Military acquisition costs: why do things cost so much?

A common issue that I’ve heard over the years has to do with costs and waste in Military programs, it’s a mocked meme, “that screwdriver cost $500” and so on. You’ve been sold a mistruth by people that are either bad at their jobs (and don’t know what’s going on), or by intentional frauds deceiving you.

NOTE: I am no fan of Govt. spending: things are too expensive, and there is a lot of waste and fraud in these systems. But I hope that this article will explain what’s really happening: the answers are not what most have been lead to believe.

Lets create a fictitious Military Airplane to help you understand where the money goes.

Fixed costs

  1. Planes cost a fortune to design: there are millions of parts on an airplane, and they almost all have to be designed custom (because there is no plane like this one before), even the engines usually aren’t shared with commercial planes (which don’t have afterburners, etc), though some of the designs can be borrowed. 
  2. Almost everything on the plane needs a redundant (backup) system (in case one fails), and that backup has to be of a different systems entirely (e.g. if you had two hydraulic lines next to each other, and one is severed, probably both would be, and the plane comes down). So they tend to use unique systems for each (and separate routing): electrical, optical, hydraulic or mechanical backups so there is not a single point of failure. This increases costs.
  3. Then they spec and document all the parts included in the aircraft to ridiculous standards (Mil. Spec. / Military Specification). These are incredibly difficult, and thus expensive: but everything will be documented, and tolerances much be high as lives and national defense are at stake, so they (we) pay the price for that. Other countries just tolerate higher failure and maintenance rates. 
  4. You not only need to make the plane, but you need to teach crews how to fly the plane, and to teach maintenance folks how to maintain the plane. So they also have to create trainers: fragments of the plane with simulation hardware behind it, to teach these folks. Those trainers are expesnive to make and design, but they’re less expensive than having the crews make mistakes on the real aircraft. (And it doesn’t tie up expensive aircraft being repaired or replaced, because someone broke it, in training). And remember the documentation that goes along with these trainers. (You have to maintain the training equipment as well).
  5. In order to make all the parts on the plane, you have to create custom tools to make the custom parts (tooling). Not just special pliers and such, but jigs and stamps, and molds and so on, in order to make it easier (less expensive) to make every part of the plane. So you don’t buy a plane’s wing, you pay for making the jig that allows you to make the planes wing, and another set of tools to attach it. And then the wing itself. 
  6. Politics: you can’t just build the plane in one place — if you do, you have a contract where one state/city is benefiting, and no other congressman/senator is seeing the benefits of that spending. (Thus they have no incentive to vote for it). So any prime contractor has to have parts of the plane made by sub-contractors, all over the nation (in as many districts as possible), so that enough congressman voting on the product get a piece of the pie, and have a vested interest in not cancelling the program (as they’d be hurting voters in their district). This costs more, but is required or politicians will use cutting that program as a lever to twist the other politicians arms (whose districts would be hurt). So spreading the program reduces the risks. 

All these are called fixed costs. You pay them up front, whether you’re building 10 planes or 1,000 planes. Thus the more planes/cars or widgets that use these tools, the more their costs are amortized (spread out). Car runs are usually 100,000+ cars, and they share parts with other cars, this gets costs down. Planes have 100x as many parts, and a smaller fraction as many planes to divide those costs over. Cars can share parts, planes rarely can. So it’s the up front setup and design costs that are 80% or more of the costs of a program. 

So let’s say our plane is $100B to do all that R&D and tooling (the fixed up-front costs required to even get started). And you are finally ready to make 350 planes. Everything is ready to go, it’s taken years (a decade) to design, and now the Dog and Pony show starts.

Politicians and the Media

So pretend we’ve spent most of the $100B on the fixed up-front costs, and are now ready to build the plane that the Air Force specced out. The real cost of each plane is $50M each (to actually make each plane in parts+labor). The Air Force figures out that they need 350 airplanes to achieve their roles, and enough parts for an extra 50. So the total cost of the project would be $120B ($100B fixed costs + 400 planes x $50M each). 

Now the way the contracts are written, they have to spread the cost of everything out over the entire contract (every part). Each plane has a real cost of $50M to make, but you have to amortize the fixed costs across every plane, thus it looks like each plane cost you $342M ($120B / 350 planes). That’s the contract cost for each plane, but only because that’s how much it will cost to pay off the contract, over that many planes — not because that’s how much it costs for that plane ($50M). If you made 1,000 planes, the per plane cost of the contract would be $150M/plane, and 2,000 planes it would be $100M/plane, and so on.

Then they have to divide that plane’s costs, over all the parts on that plane — and THAT’s where the $400 toilet seat (or screwdriver) comes from. If you broke that $400 down, it would be $30 for the seat, $370 for the documentation, R&D, testing, tooling, and other costs for the entire airplane, spread out over that part. If you used government contract account to calculate everything in your house, the cost of toilet paper would be $136 per wipe (assuming $100K salary and 2 wipes per day). But we don’t do math like that, because it’s absurd. But I digress. 

Now after 8 years of development, you’ve spent $80B, have a few planes, designed, built and tested, and you’re ready to go into mass production of the planes you tooled up for. And some politician(s) will try to make a name for themselves — lets call him Senator Dumbass (D) from California (pronounce Du-maw!).

Sen. Dumbass want to look good for his military hating base, and thus anything he can do to cut their budget will make him look good (and there aren’t enough sub-contractors in his district to stop him, since they all fled the anti-manufacturing state). So he sells the gullible on an idea, “This contract is costing too much and over budget, and we don’t need 350 planes! We can get by with only 175 planes! That will save $60B he says! (Half as many planes, should be half the cost of the program). The media champions his cause, the gullible cheerlead his brilliant idea, and the program is cut in half.

