In a previous post, I discussed how it is critical to separate the idea of corporate welfare from efficient organizational design. In this post, I briefly want to introduce the main problem confounding the analysis of corporate welfare: the sustained attack on Government in-house capabilities.
The demise of in-house knowledge.
When the Air Force sought independence from the Army after WWII, its leaders claimed that by employing practices from private industry, they would become far more efficient than the messy Army. The argument succeeded, and it affected Air Force management. Already by 1953, the Air Force outsourced 90% of its R&D budget. By contrast, the Army and Navy performed substantial amounts of work in-house. For the Navy, it outsourced 65% of research, though the figure for development was somewhat less at 40%. The Army also had robust development capabilities, which centered around its six arsenals under the Ordnance Department.
The Air Force led the way toward the acquisition process we now have. They were the first to align their budgeting process by program, whereas the Army and Navy continued to plan organically according to their technical services and bureaus. In the Air Force, it was the general staff that planned and coordinated projects — even though for most of the 1950s it was mostly an unplanned process in reality. But the formal process had the staff plan programs and assign a manager to a systems program office to go execute the program. The manager was expected to outsource the program out to a single prime contractor and ensure delivery.
The Army and Navy lagged behind the Air Force until Robert McNamara’s sweeping reforms of the 1960s. He forced the hand of the Army and Navy to conform to the concept of the program budget, system project office, and single prime contractor. The Army bureaus lost their statutory role in 1962 and later scaled back operations, including shutting down Springfield, Watertown, and Frankford arsenals. A similar fate befell the Navy bureaus in 1966, though their remnants continued to be an important source of innovation in missiles, rockets, and lasers.
Effect on contracting.
The transfer of weapon systems expenditures to organizations external to the DOD brought issues of contracting to the fore. When the actual operations of doing experiments or bending metal occur in-house, the executive may act very much like a military commander in the field. He can express his desires, or lay out his “demand function,” and command action be taken. Depending on how he judges the resulting action when compared with his updated expectations, the executive or commander can reward or punish his subordinates. This method of administrative control is often called after-the-fact control.
However, when defense executives seek production from the open market, whether it be firms, universities, or non-profits, they must use market exchange mechanisms characterized by contracts. Perhaps the most important aspect of contracts is the fact that they must be fully specified before-the-fact to ensure that the responsibilities of both parties are clearly stated and evaluated.
The uncertainty of R&D contracts made them legally ambiguous because terms could not always be met in the manner pre-specified. There are two general tactics for resolving the problems of uncertainty in defense contracts:
First, the in-house method of old.
Before the 1960s, the Army and Navy often coordinated weapon systems development themselves, in-house. For example, the Army Jupiter program and the Navy Polaris/Fleet Ballistic Missile programs were run in-house. For the latter program, the Navy contracted separately with several subsystem developers, often competing on the same subsystem. The Navy also employed the assistance of a private integrator, but they didn’t supplant Navy decision rights.
With large amounts of in-house talent, the Navy was able to create a stunning success in short order — creating a nuclear submarine that can also launch nuclear intermediate range ballistic missiles. With the Navy as a focal point, it could coordinate disagreements between contractors and resolve certain issues of technical uncertainty.
One benefit of the Navy’s technical talent was its ability to evaluate contract requirements and price. Such evaluation cannot be performed in a Sears catalog (or even on Amazon). It seems to take many years of experience to eyeball a fair deal, and then follow that up with assessing the deliverable.
Perhaps more importantly, the Navy did not try to fully plan the program development. It broke down the effort into smaller sub-tasks in dimensions of time and system components. This made individual contracts easier to evaluate and price. It also avoided locking in program decisions over many years until uncertainties were resolved. The first Polaris test, for example, only sought half the range and accuracy as its requirement. It learned from there.
The combination of breaking programs down into smaller tasks, and having ample in-house talent, allowed the Navy to contract with control exerted after-the-fact. Instead of taking firms to court for contractual default, the bureaus punished firms which did not expend resources in an appropriate way when judged after-the-fact by not awarding them future work. The repeated interactions between a diverse set of government and industry participants led to significant reputational effects.
