We propose a new explanation for the WTP/WTA disparity in experiments and surveys. Uncertainty, irreversibility and limited learning opportunities can generate commitment costs, driving a wedge between WTP and WTA. We present experimental evidence that supports our hypothesis.