| Many projects in new areas are characterized by unforeseeable uncertainty (i.e., inability to recognize all influence variables) and complexity (i.e., interactions between many variables, which make the overall performance difficult to estimate). Research has identified two approaches to project management under these conditions: Selectionism (pursuing independently multiple approaches and picking the best one ex post ), and Trial & Error Learning (flexible adjustment to new information as it emerges).; The first part of this thesis examines under what conditions which of the two approaches is more efficient. We find that trial & error learning is more efficient, if the team cannot test the selectionist trials in a realistic user environment. Further, presence of unforeseeable uncertainty can reverse the established result of computational optimization that the optimal number of parallel trials increases in complexity. Unforeseen factors make the trials less informative as complexity grows, and thus the optimal number of trials may decrease.; The rest of the thesis deals with setting incentives for project managers under such conditions. Trial & error learning involves actions after the unforeseeable uncertainty has resolved. In such situations, actions and targets planned at the beginning may no longer be optimal at the time of execution. The thesis constructs relational contracts between the firm and a project manager that re-instate optimal incentives by a priori defining time points and aspects of contract adjustment. Depending on what each party learns, the optimal contracts reflect different recommendations in the literature (process monitoring, downside protection, upward rewards).; Under selectionism actions are taken before the unforeseeable uncertainty resolves. Thus project managers have to choose actions based on their belief about the effectiveness of these actions. The thesis analyzes whether different types of contracts can set optimal incentives for effort in presence of unforeseeable uncertainty. Results stress the need to monitor the process or the effect of unforeseen events whenever possible. Otherwise, the firm can set optimal incentives for effort only under very restrictive assumptions. Further, results suggest that encouraging project managers to pursue more risky strategies can be an attractive alternative for the firm to increase the expected payoff. |
