Why the Military Can’t Get the Figures Right

Twenty years ago, Alain C. Enthoven was one of America’s most controversial intellectuals in the field of military affairs. He had gone to the Pentagon in 1961 to act as a civilian adviser to Secretary of Defense Robert McNamara. By training, he was an economist, with degrees from Stanford (1952), Oxford (1954), and MIT (1956). From 1956 to 1960, he worked at the RAND Corporation, doing contract research for the Air Force.

Enthoven and the other young advisers who joined the Department of Defense at the beginning of the Kennedy administration were derisively called the Whiz Kids. They tried to apply statistical analysis to problems that had traditionally been resolved by a mixture of intuitive reasoning, inter-service log-rolling, and congressional politics.

Partly because of their youth and their occasional lack of tact, partly because their methods seemed uncongenial, the Whiz Kids made enemies. But, despite the hostility aroused by their personalities and their methods, they contributed to important decisions on military spending during the Kennedy and Johnson years. In 1971, Enthoven and K. Wayne Smith wrote How Much Is Enough ?, a book that deals with questions of military expenditure.

Shortly before leaving the Pentagon, in 1968, Enthoven joined the board of directors of Georgetown University. There his discovery that the cost of operating the medical center was one of the university’s major financial problems led him to specialize in the economics of health care. This interview took place in his office at Stanford’s Graduate School of Business, where the fifty-five-year-old economist has taught since 1973.

Your work at the Department of Defense and the work you’re currently doing in health care analysis both demand that you balance the opinions of people who are considered to be professionals against the opinions of independent analysts who come from other disciplines. How do you resolve that problem?

When I got involved in health care, I realized that doctors were a lot like admirals. In both groups you have a technological elite of highly trained people with a very distinct point of view. They go around in white coats and they say: “What I am doing is too important to be contaminated with considerations of money. Your job is just to get me the money so I can go on doing this important and wonderful thing.”

And in both cases I had the reaction that when it gets so expensive, considerations of value for money have to be brought in. It’s a very exciting challenge to figure out how to do that. It’s not simple or obvious or easy.

Among doctors and admirals, don’t you find an attitude of “Leave it to us—the professionals—we know best”?

Yes. What they don’t realize is the limited perspective of their professional interests. A man who has spent his whole life flying and developing bombers is very likely to think in terms of more bombers, even after technology has changed and bombers are no longer an effective use of resources. Despite the fact that historians have written about the Maginot mentality and about generals preparing for the last war, I think generals are just as smart as other professionals. In any profession, including the academic, there is a strong preference for the particular means you have been trained in; so it is with the military, just as it is with physicians and academics.

When some new technology comes along like the Polaris submarine, the admiral and the general sitting at the bargaining table are in charge of aircraft carriers and bombers, respectively. They like aircraft carriers and bombers, but they don’t like this new technology, which, because it is small and new, isn’t represented at the bargaining table.

So, you get people who are inherently unwilling to adapt to new technologies that upset traditional roles and missions and the balance of power. They’re delighted to have a new technology that makes bombers fly higher, faster, and farther, but they’re not interested in a technology that makes bombers obsolete.

Could you tell me something about the kind of thinking you were employed to do at the Defense Department? For example, how was it decided to cancel the B-70 bomber program and go ahead with the Polaris submarine?

The Air Force people, led by General [Curtis] LeMay, wanted the B-70 because, through their whole careers, they had developed one bomber after another that could fly higher, faster, and farther than its predecessor. The enemy of the bomber was the fighter plane, and the bomber that could fly higher and faster would be more likely to be able to outrun the opposing fighter planes. And if it could fly farther, its bases could be farther from enemy lines and therefore less likely to be bombed. You might say that a whole professional lifetime of experience told these people that to extend this line of thinking was the appropriate thing to do.

Our opposition to the B-70 was based primarily on strategicair-power studies done at the RAND Corporation. These studies clarified the point that our main purpose in the nuclear age should not be to threaten a first strike as a way of coercing the Russians, but rather to be able to survive an attack and inflict unacceptable damage in retaliation as a way of deterring attack.

The first problem with the B-70 or with today’s version, the B-I, is that bombers are stationed on relatively few air bases. These bombers are concentrated and vulnerable and can be knocked out by enemy intercontinental missiles. Rather than deterring attack, they’re more likely to invite attack.

