Causal network methods for integrated project portfolio risk analysis

Corporate portfolio risk analysis is of primary concern for many organizations, as the success of strategic objectives greatly depends on an accurate risk assessment. Current risk analysis methods typically involve statistical models of risk with varying levels of complexity. Though, as risk events are often rare, sufficient data is often not available for statistical models. Other methods are the so-called expert models, which involve subjective estimates of risk based on experience and intuition. However, experience and intuition are often insufficient for expert models as well. Furthermore, neither of these approaches reflects the general information available on projects, both expert opinions and the observed data. The goal of this dissertation is to develop a general corporate portfolio risk analysis methodology that identifies theoretical causal relationships and integrates expert opinions with the observed data. The proposed conceptual framework takes a resource-based view, where risk is identified and measured in terms of the uncertainty associated with project resources. The methodological framework utilizes causal networks to model risk and the associated consequences. This research contributes to the field of risk analysis in two primary ways. First, this research introduces a new general theory of corporate portfolio risk analysis. This theoretical framework supports risk-based decision making whether through a formal analysis or heuristic measures. Second, this research applies the causal network methodology to the problem of project risk analysis. This methodological framework provides the ability to model risk events throughout the project life-cycle. Furthermore, this framework identifies risk-based dependencies given varying levels of information, and promotes organizational learning by identifying which project information is more or less valuable to the organization.