apollo 204 fire, and operate the Mississippi Test Facility for Marshall. As with Bellcomm, NASA wrote a hardware ban into the GE contract, forbidding the company to participate as a prime contractor or first-tier subcontractor for Apollo launch vehicle or spacecraft stages or as a supplier of nonstandard components for launch vehicles or spacecraft, save at the discretion of the Administrator or Deputy Administrator. However, this restriction did not prevent GE from manufacturing much of its own checkout equipment, did not apply to existing contracts, and did not bar GE from bidding on "standard" items in managing the Mississippi Test Facility.79

[90] Each of these contracts involved some kind of "service." Systems engineering requires that the environment within which systems operate be known, that the objectives for which the systems are designed be understood, and that the resources available for the system be adequate to meet its goals. Integration, on the other hand, means fitting things together, whether components into a subsystem, subsystems into systems, and the system-for example, the first stage of a launch vehicle-into the total launch vehicle.

But "integration" and "evaluation" are not synonymous.

 

Technical integration is the act of making two or more elements, such as a launch vehicle and a spacecraft, fit and work together.... Technical integration must be performed between the elements for which an organization is responsible, as well as the major elements between different organizations.
Technical evaluation consists of examining the mission requirements specifications and the hardware capabilities . . . to determine that the hardware will work . . . to accomplish its role in the performance of the pre-established mission.80

 

Boeing, which already had the contract for the first stage of Saturn V as well as for integrating all three stages of the booster, was now being asked (according to Webb) to certify that "the whole unit, vehicle and payload, does function together, is compatible, and is ready for flight."81 Boeing was called in because there was no time to build an in-house capability; in any case, the company's role was "advisory." NASA's original intention-to award the contract to North American Aviation-was dropped after the Apollo fire. Boeing was chosen although the contract, first announced by Webb in congressional testimony on 9 May 1967, was not made definitive until 30 May 1968. NASA settled on Boeing because of the company's experience in the Saturn and Minuteman programs, its management of Lunar Orbiter, and the management abilities of senior Boeing officials.82 Also important, although unmentioned, was the familiarity of OMSF staff, several of whom had worked for Boeing, with the company's mode of operations.83

Bellcomm and TIL were the "upstream" and "downstream" sides of Apollo systems engineering. Put simply, Bellcomm mapped the requirements for lunar missions; Boeing implemented the requirements in many ways-by evaluating the adequacy of hardware, analyzing proposed changes in flight equipment, and preparing backup material for the flight readiness reviews held a month prior to each Apollo launch. The TIE contract was as significant for what it forbade as for what it sanctioned. Boeing personnel would not supervise other NASA or contractor personnel and would only perform work that was not being handled by a specific Apollo prime contractor. Again and again, NASA management insisted that the TIE organization was to "assist," "advise," "recommend," and "support" Apollo.84 Like those with Bellcomm and GE, the Boeing contract was established in response to a program whose complexity and urgency precluded a gradual buildup of the needed skills.

Now to return to the earlier question, to what extent did the three contractor organizations assume functions that NASA should not have delegated? This question is distinct from inquiring into the contractors' effectiveness. In fact, they [91] made substantial contributions to Apollo. When Finger visited Houston in August 1968, Manned Spacecraft Center Director Robert Gilruth told him that "the best return MSC could get from the [TIE] contract was additional engineering support which was sorely needed in many areas of the Apollo program. . . spacecraft sneak circuit analyses+++ were a major Boeing contribution. It was felt all along that these analyses should be made, but personnel had not been available to do so."85 Another example of work done under the TIE contract was the analysis of the "pogo" problem, the oscillations of Saturn V after lift-off, and its effect on the Apollo spacecraft. Similarly, Bellcomm's work in systems engineering was more than satisfactory, judging by the extensions of the original contract.86

These contracts violated neither Government policy nor the spirit of the Bell report, but they tended to compound the confusion between public and private sectors that the Bell task force had tried to dispel. NASA managers could respond, with considerable justice, that their policy of using support contracts was fully disclosed, that the agency had no alternative to contracting out most of its development work to industry, and that the final word in policy matters remained with civil service employees.

One can also acquit NASA of the charge that, by allowing contractors such as Bellcomm to define problems, the contractors set the terms on which decisions were made. The charge made by one critic of NASA, that Bellcomm in effect provided the rationale for the decision to make lunar orbit rendezvous the Apollo mission mode, does not bear close analysis.87 A significant portion of the NASA community, including the entire Space Task Group, had come to favor lunar orbit rendezvous by the end of 1961, almost three months before the Bellcomm contract was let.88 Also, the decision of the Manned Space Flight Management Council in June 1962 to adopt lunar orbit rendezvous was heavily influenced by nontechnical influences-for example, the desire for unanimity and the consensus that the Saturn vehicle should be used-that were peripheral to Bellcomm's assignment.

