In this article we study the implication of thresholds in preferences. To model this we extend the basic model of John and Pecchenino (1994) by allowing the current level of environmental quality to have a discrete impact on how an agent trades o ff future consumption and environmental quality. In other words, we endogenize the semi-elasticity of utility based on a step function. We motivate the existence of the threshold based on research from political science, from arguments based on regulation and standards, cultural economics as well as ecological economics. Our results are that the location of the threshold determines both the potential steady states as well as the dynamics. For low (high) thresholds, environmental quality converges to a low (high) steady state. For intermediate levels it converges to a stable p-cycle, with environmental quality being asymptotically bounded below and above by the low and high steady state. We discuss implications for intergenerational equity and policy making. As policy implications we study shifts in the threshold. Our results are that, in case it is costless to shift the threshold, it is always worthwhile to do so. If it is costly to change the threshold, then it is worthwhile to change the threshold if the threshold originally was sufficiently low. Lump-sum taxes may lead to a development trap and should be avoided if there are uncertainties about the threshold or the eff ectiveness of the policy.