간행물

기본연구 update

  • : 정보통신정책연구원
  • : 사회과학분야  >  경영학
  • :
  • :
  • : 연속간행물
  • : 부정기
  • :
  • :
  • :

수록정보
수록범위 : 1985권1호(1985)~2021권9호(2021) |수록논문 수 : 947
기본연구
2021권9호(2021년 12월) 수록논문
최근 권호 논문
| | | |

1매몰비용 효과를 고려한 경매방식 연구

저자 : 김희천

발행기관 : 정보통신정책연구원 간행물 : 기본연구 2021권 9호 발행 연도 : 2021 페이지 : pp. 1-165 (165 pages)

다운로드

(기관인증 필요)

키워드 보기
초록보기

■ Purpose of Research
As a market mechanism that allows bidders or consumers to evaluate and value the price of any objectives, the auction was widely adopted for several fields. Research on auction mechanisms is also developed as one of the most prominent topics in the field of theoretical and experimental economics.
In this study, we designed an auction experiment that describes a spectrum auction under a reassignment situation. Especially, we considered the sequential auction environment in which bidders compete for the same virtual objective(licenses for spectrum band). Given that the only bidder is awarded for each objective, such a winner, who paid for the price in the first bidding, may consider the price as a sunk cost. Considering that such a bidding payment acts as (a sort of) sunk cost, we tried to test the effect of variables that can affect the sunk cost: the accountability(or responsibility), observability, size of sunk cost, and endowment effect. In detail, we designed the auction experiment that can explicitly involve and separate the accountability and/or observability in each combination of treatment sessions. Two variables, which were appointed as the major factors that stimulate an individual's motivation to justify previous decision making (Staw, 1981), were not tested for their effectiveness at the same place under the competitive environment. It is widely accounted that the observability, whether the bidder's own decision and its outcome are observed or monitored by the other competitors, may derive more dense competition in the auction. Such an outcome is explained by several factors like anticipated regret (Filiz-Ozbay and Ozbay, 2008), reputation effect, or utility from winning. On the while, accountability or responsibility characterizes whether the bidder's own previous bidding decision and its outcome affect the following bidding process for their own sake. In other words, when a bidder won the prize by her own decision, the bidder may feel responsible for the outcome and paid the price of the prize. If not, that is, the awarded prize in the first bidding is not driven by her own decision, then the bidder is less likely to feel that the followed outcome and the paid price are her responsibility. Such responsibility (or accountability) is expected to work the decision maker's action to derive a higher bidding price for the same prize in the followed bidding. Responsibility is most likely to be involved with the sunk cost, but we consider the case in which sunk cost is NOT involved with responsibility. We considered all the paid prices in the previous bidding as sunk costs if they are paid from the participant's own ledger, which is different from the endowment that is given to participants without any paid price. However, most previous research handled such two variables in the same basket. That is, there is no previous research about the effectiveness of two variables when they are handled individually. For example, we can easily expect the outcome when observability and responsibility factors are imposed at the same time. However, when we consider the case in which only observability is imposed while responsibility is turned off, we cannot expect whether observability can work as if it were in the all-you-can-have session. To answer such a question, we designed an experiment that can explicitly separate such factors. By having an answer to such questions, we may have a better understanding of how the sunk cost effect will affect the outcome of strategic decisionmaking in the competitive environment and can improve the design of the auction mechanism to fit its policy implication.
■ Structure and Scope of Research
In the first section, we will discuss the goal and purpose of this research and study the previous literature on the sunk cost effect, endowment effect, and experimental auction. In the following section, we will explain the design of the experiment and set the hypothesis we want to test from the current conduct of experiments. The main hypotheses we want to examine is the following:
H1) Bidding price for the prize is high when it involves a high sunk cost from the previous bidding.
H2) Bidding price for the prize is higher when it involves sunk cost than when it involves the same amount of endowment.
H3) Bidding price for the prize is higher when it involves responsibility than when it is not.
H4) Bidding price for the prize is higher when it involves observability than when it is not.
H5) Bidding price for the prize is higher when the bidder has a higher wealth
■ Main Outcome
We experimented with paper and pencil and recruited 113 participants for the sessions. For each session, we varied the treatment of variables. The following table summarizes details of treatments and recruited participants.
We excluded data from two pilot sessions because they contain some possible contamination from errors in experiment operation. Except for them, we have four sessions with responsibility (51 effective subjects) and without responsibility (45 effective subjects) respectively. We also conducted four sessions with observability(48 effective subjects) and without observability(48 effective subjects) respectively. Endowments treatment is varied by distributing the different endowment coupons in the same session. By doing that, 82 subjects were imposed with endowment and 41 were not.
The result of the experiment is as the following. The size of sunk cost is measured by the paid price in the first bidding. Sunk cost showed the most statistical significance(p-value: 0.0179) among all possible variables in the second bidding. While the responsibility factor did not show a statistically significant effect (p-value: 0.141881) on the winning price of the second bidding, the direction of its effect is observed as a positive (615.7194). Observability showed statistical significance in both biddings, but it showed more effectiveness in the first bidding in terms of size(981.33) and statistical significance (p-value: 0.0125). Endowment also showed effectiveness(p-value: 0.0383) in the first bidding, but the effectiveness is not reached in the second bidding (p-value: 0.303718, when pooled). Size of wealth is measured by the revenue that each participant was holding at the end of the first bidding, excluding the price they paid in the first bidding. It also showed a statistically significant effect, but the size of the wealth effect is observed as smaller than that of sunk cost effect.

1
권호별 보기

내가 찾은 최근 검색어

최근 열람 자료

맞춤 논문

보관함

내 보관함
공유한 보관함

1:1문의

닫기