Are Sunk Costs Irrelevant? Evidence from Playing Time in the National Basketball Association

Abstract

The importance of sunk costs in economic decision making is a major source of disagreement between neoclassical and behavioral economists. We provide a direct test of the role played by sunk costs using evidence from the National Basketball Association. Because teams have a greater financial and psychic stake in players chosen in the lottery portion of the draft and in the draft’s first round, behavioral economists predict that teams are more committed to them than to players selected later in the draft and give these players more playing time. Neoclassical economists would assert that current performance should be all that matters when allocating playing time. Our study builds on previous studies in two ways. First, we use a more appropriate data set than earlier studies, one that captures total available playing time more accurately than previous studies by accounting for time lost due to injury, suspension, and other exogenous factors. Second and more importantly, we use a more appropriate estimation technique --- regression discontinuity (RD) --- to show how behavior changes when a player's draft position crosses the threshold between lottery and non-lottery or first- and second-round status. Our RD estimates provide little evidence that teams allocate more playing time to more highly-drafted players than to otherwise identical teammates. We conclude that the neoclassical model more appropriately captures behavior.