Tom Verducci of SI.com hammers the management of pitching in MLB, particularly the role of "closers," in "With More Closers Breaking Down, It's Time to Rethink the Modern Bullpen." In spite of extensive medical/training resources devoted to them, relievers (and starting pitchers) are pitching less and less but breaking down more and more. The piece includes a variety of data describing these trends to which Verducci adds:
It's not just that they break down; Wilson, Soria, Madson, Bailey and Farnsworth will earn $30.2 million combined this year, whether they pitch or not. It's that paying a guy $12.5 million to throw 60 innings ... is a waste of a great arm.
What does this say about managerial practice? Some decisions (say, outfielder position) seems to fit a model where each manager makes incremental decisions to maximize their player assets subject to existing conditions (say, opposing hitters, or more generally player/budgetary limits). In other areas, however, managers shift practice more like an on/off switch. When the Dodgers, Indians, and Giants innovated with black players in the late 1940s, the successful innovation caught on with other teams. Billy Beane increased the use of data analytic methods and after a decade or so, these methods became widespread. This "switching" fits the copy-cat managers who saw Tony LaRussa succeed by using Dennis Eckersley in one-inning segments. The trouble is that while some of the advantages of using the new decision are akin to using a new technology with the potential benefits common to all teams (e.g. black players), in other cases like relief-pitcher decisions, the resources and, therefore, the soundness of the practices differs across teams.
Is this a rejection of the typical way economists think about decision making and management -- an endorsement of "behavioralist" ideas? I think it more refinement than rejection. Yes, these episodes suggest an amendment the understanding of decisions under uncertainty, taking into account the application of successful decision templates around them. When problems are "ill-structured," -- they cannot be fully written out in a solvable mathematical form -- cutting-and-pasting the methods of others is one means searching for solutions.
In addition to the uncertainty issue, politics also constrains decisions. Friedman's discussion of "tyranny of the status quo" and similar analyses of "organizational inertia" describes these forces that build up that lobby for continuance of current practice. In the case of using black players, these kinds of constraints are fairly obvious, at least in retrospect. With relief pitchers, they are more subtle. Closers and agents who are compensated by saves evolve into an interest group protecting current practice. Starters and their agents who, rightly or wrongly, link pitch counts to longevity lobby for limits. Media commentators not only echo these concerns but serve as skeptics for any "new" practice. Together with the uncertainty problem, these "political" constraints encourage highly discrete or "lumpy" jumps in managerial practices rather than continuous and incremental ones. Eventually someone takes the step, succeeds, and resets practices. This lumpiness is implicit in a couple of questions in Tom Verducci's article:
Is anybody watching the Tampa Bay Rays? They don't have the money to waste nor do they waste a valuable young starter in a closing role. The team with the fourth best record in baseball since 2008 has done just fine with five different pitchers leading the team in saves over those five years.
And yet the universally accepted system is a failure when it comes to reducing the rate of injuries. What can change it? A maverick organization. (The Rangers and Giants are loosening pitch count restrictions in the minors, but the evidence is not yet very apparent in the majors.) A maverick manager. (Why won't somebody use a closer -- say Sean Marshall or Aroldis Chapman in Cincinnati -- in the manner of a 1980s closer such as Jeff Reardon?