18 Feb 2026

Civil Works Megaprojects: Cognitive Biases and Lessons for Effective Risk Mitigation

Abstract: U.S. Army Corps of Engineers (USACE) megaprojects are critical infrastructure investments that often experience cost overruns, schedule delays, and performance issues, with over 35 % of USACE dam structures rated in poor condition. This paper presents the first quantitative case study of past USACE Lock and Dam megaprojects—including the Charleroi, Chickamauga, Olmsted, Soo, and Montgomery projects —to identify patterns of risk mismanagement that contribute to these challenges. Using Bayesian methods and quantitative analysis on the risk registers of these projects, the study demonstrates that over-optimism in initial risk assessments, particularly the underestimation of risk likelihoods, can diminish the effectiveness of subsequent mitigation efforts by 25 % on average. The analysis also finds that risk managers tend to prioritize mitigation for risks where their confidence in assessment is highest, leaving higher-uncertainty risks less addressed. Based on these findings, the paper offers evidence-based recommendations for implementing structured risk governance frameworks, such as reference-class forecasting and external risk validation. This study advances project management knowledge by providing the first quantitative evidence of cognitive biases shaping risk register practices in civil works megaprojects. Incorporating such insights into future risk assessment and decision support methods can support critical infrastructure management and performance.