Thursday, June 28, 2012

Shipbuilding Costs - Impact of Low Volume

The Navy is clearly on a fiscally unsustainable path as regards shipbuilding.  They simply can’t afford to keep building ever more expensive ships on what amounts to a fixed budget, at least for the foreseeable future.  But why do ships cost as much as they do?  Why should a destroyer cost $2B-$4B?  Why does a carrier cost $14B, now?

The easy answer that most people give is bureaucratic inefficiency and waste, both government and industrial.  Other reasons include changing requirements during construction (the concurrency that was discussed in a previous post), overuse of MILSPEC, overly complex win-the-war-singlehanded designs, lack of in-house oversight, and so on.  All these reasons undoubtedly contribute to high costs, however, there's one other contributing factor that I believe is far more significant and that is construction volume.  Ships have gotten so expensive that we only build a few each year.  That's not much volume for the shipyards.  So, instead of spreading their overhead costs and profit over many ships, each shipyard has to roll all their costs as well as profit into the one or two ships they build.

Typical profit margins for "stuff" that you buy commercially in everyday life is 10%-20%.  So, 10-20 cents of each dollar you pay goes to profit for the company that makes the item.  That's fair.  What if the company(s) that make soap knew they were only going to sell two bars for the year?  In order to stay in business, they would have to roll all of their production costs, overhead, and profit into those two bars.  Each bar of soap would cost hundreds of thousands of dollars.

Now, that was a bit of a ridiculous example but it illustrates the point.  Shipyards used to build many ships per year during WWII, several per year during the 600-ship fleet days, and only a couple per year now.  And yet, all the same costs and profits (adjusted for inflation, of course) now get rolled into the couple of ships instead of many.

Ford Class,   Low Volume = High Cost

If the shipyard is only going to build one carrier every five years then it has got to roll five years worth of costs and profits into the price.  And guess what?  We're probably going to a seven year cycle for carriers with the next Fords already looking at intentional two year delays.  Those extra two years worth of costs and profits still have to be included or else the shipyard goes out of business.  Instead of a carrier costing five years of shipyard costs and profits, now they have to cost seven years of shipyard costs and profits.  For the Navy and government, a seven year cycle lowers the yearly costs a bit but drastically raises the total costs.

Here’s some official testimony on the subject.  Ms. Eaglen testifying in front of the House Armed Services Committee said,

“I think another thing we need to do is be honest about what stretching out programs does to the costs.  So, for example, take [the] Joint Strike Fighter program as part of the 2013 budget requests.  The overall buy wasn’t cut but a lot of the planes were shifted into the out years.  The Pentagon said this morning that costs [Ms. Eaglen’s verbal emphasis] $6B.  That’s not a savings, that’s a cost.  But it was done in order to meet a budget target of savings.  … unless we’re honest about what stretching out of these programs, including carriers, is really doing to the cost of them by making them actually rise over time …”
We were going to build 15 carriers and now we're only going to build 10.  We were going to build 32 DDG-1000 and now only 3.  We were going to build 55 LCS and now only (?).  We were going to build (?) CG(X) and now none.  And so on.

In effect, the government is subsidizing the shipyards by paying higher than justified prices in order to keep the shipyards in business.  There is an aspect of national strategic/security concerns at work here and that's understandable.  We don't want to find ourselves with no capable shipyards.  However, the act of subsidizing ensures that shipbuilding costs are higher than they should be.

I'm beating this point to death and I'll stop now.  I think this may be one of the more significant reasons for cost increases over and above inflation but I don't have actual numbers to prove it.  Consider, though, the carriers which have had the largest price increases over inflation.  We're stretching fewer buys over longer periods.  Contrast that with the Burkes which have had the most buys over the shortest periods.  I don't think it's coincidence that the Burkes have had the smallest price increases over inflation (possibly no increase).

For what it's worth, here's my ranking of the sources of cost growth in shipbuilding and the relative impact I think they have (in percent).  Just guesses and opinions on my part!

1.  As just discussed, limited volume - 45%
2.  Changing requirements during construction (concurrency) - 30%
3.  Overuse of MILSPEC - 20%
4.  Simple inefficiency - 5%

What can we do?  Build more ships by building smaller, less capable, and therefore less expensive ships.  Also, upgrade existing ships.  Shipyards don't have to just build new ships, they can rebuild existing ones.  Instead of early retiring good ships, let's drastically upgrade them.  It will still be way cheaper than new construction and will provide more work for the yards and more projects for the yard's costs and profits to be spread over.


  1. There is a push to what looks like savings but is actucally hurting us.

    Hell thats not even counting the cost in replacment parts for the Zumwalts. Parts are going to be more expensive for 3 than 32.

    Dont build ships geared for the-master of nothing, OK at a couple things, but listed as great as everything- build ships that are OK at one thing Good at another. Do one thing real well even if that limits what they can be used for and something else OK.

    Then you build alot of them. This cuts down on cost and complication. Use a similar hull on multiple platforms as long as it doesnt hurt the ships overall mission.

  2. James, you make an excellent point about the additional costs of limited quantity runs like the Zumwalts. Parts, maintenance, training pipelines (to an extent), etc. all become "one-of-a-kind" costs. Even seemingly simple things like the reams of documentation assoicated with a ship (manuals, blueprints, operating envelopes, operating procedures, etc.) all have to be generated at significant cost and yet only apply to three ships out of the entire fleet!

  3. See GAO-09-322.

  4. How expensive is operating such a platform in comparison?