We are developing the social individualist meta-context for the future. From the very serious to the extremely frivolous... lets see what is on the mind of the Samizdata people.

Samizdata, derived from Samizdat /n. - a system of clandestine publication of banned literature in the USSR [Russ.,= self-publishing house]

No wonder defence spending is such a nightmare

I look at this item over at Wired, the technology/culture publication, and think that this is all very geeky, very Sci Fi and very clever, but it also makes me think, as a commenter does on the article, that it is hardly surprising that defence procurement costs are so high, and getting higher. Which is possibly not very smart if government budgets are under so much strain.

26 comments to No wonder defence spending is such a nightmare

  • I’m no great fan of ‘virtual reality’. However, this work advertises itself as a lower-cost way of more fully checking out the human interface of new technologies (and also, presumably, manual procedures).

    Is it purely defence oriented?

    Back in 1988 (yes, that is over 22 years ago), I remember having a personal demonstration (in the USA) of something similar but simpler, whilst consulting on the potential contribution of ‘computational chemistry’. This was to feel the electro-quantum-mechanical forcefield, (primarily) as a medication molecule moving around a protein.

    What’s not to like, providing it does really deliver the knowledge at lower cost, faster timescale, greater accuracy, etc?

    And that will depend on each individual investigation and the sum of those chosen to use the facility.

    Best regards

  • JP,
    I must disagree. Though defence procurement is so pricey because of avarice, vested interests, misplaced nationalism, political lobbying & pork* and the vicious circle this results in which pushes unit costs up which pushes numbers bought down which…

    *Note whenever a new fighter plane or whatever is bought the first thing the pols will say is how many jobs the project “creates”. Really bad in the US where 50 states are after a slice of Sæhrímnir.

  • Johnathan Pearce

    NickM, I don’t think I disagree; I would be happy if the whizzbang tech. as shown here really did mean that the procurement costs could go down like a rock, but you and I are cynical enough to know that is not how things work.

    The procurement model used by the Western military – with the odd exception – and of course, NASA – is a sort of “cost-plus” one in many cases. This may, of course, be changing. Sometimes big, expensive projects to build hundreds of things like the Raptor (which I personally think looks a bit ugly) get held up or put on partial holds, as Gates, the outgoing US defence sec, has done.

    Some defence projects are of course great value for money as the basic designs go back for decades and can be endlessly modified. The RAF jump-jet harrier being one. The Jaguar strike aircraft lasted for years, as did the Phantom.

  • JP,
    Phantom… Yes… except consider how the RAF in particular got Phantoms and then Tornado. A brilliant case study in ineptitude and politicking.

    The Raptor ugly? I suppose it doesn’t matter because most of it’s kills will never even know it was there.

    JSF could be considered ugly. It looks like the F-22’s shorter, fatter kid brother.

    Probably the current uber-case of defense procurement having nothing much to do with capability and all to do with politics is the unholy row in the US over B767 or A330 tankers. This still rattles on ad-nauseum whilst the USAF has antediluvian KC-135s first orderd by Curtis LeMay! They must cost a fortune to run.

  • Steven Rockwell

    I think the calls to revamp defence spending and procurement are right on the money, but I also think it’s a horribly complex system that really can’t be swept aside for some soundbite plan.

    Part of the reason procurement is so expensive, especially for complex weapons and aircraft is simply to keep those few businesses in business. If the Air Force says it wants a new fighter plane and three or four deisgners compete for it, they might each spend a couple billion in R&D just to compete. A few failed bids will put any company out of business and there aren’t all that many companies that can do certain things in the first place. So the winner gets the contract, which on its face is way overpriced, but that winning bid will have to sustain the company when it doesn’t get the next couple of contracts. The free marketeer in mee says “so what, let’m fail and go under” but realistically, some business are essential to the military and simply can’t go out without harming the nation. I’m reminded when the Navy kept buying a single submarine at a time that it really didn’t need just to make sure General Dynamics submarine division kept it’s skilled workers, because they couldn’t be replaced if they lost their jobs.

    Then there’s the military’s end of this whole equation. It used to be they’d simply say “we need a fighter plane that can do X,Y, and Z. Show us what’s you’ve got.” Nowadays the specifications sent out to bidders for a single piece of equipment can be several inches thick, much less for something like a ship or a plane. And like all beuacracies, the military is incredibly top-heavy with committees, boards, meetings, and internal politics. It’s nigh on impossible to nail down someone with the power to say “yes or no” for a project, especially one that has grown to take on a life of it’s own, even more so when the military constantly changes what they want. Add in generals and admirals trying to justify their billets (and budgets) and politicians looking to keep jobs in their district and there we are.

