That's the suggestion made by Leo Blanken (of the Naval Postgraduate School) and Jason Lepore (of Cal Poly) in a paper I read on the flight home from Kansas City. As they put it, "the manner in which one measures progress incentivizes the behavior of those who are conducting the war."
For example, they say, the use of the "body count" in Vietnam "incentivized large-scale killing and destruction, which worked against the goal of building a viable political regime in the South."
But I am not sure I agree with their assumption that the "principal" (the policymaker back in Washington) "possesses more strategic information about the conflict" than does the "agent" (the commander in the field). Looking at Iraq, I would say that with the first three commanders in Iraq, neither side had more strategic information. Then, when Petraeus took over, he actually knew more strategically than his bosses (Gen. Peter Pace, Defense Secretary Rumsfeld, and President Bush) did.
BTW, if you plan to read this paper, it helps to like math.
Thomas E. Ricks covered the U.S. military for the Washington Post from 2000 through 2008.