This is coming straight from the GAO:
According to some Department of Defense (DOD) and industry experts, consolidation of the defense industry along with a shift in prime-contractor business models has resulted in prime contractors subcontracting more work on the production of weapon systems and concentrating instead on systems integration. Based on some estimates, 60 to 70 percent of work on defense contracts is now done by subcontractors, with certain industries aiming to outsource up to 80 percent of the work. At the same time, there is evidence that subcontractor performance may contribute to cost and schedule delays on weapon system programs. Congress has raised questions about the extent to which primes are awarding subcontracts competitively and about the government's insight into the process prime contractors use for determining what work to make in-house and what work should be bought from subcontractors (make-or-buy decisions). In the 2009 Weapon Systems Acquisition Reform Act (WSARA), Congress directed DOD, as part of efforts to improve competition throughout the life cycle of major defense programs, to ensure that contractors' make-or-buy decisions are fair and objective. Specifically, the Secretary of Defense was directed to require prime contractors to give full and fair consideration to qualified sources other than the prime contractor for the development or construction of major subsystems and components of major weapon systems. These actions were to be taken by November 22, 2009. Congress also directed DOD to revise its acquisition regulation regarding organizational conflicts of interest (OCI). In response to both of these requirements, DOD has drafted revisions to its acquisition regulation that are pending final approval. The 2010 National Defense Authorization Act required us to study the structure and management of major subcontracts under contracts for the acquisition of selected major weapon systems. In response to this mandate and given the reliance on subcontracts and the possible implications for government oversight, we (1) examined how government and prime contractors defined "major" subcontract, and the number and value of those considered major, (2) analyzed prime contractors' approach to selecting and managing major subcontractors, (3) examined the extent to which the government has visibility into major subcontracts and, finally, (4) examined how potential OCIs are addressed and the government's role in selecting the approach chosen.
While the Federal Acquisition Regulation (FAR) provides dollar thresholds for reporting on subcontracts, there is no set definition for major subcontract. Prime contractors and various government entities (program and contracting offices, as well as DCMA and DCAA) define "major" subcontract differently, for example based on a certain dollar value or on the criticality of the item being purchased. These differing definitions could affect levels of government or prime contractor insight, particularly with regard to contractor subsidiaries or affiliates, which the primes did not always consider to be subcontractors. The number and value of major subcontracts also varied considerably among programs in our review--particularly when affiliates and subsidiaries were included. One program had as many as 364 major subcontracts, representing 58 percent of the total value of all subcontracts. Another program had 13 major subcontracts, but they made up over 90 percent of the total value of all subcontracts. Prime contractors did not always include subsidiaries or affiliates in their definitions of major subcontract, even when they were managed as such. (2) Prime contractors in our review told us they structure their subcontracts to provide the required items, while reducing their risk exposure and maximizing their profit potential. For example, prime contractors generally attempted to shift cost risk onto their subcontractors through the use of fixed-price subcontracts even when their own contract with the government was cost-reimbursement. Primes also use make-or-buy processes to define what products and services must be retained internally to exploit their core competencies, and what should be outsourced to qualified suppliers to achieve cost efficiency. While primes define work performed by affiliates as part of their company's core competency--that is "make" activities--they often select and manage affiliates using similar methods to those used with external subcontractors. (3) The FAR emphasizes the prime contractor's responsibility in managing its subcontractors. Officials in our case studies underscored the limited role of the government in selecting and managing subcontracts. Prior to contract award, the government's visibility into subcontracts is restricted to the minimum amount of information necessary to determine that subcontract costs are fair and reasonable. To a great extent, the prime contract approach has implications on the degree of government's visibility into subcontract costs. For prime contracts awarded competitively, programs generally rely on the prime contractor to evaluate subcontractor proposals (even when subcontracts are not awarded competitively). For prime contracts awarded noncompetitively, the government has greater visibility into subcontract costs by validating noncompetitively awarded subcontractors' cost and pricing data over certain thresholds. Four of the six prime contracts we reviewed were competed in the development phase, and all six low-rate production contracts are planned to be awarded on a sole-source basis. In addition, when the prime contractor designates a subcontract as a commercial item, the government's visibility is significantly limited, including DCMA's postaward monitoring. (4) Most programs we reviewed displayed limited concern about the potential for organizational conflicts of interest (OCI) in their contracts. Although the FAR requires contracting officers to identify and evaluate potential OCIs, government officials told us that it was the prime contractors' responsibility to identify an OCI at both the prime and subcontract level. Some contracting officers also stated that they assume the OCI clauses are included in any support contracts, such as for testing and evaluation, and that they saw no need to include the clause in the weapon system contract itself. Consequently, most programs did not use contract clauses as a means to prevent an OCI.