DOD's acquisition environment does not provide incentives to prime contractors to use best practices to efficiently build high-quality weapon systems. The department faces challenges setting achievable requirements for systems development and providing effective oversight during the development process. In conducting systems development, DOD generally pays the allowable costs incurred for the contractor's best efforts and accepts most of the financial risks associated with development because of technical uncertainties. However, DOD and its contractors often enter into development contracts before requirements have been analyzed with disciplined systems engineering practices. This introduces significant cost and schedule risk to a development program, risk that is not borne by the prime contractor, but by DOD. Contractors have little incentive to utilize the best systems engineering, manufacturing, and supplier quality practices to control costs. DOD also has limited oversight of prime contractor activities and does not aggregate quality data in a manner that helps decision makers assess or identify systemic quality problems. In contrast, commercial companies we visited operate in an environment that requires their own investment of significant funds to develop new products before they are able to sell them and recoup that investment. This high-cost environment creates incentives for reasonable requirements that have been analyzed and proven achievable, the use of best practices, and continuous improvement in systems engineering, manufacturing, and supplier quality activities. -- GAO Report 08294 (emphasis mine)
It's not just DoD product managers that can be led down the garden path to the "easy money" of low/no oversight and outsourcing critical tasks.
To save money and minimize its financial risk, Boeing outsourced not just production, but design and engineering in ways it had never done before.
"Boeing had this colonial model," says Ray Goforth, executive director of the engineers' union at Boeing. "They were going to be the colonial power, and they would have these colonies, and they would tell them what they needed, and the colonies would deliver the parts and they would reap the profit. It was really this arrogant vision of how this was going to work." -- NPR, "Boeing's Dreamliner..."
If your core competency is designing and manufacturing huge commercial aircraft, why would it be a good long-term solution to farm big chunks of that out? It seems like you are just sowing the seeds that will grow into future competitors. Or in the DoD's case, farming out all of the engineering design and production expertise, the seeds grow into in-house incompetence and "poor outcomes", as the GAO would say.
Another interesting snippet, from that same GAO report (emphasis mine):
... the Joint Air-to-Surface Standoff Missile program, developed by Lockheed Martin, has experienced a number of flight test failures that have underscored product reliability as a significant problem. Ground testing, which prime contractor officials said could have identified most of the failure modes observed in flight testing, did not occur initially. Prime contractor officials indicated that ground testing was not considered necessary because the program was a spin-off of a previous missile program and there was an urgent need for the new missile.
It's easy to convince ourselves that Nature should care about our sense of urgency, but how badly we need a thing doesn't change how much testing or design work is needed to successfully produce a thing. Urgency may affect our risk tolerance, but the urgency of bureaucrats is generally manufactured to satisfy political and budgetary pressures that may incidentally involve national security, weapon's reliability or troop survivability.
I think Richard Feynman put it best, at a commencement address
The first principle is that you must not fool yourself, and you are the easiest person to fool.
and in his work concerning the Challenger disaster
Reality must take precedence over public relations, for nature cannot be fooled.