We came to prefer weapons like the Polaris submarine, because that couldn’t be knocked out in a surprise attack, and the Minuteman ICBM, based in dispersed concrete-andsteel silos so that the Soviets would have had to use several missiles to destroy one of our Minutemen. As long as we had about as many missiles as they did, enough of our forces could survive to strike back.

The B-70 was very vulnerable on the ground, and it was extremely expensive compared with ballistic missiles.

What did the Air Force say in rebuttal?

They fought back mainly with political power—with appeals to their supporters in the Congress and to people who were thinking in the same terms as they were. They would say, for example, that the great thing about the manned bomber is that you could launch it subject to warning and then recall it if it was not a real attack. Our reply to that was, “The trouble is that you have to launch it on warning, because it’s so vulnerable; that’s not a virtue, it’s a by-product of its deficiency.”

In the case of the B-70, that’s why the Air Force ultimately lost. Gradually it became apparent that the bomber was terribly costly and definitely not worth it. The B-70 would be vulnerable in the air—to surface-to-air missiles—as well as on the ground.

With the Polaris submarine, the key decision, made by President Kennedy and Robert McNamara right at the beginning of their administrations, was to move very quickly to get into a position of having survivable retaliatory power. So we accelerated the production of the Polaris and the Minuteman missile and, at the same time, to help pay for it, phased out the B-47 bombers, which we perceived as being very vulnerable to a Russian missile attack.

The question “How much is enough?” was essentially a “flat-of-the-curve” argument. If you plot a curve that shows, on the horizontal axis, dollars spent, and, on the vertical axis, targets destroyed, that curve gets very flat as more dollars are spent. More missiles don’t get you more target destruction, because you’ve shot at everything that’s worth shooting at.

On a number of occasions, you found yourself arguing against what the Air Force wanted, and some Air Force people felt surprised and hurt. But you had been employed to do the same kind of analysis at RAND, which was an Air Force think tank. After they’d been paying you to do this for years, how did it happen that they were surprised?

To the credit of General LeMay and other Air Force leaders, they had left RAND people with a lot of freedom to call the shots as they saw them; at RAND in those days, we were not strongly constrained to follow the Air Force party line. For example, a RAND study of the Polaris weapon system in the late fifties said to the Air Force, in effect, “This is a good weapon system—don’t fight it, because it’s really an important contribution to national defense.”

As I look back on it now, I can see that part of the price of that freedom was that we were giving confidential advice to the Air Force. We didn’t talk to the press, nor did we talk to the Army or the Navy. Our thinking was not even widely disseminated in the Air Force, although it was available at headquarters. I think the surprising thing, for them and for us, was when people like myself who had been working for RAND suddenly found themselves working for the Secretary of Defense.

The immediate precedent for the work you did at the Defense Department was the civilian research on military problems carried out during World War II, first by the British and then by the Americans. What are the earlier precedents? Did George Washington and Abraham Lincoln have the kind of help American leaders have had since World War II?

I don’t doubt that there have been many applications of “quantitative common sense” down through the ages. But the idea of independent civilian expertise on military topics, and the use of advanced methods of quantitative analysis on broad policy issues, were quite new with the RAND Corporation and McNamara’s Pentagon.

One of the first things you did at the Defense Department was to try to make it possible to compare the estimated costs of a project with the actual costs. Since your time, has there been any useful passing along of wisdom on those problems, or are people just as astonished about huge cost overruns as they were in the past?

The first thing to understand is that it’s not really a matter of astonishment. The game was, and unfortunately still is, to understate the costs in order to get the weapon sold—to tell the Congress that it will cost a billion dollars when you know it will cost ten billion.

We tried to improve methods of cost estimation so that we would have our own independent estimates before making a decision to go ahead with something. On some weapon systems, like the B-70 or the antiballistic missile system, our own cost estimates were considerably higher than what the military service in question was estimating. We informed the secretary of our estimates; a weapon system that might be a good deal at five billion might be a very bad deal at forty billion.

