通知公告
首页»
通知公告
通知公告
03.30经济系Seminar：Reference Dependence under Risk: An Experimental Examination
发布时间：2015-03-19
浏览量：
主 题：Reference Dependence under Risk: An Experimental Examination
时 间：2015年3月30日（星期一）14:00-15:30
地 点：后主楼1610
主讲人：闫巨博，康奈尔大学应用经济与管理学院
主持人：何浩然，北京师范大学经济与工商管理学院
主讲人简介：
闫巨博，康奈尔大学应用经济与管理学院博士候选人。研究领域包括行为经济学，实验经济学以及能源经济学。近期的研究主要为风险与不确定条件下的决策问题以及公共物品的提供。采用的研究方法则包括实验室实验以及对既有数据的检验。论文发表在Environmental and Resource Economics、Resource and Energy Economics等国际重要期刊上。
内容摘要:
The critical feature of prospect theory, loss aversion, has been extensively tested in both laboratory and the field. The other important feature, probability weighting, has been less emphasized in economic literature. This paper studies reference dependent behavior under risk using a controlled laboratory experiment and demonstrates the need of probability weighting in explaining the observed behavior. The paper also suggests a behavioral model that explains reference dependence under risk using loss aversion and rank-dependent probability weighting. The behavioral model used in the estimation is flexible enough to accommodate both expected utility theory and prospect theory as special cases. Certainty equivalents for either lottery tickets for a $5 gain or insurance policies to protect against a $5 loss are elicited. On one hand, between subject variation in elicitation mode (Willingness To Accept vs. Willingness To Pay) provides two different values for each lottery ticket or insurance policy which enables further exploration of reference dependent behavior under different probabilities. On the other hand, within subject variation of probabilities allows an examination of respond to risk changes. A specific functional form derived from previous studies is used to estimate a single loss aversion parameter and separate decision weights for each of nine probabilities between zero and one. The resulting nonparametric probability weighting function is consistent with the inverse S shaped curve in literature and the loss aversion parameter is also similar to prior estimates supporting prospect theory. The experiment is then replicated using risky prospects of obtaining or losing a coffee mug with a university logo. With these data, one can explore the formation of reference points in the different treatments of the experiment. Consistency between the two sets of estimated parameters again supports the behavioral model with loss aversion and rank-dependent probability weighting.
时 间：2015年3月30日（星期一）14:00-15:30
地 点：后主楼1610
主讲人：闫巨博，康奈尔大学应用经济与管理学院
主持人：何浩然，北京师范大学经济与工商管理学院
主讲人简介：
闫巨博，康奈尔大学应用经济与管理学院博士候选人。研究领域包括行为经济学，实验经济学以及能源经济学。近期的研究主要为风险与不确定条件下的决策问题以及公共物品的提供。采用的研究方法则包括实验室实验以及对既有数据的检验。论文发表在Environmental and Resource Economics、Resource and Energy Economics等国际重要期刊上。
内容摘要:
The critical feature of prospect theory, loss aversion, has been extensively tested in both laboratory and the field. The other important feature, probability weighting, has been less emphasized in economic literature. This paper studies reference dependent behavior under risk using a controlled laboratory experiment and demonstrates the need of probability weighting in explaining the observed behavior. The paper also suggests a behavioral model that explains reference dependence under risk using loss aversion and rank-dependent probability weighting. The behavioral model used in the estimation is flexible enough to accommodate both expected utility theory and prospect theory as special cases. Certainty equivalents for either lottery tickets for a $5 gain or insurance policies to protect against a $5 loss are elicited. On one hand, between subject variation in elicitation mode (Willingness To Accept vs. Willingness To Pay) provides two different values for each lottery ticket or insurance policy which enables further exploration of reference dependent behavior under different probabilities. On the other hand, within subject variation of probabilities allows an examination of respond to risk changes. A specific functional form derived from previous studies is used to estimate a single loss aversion parameter and separate decision weights for each of nine probabilities between zero and one. The resulting nonparametric probability weighting function is consistent with the inverse S shaped curve in literature and the loss aversion parameter is also similar to prior estimates supporting prospect theory. The experiment is then replicated using risky prospects of obtaining or losing a coffee mug with a university logo. With these data, one can explore the formation of reference points in the different treatments of the experiment. Consistency between the two sets of estimated parameters again supports the behavioral model with loss aversion and rank-dependent probability weighting.