Rethinking the Sunk Cost "Fallacy": A Commitment Device against Present-Bias

The sunk cost effect -- the tendency to persist with projects after incurring irrecoverable costs -- is traditionally viewed as a decision-making fallacy violating economic rationality. This paper challenges that interpretation by showing that sunk cost sensitivity can raise the welfare of present-biased individuals by serving as a commitment device. A simple three-period investment model shows that present-biased agents who experience psychological costs from abandoning sunk investments are more likely to continue profitable projects they would otherwise abandon.

While this mechanism can also lead to inefficient continuation -- the classic sunk cost fallacy -- the net effect is increased expected welfare. Optimal psychological costs are strictly positive and further improve outcomes as they increase in the initial investment. Unlike canonical models of self-control, which rely on deliberate commitment devices or choice-set restrictions, the mechanism here operates through a naturally occurring psychological response and does not require conscious awareness.

In fact, in the model it is naive agents who fail to anticipate their time inconsistency who benefit more than sophisticated agents. These findings suggest that seemingly irrational biases may function as adaptive mechanisms that partially correct for other cognitive limitations. They imply a testable relationship between present bias intensity and sunk cost susceptibility, pointing to directions for future empirical research. Because the model relies only on present bias and sunk cost sensitivity, its insights extend to a broad class of decision problems, with policy implications for areas where present bias undermines long-term goals.