There have been a few studies done that will take a players weighted averages from surrounding seasons to determine what he should have done and then compare it to how he did.
For instance, in 2001, Luis Gonzalez performed 40 runs better than his weighted average of the '99, '00, '02, '03 seasons. The idea is that this increase could then be attributed to the manager.
I don't like these stats because there are a lot of variables that could influence the variance. For instance, Mark Grace, who primarily batted in the 5 spot behind Luis Gonzalez in the 4 spot had a 2.6 WAR in 2001, followed by -0.1 & -0.6 WAR seasons.
There are to many variables that determine a teams success to reliably discuss the importance of a manager using metrics.