What really happened? Remember, $100B was up front costs, that doesn’t go away, no matter how many planes you build. He actually only saved $50M per plane x 175 planes or $8.75B dollars, not the $60B he said he did. So now the total cost of the project would be $111.25B ($100B fixed costs + 225 planes x $50M each) — and the Air-Force got half as many planes as they needed to be effective in the missions it was designed for. 

Now Consumer Reporter Dan Rathernot comes along, and reports what? Why the program is $111B for 175 planes, why that’s $634M per airplane! Originally, it was supposed to only be $342M/airplane by original calculations — that means each plane is costing nearly twice as much (85% more) than originally promised! Those bastards! He does a exposé after exposé on how over-budget the plane is, how incompetent the prime contractor is, and so on.

Plus since it’s not yet finished (and debugged), it is still under-performing the plane we already have (as it will until we have a few years of training, and at least the version .1 or .2 upgrades on the older plane). Every plane takes an upgrade or three to catch up to the prior plane which is mature and in it’s 8th or 9th refinement. He sells this as a disaster. We should obviously cut this program completely, and spend it on social programs instead. 

So the media, and the anti-military politicians colluded to defraud the public. (Even if there’s not a conscious conspiracy, it still works the same). 

More to the story. There’s another thing that’s going on, these programs take 5-20 years to come to fruition, so inflation matters. The cost of the program is going up with inflation. So the initial contract was $120B, but after 10 years of inflation, that’s $160B in adjusted dollars. So the actual dollar amount is over original promised price, but NOT if you adjusted for inflation over that time. On top of that, the Air-Force often changes requirements while in project, and those have to get priced in. So programs usually do get higher over time, but it is because of inflation, changes to the requirements, and most of all: fewer units to amortize the costs over. The reporters never remember to explain that. 

 So for example the F-35  started in 1993 (or before), as the Joint Strike Fighter program, and Lockheed Martin’s X-35 was awarded the contract in 2001 (after beating out the Boeing X-32), when completed in 2037, that means the program was spread out over 36 years. Using the last 36 years as a guide, the inflation costs alone will mean that smarmy politicians/media can claim it’s 189% over the original budget. Not counting any cost overruns due to changes in the program that come in over next 20+ years.

Sen. Dumbass convinced the public that he saved the public $60B — but all he did was halve the number of airplanes we got and decrease the odds of being successful at the planned mission. And Reporter Dan Rathernot deceived the public that the program is nearly twice over budget, and convinced people that the plane is a disaster. Since both Dumbass and Rathernot got a lot of attention (and screwed over the public and military), it’s a win-win. And they’ll both go back to that well until it’s completely dry.

Sen Dumbass promotes how he wants to kill the program entirely and save $100B on killing the remainder of program, and spend half that on midnight basketball and urban development (in his district). Of course Dan Rathernot loves that idea and champions it on TV, and eventually they get the plane scrapped (the Air-Force sees limited value because the number of planes is cut too much, and the manufacturers of the plane are not allowed to defend the plane based on contract terms). And the social program gets passed, because it looks like it’s free (found money). Everyone wins: except the public. 


What really happened?

  • We spent $100B up front costs, and have to pay off the prime contractor that part no matter what (that’s in the contract).
  • We built 5-20 planes (some prototypes, some early ones), that are too few to be useful for anything, so we have to pay to have them mothballed in the desert (to never be used)
  • All that design and documentation, tooling, etc., is scrapped — even if the new plane would have cost less to maintain, and been far more effective than the old plane, there’s not enough of them to be useful.
  • No other country can buy those planes, because of the PR boondoggle, so the contractor is screwed over
  • Instead of having a new plane, we have to create a $25B contract to upgrade and maintain the older planes for another 10 years (and those planes cost more per year to maintain than the new plane would have), and the Air-Force needs to start a new $200B program for the next generation plane to replace those aging designs/airframes that we no longer have a replacement for. 
  • the public got saddled with $100B debt on the plane, and a $120B on midnight basketball social program, which eventually goes twice over budget and increases gang violence and crime over the next 5-10 years. The media won’t report on that until everyone forgot who championed it in the first place. 

Thus in the end, we end up with $250B in losses — while they’re told by the media (Dan Rathernot) and the politicians (Sen Dumbass) is that they somehow saved us $100B and got us this great social program for free. Anyone with a clue beats their head against walls in frustration at what really happened.

I worked aerospace for about a decade, and saw how these contracts happened. And all of that stuff is really what happens: though there’s more complexity and interactions going on, this gets you the basics.

NOTE: I’m not saying you should support every military program, or that they’re all good. Many are dogs that should have been cancelled before they were started. I think the F-35 is a mess: just not as big of a disaster as some are claiming. So there are plenty of problems in procurement, or whether we need all these systems in the first place. But I’m trying to explain how fucked up the media and politicians are, on the systems we do need (as well as the ones we do not). 

Why haven’t you been told this before? Because Dan Rathernot and Sen Dumbass (D), both get money and attention (or votes), by not explaining what really went on. They and their network aren’t going to make them look bad. Plus it’s not easy to explain: people have to care about amortization costs and fixed costs, and how do you explain that in a 45 second blurb between an advertisement and sports reels? Easier to perpetuate the myth, than to explain what’s really going on.  So the myth continues. 

But every time you hear some politicians or media talking-head complaining about a $500 screwdriver, you now know they’re either a fool or a fraud. Someone who doesn’t know what they’re talking about, or trusting that his base won’t. 

Written: 1997.08.15
Edited: 2016.02.19