The second argument usually used against setting up a second source is that the efficiencies of the first source would have to be scaled back because of the reduced volume to be produced. There is some credibility to this argument if the initial production line has been scaled to a certain level (that is, full production by a sole source) and if the high overhead and inefficiencies of the initial source are allowed to continue. Under these circumstances, costs will go up – since those large overhead costs have to be absorbed by the smaller quantities being produced.
However, if the production is initially scaled under the assumption that there will be a split buy, if after each bid the two firms scale their overhead to the size of their productions, and if the bids that each submits are based on efficient operations for economic production rates, the it is possible to have efficiency at smaller quantities. For example, when the Navy decided to scale back on the production of the Trident D-5 strategic missile system at Lockheed Martin from a rate of sixty per year to twelve per year, the unit cost of these large, sophisticated strategic missile systems fell.
That was from Jacques Gansler’s excellent book, “Democracy’s Arsenal.” In the book, Gansler showed cost curves which explicitly assumed that there were decreasing returns to scale over most relevant quantities. Said differently, production is more efficient the fewer units are produced. This would explode the heads of many DOD analysts.
I wish Gansler went into more detail about the precise mechanisms that are at work here, particularly on the Trident D-5 example. I personally don’t buy much into the whole “size your overhead right and it works” story. I think there are simply bottlenecks when it comes to skilled labor, specialized capital/supply chains. For example, Lockheed had to ramp up 3,000 employees on the F-35 and that contributed to a stall or even bump up on the learning curve. They encountered the scaling problem without even getting to full rate production yet.
This is why defense planners are so careful about making sure a follow-on program is ready for production when the legacy winds down. But that inevitably leads to planning problems — namely predicting how technology, regulations, and other matters will progress over the course of decades.
Another issue is that complex systems and intangible assets (e.g., software) require intimate communication between employees. No longer can you add new workers with little training and zero communication to an assembly line as though it were the industrial era. As you scale people up linearly, the number of interconnections that must be maintained so that everyone is on the same page grows geometrically. The communication problem is very real, and is why we tend to see small but highly competent teams in software (two-pizza teams).
And then, of course, there are the benefits of competition which is made possible by lower program quantities:
….when Navy Secretary Lehman said that he might buy F-14 fighters instead of the F-18s, which had been significantly increasing in price, the price of the latter dropped dramatically.
Leave a Reply