Factors Influencing Managers' Decisions to Continue/Discontinue Capital Budgeting Projects

Normative decision theories stipulate that rational economic decisions (including investment decisions) are made based on the assessment of the expected outcomes. Accordingly, if available feedback information about an investment project indicates at any given point in time that the expected outcome would be negative, one would expect that such an investment project would be immediately terminated. This implies that any prior funds spent on an investment project that cannot be recovered (known as sunk costs) are considered irrelevant to the decision to terminate. However, the decision-making literature provides instances of entrapment or escalation where decision makers were influenced by sunk costs in their decisions.The psychology literature provides two basic explanations for this entrapment or escalation behaviour. The first explanation attributes the escalation behaviour to the notion of self-justification and asserts that a decision-maker who is associated with the initial decision to invest in a project will elect to commit additional resources to the investment project as a way to justify past decisions, even when the available information indicates that the project should be terminated. The second explanation is rooted in prospect theory, which postulates that decision-makers tend to evaluate choices as gains and losses in relation to a neutral reference point and that decision-makers are generally risk seeking when faced with possible losses. That is, a risk-seeking person would reject a certain loss in favour of a gamble with equal or even "lower" expected value. Thus, for commitment escalation, the theory asserts that decision makers would view sunk costs as a sure loss and they would prefer to engage in the escalation of prior commitments with the possibility of incurring greater losses or recovering past investments.Agency theory provides a third explanation for escalation of commitments. When a manager posses private information on the state of the project, abandoning a potentially failing project would reveal this state and adversely affect the manager's reputation while further commitment of resources protects the manager's reputation. This paper extends previous work on the escalation of commitment by considering the effects of two additional variables that usually constrain managers' decisions. These variables are managers' potential liability for committing additional resources to a potentially failing project, and the level of credibility of a project's prospective information that the manager obtains.A between-subject experimental design is used where the three research variables (managers' responsibility, potential liability, and credibility of information about the prospects of a hypothetical project) are factorially crossed to obtain eight sets of experimental conditions (a 2x2x2 ANOVA design)The results indicate significant main effects of initial responsibility, potential liability, and information credibility. In addition, there were significant two-way interaction effects between initial responsibility and potential liability and information credibility.

规范决策理论（Normative Decision Theories）规定，理性经济决策（包括投资决策）的制定基于对预期结果的评估。因此，若某投资项目的可用反馈信息在任何特定时间点显示其预期结果为负，则该投资项目应立即终止。这意味着，投资项目中已投入且无法收回的前期资金（即沉没成本，sunk costs）与终止决策无关。

然而，决策领域文献记载了决策者受沉没成本影响的陷入或升级现象（entrapment or escalation）。心理学文献针对这种陷入或升级行为提供了两种基本解释。第一种解释将升级行为归因于自我辩护（self-justification）的概念，认为与项目初始投资决策相关的决策者会选择投入更多资源，以此为过往决策辩护——即便现有信息显示项目应终止。第二种解释源于前景理论（prospect theory），该理论假定决策者倾向于以中性参考点为基准评估收益与损失，且在面临潜在损失时通常表现为风险寻求。也就是说，风险寻求者会拒绝确定损失，转而选择预期价值相等甚至更低的赌博。因此，对于承诺升级，该理论认为决策者会将沉没成本视为确定损失，进而倾向于继续升级前期承诺——即便可能承担更大损失，或寄望于挽回前期投资。

代理理论（Agency Theory）为承诺升级提供了第三种解释。当管理者掌握项目状态的私有信息时，放弃潜在失败项目会暴露该状态并损害其声誉，而继续投入资源则可维护其声誉。本文通过考虑两个通常约束管理者决策的额外变量，拓展了承诺升级领域的现有研究。这两个变量分别是：管理者为潜在失败项目继续投入资源的潜在责任，以及管理者获取的项目前瞻性信息的可信度。

研究采用被试间实验设计，将三个研究变量（管理者责任、潜在责任、假设项目前瞻性信息的可信度）进行因子交叉，得到八组实验条件（即2×2×2方差分析设计）。

结果显示，初始责任、潜在责任及信息可信度均存在显著主效应；此外，初始责任与潜在责任、初始责任与信息可信度之间存在显著双向交互效应。