McNamara spent a lot of effort on what was called totalpackage procurement, and the C-5A transport plane was the big trial of that idea. We would have a competition and make a complete contract to develop and build the airplane to certain specifications, with a price agreed upon in advance. We made allowances for inflation, but basically the real costs were specified in advance, and the supplier was taking a risk.

What happened was that the bid Lockheed submitted was far below what it actually cost them to produce the plane. They expected that the contract would be renegotiated, and when it wasn’t, they ran into a situation where they would be losing a lot of money or even going broke. So they and the rest of the defense industry orchestrated a campaign against that kind of contracting; they wanted to go back to the old cost-plus contract. Eventually they were able to prevail; now we don’t have much competitive contracting.

Was anything learned during that period to help us control costs better?

That we should have competition and that we should write firm contracts. Unfortunately, the Congress didn’t learn the lesson at that time. Perhaps now that it sees the bad experiences we’re having, it will get interested in controlling costs again.

I do believe that the kind of contracting we were trying to do with the C-5A was the right idea. At the time, when Lockheed got into trouble, the defense industry said, in effect, “This is impossible; there are too many unknowns; you can’t ask a contractor to write a firm contract for such a plane.”

In answer to that, 1 would note that Boeing, having lost the contract for the C-5A—which 1 think they lost because they bid much more honestly on the price—then decided to develop the 747, an airplane of similar size and technology. They risked their own money on it, and they managed to produce it for the amount they estimated and in the time they estimated. So you can’t say it would have been impossible for Lockheed to estimate the costs accurately and then produce the plane for what they said.

In the book that you wrote about your experience at the Defense Department, you said the single greatest need in that field was to create an independent force, outside the professional military, that could look at military decisions in relation to cost and other considerations. What progress has been made on that?

It’s not very encouraging. When Nixon came in as president, instead of continuing this work, he announced that he was going to “root out the whiz-kid approach from the Pentagon,” which meant in effect to turn the decision-making over to the professional military. There have been ups and downs since then. The tragedy is that there is no powerful interest group in favor of cost effectiveness in national defense.

Some people want to spend less on defense because they want to see money spent on welfare, or health care, or education, or something else. But they don’t really care how the money is spent on defense; they just want to spend less. Other people, led by [Secretary of Defense] Caspar Weinberger these days, want to spend more and, I think it’s fair to say, don’t seem to care very much how it’s spent. While it is definitely in the interest of the American people in general for there to be a tough scrutiny of the cost effectiveness of every defense program, it’s often not in the interest of any individual strongly enough to back it. And for the typical member of Congress, cost-effectiveness analysis could be like a loose cannon rolling around on the deck. We might come along and analyze some program in his district, and want to cut it back or shut it down.

Members of Congress don’t like that.

Exactly. Similarly with bases—we closed a lot of bases in those years. We found that there were Indian forts in South Dakota that hadn’t seen action in a hundred years. And we wanted to get the soldiers out of the Indian fort and up into Germany, where we needed them.

Well, the problem we found was that there’s a man who owns a drugstore and another man who owns a gas station across the street from the Indian fort. When we talk about closing the fort, their livelihoods and the value of their property are threatened. They write to their senator and say, “I’m going to donate a thousand dollars to whoever runs against you in the next election if you let them close that Indian fort.” So, the senator gets all upset and sees the real problem of national defense not as “How do we have a cost-effective, efficient defense program?” but, rather, as “How do I keep that Indian fort in my district?” or “How do I keep that airplane plant?”

One of the most striking pieces of research you did was a comparison of NATO and Warsaw Pact forces on the ground. Our rationale for deploying tactical nuclear weapons in Europe is that the other side has a vastly larger number of men and tanks and guns, so we need a superweapon, or else we will be overwhelmed by sheer numbers. Your study showed that there is approximate parity between the two sides if you count men under arms, and tanks, and guns. But we’re still worried about these vast Asiatic hordes that overwhelm us in divisions. How did your research findings get lost?

Partly, it happens because the military is used to counting units, like divisions, and not counting actual things, like guns and men and so forth. Then there is a widespread belief that, when in doubt, the safest thing is to overestimate the enemy’s strength. And yet in this case, the overestimation of Russian strength is counterproductive, because that’s the reason we haven’t increased the size, strength, and readiness of our conventional forces. People believe the situation is hopeless because they’ve greatly overstated the size of the Russian forces.