The problems generated by these contracts were more subtle. In all three cases, there was a tendency for support contractors to continue working for NASA after their original reason for being had ended. Bellcomm, which had been set up to do systems engineering for Apollo, went on to do studies for Apollo Applications and the space shuttle, and in 1968-1969 evolved into an in-house entity like the RAND Corporation for OMSF; it had the manpower and the budget to prepare a post-Apollo plan for NASA and to participate intensively in the planning review sessions held throughout 1969.89 Moreover, the very functions of evaluation and review for which Bellcomm was created and TIE was negotiated were those for which, it could be argued, an in-house force was needed. Support contractors were used for three reasons: direct support, "added assurance that all possible problems have been identified and solved," and to provide checks and balances, that is, to check on the centers before the commitment to fabricate was made."90 These [92] reasons, particularly the latter two, might seem to apply with greater force to headquarters, both to top management and the program offices.

In addition, hiring contractors to do what might have been done in-house was welcomed neither by other prime contractors nor by the centers. According to the official history of Apollo, the GE contract "was seen in some circles as a matter of GE telling government officials how to do their jobs." And one GE official conceded that at Houston "they didn't want us. There were two things against us down there . . . it was a Headquarters contract, and it was decreed that the centers shall use GE for certain things; and . . . they considered us Headquarters spies."91

The most serious charge against support contracts was that they caused NASA to depend on industry for what could have been done by agency employees. When Mueller wrote to his center directors at the end of 1968 to ask their opinion on phasing out engineering support contracts, the three directors were unanimous. Gilruth put the matter with special force:

 

We have agreed with the "surge tank" philosophy and have recognized that an adequate in-house manning of Civil Service personnel was out of the question. However, we think that this in-house capability should have been greater and have, from time to time, requested a larger in-house manning. We agree that the most effective management of future programs calls for greater in-house engineering capability. We have been very weak in systems engineering, analysis, and trajectory work. We have developed practically no engineering competence in the field of Reliability and Quality Assurance. We have probably been too dependent on Philco, IBM, and TRW in operation of the Mission Control Center and in operations generally....
 
We are studying intensely what we can do to reduce our dependence on engineering support contractors.... We are sure that it will not be practical to eliminate outside engineering support for several years. However, we do plan stronger program level systems groups and well defined tasks for the outside engineering support.92

 

By 1969 NASA was faced with declining personnel ceilings. Thus the attempt to strengthen the in-house staff meant, paradoxically, that service contractors would remain indispensable for several more years (table 4-4).

The case for service contracts rested on one powerful argument that was never adequately refuted: An agency with such urgent and unique assignments could not have done the job with its in-house staff alone. That case was not always well-presented, as indicated by continued skirmishing between NASA and the House Government Operations Committee as to whether TIE was or was not a support contract, whether the Boeing personnel ceiling at headquarters was 75 or 300 persons, whether these figures were for "direct" or "indirect" personnel, and whether or not Boeing was being paid twice for its work on Apollo.93 Even those like Gilruth who thought that support contracting had gone entirely too far conceded that Apollo had made such contracts unavoidable. Nor need one conclude that NASA abdicated its responsibility to manage. In the Apollo program, NASA possessed a far greater depth of experience and talent than the Air Force's Western Development Division or the Special Projects Office that developed the Navy's Polaris.94 NASA acted to control its programs in the following ways:

 


[93] Table 4-4.-Nonpersonal service contracts: Number of Government and contract personnel, 30 June 1962-31 May 1968.

Date

Government

Contractor

.

30 June 1962

23 511

3 525

30 June 1963

29 488

6 758

30 June 1964

31 957

10 200

30 June 1965

34 049

Not Available

30 June 1966

35 708

26 962

30 June 1967

35 860

29 267

1 Nov. 1967

34 281

33 768

31 May 1968

33 202

31 511

Sources: 1964 and 1966 program reviews, Procurement Programs (27 Oct 1964, fig. 26; 31 Oct. 1966, fig. 48); and hearing of House Science and Astronautics Committee, Subcommittee on NASA Oversight, 90th Cong., 2d sess. (11 July 1968), p. 9.

