Economic studies in project escalation and project selection

Escalation is generally defined as continued commitment of resources following negative feedback. Chapter 1 presents a literature review. In Chapter 2, a principal-agent model of asymmetric information uses escalation behavior to link managerial turnover and project outcomes. Managers escalate to conceal mistakes and protect reputation. Turnover serves to prevent the revelation of mistakes and to avoid punishment. The model is tested using a large current dataset with the data on managerial turnover and project discontinuations. Evidence is found that turnover among top executives significantly increases the likelihood that the firm subsequently discontinues operations. Length of tenure at firm is found to reduce escalation. Results on the effect of the specific reasons for managerial departure on the project discontinuations are reported. Chapter 3 focuses on the role of adoption reversibility and switching costs in technology adoption patterns in the presence of network externalities and uncertainty. Network externalities exist when a product\u27s benefits include the intrinsic payoff and a payoff from the network of other adopters. Irreversible adoption gives advantage to the safer tested technology. Choosing the untested technology creates the risk of being stranded without a network of followers if the payoff is inferior. In the case of low switching costs, the adoption is reversible and the gain of sampling from risky distributions first is observed. Chapter 4 views escalation through real option theory. The model yields behavior consistent with case study evidence. It makes escalation a product of solely the objective project characteristics. A project with negative “myopic” NPV may be beneficial for the firm once the option value of continuing for one more period and making the stopping decision later is incorporated into the analysis. Chapter 5 applies bandit process studies to the issue of information technology project failures. The bandit perspective shows that managers choosing a risky project with high potential reward before a safer one are behaving in the firm\u27s interest. A high proportion of risky projects adopted leads to a high number of project failures. The bandit approach supports escalation prevention studies that advocate evaluating decision-makers on the basis of their decision processes, rather than specific outcomes