Economies of Scale

July 11, 2014Filed under techMarkdown source

I was reading through an interesting Ars Technica article on the new Long Range Strike Bomber (LRS-B) proposal the Air Force is soliciting. It’s generally interesting to me in part because I’ve worked on a related project in the past, and we talked fairly often about how the LRS-B program might impact it. The article is worth your time. This quote from Robert Gates in the middle of the article, which touches on the program the LRS-B would replace, caught my attention, though:

What we must not do is repeat what happened with our last manned bomber. By the time the research, development, and requirements processes ran their course, the aircraft, despite its great capability, turned out to be so expensive—$2 billion each in the case of the B-2—that less than one-sixth of the planned fleet of 132 was ever built.

Looking ahead, it makes little sense to pursue a future bomber—a prospective B-3, if you will—in a way that repeats this history. We must avoid a situation in which the loss of even one aircraft—by accident, or in combat—results in a loss of a significant portion of the fleet, a national disaster akin to the sinking of a capital ship. This scenario raises our costs of action and shrinks our strategic options, when we should be looking to the kind of weapons systems that limit the costs of action and expand our options.

Now, in one sense, Gates was absolutely right. On the other hand, he seems to have committed a classic blunder in dealing with these kinds of costs: economies of scale matter. Part of the reason the per-unit price of the B-2 was so high was precisely that we only bought 20 of them. While the units were individually expensive to manufacture and maintain, because of unique materials used in their construction and so on, they were much more expensive to manufacture in small numbers than they would have been in large numbers. There are basically two reasons for this:

  1. The manufacturing process couldn’t do what it does best (turn out large numbers of standardized parts and thereby reduce costs).
  2. The costs of development—research, software development, etc.—were all distributed over a much smaller pool than they would have been had the government purchased more aircraft.

This second point is incredibly important to understand. It is certainly true that the absolute cost of buying 132 B-2s would have been high, possibly astronomically and unaffordably high. What it would not have been is $264 billion. Even assuming that manufacture costs were fully half of the cost-per- plane (almost certainly not the case), it would have been barely over half that. Assume that the B-2 cost $1B per plane to build, and that the other $10B was research. Well, that’s still an expensive plan… but the total cost is something like $144B, not $264B. Those economies of scale matter.

This same reality is a point made later in the article by another commentator, but I couldn’t let it go. Things like this drive me nuts, because they’re such a common failing in our political discourse. Ignorance of basic economics from the people making decisions with this kind of economic impact is profoundly unhelpful.