 

1. By determining the conditions under which contracting would be necessary.
2. By anticipating problems before the contractor did, who would then do whatever detailed engineering was necessary.
3. By conducting enough of the effort internally to acquire the expertise to judge and evaluate contract performance and, if required, to provide support or actually take over contractor work in order to complete the program.
4. By terminating or phasing out the contract, although this was as much a managerial as a technical decision.

 

In sum, several of the most controversial NASA contracts were for work that might have been performed in-house, but the contractor's involvement gave assurance that everything had been double and triple checked. Faced with ambiguous guidelines, NASA officials believed that resort to the private sector was inevitable and that the question of whether a task was covered in-house or by contract was less important than the knowledge that the capability would be there when needed.

 

SELECTION CRITERIA AND THEIR EFFECTIVENESS

 

Before turning to NASA's source selection procedure, something must be added about the agency's policy of fostering maximum competition in the selection of contractors This must be understood in a highly qualified sense. For NASA, competition was strictly a means to an end. The nature of the products for which NASA sought suppliers precluded competition in the classical sense for several [94] reasons: Only a small number of firms were able to compete for major R&D contracts because of the prohibitive cost of entry; the firms were concentrated in certain geographic areas; and, most of all, unique uncertainties are inherent in the development of sophisticated space and weapon systems. As Peck and Scherer observed in their study of the weapons acquisition process, in soliciting for R&D work "the seller does not offer a finished product which the buyer can either accept or reject. Rather, the government pays development costs before it knows what the ultimate performance of the product or its desirability relative to other products will be."95 Two additional considerations that affected the NASA acquisition process were the absence of follow-on production and the uncertainties of the future of the space program in 1960-1961 and again in 1967, when funding for most NASA programs, especially the newer ones, had begun to decline.

Thus NASA had to take its suppliers where it could find them. Despite pressure from regional associations and from Congress-year after year, NASA authorization acts included a clause urging NASA to distribute research funds on a wider geographical basis-NASA placed most of its prime contracts in those areas where the capability already existed. The NASA position, as stated by Webb and Hilburn, was that

 

to base the award of contracts on geographical considerations, rather than on competition for all companies regardless of location, would be inconsistent with the statutory procurement authority currently applicable to NASA. Moreover, limiting competition to geographical areas might mean that the company with the best capability for a project of importance would not be awarded a contract because of its location.96

 

In fact, NASA prime contracts were concentrated in a few regions. Between fiscal years 1962 and 1969, 60 percent of NASA R&D dollars were placed in only three States: California (42.7 percent), New York (10.2 percent), and Louisiana (7.2 percent).97 However, the sheer scope of NASA contracting activity led to a substantial dispersal of funds, principally because 40-50 percent of NASA procurement dollars for major R&D contracts were subcontracted. For example, in FY 1969 subcontracts on fifty-eight major prime contracts amounted to $418 million. Of this amount, only $120 million (29 percent) remained with the States to which the prime contracts were awarded, while the remainder was redistributed elsewhere. A further breakdown illustrates how extensive this redistribution really was. Twelve States received more in subcontracts than they did in prime contracts, while seventeen of eighteen States with NASA prime contracts transferred some work to other States. Indeed, many subcontracts were awarded to companies that were probably incapable of becoming prime contractors themselves. Thus, while the dispersal of contract dollars was incidental to its mission, NASA did encourage its prime contractors to award subcontracts on a broad geographical basis.

Analysis of the source selection procedure is in terms of the following questions. First, what did NASA mean by "competition," and what was the actual degree of competition in bidding on NASA prime contracts? Second, how much [95] weight did officials attach to the various criteria for source selection? Third, how efficient was the selection process itself? Finally, how did NASA management regard the procedure; in particular, how far did they believe it was necessary to revise, overhaul, or do away with it entirely ? For NASA, competition was between firms rather than between regions, and its function was to provide the agency with a service, component, or major system. It was in NASA's interest to promote competition because the more competition there was, the less NASA would have to depend on any single contractor; the more entries there were into the aerospace industry, the more likely the industry was to generate technical and managerial innovations; and by spreading NASA funds as widely as possible-as far as this was compatible with its mission-the agency would secure a nationwide base of support for its programs. The extent of competition in NASA procurements is shown in table 4-5.

