James Roherty’s classic 1970 book, The Decisions of Robert S McNamara has a pair of interesting discussions on the takeover of systems analysis on military programs (which continues to this day). He first quotes RAND analyst ES Quade:
[The cost-effectiveness analyst] is likely to be forced to deal with problems in which the difficulty lies in deciding what out to be done, not simply in how to do it…. The total analysis is thus a more complex and less neat and tidy procedure which is seldom suitable for a quantitative optimization over the whole problem…. The situation is not like an empirical science, which starts with observed facts, but more like that of mathematics, where the results take any ‘validity’ they might have in the real world from the assumptions… it is important that the assumptions be the ‘right’ assumptions.
This is the classic view of the systems analysts. Every valid specification of a weapon system can be derived from a set of assumptions, just like mathematical proofs derive from their axioms. That allows people schooled in quantitative analysis to be in the best place to make “objective” decisions on weapons, rather than an interaction of technologists and military operators who have “parochial” interests.
(In my view, parochialism is a condescending way of saying the military services have inarticulate knowledge, domain expertise, and genuine difference of opinions based on that specialization.)
The “assumptions” which so often drive outcomes of systems analyses, including famed cases like strategic bombers in the 1940s and nuclear carriers in the 1960s, are often hidden from view and debate. Moreover, even if one were to presume the “optimal” assumptions, the mathematical process cannot provide the optimal specification that maximizes those assumptions.
Consistency and completeness in arithmetic has been proven for first order logic, such as found in the statement “every requirement has a technical solution.” However, Gerhard Gentzen and Kurt Gödel proved arithmetic cannot be both consistent and complete for second-order logic such as found in the statement “every set of requirements has an optimal technical solution.”
Anyway, to contrast ES Quade’s take on systems analysis requiring assumptions, Roherty later quotes the articulate views of Captain Stanley Barnes:
The key element of all analyses is the use of assumptions. The key assumptions, which can swing results one way or another, have a way of being removed from conclusions to such distances that their criticality is forgotten… The question of assumptions highlights an even larger problem. Military concepts of problem solving are much more than a logical process alone. In the same way that systems analysis has a major contribution to make to problem solving, military concepts of accountability should influence decision making. The basic precept of military professionalism is accountability for some part of national security. Operationally this is direct and inescapable. In other matters, it takes the form of service reputation for professionalism, and its consequences are highly effective constraints. In evaluating the defense planning organization, the first question should be – where, short of the highest levels, does accountability commensurate with authority, restrain incompetence, arrogance, and all other human frailties which can inhibit responsible and objective performance of duty.
I think Capt. Barnes is correct that there is an inherent tradeoff between reliance on (1) a logical process of quantitative analysis; and (2) an interpersonal process of duty and responsibility. One focuses attention on the weapon system itself, requires numerous predictions are total lifecycle costs and benefits fraught with error. The other focuses attention on building an institutional culture of personal accountability where specific technical details are delegated and reviewed.
The construct of budgets is critical in this debate because budgets are essentially a contract between policy-maker and administrator. The systems analysis method of choice is represented by a PPBE program budget. The traditional military method of choice is represented by a portfolio-based organizational budget.
Leave a Reply