    Granted, we shouldn’t be subjected to 300 dollar hammers and 20 dollar boxes of pens, but even if we did away with the whole procurement process for off-the-shelf items, we’re still stuck with the mentality that a budget must be spent this year because otherwise next year someone above you will figure you don’t need as much. And we haven’t even gotten into the waste or the distribution process and flaws.

  • Kelly Johnson developed the P-80 in six months. Earl Bakken designed and prototyped the first Medtronic pacemaker in less than a month. He gave it to C. Walton Lillehei to check out, and the next morning found it in use keeping a patient alive.

    This sort of thing cannot be allowed. Nor will the government allow it.

  • JP- it is not the design tools from various contractors which drive up the costs of military procurement. What would you have the Lockheed engineers do, go back to slide rules?

    This virtual reality thing might be a gimmick, but I suspect it has actual practical use in helping engineers understand the 3D spaces they will be designing for.

    During my time at Martin Marietta (which was acquired by Lockheed), an engineer I know designed a wonderful robotic arm for the space shuttle. Wonderful, until we discovered when trying to build the physical prototype that one of the joints on it did not allow the arm to bend as much as it was supposed to. Extra money spent on 3D modeling software that understands solids would have saved money in that case.

    Ellen- the days of designing an advanced aircraft in 6 months are over.

  • Darryl – If you’ll forgive me some worship of Kelly Johnson, he and the Skunk Works got the contract for the U-2 in 1955, and it was flying missions by the middle of 1956. They got the contract for the A-12 late in 1959, and the first flight was in April of 1962. (The plane was later modified to have a second seat, and became the SR-71 blackbird.) Johnson took requests from fighter pilots in Korea for something “small, simple, with excellent performance”. The first flight of the F-104 was in March of 1954, some 28 months later.

    I do realize that planes are more complicated today. So, also, is the procurement and design process. And that was my point: the bureaucrats make the engineers fight them for every inch of progress.

    As a contrast: consider the SpaceX Falcon rockets. The company was founded June 2002. The Falcon 1 made its first successful orbital launch Sept. 2008. The Falcon 9 made two successful flights in 2010, and the second one launched the Dragon capsule, which deorbited and was recovered according to plan.

    Do you think NASA could have done that with as little money and time? NASA has bureaucrats, not to mention the entire Federal government on its back. The Ares program was started in 2005, it’s had one prototype launch, it’s probably canceled, and even if it’s canceled, NASA still has to spend $500 million more keeping the program alive in case Congress changes its mind.

    Bureaucrats want to control every little detail. When they’re put in charge of engineers, the results are both ugly and expensive.

  • Ellen,
    We shall bow at the shrine of Kelly together. You know of the L-133? That’s what Johnson wanted to build and not the P-80. With an afterburner and everything.

    As to complexity. Yeah, but then the design tools are way more sophisticated too. So is the manufacturing technology. It should if not even out but go some way. 20 years lead time for a fighter is ludicrous.

  • Laird

    While you’re genuflecting to Kelly Johnson, you should also give a nod to his protegé Ben Rich. His book “Skunk Works” tells (among other things) the story of the stealth bomber, and is a great read. At the end he gives some very sensible suggestions on how to improve the procurement process and reduce the astronomical costs of aircraft development. It’s unfortunate that no one listens to him.

  • Laird

    Oh, and back to the topic at hand, I very much doubt that this VR system is worth anything close to what it cost to build. All it seems to do is virtually replicate the dimensions of the spacecraft (or whatever), so the designers can get a better sense of how it feels to be inside. Big deal. A non-functioning mockup built to scale would be every bit as good and much cheaper. This is just a hugely expensive toy for the geeks. I give it thumbs down.

  • Laird:

    Oh, and back to the topic at hand, I very much doubt that this VR system is worth anything close to what it cost to build. All it seems to do is virtually replicate the dimensions of the spacecraft (or whatever), so the designers can get a better sense of how it feels to be inside. Big deal. A non-functioning mockup built to scale would be every bit as good and much cheaper. This is just a hugely expensive toy for the geeks. I give it thumbs down.

    Big deal? You build one. In your cost estimates, include the expense of managing classified material. Also include a marketing budget, because a toy like this is obviously designed to be show and tell, as well as be a functional design checker.

    Do you know what it cost to build? Do you think Lockheed just bleeds money on geek masterbatoriums because they can? I would like to see the pricetag before I give a thumbs up or down on its return on investment quotient.