Second, the program office method.
The Air Force took a very different approach. It relied on a systems analysis of alternative designs, which would inform the optimal requirements/cost of a new program. That kind of optimization exercise required the program lifecycle be costed (if not through sustainment, then at least through production). That meant predicting the course of technology over the next several years.
In order to avoid wasteful expenditures in massive programs, the Air Force sought to use analysis to direct its managers to single-best plans. As the Air Staff determined what kind of system was needed, it whittled down the tradespace in terms of specifications and cost. When that authorized program was given to a program manager, he was expected to engage in another long planning period, this time for the development contract.
In contrast to the Army and Navy, the Air Force looked to put the totality of development work into one contract. This created a premium on the accuracy of plans, which, to the Air Force, meant increasingly fine levels of detail. Here’s the problem with that approach:
Any attempt to schedule an entire R&D program at one time is likely to lead to inefficiency, either because plans for the later stages may have to be scrapped and remade on the basis of information yielded by early tests, or because, in pursuing premature plans, a development program may fail to profit from new information gained along the way. Either case will cause delays, or raise costs, or both.
Results.
The Air Force thought it could handle uncertainty through more analysis and longer lists of requirements. Before McNamara, the Army and Navy primarily handled uncertainty by building in-house talent and partitioning program tasks into smaller bits.
Now, what does this all have to do with corporate welfare? Well, two things:
(1) When you have smaller contracts and smaller tasks you can avoid the “too-big-to-fail” problem where poor performance on one contract can bankrupt the company and kill the program. (When the DOD program is vital to national security, the DOD will always be more risk averse than the industry execs with golden parachutes.)
(2) In-house technical knowledge can use qualitative as well as quantitative means for evaluating contractor performance, whereas the Air Force method relies solely on the quantitative merits as described by contract requirements.
More briefly, Government in-house capabilities avoids problems of (1) dis-economies of scale; and (2) extreme information asymmetry. The contractors have all the information and knowledge. The Government basically looks to make sure contractors followed “the process” and provided more data than anyone can look at. This is the situation that leads to corporate welfare.
It sounds easy in the market. If a company screwed you over on a purchase, you don’t deal with them again. But what happens when you make a deal with the company for something of enormous importance to be delivered in 5 or 10 years?
Well, first, in the DOD, the person evaluating the contractor 5 or 10 years later is not you. It’ll be some other person completely.
And second, as deficiencies and cost growth becomes apparent, it is only done so incrementally, each time causing you to fall prey to the sunk cost fallacy.
If you knew at the first change order that the whole thing would be unworkable — you could perhaps exercise some clause and cancel the contract at little time/cost wasted. But if you lacked in-house knowledge to make that evaluation, then you’d buy what the contractor is selling in terms of explanations. You’d continue along the sunk cost fallacy, seeing costs creep up even while it was obvious to anyone looking that the program was in trouble within a year.
(Note: contractors are often way behind on even front end tasks, and yet somehow schedule growth isn’t reported until much later, roughly 60-70% complete.)
Anyway, the length and scale of the tasks being contracted out confounds the responsibility for failure. And in many cases, the huge scale of defense programs is artificial. “World saving” joint-service programs are encouraged by Congress and others due to a misunderstanding of technology development. They have delusions of huge economies of scale in production and sustainment which rarely materialize. They neglect the role of discovery and the growth of knowledge.
Sometimes scale can’t be avoided in a program. And that’s precisely where Government in-house knowledge helps the most. Our biggest programs were all lead by the Government, including the nuclear explosives, ballistic missiles, the nuclear navy the moon landing, and even GPS. The development of these systems is not associated with a single prime contractor executing to a fully-planned project. That organizational design, I think, also helps Government officials identify poor contractor performance without encountering problems from “too-big-to-fail.”
I hope this has been useful context for thinking about how the decline of Government in-house capabilities has produced symptoms that look like corporate welfare.
Leave a Reply