The entire purpose of NASA's selection process was to procure the best source, one that would be compatible with the agency's budget, the armed services procurement regulations, and external forces that impinged on the agency's selection procedures, such as Congress, industry associations, the Executive Office, the President's Committee on Equal Employment Opportunity, the Small Business Administration, and the like. The criteria NASA used to evaluate potential sources varied depending on the size and complexity of the system to be developed, the point in NASA's history when it issued requests for proposals, and the agency's desire to build a capability rather than to see it concentrated in one or two firms. At the risk of oversimplification, it can be said that NASA officials weighted selection criteria according to the context. One consideration in many early procurements was the need to attract competition in a specific area. This was the justification given by NASA for awarding the contract for the S-II stage of Saturn to North American Aviation rather than to Douglas, which already had the contract to design and build the S-IV stage.98 By placing the S-II contract with North American, NASA would have available three firms-North American, Douglas, and the Convair division of General Dynamics-capable of advancing the state of the art in liquid-hydrogen technology.

The data reveal that NASA chose competitively more frequently in the late 1950s and early 1960s than it was to do later. Between 1958 and 1960 NASA held source selection competitions for eighteen of twenty-one major programs, including Rover, Ranger, Surveyor, Mercury, Nimbus, and the Orbiting Astronomical Observatory.99 But during the period 1958-1965, in sixteen of seventy-six cases involving contracts of $5 million and over, a source evaluation board was not converted' either because there was only one firm with the requisite capability or because the action was a follow-on to an existing contract (table 4-6). Indeed, by the end of 1965, non-SEB actions constituted almost one-third of all contracts for $5 million and over.100 It may be that there were fewer new programs on which to bid or that the high cost of entry served to discourage prospective contractors.

It also seems likely that there was less competition in the mid-1970s than there had been a decade earlier for at least three reasons. First, while it was....

 


[96] Table 4-5.-Competition in NASA awards to business firms, FY 1961-1968.

Award

1961

1962

1963

1964

1965

1966

1967

1968

.

Total (in billions)

$O.423

$1.O30

$2.261

$3.521

$4.141

$4.088

$3.864

$3.447

Competitive-Total

0.276

0.565

1.302

2.119

2.630

2.693

2.698

2.308

Advertised

0.038

0.064

0.106

0.134

0.169

0.111

0.081

0.057

Negotiated

0.238

0.501

1.195

1.985

2.461

2.582

2.617

2.251

Noncompetitive-Total

0.146

0.464

0.959

1.401

1.511

1.395

1.766

1.139

Follow-on after competition

(1)

(1)

0.255 2

0.4952

0.504

0.373

0.347

0.336

Other noncompetitive

0.146

0.464

0.704

0.906

1.007

1.022

0.819

0.803

1 Data not compiled, included in other noncompetitive.

2 Follow-on after competitive procurements of less than $25 000 are included in other noncompetitive procurements.

Sources: NASA Procurement Reports for 1964, p. 4-2, and for 1968, p. 14.

 


[97] Table 4-6.-Breakdown of SEB and non-SEB cases, 1958-1965.

Time Period

SEB

Non-SEB

Percentage

Total

.

12/58-2/61

10

1

9

11

3/61-2/62

10

2

17

12

3/62-3/64

19

3

14

22

4/64-12/65

21

10

32

31

Source: Joseph Fernandez, "The Origin, Evolution, and Operation of the NASA Contractor Source Evaluation Board Process,' unpublished M.Sc. thesis (Massachusetts Institute of Technology, June 1966), p. 23.

 

.....expensive to enter the space business, it was even more costly to stay in. Thus Grumman, which was NASA's number two prime contractor during the late 1960s, virtually withdrew from space systems after completing its work on the Orbiting Astronomical Observatory and the lunar module, both of which were plagued with overruns and technical difficulties. Second, if a company supplied a recurring service or standardized items such as the stages of a launch vehicle, then it was generally more expensive for NASA to go to an alternate source. This, as indicated above, was the justification given by NASA for negotiating only with McDonnell-Douglas for ten improved Delta vehicles, but the same rationale applied to the Centaur stages supplied by General Dynamics or the GE Apollo engineering support contract, which Houston extended to cover the Skylab and Apollo-Soyuz missions. Third, the structure of the aerospace industry changed in the 1960s in ways that tended to reduce competition further. McDonnell Aircraft and Douglas Aircraft, two of NASA's biggest contractors, merged in 1967. Similarly, companies like TRW or North American Aviation, which merged with Rockwell-Standard in 1967, were able to remain and thrive in the space business because they were large and diversified corporations that were not overly dependent on NASA contracts. Under these conditions, newer and smaller firms could enter the space industry only (if at all) as subcontractors-a situation not without danger for NASA's bargaining position.

