概率水平与模糊性对短视损失规避的影响

The Effects of Probability and Ambiguity on the Myopic Loss Aversion

短视损失规避(MLA)是指如果投资者频繁地评估其投资回报,那么他在投资时就会减少对风险资产的投入。本文采用实验法探讨了任务概率水平与模糊性对MLA的影响。结果 MLA只存在于中等损失概率中,说明该现象并非普遍存在。一旦概率变得模糊,在高损失概率任务中会出现MLA,而在低损失概率任务中则出现反向的MLA,说明模糊性能扩展"短视"起作用的概率范围。

The concept of myopic loss aversion （MLA）, introduced by Benartzi and Thaler （1995）, is important in behavioral economics. In the context of financial decision making, MLA implies that the more frequently returns of investments are evaluated, the lower is the average input in risky assets. Since it was put forward, researchers had been focusing on its conditions and influencing factors. A common feature of all the literature was that investment decisions were taken under risk, since individuals knew the probabilities of the tasks they were betting upon. However, it was almost impossible to know precisely the probabilities in real markets. Based on the contradiction, the present study discussed the influence of task probability on MLA, including probability level and ambiguity. The assumptions of the present paper were as follows： H1, task probability level had a remarkable influence on MLA; H2, probability ambiguity also had a significant influence on MLA. Experiment 1 used a random sampling method, employing 69 undergraduate students from a domestic university. Besides, the experimental process was presented on the computer monitor, and participants would make risk decision one by one in a quiet laboratory without distractions. The experiment 1 adopted a 2 （evaluation frequency： high vs. low） x 3 （probability level： high vs. medium vs. low） mixed design. In Experiment 2, the win probability was replaced by a letter ＂p＂ while the loss probability was replaced by ＂l-p＂. The study expanded the findings of experiment 1 with a 2 （valuation frequency： high vs. low） x 2 （probability level： high vs. low） between-subject design. A total of 380 undergraduate students were recruited. The other processes were the same as experiment 1. The results showed that both experiments did not find the main effects of evaluation frequency on MLA （in experiment 1： F（1,134） = 1.067,p = .305; in experiment 2： F（1,330） = .000,p = .985）, while the main effects of probability level were all remarkable （in experiment 1： F（2, 134） = 60.306, p 〈 .001, r/2 = .684; in experiment 2： F（1,330） = 12.098,p = .001, 1/2= .187）. In experiment 1, the interaction effect of the valuation frequency and probability level was significant （F（2, 134） = 4.122,p 〈 .05, r/2 = .243）. The simple effect analysis showed that MLA only occurred in the medium loss probability task, while did not exist in both high and low loss probability task. Experiment 2 showed that the interaction effect of such two variables was also significant （F（1,330） = 10.253, p = .001, ~2 =. 173）. The simple effect analysis found if the probability was ambiguous, individuals would exhibit MLA in the high loss probability task and reverse MLA in the low loss probability task. These results indicated that the impact of MLA was stronger under ambiguity. In brief, the MLA was not a general phenomenon in markets, and whether it appeared depended on the rates of financial assets. Specifically, it only occurred in the medium loss probability. However, the impact of MLA would be greater when the probability was ambiguous, which suggested us gaining more information before decision-making to weaken the negative side of such impact.