Incentives are usually expected to increase motivation and cognitive control and to thereby improve performance. A small but growing number of studies have begun to investigate whether the effects of incentive on cognitive performance differ for younger vs. older adults. Most have used attention and cognitive control paradigms, trialwise implementation of incentive condition, and gain incentives (reward), with only a very few investigating the effects of loss incentives. The present study takes a complementary approach: We tested younger and older adults in a working memory paradigm with loss incentives implemented session-wide (between subjects). We also included self-report measures to ask how loss incentive affected participants' perceptions of the mental demand of the task, as well as their perceived effort, frustration, motivation, distraction, and metacognitive judgments of how well they had performed. This allowed us to test the disparate predictions of different theoretical views: the intuitive hypothesis that incentive should increase motivation and performance, the motivational shift proposal that older adults are especially motivated to avoid losses (Freund and Ebner, 2005), a heuristic "positivity effect" perspective that older adults ignore losses (Brassen et al., 2012; Williams et al., 2017), and a more nuanced view that suggests that when negative information is unavoidable and increases perceived costs, older adults may instead disengage from the situation (Charles, 2010; Hess, 2014). The results seemed most consistent with the more nuanced view of the positivity effect. While neither group showed incentive-related performance differences, both younger and older adults reported greater perceived demand and frustration under loss incentive, especially in the most challenging conditions. Loss incentive increased the accuracy of immediate metacognitive judgments, but reduced the accuracy of later, more global judgments of competency for older adults. Self-report measures suggested that the loss incentive manipulation was distracting to young adults and demotivating for older adults. The results suggest a need for caution in generalizing from existing studies to everyday life, and that additional studies parameterizing critical aspects of task design and incentive manipulation are needed to fully understand how incentives affect cognition and motivation in younger and older adults.