    I think we agree, just from different angles. Big ticket military procurement is expensive and time-consuming because it is a political process, not an engineering process.

  • Laird

    Darryl, I said build a non-functioning mockup. That means no “classified materials”, just plastic and plywood. And we haven’t even talked about the cost of programming whatever is being modelled. Unless this toy does something more than what is shown in the video (i.e., something beyond just giving the operator a feel for interior dimensions) I stand by my criticism.

  • Jeff

    Laird, as an industry insider who has seen and touched the wooden wing mockup of one of the earlier airbusses at Filton, walked around the huge wood-working shop, and smelt the sawdust; let me assure you that they are not cheap. Not chap at all. There is a very hard headed business reason that a huge amount of the design and planning work on the A380 moved to VR based work. Including mocking up in vr of entire assembly operations. Try building an entire production line in plywood.

    One Vignette to illustrate what engineers think of this stuff: Several years ago a bunch of engineers were on a visit from airbus to evaluate new design tools, looking at their own engineering model from their CAD systems loaded into a VR system. 15 minutes in the evaluation of the VR system was all but out of the window. These guys were to all intents and purposes doing a design review of their models.

    What is often not recognized by people who have not used a CAD system in anger is how little of an entire design can be held in a full blown cad system at one time. Even a moderately sized sub-assembly is pushing things. This is due to the vast amount of interrelated information that is kept in a high end CAD model. Change one small design feature, the diameter of a bolt say, and the knock on effects can ripple through an entire assembly, as design rules are enforced. A related issue is that different CAD software is optimized for different uses. You don’t use the same software to design a car engine as you do to design the shape of the body shell, and CAD systems do not tend to read data from their competitors at all well. The end result is that for any moderately complex design you do not have a CAD model, you have several: and they don’t all work well together. The VR systems strip out extraneous info and fuse models from disparate CAD systems, to the point where I can have a geometric representation of an entire A380, down to individual bolts, on this laptop, and animate it at an OK frame rate. Not only that, but I can apply high level checks of design rules between assemblies, directing design review effort to areas where there might be problems. This all saves time and money.

    Back to that Lockheed video. What you saw was only the tip of the iceberg. An initial work though to quickly identify problems early ( and thus cheaply), and gather data for the next stage. What comes after is that the motion capture is stored, and then applied to a whole range of statistically accurate biometric models of US air force personnel, to prove that this procedure is safe, and achievable by any likely aircraft technician. Depending on the quality of the biometrics, this modeling will be down to the working out what force is required to move a part, given its mass and moment of inertia (try getting that right in plywood!), and whether an individual in that posture can be expected to be able to perform that action. Dozens of analysis with different bodyforms, from the smallest female to the largest male will likely be done. Then after every design change the scenarios would be run again, automatically, as part of the change request, to make sure that no change has compromised any of the maintenance procedures. This automation of human-intensive tasks that have to be repeated many times in the design process is what really saves money.


  • Johnathan Pearce

    I may have been a bit dismissive of the VR idea, but I am not all that surprised at Laird’s own skeptical take on the matter. An awful lot of this tech is presented as a cost-cutting measure, and yet the bills keep rising.

    The jury is out, I am afraid, on this one.

  • No no no.
    This stuff makes the really costly stuff cheaper.

    You have to look at the suppliers who con un-knowing Defence bureaucrats out of vast sums, a little at a time, for things like…

    5/8″ roller bearings (from Barden)….at £64 + vat (EACH) I have lots of them here, in sealed bags, they are so expensive that I can’t bear to open them to the air…
    a simple AMD 16-pin pic chip, for £86
    An electrolytic capacitor (OK it’s tantalum) for £6.80 per capacitor
    A 1/4 watt carbon resistor, 5% tolerance, for £0.42 EACH
    A pair of pliers, which (it is said…?) was to cost the Pentagon $33,000….until some sharp-eyed clerk spotted what it was….

    I can buy most of this stuff from China on ebay for….about £2.99.

  • llamas

    Ellen wrote:

    ‘Kelly Johnson developed the P-80 in six months.”

    Yes, he did, or rather, his team did, and it was a complete dog of an airplane as originally built, with an astronomical failure rate and a distressing habit of killing its pilots in the process. General Yeager (who was there) has written quite descriptively of the issues involved in just keeping the early models flying at all. If any current development of military aircraft had the failure rate of the P80 at the same stage of development, it would be cancelled out of hand as a hopeless failure.

