Abstract
"Recent applications have demonstrated the crucial role of decreasing absolute ambiguity aversion in financial and saving decisions. Yet, most ambiguity models predict that ambiguity aversion remains constant when individuals become better off overall. We propose the first tests of constant absolute and relative ambiguity aversion, using simple variations of the Ellsberg paradoxes. Our tests are axiomatically founded and grounded in the theoretical literature. We implemented these tests in an experiment. Our results call for the use of ambiguity models that can accommodate decreasing aversion toward ambiguity."