Integration of VaR and expected utility under departures from normality
建立了期望效用风险规避与VaR置信水平之间的映射关系，使两者在正态和非正态分布下能一致应用，并通过农场收入案例展示了非正态分布下的方法。
Abstract This article identifies the level of the expected utility (EU) risk aversion and Value‐at‐Risk (VaR) confidence level that yield the same choice from a given distribution of outcomes, and thus allow for consistent application of the two criteria. The result for a given distribution is an explicit mapping between risk aversion under EU and VaR, for both normal and nonnormal distributions. The Cornish–Fisher expansion is used to establish adjusted mean‐deviates for nonnormal outcome distributions and the investor's preference function is expanded to include elements for variance, skewness, and excess kurtosis. A farm‐level application with nonnormal revenue distribution illustrates these approaches.