On the whole, NASA did not give preponderant weight to any single criterion in the selection process. Cost estimates, for example, were apt to mislead either because of underbidding by the contractor or because neither side had realistically appraised the costs involved. Nor was straight technical competence sufficient by itself. In the mid-1960s many top officials concluded that a company could not take on a complex, difficult program without at least demonstrating by past performance that its organization had adequate managerial skills to get the work done. The cost overruns and slippages that plagued so many projects in 1962-l964 - the Centaur launch vehicle (whose prime contractor was General Dynamics)' the RL-10 engine (Pratt and Whitney), Gemini (McDonnell), and [98] the S-IV and S-IVB stages of Saturn (Douglas)-were problems of managerial as much as technical competence. And it was largely as a byproduct of management failures that NASA, in 1964-1965, moved to the principle that

 

greater emphasis should be placed on company management flexibility and its ability to marshal! the management plus technical competence necessary to perform the contract rather than accept a simple technical proposal and evaluate it on technical merits with only secondary evaluation of the organizational factors involved in getting the job done.101

 

All things being equal-which they seldom were-NASA officials would give as much weight to business as to technical criteria: availability of manpower and facilities, preparation of a detailed management plan, understanding of the project management approach, and understanding of the form of the contract that NASA desired to write.102

A proposal that was both technically and managerially sound and that widened the scope of competition stood a better chance of being accepted than one whose only virtue was a low bid. Douglas, for example, won the S-IV contract although its proposal was $5.44 million higher than Convair's. In this case, at least, Douglas won because of Glennan's desire to promote "at least limited competition as we start off this new technology . . . [otherwise] it might well be the case that subsequent competitions for the other stages of Saturn would become quite one-sided." 103 The history of the Lunar Orbiter contract is a more straightforward case of a contract awarded for managerial as well as technical criteria. Boeing, which received the contract, submitted the highest bid of all five proposals. Aside from the considerable technical merits of its proposal, Boeing impressed the selecting officials because of (1) its past experience with Minuteman, the B-52, and Bomarc, and its willingness to use developed components; (2) its ability to assemble a strong project management team, owing, in part, to the phasing out or termination of the Bomarc and Dyna-Soar programs; (3) Boeing's link with Eastman-Kodak and RCA as subcontractors; and (4) the desire of NASA management to encourage Boeing's entry into the design of spacecraft systems. In such cases, what counted in the final evaluation was less the offeror's actual experience in NASA programs than the potential revealed by its handling of other kinds of major systems work.104

As selection procedures became more elaborate, they tended to slow down the entire acquisition process. The matter was thought serious enough to prompt a "procurement lead-time" study in 1968-1969, whose principal finding was the existence of major delays in contracts processing. It took an average of 420 days to process a contract involving a procurement plan, 3 months for headquarters to review the plan, and 47 days for headquarters to approve a negotiated contract.105 It may be that all this merely signified the advent of bureaucracy and red tape. Or it may have reflected the learning process NASA had undergone during the preceding decade in negotiating with its prime contractors. Or, finally, what seemed to be a delay in the acquisition process may have really masked delays in the decision to fund new programs. As Finger observed, NASA had gone from "a [99] large, fast-moving program aimed at a clearly identified national objective to a situation in which reductions have been made in our program budget and, therefore, to a requirement for a very detailed examination of every element of our program before final approval is given to proceed. There is no question . . . that the uncertainty of the AAP [Apollo Applications] program scope and schedule and the delay in making firm AAP program decisions were reflected in long procurement lead times." 106 In addition, the NASA budget was tied to the planning programming-budgeting cycle introduced by McNamara at DOD in 1961 and imposed on many civilian agencies in 1965. One effect of the new system was that it required that NASA submit its projected cost estimates to Congress earlier than good estimates could be produced. By 1969 NASA was some six years beyond the time when Congress had, so to speak, agreed to hand the agency a blank check. The uncertainty as to the future of NASA programs, the parceling of the NASA budget for review by many congressional committees, and the elaborate reporting requirements engendered by planning-programming-budgeting meant that NASA could not let contracts on the assumption that a program funded for one year would automatically be funded thereafter. Program planning and the acquisition cycle were inextricably linked. As one changed, so did the other.

But there was another, more technical, reason for the protraction of the acquisition process. As a large research and development organization, NASA had good reasons to make its regulations as specific and as uniform as possible. Kaufman's observation concerning organizational change applies here.

