Costly switching from a status quo

We axiomatically characterize a theory of status quo-dependent choice where an agent faces switching costs that depend upon both the status quo and the alternative he switches to. In a choice problem with a status quo, the agent chooses the alternatives that yield the highest utility net of switching cost. This generates status quo bias and also allows for a wide range of reference effects. We examine the behavior of such agents in Prisoner's Dilemma (PD) games. In a single PD game, switching costs can lead to cooperation. However, across different PD games, it is not "anything goes" and instead we derive necessary and sufficient conditions for cooperation rates to be consistent with our model. We then verify that these conditions are satisfied by Charness et al.'s (2016) experimental data. We also perform a similar analysis for other theories such as models of status quo bias, magical thinking, inequity aversion, and fairness; and find that these theories make either invalidated or looser predictions. (C) 2018 Elsevier B.V. All rights reserved.