A correspondent e-mails to point out several lasting political legacies of Robert McNamara: one was the 1986 Goldwater-Nichols Act (GNA), which codified a major change in the command structure of the military, taking away power from the service chiefs and giving it to the civilians -- the President, the Secretary of Defense (who were given direct line authority over theater commanders) and away from the military bureaucracy, which had grown either too powerful or too parochial. Concepts like "jointness" predominated. But so did the give and take that might have led -- or might lead to -- better decisions. If the Chairman of the Joint Chiefs of Staff doesn't have to listen to the generals and is politically responsive to the president, then dissenting views -- think Eric Shinseki's at the outset of the Iraq war -- are formally given no chance to be heard. The McNamara model implies a strong secretary of defense whose principal political task is to watchdog the service chiefs and correct for their parochialism. Political appointees gained enormous informal power under McNamara and then formally under the GNA. (One underlying assumption of GNA is that previous defense secretaries were too weak to resist the demands of the service chiefs; McNamara couldn't significantly reduce manpower requests of the Army and Air Force in his later Vietnam years even though he grew increasingly convinced that the approach was wrong.)
Legacy two: "Systems analysis" -- the introduction of economic modeling for determining courses of actions, tactics, strategies and weapon systems to buy. McNamara's approach (the guy was trained as a statistician) transformed military decision-making. Today, most Pentagon decisions are informed by this approach.