    It took the best part of 3 years to iron out all of the dreadful problems on the early prototypes and develop the -B and -C models to a state where they were reliable enough for serious deployment.

    Apocryphal tales of instant development of the most amazingly-complex products almost-always overlook the reality of their actual reduction to a device that’s useable in the real world. A sketch on a cocktail napkin, or a single working prototype, may lead the average person to think that the process of design and dvelopment is complete – alas, it is not so. The same goes for Earl Bakken’s first pacemaker design – it was a long, long, looong way frm being a viable product, as the subsequent trials and tribulations of Medtronics shows. First functional demonstration of the concept does not equal finished product.

    For an exellent discussion of why military procurement makes everything so complex and costly, I suggest (as before) Edwin Luttwak’s book ‘The Pentagon and the Art of War’ – 25 years old and still 99.875% accurate.



  • llamas

    @ David Davis :

    ‘I can buy most of this stuff from China on ebay for….about £2.99.’

    Er – no, you can’t.

    What you can buy on eBay looks like what you have, may be dimensionally similar, but lacks the supporting documentation and infrastructure that proves that it is, what it is specified to be.

    I hear this all the time – but I can buy that for $XX less on eBay! Sorry – not if you want to guarantee the performance that was originally designed. When you go to buying Chinese parts that come with certifications and calibration reports and RoHS compliance and all the rest of it – suddenly, the price goes up quite a bit.

    Now, we could debate forever about whether some or all of the specifications that surround a piece of militray kit are really required for function – especially in the US, where many of those requirements are social or political and have nothing to do with meeting militray goals – but the fact remains that if you want something that is guaranteed to meet all of its specifications – you have to pay for it.

    A 3″ 28-gauge hypodermic needle, for applying adhesives, from EFD Inc – 3¢ each.

    The physically-identical needle, but for applying anaesthetics – 85¢ each. Because that needle comes certified as being sterile, made from the specified materials, sealed, autoclaveable and all the rest of it. The material cost is identical – it’s the processing and system cost you’re paying for.



  • Laird

    Jeff, I appreciate your post, and I defer to your experience and expertise. You’ve mollified my skepticism somewhat; I’ll back off on my initial reaction from “thumbs down” to “neutral pending additional information.” (I’m not quite ready to move all the way to “wild enthusiasm,” though!)

  • Kim du Toit

    “I can buy most of this stuff from China on ebay for… about £2.99.”

    …and then you’ll discover why the “lowest bidder” mentality costs lives.

    The term “mil-spec” often gets sneers from people who don’t know what’s involved.

    Here’s a simple case in point. Some time ago, my old website’s readers got together to buy new sniper scopes for some soldiers about to ship out to Iraq. Their old scopes were Vietnam-era, and were completely worn out.

    So we got them brand-new, state-of-the-art, best-in-the-business, top-rated civilian scopes. Cost: about $1,800 each. Mil-spec scopes typically cost about $5,000 each, so we thought we were doing it properly.

    All three of the new scopes were broken within a year, because of the stresses and strains from typical military use. So we had to buy them new ones, again. And again, less than a year after that. (Remember: you cannot get any better scopes on the civilian market than NightForce –forget Swarovski and Zeiss; they break even quicker.)

    And these were replacing scopes that first saw service in the mid-1970s, finally wearing out in the early 2000s.

    As Llamas said above, you can always get something cheaper. But it’s not what you want to carry in the field when lives are at stake, and breakage is fatal.

  • Kevin B

    Sounds like Captain Vimes’ theory of boots. Poor people can only afford cheap boots which lasted less than a year and needed cardboard in the soles, while rich people could afford boots that lasted generations. Still, Vimes could tell where he was in Ankh Morporprk’s foggy streets just by the feel of the cobbles through his cheap soles.

    And on the subject of getting planes built in a short time, here’s a vid I found in a comment thread on the SOTU speech.

    America’s First Jet Flight, October 1942 (Link) complete with a look at another aviation pioneer, Larry Bell. I shall resist any snark on the probable cost of the doorbell plunger in today’s procurement environment.

  • llamas

    The military is different.

    I have a friend who runs a small manufacturing shop, that does a lot of business making small accessory bits for armored vehicles. TAC-COM is just around the corner. Control boxes, consoles, switch housings, that sort of thing.

    He keeps a ‘black museum’ of parts that have made their way back to him as being insufficiently rugged.

    “Failed after having full can of .50-caliber ammunition dropped on it.”

    “Legend became illegible – suspect repeated exposure to insect repellent.”

