A Lot of Ambiguity
We consider a risk averse decision maker who dislikes ambiguity as in the Ellsberg urns and compare the certainty equivalent of this gamble with the certainty equivalent of the anchoring probabilistic lottery. We deal first with the Choquet EU model and show that under some conditions on the capacity $
u$, when independent ambiguous gambles are repeated and the expected value of the anchoring lottery is zero, the difference between the average ambiguous and risky certainty equivalents converges to zero. When the parallel expected value is positive, we show that if the average certainty equivalent of the risky lottery is non-negative, then so is the limit of the average value for the ambiguous model. These results do not extend to the maxmin model, but under some conditions they extend to the smooth recursive model.
(with Zvi Safra, WBS)