 

The more experience an organization acquires, the more numerous it discovers the complexities and the ambiguities of its work to be, and the more its leaders feel obliged to clarify policies and refine procedures.... In quest of the best and most up-to-date technical performance, organizations in this way generate still more sets of specifications governing the actions of their members.107
 

By about 1966 the procurement cycle showed signs of becoming a lengthy procedure, owing to requirements imposed on NASA from within the agency and by other Federal agencies. NASA procurement regulations stipulated that all contracts negotiated by the centers above a specified dollar value (as well as contracts for individual facilities) had to be reviewed and approved by headquarters. Contractors were encouraged (or required) to submit cost reduction plans, identify new technology, and comply with the Government's equal employment opportunity (EEO) program. Requirements imposed on NASA from outside multiplied throughout the 1960s. NASA was designated a "predominant interest agency" for investigating EEO complaints against more than 200 prime contractors and subcontractors at some 450 facilities.108 The agency was also assigned full responsibility for setting aside (i.e., reserving) competitions for small business, rather than determining them jointly with the Small Business Administration. In fact, there was a substantial area of public policy that NASA, along with all other Federal agencies, was required to implement through the procurement process: the Buy American Act, the Fair Labor Standards Act, the Work Hours Standard [100] Act of 1962, the Copeland Anti-Kickback Act, the rulings of the U.S. Comptroller General, the National Environmental Policy Act of 1969, and the like. The lengthening of the procurement cycle was the sum total of forces at work not only within NASA but throughout the Federal community. There was the proliferation of forms and review procedures that are one sign of Parkinson's law at work. More significantly, perhaps, the cycle lengthened because, in the course of planning programs, NASA program managers had come to understand the acquisition process better and to know what to expect of their contractors.

At the end of the 1960s NASA management looked at its selection process and found it to be good despite complaints that procurement lead time was too great; that losers in competition did not know why they had lost or winners why they had won: that requests for proposals were poorly drawn, in part because they were prepared by different people, "so that the need for all the material is seldom evident to one person"; and that a severe case of "echelonitis" existed between the program offices and senior management."109 These were defects in a system that most of NASA considered to be basically sound. NASA officials dismissed out of hand the fundamental criticism of the process-that it was essentially noncompetitive. This thesis, forcefully stated by Prof. Edward B. Roberts of the Massachusetts Institute of Technology, was that the process underlying R&D awards was "informal and highly selective" and that the formal acquisition process succeeded only "in increasing the costs of research and development, adding time delays, and producing other damaging effects on government-sponsored research and development.")110 Part of this critique was only a more forceful restatement of what NASA procurement officers were saying privately. But they were not prepared to go the whole route and concede with Roberts that many large R&D contracts were in effect preselected and that the weightings given to proposals only amounted to "after-the-fact representations of general agreements . . . justifications for decisions, rather than causes."111

NASA's position may be summarized as follows. The NASA task force that studied the NASA acquisition process in 1970-1971 concluded that it "sufficiently met NASA's needs and objectives."112 The task force preferred to judge by results. NASA had tapped-to a degree had created-an industrial capacity adequate to carry out its programs. It had created safeguards, such as the recompetition of support contracts and phased project planning, to prevent the agency from becoming captive to its contractors. NASA would concede the delays, the vague criteria, the unnecessary review levels, and the tendency to use technical criteria when the contractor as a whole was to be judged. But to abandon formal source solicitation would have been politically undesirable, not calculated to promote the agency's goals, and more likely to leave particular firms in sole-source positions for given areas than would otherwise be the case.

 

USE OF INCENTIVE PROVISIONS

 

NASA introduced incentive provisions into its procurement system when the entire subject of R&D contracting was under review by Congress and the Bureau [101] of the Budget. The time was ripe, judging from the circumstances that led to the introduction of incentives: the notorious inefficiencies of CPFF contracts, with cost prediction errors sometimes mounting to more than 200 percent; the recommendations of the Bell report; the March 1962 revisions of the armed services procurement regulations that authorized DOD to make more use of incentives; the pressures within NASA to use incentives, especially in the development phases of large systems contracts; and a series of meetings in December 1961 involving Webb, the chairman of the Atomic Energy Commission, and contractor representatives, in which they discussed the rationale for the cost-plus-award-fee (CPAF) concept. Once the decision was made to write incentive provisions for new contracts and to convert older ones from CPFF, NASA moved quickly. NASA issued one incentive contract in FY 1961, six in 1962, ten in 1963, and thirty-four in 1964. By the beginning of FY 1967 NASA was managing some 200 incentive contracts (figure 4-1).113 The selection of contracts for conversion was made by top management, the program offices, and the staff offices, especially the Office of Procurement. However, the criteria for applying incentive arrangements in each situation had to be developed within the centers.