Moreover, some people think that overestimating the enemy is good, because it’s a way of getting more forces. And many people believe that if I’m saying the Russians aren’t so big, it must be that I’m trying to disarm America. That is just wrong. I personally believed then, and still believe, that the Russians are extremely dangerous. They’re a real menace to us and to our European allies. In fact, paradoxically, that’s one of the costs of overestimating them—they have more intimidating power if we overestimate them. I think one of the reasons that Khrushchev, in 1961, was able to be so belligerent in the United Nations, and toward Kennedy, and so forth, leading up to the Cuban missile crisis, is that he perceived that we felt ourselves weak. And we felt ourselves weak because we were overestimating the Russians.

I’m all in favor of a very strong national defense; I’m not particularly interested in seeing us cut defense spending now, but in seeing that we spend the money wisely and make decisions based on accurate information. With a fairly modest increase in NATO defense spending and the effort to spend the money efficiently—which means to close unneeded bases and cancel ineffective programs—we would be able to have conventional forces that equal the Russians’ in strength and effectiveness. Unfortunately, a lot of people see that as upsetting their vested interests, and so they prefer to go on with the myth that we are hopelessly outnumbered and therefore don’t have to make any serious effort to improve our conventional forces.

The analyses you did during your years at the Pentagon were based on intelligence estimates, some of them probably close to the truth, others highly inaccurate. How did you and your colleagues feel about basing your analysis on data that must have seemed very conjectural at times?

One of the things I learned to do was to get my analysts cleared for the basic intelligence data. I got them to study the data carefully and analyze it themselves. When it came to uncertainties about the future, we tried to recognize them explicitly in our analyses. Secretary McNamara asked us to develop plans to respond to the most probable threats, but to combine them with contingency plans for responding to “greaterthan-expected threats,” should they emerge.

For example, in the early 1960s, McNamara approved the construction of a whole additional factory to build Minuteman ICBMs that we expected never to use but that we bought as insurance in case the Soviet force turned out to be much greater than we expected. Building the factory would save lead time and increase our flexibility to respond. But that was still a lot cheaper than building missiles to oppose a threat that might not materialize.

So, the answers to conjectural data include working hard to get better data, buying some insurance, and designing a good deal of flexibility into one’s posture to hedge against uncertainties and to be able to adapt to unexpected developments.

You’ve written that, until fairly late in our involvement in Vietnam, the military didn’t have a really good information-gathering system there that would have enabled you to find out the actual results of their efforts. Have we made progress since then?

I haven’t been closely involved in defense matters since I left the Defense Department in January of 1969, but to the best of my knowledge, the answer to that is negative. At that time there was no serious interest in evaluating weapons and operations, and stopping the ones that were ineffective and building on the ones that were effective. General Westmoreland just didn’t think that hard-nosed scientific analysis of the effectiveness of different operations was information he needed, any more than a lot of surgeons feel that a very careful evaluation of surgery is needed. They’re afraid that this kind of analysis might show that what they’re doing is ineffective, and give them information they don’t want to know.

It’s clear that we’re deploying a lot of very costly and ineffective weapon systems, because the services that sponsor them, and the contractors, just don’t want a real evaluation. That’s something the Congress ought to work on—to strengthen the function of independent evaluation in the office of the secretary of defense. However, to make that work you’ve got to have a secretary of defense who wants it and believes in it.

Looking back, are there assumptions of your own that have changed, or are there things you would have done differently on the basis of what you’ve learned since those years?

I don’t regret anything I did. I know that I made some generals and some members of Congress angry, because I fought or canceled some of their favorite programs. But the basic lines of what we tried to do were right, and most of the decisions we made were good ones. I learned a lot about the power of politics. Congressmen want to get reelected, and that may mean fighting for a base, or production, in their district, even though it’s not efficient from a military point of view.

I used to get very upset and angry about that, feeling that Congress was trying to make the defense program into a pork barrel. I guess with increasing age I feel more philosophical about it; I would rather that the legislature had to get reelected than that it didn’t have to get reelected. For all the problems of democracy, the alternatives appear worse. But I do still deplore the fact that our system does not provide much support for the efficient and effective use of resources for national defense.