What Static Elicitations Recover When Choices Change Future Opportunities

Researchers estimate behavioral parameters by inverting static, payoff-only elicitation equations. We show that when the alternatives also change the subject's future opportunities, the recovered coefficient absorbs their value, and derive the exact adjustment: the estimate equals the primitive plus the omitted continuation-value difference, discounted to the divergence date and divided by the equation's sensitivity. One unrestricted elicitation cannot separate them; across arms the coefficients form a linear system determining when the primitive is point identified and how many arms suffice. The continuation problem enters through a sufficient statistic whose content depends on the evaluator: a mean under expected value, a mean and one premium under maxmin or fixed curvature, the full distribution when curvature is unrestricted. We study two applications: a continuation-neutral arm recovers loss aversion, while in monetary timing tasks liquidity values defeat identification of the discount factor; rewards consumed on receipt, with date-stable utility, restore it.