The basic criterion in applying incentives was to not use them before both sides had arrived at a clear definition of requirements. In other words, project definition and contract negotiation went hand in hand. As a corollary, most of the centers, when questioned by headquarters procurement officials in 1964, recommended that NASA use a phased approach to R&D contracts. Even before the introduction of phased project planning, center officials had instituted a phased contracting cycle. They began with a CPFF contract for research and initial development, converted to cost-plus-incentive-fee as requirements became firmer, and to fixed-price-incentive as cost prediction became more accurate. It was essential, however, that both NASA and its contractors understand how incentives worked. Or, what amounted to the same thing, "you've got to negotiate the contract all the way through or you're hung."', 114 It was the lack of experience in handling incentives that led to most of the problems, such as placing incentives on final mission performance without including incentives for important technical subgoals.

Granted that early definition was imperative, what improvements did NASA program managers have in mind? Consider, for example, the Office of Manned Space Flight (OMSF), which was spending more than two-thirds of NASA funds obligated for R&D. Meeting in executive session, the OMSF Management Council resolved to give primary emphasis to schedule, then to cost, and third, to performance.115 In the opinion of Council members, emphasis on schedule was the best or even the only way to force both parties to define the end product because there would be "a premium on getting the specifications correctly written in the first place, because the Government personnel realize that the contractor is going to insist that they accept the hardware based on these specifications in order to earn the schedule incentive fee."116 For OMSF, there was good reason to stress schedule over costs. Schedule and cost control were very closely related; Mueller....

 


[
102]

Figure 4-1.-Number of incentive contracts under administration by type of contract, FY 1961-1966.

Figure 4-1.-Number of incentive contracts under administration by type of contract, FY 1961-1966.

 

...believed that the only way to keep costs down was to keep on schedule. The use of schedule incentives worked better than any other means to create "hard" communications between the project manager and the hardware producer, to get the contractor to deliver its hardware in order that NASA might conduct early ground testing, and to give OMSF added insurance against unforeseen difficulties Not that schedule and performance were identical; as one program manager put it, "There's no point in delivering a product that's unsatisfactory on time."' 117 To improve performance NASA and the contractor had to define the discrete tasks or work packages, into which the work was divided; without such definition neither side could communicate with the other.

[103] Almost as soon as NASA introduced incentives, it began to study their effectiveness Just before the Hilburn task force submitted its reports on schedule slippages, the Office of Procurement asked the centers to report on their experience with incentives and subsequently contracted with the management consulting firm of Booz, Allen and Hamilton to study the effectiveness of NASA incentive contracts. The final report was submitted in August 1966, and it furnished a detailed balance sheet of the benefits and disadvantages of incentives. 118 It did not, perhaps, provide conclusive evidence on their long-term results. But it was based on a study of fifteen contracts that totaled $1.5 billion, or 62 percent by dollar value of NASA incentive contracts at the time the study began in July 1965. 119

The task force discovered that, properly applied, incentives were more effective than CPFF in improving delivery on schedule and performance according to specifications. Incentives did not cost more to administer than CPFF and reduced the need for day-to-day surveillance of contractor operations. Finally, incentives led to better program definition, although more as a byproduct than as a stated objective.120 On the other hand, there was little apparent relation between the intensity of contractor motivation and the use of incentives. Typically, the contractor was under pressure to secure new business or follow-on work or, in the case of the Manned Space Flight programs, to perform well in the face of the publicity that these programs, above all others, engendered.

The task force also concluded that in at least four cases incentives were introduced under pressure from headquarters and that in three cases incentives were used to control cost growth. The center personnel had not been enthusiastic about incentives. They believed incentives required more controls, not less, and lengthened, rather than shortened, the overall procurement cycle. Incentives placed a heavier administrative burden on the technical staff. And by increasing the element of risk, incentives prompted the contractor to ask for higher fees-say, 8 percent rather than 7 percent. Only recognition that incentives had improved contractor performance in a number of cases caused center employees to change their minds. Also, by the end of FY 1964 almost 700 center personnel had received some training in the management of incentives. By then-and certainly by the conclusion of the Booz, Allen and Hamilton study-NASA procurement policy had moved beyond the experimental stage. By 1966 NASA was able to define with some precision four "desired conditions" to be considered in assigning incentive contracts the procurement should require no significant technological breakthroughs; there should be no overriding need for intensive technical direction during the life of the contract; it should be possible to estimate the cost of getting the job done; and, within constraints, the contractor should be given almost complete freedom to perform.++++ 121 Moreover, NASA had at its disposal a variety [104] of techniques that ranged from incentives directed to a single variable (cost, schedule, performance), to multiple incentives, to contracts employing interdependency, that is, "the amount of penalty or reward . . . under any one incentive element in a multiple-incentive contract will vary according to the achievements under the other incentive elements."122

