WEAK ORDER STABILITY UNDER RISK

The assessment of a Bernoulli utility function is often accomplished b y an approximation process which makes use of ever more information on the preference as it is elicited from a decision maker. Simultaneous approximations of the utility function and the underlying distribution s are generally not sufficient for the expectations to converge to the expected value of the respective limits. We give various sufficient s tability conditions for the expected utility. All these conditions con tain some ''uniformness.'' (C) 1995 John Wiley & Sons, Inc.