There are internal economies of scale when average unit costs fall as production increases. One reason this might happen is because it might allow the firm’s workforce to specialize further, increasing their productivity. Businesses with high fixed costs, like research, development, and testing, are also typically enjoy returns to scale. Industries with internal returns to scale are characterized by relatively few and large firms. The armaments industry — e.g. Lockheed Martin, General Dynamics, Northrop Grumman, BAE, Raytheon, et cetera — is one such industry.
The single largest consumer of weapons, especially military-grade arms and vehicles, are national governments. Only select countries, for example, are allowed to field M1 Abrams tanks. Like a good deal of government projects, armament development programs are typically very expensive, and suffer excessive overruns. The F-35 is approximately $160 million over budget. The RAH-66 Commanche consumed $6.9 billion before being canceled. The V-22 Osprey ended up costing 2068% more than originally projected. Yea, military procurement is extremely inefficient. (Actually, for a scientific look into the relationship between the U.S. government and weapons manufacturers, see Arms, Politics, and the Economy, edited by Robert Higgs.)
Longer periods of research and development, more prototypes, testing, and other pre-production steps mean higher fixed costs. Internal economies of scale not only mean that more production will lower average cost per unit, but it also means greater profits for the firm — that high fixed cost is spread over a greater number of units of output. As long as marginal costs are below average costs, there is a strong incentive to keep producing.
For a firm looking to sell their weapon system to the U.S. government, they have to persuade the latter to buy it. If selling more equates with making a higher profit, there will be some amount of money the company will be willing to spend on lobbying. The greater the amount of output necessary to reach “minimum efficient scale” (where average costs are lowest — at some point, marginal costs will pass long-run average cost, and the latter will begin to increase again), the more sales the firm will want to persuade the government to make. The higher the fixed costs, which include budget overruns, the more the contractor will want to sell, the more it will lobby.
Cost overruns are a good signal for waste. When waste attracts more waste, which in this case is buying more military equipment, we call this a perverse incentive. Is the process I sketch out here one such way this manifest itself in the real world?