And yet, while NASA had learned much in handling incentives, it had not reaped the full benefit of their use. NASA employees continued to monitor contracts even when incentives made such surveillance unnecessary. Also, while NASA had improved its method for processing contract changes, it had not done enough to reduce the number of changes. The Booz, Allen and Hamilton team concluded that the benefits of incentives were not yet well understood within NASA. Incentives were used, so to speak, to communicate NASA objectives, but not to convey the relative importance of cost, performance, and schedule. On the other hand, the task force members were well aware that the structure of NASA contracts could go only so far to eliminate bottlenecks. As they put it, "certain objectives . . . appear to be largely beyond the control or 'influence' of the contract structure," especially cost growth.123 Incentives might reduce but they could not eliminate the technical uncertainties dogging most R&D programs. Nor was the sort of technical definition required by incentives already in use available when it was needed most-at the beginning of the program. A contract designed to cover everything from early development phases to small-quantity production was not flexible enough for the kind of program (which covered most of NASA's R&D work) where the end item itself changed over the life of the program. The contradiction between fixed targets and changing programs remained insoluble.124

In sum, the recommendations of the Booz, Allen and Hamilton study group were directed toward attainable goals. Incentives were not to be used as a cure for unavoidable uncertainties, nor could they compensate for the lack of program definition that their use presupposed. Within these assumptions, the study group recommended continued use of incentive contracts for R&D and support services; use of a phased approach in contracting, in which the final fee would reflect the contractor's accomplishment in all phases; introduction of "steeper" (higher) incentives as technical definition improved; and NASA's implementation of additional case studies on recent contracts.125 The basic conclusion, however, was that NASA's use of incentives had been justified by the results. And though the study group did not say so explicitly, its drift was that incentives were the most effective way in which NASA could simulate market conditions in what was essentially a nonmarket environment.

The conclusions of the Booz, Allen and Hamilton report suggest a final observation. In the mid-1960s NASA was moving toward more rather than less technical direction of programs. The introduction of incentives and their widespread use influenced and was affected by the establishment of phased project planning guidelines. It was the essence of phased project planning that it kept open many technical options, allowing management to intervene at almost any point before the final developmental phase. And by recommending that incentives [105] be phased to stress those subgoals that collectively made up the program objective, Booz, Allen and Hamilton clarified what has been implicit throughout this discussion: While procurement and program planning were two ways of regarding a unified acquisition process, the latter was logically prior to the former.


* An autonomous Headquarters Procurement Branch was established in July 1961.

** See appendix D for definitions of contrails authorized by NASA.

*** In one case NASA departed from the principle of using DOD capability. From 1962 to 1967 NASA maintained a contract management unit at North American Aviation's plant in Downey California, to represent NASA on plantwide matters and to provide day-to-day support for the Apollo command and service modules and the second stage of the Saturn V.

**** NASA programs were canceled or delayed for budgetary reasons, either because the Bureau of the Budget disallowed them, Congress would not fund them, or NASA sacrificed them on the altar of Apollo. Examples include Voyager, a Mars mission canceled in 1968, the Advanced Orbiting Solar Observatory, and the large solid motor.

+ From March 1962 to its termination in May 1972, the Bellcomm contract (NASW-417) and its extensions cost NASA about $91 million. At the end of FY 1968, the GE (NASW-410) and TIE (NASW-1650) contracts had run to $670 million and $52 million respectively.

++ However, in a personal communication to the author, Dr. Seamans has noted that "we originally planned to carry this effort out primarily within NASA as part of the Apollo office When we had difficulty recruiting the caliber of people we wanted and obtaining sufficient manpower slots from the Office of Management and Budget (sic), we went to AT&T and asked them to set up a dedicated organization, which became Bellcomm."

+++ Sneak circuits are inadvertent electrical paths.

++++ One former NASA official gave an example of the sort of freedom that NASA gave to its prime contractors. One year NASA found that Boeing needed $10 million for Lunar Orbiter that was not programmed in the budget. To keep from breaking the contract and, more importantly, from upsetting Boeing's management plans, NASA reprogrammed the $10 million from other projects.


previousindexnext