    “Graphic not clear when using NVGs.”

    “Repeated impact causes unacceptable damage to PASGT helmet.”

    And so forth.

    This stuff has to work even though it gets a daily battering from boots, dirt, diesel fuel and puke, dished out by an endless stream of soldiers whose very last concern is the careful handling of the equipment Uncle Sam supplies. They tend to keep it clean, to be sure, but they beat the snot out of it when in use.

    The militray environemnt is so much more dmeanding that it’s no surpise that their stuff costs more. I’m not saying that that explains all of the added costs of military procurement, far from it, but it does cover a large part of it.

    I can think of only a couple of pieces of civilian equipment which have recently transitioned into the US military successfully more-or-less unmodified. I don’t mean 3-ring binders or canteen trays, I mean pieces of warfighting equipment. Anyone care to try and name as many as a half-dozen such?

  • yup

    Llamas, Kim and Laird have it right. Mil-spec equipment often has a decades-long product life, must consistently meet rigorous standards and sees use that civilian equipment can’t begin to tolerate. But there’s another factor at work here too.

    Those quickly tossed out prototypes from the skunk works – they were analog planes. From the F16 onward military planes have been fly-by-wire – i.e. highly software-intensive. A very large number of lines of high level real time code intensive – orders of magnitude larger than anything most software engineering teams ever must design, code, test and integrate. That code must keep an airframe aloft that would sink like a rock, not glide to earth, if power is lost. That code controls wing surfaces, manages electromagnetic protections, coordinates defensive and offensive systems, keeps crew alive and breathing … and it does it in coordination with specialized electronics and the other hardware.

    The recent Airbus problems only hint at the complexity involved in designing, building and testing such planes. Airbus requirements don’t come close to the performance extremes demanded of a military plane – much less a plane that, say, must be stealthy and serve both as a fighter and bomber, with versions that can take off/land from carriers as well as air fields.

    One last point re: the VR system. What is originally designed to simplify human factors validation may well serve as the platform for training maintenance and crew later.

  • I for one welcome this VR stuff. Just from the video it looks like the engineers are seeing a little of what it takes to work on the planes (and that seems to be the intent, at least in part). My dad is a contractor (in construction, etc), and he has complained for years that the architects and engineers, while designing things that might be structurally sound, don’t take into account the people that have to actually build and maintain what they are designing, and in general make things hell for the guys in the field – and even at times designing things that are next-to-impossible to construct in the intended manner. And I’ve heard similar complaints from mechanics as well.

  • Tedd


    So the winner gets the contract, which on its face is way overpriced, but that winning bid will have to sustain the company when it doesn’t get the next couple of contracts.

    That’s not how it works, though — at least, not for large contracts like a jet. The winning bidder becomes the primary contractor, but the Pentagon farms out substantial portions of the work to other contractors, to help spread around the cash flow. They do that because of exactly the problem you described: that there would soon not be enough companies around for a healthy bidding process if they didn’t.

    For what it’s worth, I suspect the VR system is incredibly good value for the money. I didn’t see anything in that video that looked even remotely expensive, and the potential payoff is huge. Military aircraft are very complex to build, making production efficiency a major factor in project cost. Anything that helps move the aircraft through the process faster is worth a very large R&D investment.

    VR in one form or another is clearly the way of the future in aircraft design. Already, flight testing has become primarily an exercise in confirming the data generated by the computer models. The F-22 performed within 2 percent of model data on every parameter throughout the flight test phase. Imagine how much it’s worth to prevent just one flight test accident!

  • Paul Marks

    Four billion Pounds on the new Nimrod radar aircraft – they still do not work properly, and the government spends vasts sums of money SCAPPING them.

    Anyone who thinks the American military system is wasteful and worthless (and it does indeed have very serious problems indeed) should have a look at the British system.

    Well actually you should not – because you will scream too much.

    The United Kingdom spends a lot of money on defence and has brave and professional soldiers, sailors and airmen.

    But the army, navy and airforce are falling apart – they are becomming nonfunctional.

    The Royal Navy and Royal Air Force have virtually gone already. And the army will follow.

    Things are that serious.

    And (just to give one example) in the South Atlantic we face Argentina in alliance with Brazil (before you laugh – look up their combined forces).

    “But America will save us”.

    These days “America” is President Barack Obama (I hate even writing the word “President” before that person’s name, but he is President).


    You would not think so from the fact that British book shops (from Waterstones to Tesco – not that there is much difference) only put pro Obama books on their shelves, but he does.