On Balance: Systems Analysis, Cost-Effectiveness Analysis, Benefit-Cost Analysis, and Government Decisions
Rob Moore invited me to share further reflections on BCA in practice, lessons from 50 years ago but still relevant today.
First, be sure to get the basic facts surrounding your problem right. This may sound obvious, but it isn’t, and getting the basic facts right may prove to be very difficult. In the 1950s, the U.S. Government and our NATO Allies accepted as fact that NATO land and tactical air forces were hopelessly outnumbered by the forces of the USSR and its Warsaw Pact allies, so outnumbered that in the event of a Soviet attack, our only defense would be first use of nuclear weapons. The widely accepted numbers were 175 divisions on their side v. 25 on ours. This was an extremely dangerous and unnecessary strategy. John F. Kennedy criticized this as confronting the President with “a choice between Suicide or Surrender.” And when McNamara became Secretary of Defense, Kennedy ordered “Give me some better alternatives”.
Of course, this was a huge exaggeration of Soviet and Warsaw Pact capabilities. But to debunk this myth took years of effort. K. Wayne Smith and I told the story in our book How Much is Enough?, published by RAND. And economic analysis proved to be a most valuable tool. For one example, the Joint Chiefs were counting the MIG 21, the Soviet front line tactical fighter as equally capable as our F-4. The CIA had acquired a MIG-21 from a defector and obtained an estimate that if we produced it in our factories, it would cost about a third the cost of an F-4. The F-4 could carry a much larger payload and deliver bombs with greater accuracy because of superior electronics the MIG-21 did not have. I remember McNamara explaining to the Chiefs if a MIG 21 is as good as an F-4 we are buying the wrong plane. So getting the basic facts right can take a lot of effort and willingness to challenge the accepted conventional wisdom.
Next, when analyzing a problem, always start with the grand totals, the big picture, so that your problem can be situated in the relevant context. In the Systems Analysis Office (now Cost Assessment and Program Evaluation or CAPE) we used to call this McNamara’s first law of analysis. He said this in the context of a study of the air battle in central Europe in which the whole Warsaw Pact air forces were in combat with a faction of the NATO forces. We changed the outcome by assuming that NATO would commit more of its forces.
A tool that I found useful is what Paul Samuelson called “the generalized law of diminishing marginal returns”, something familiar to every economist. In the situation in which I first recognized and used it, we had already committed to the deployment of 1000 Minuteman ICBMs, each in a concrete and steel underground silo. How many more should we buy? The Joint Chiefs were recommending a total of 2400. We also were deploying a fleet of ballistic missile-carrying submarines. How much was enough? I plotted the curve of targets destroyed versus Minutemen deployed, and, under a wide range of assumptions, it got pretty flat above 1000. So the Secretary decided to stop at 1000. I remember an admiral coming to me and saying “Dr. Enthoven, I want you to know that our program is not on the flat of the curve.” I replied “We’ll have to do the calculations and see.”
President Johnson liked our work product because our analyses served as the basis for the recommendations McNamara provided him. So in 1966, he directed that all the departments in the executive branch should have a similar office. At least in some cases, the recipients of the order were at a loss as to what it meant and what to do. (Although in some cases such as DHHS, they hired a RAND alumnus to head the office who was a great success.) I was reminded that the Office of the Assistant Secretary of Defense for Systems Analysis in the Defense Department had the benefit of about 10 years of research on these issues at RAND learning how to identify the issues and to analyze them. We developed and argued out the principles of what we then called Systems Analysis. (Systems Analysis, short for Weapon Systems Analysis, was a discipline-neutral term of art that reflected the fact that the problems called for collaboration of scholars from several different disciplines.) And, of course, the success depended a great deal on the intellect of Robert McNamara who could and did ask many good and penetrating questions and who could spot weaknesses in analyses and demand improvements.
In later years when I switched my attention from National Defense to health care, I could find many “flat of the curve situations.” With the help of a grant from the Henry Kaiser Family Foundation, I created a course on “Analysis of the Costs, Risks and Benefits of Medical Technology” in the Stanford Graduate School of Business, in collaboration with the Medical School, which has continued to this day, about 40 years later. With all the understandable concern over the costs of Medicare, Medicaid, and Health Insurance for public employees, development of BCA for medical technology seems timely and likely to grow in importance.