의사결정 상황에서 매몰원가효과에 관한 실험연구

In an uncertain decision making sunk costs are irrelevant costs. So it is instructed in management accounting that sunk costs should be neglected in decision making. Nevertheless in real decision making they are used as if relevant costs. That is called as sunk cost effect. Even the reason of the effect is explained by various theories, the prospect theory of Kahneman and Tversky (1979) gives the most powerful explaining. That is, the sunk cost effect is explained by the value function of the prospect theory. According to the value function, under the situation in which there is some sunk cost that is initial expenditure or investment, since the reduction of the value of additional loss is relatively small but the increase of the value of additional gain is relatively big, additional expenditure or investment is executed. Several experimental studies were conducted on sunk cost effect. Especially Arkes and Blumer (1985) conducted earlier experiment to test the sunk cost effect. They conducted 10 experiments using various scenarios and found the existence of the sunk cost effect. The purpose of this study is that following hypotheses are tested in Korea. Hypotheses are as followings: (1) Sunk cost effect exits in decision making. (2) Sunk cost effect is revealed stronger as the amount of sunk costs gets bigger. (3) The instruction on sunk cost may reduce sunk cost effect. The participants are composed of 76 undergraduate students who are major in accounting. They are assumed to have the ability to make relatively rational decisions. The participants are grouped into two groups, those are control and experimental group. The experiments are conducted three times with scenarios which are a little adjusted from those of Arkes and Blumer (1985). Hypothesis (1) is tested by all the experiments, hypothesis (2) is by the second experiment, and hypothesis (3) is by the third experiment. Scenario (1) used by the first experiment contains the concert ticket choice problem. If he neglects sunk cost, that is the ticket price which he paid, he should choose singer B concert, but he considers sunk cost he may choose singer A concert. Scenario (2) used by the second experiment contains the same problem with scenario (1), but it contains one more context that if the ticket price changes which concert he may choose. This scenario tests the amount sensitivity of sunk cost effect. Scenario (3) used by the third experiment contains continuous investment problem of car development project. The results of the experimental study are followings: Hypothesis (1) is partly supported, hypothesis (2) is supported but hypothesis (3) is not supported. That is, sunk cost effect is not always revealed, but the effect is revealed strongly as the amount of the costs gets bigger. But the instruction on sunk cost could not reduce the effect. This study was done with experimental method. So the generalization of the result is a little limited. Future research with various scenarios and contexts are suggested.