Sunk costs / Principal-agent problem / How not to

Communicated with another blogger, found out he agrees and actually planned to write about the very same conclusion. Now I need to hurry in order to save whatever originality my thought might have had. :)

Cuts of procurement problems programs usually yield surprisingly little savings. Well, experience alleviates the surprise, for there's really a pattern.
Assume a procurement program was about 200 aircraft and after delivery of 60 aircraft something changes and the program shall be cut down to 160 aircraft - a cut of remaining deliveries by 40%. The price for the remaining deliveries usually drops by far less than 40%, usually more as 10-15%, and this is not only due to lost economies of scale.

The reason for why it may even happen that such cuts don't save anything or no substantial sum is in the existing contracts; there's usually a penalty clause saying that if the government cancels or renegotiates the contract, a certain substantial penalty shall be paid.

Why would governments agree to such a clause in the first place? They're in a position of strength at the during the bargaining (assuming there's any bargaining at all). Sure, they want that piece of kit, but they can buy elsewhere, while the supplier likely needs the development money and order to get an exportabel product in the first place.

The answer is that bureaucrats and politicians in office want to rig the contract so fast that it makes cancellation difficult. This rigging is meant to reduce the freedom of action of governments in the future and doesn't serve the interests of the people at all. It merely serves the personal preferences of the bureaucrats and politicians in favour of 'their'  project. They exploit the 'sunk costs' issue to suit their personal preferences.The more the program advances, the less costs are not sunk and depending on the contract text it's even possible that a point approaches at which the contract penalty is as large as the remaining regular order. The program would then be immune to cuts because cuts would not save any funds at all.

illustration principal-agent problem, (c) "MisterX000", wikipedia

This is yet another example for how bureaucracies (and politicians) as agents of the citizens (principal) don't truly represent the interests of the people, but their own ones. It's an example for a principal-agent problem.

The solution is simple, and may actually help not only in military procurement, but in governance in general: Outlaw contract penalties and anything which has the same effect in government procurement of goods or services. Make it plain illegal and ineffective. This wouldn't neutralize penalty clauses retroactively (because of the rule of law), but it could save a country such as Germany billions of Euros per year on average.


Edit: There's an additional and substantial benefit in this proposal for countries with a tendency to cancel big programs often. An example would be the U.S.Army and USMC, which appear to be almost totally unable to bring an all-new combat vehicle or helicopter program into the production phase.
The example is that while the costs for cancellation are not initially high-profile, the corporations would likely seek to get a fixed and lower revenue as a substitute for it. This would be hig-profile, visible from start as costs.


  1. I'm not sure the proposed solution will work. Let's say I'm an aircraft manufacturing company and I wish to sell my new (sic!) airplane to a government. The price per unit I offer will include not only the manufacturing costs, but also the tooling up and R&D costs. If the contract provides for no penalties I may be bankrupt by a government cancelling the contract. In which case no one will ever sign a similar agreement with that government.

    1. There's nothing proposed that would keep the parties from negotiating a lump sum for production preparation and a per-unit price for production.

      Besides, govenrments already pay for msot R&D expenses with R&D contracts long before actual production begins. But there's nothing good to be expected from a contract penalty in a R&D contract either.

  2. The government may find the possibility of paying the lump sum upfront, before production even starts, unpalatable. And it has to be upfront and independent of the number of units ordered, else it has the same effect as penalties.
    Oh and I forgot the part about manufacturer's maintenance. It also has to be an independent contract with a lump sum upfront and a per-annum and per incident charge.

    1. No, you're confusing this.

      Governments actually DO pay lump sums up front, before production even starts. That would be no change at all.

      Maintenance costs are variable costs, not fixed costs, and thus there's no reason whatsoever to pay for them as if they were comparable to the costs of setting up a production line.
      To choose the option to extend a maintenance contract for a year over and over again is far from having any effect comparable to a contract penalty.

      Excessively long maintenance contracts with guaranteed revenues were covered by my "and anything which has the same effect in government procurement of goods or services".

    2. Do they pay enough to cover setting up the entire production line? Including subcontractors, logistics and whatnot? I don't know, but I guess not. If I'm correct the rest of the production costs pad the the per unit cost of the product. Therefore if the government were able to cancel without penalties, then the manufacturer would incur a loss.

      Maintenance costs do have a fixed component - the initial training, facilities, etc.

      What is an excessively long maintenance contract? Beyond the lifetime of the product? The only reason for a guaranteed revenue would be keeping up a maintenance capacity far in excess of peacetime requirements - which, you'd agree, would make sense in certain cases.

    3. "Do they pay enough to cover setting up the entire production line?"

      In many cases the contractor gets the money for setting it up after inflating the costs to do so, and gets paid several per cent profit on top of it.

      "Maintenance costs do have a fixed component - the initial training, facilities, etc."
      So what? Price it into the first year's maintenance and it's fine. There's no problem that would justify contract penalties.

      "What is an excessively long maintenance contract?"
      A maintenance contract which creates a monopoly for the contractor. Outsourced services should at least in theory be limited to a couple years and then there should at lest be a tender so potential competitors can put pressure on the winning bid's price.

  3. A wrote: "The government may find the possibility of paying the lump sum upfront"

    Why, this works in "civilan" industry, and the German government get's cheap money.

    ", before production even starts, unpalatable. " Why????

    "And it has to be upfront and independent of the number of units ordered"
    Yes, the government pays for R&D and production of a prototype that fulfills the specifications. What is your problem?


    Why do you think that most military procurment project go wrong? The current SOP does obviously not work.


    1. In case of no-penalties I think with large contracts the up-front sum will be enough to induce a stroke in certain politicians. And it must be payed before any product is delivered. So you have spend large sums for several years (probably) without any guarantee that you will see a product. The opposition (wherever) will have field day.

      The R&D contracts are a bit different, as their output are prototypes or even pre-production units.

      I think there's nothing special about military procurement failures. They fail for the same varied reasons that civilian projects fail. However, the budget, and hence the scale of the possible failure, is much larger.

      I've been browsing various civilian project failures from here:
      Usually management failures, engineering overextension and changing specs figure prominently in the reasons for failure.

    2. "In case of no-penalties I think with large contracts the up-front sum will be enough to induce a stroke in certain politicians."

      Well,t hat's the point. It's systemically wrong to fool them into believeing the program is cheaper than it is. The contractors already attempt that, why help them to deceive our representatives?

  4. @leo

    Of course some civilian projects fail, however, almost all of the larger military projects fail. In some cases they will do to the customer more damage than the potential enemy. That is my problem.

    I have absolutely no issues with clearly seperating the number independend first part from the rest, this gives better overall picture, i.e. it becomes clear for the tax payer which part actually caused the price, and forces politicians to think harder about the sense or nonsense of a program.