A new method to mitigate cost underestimation in transport infrastructure projects: the decomposed contingency approach

Cost underestimation remains a recurring concern in transport infrastructure projects, undermining funding reliability and public trust. Conventional approaches to cost contingency – Quantitative Risk Analysis (QRA) and Reference Class Forecasting (RCF) – treat uncertainty as uniform. QRA compresses epistemic uncertainty into probability distributions, creating a false sense of precision, while RCF relies on historical uplifts that combine risks, aleatory, and epistemic uncertainties into a single adjustment. In this article, we introduce a new method, the decomposed contingency approach, which separates these components: risk (probabilistic events), aleatory uncertainty (irreducible variability) and epistemic uncertainty (limited information). Epistemic uncertainty is made explicit through smart heuristics: simple, task-specific decision strategies informed by ecological rationality. Applying a structured six-step process to a hypothetical suburban rail extension, we show how the proposed approach can enhance transparency, reduce double-counting and position RCF as a governance benchmark rather than an additive adjustment. We suggest that this decomposition method provides policymakers with a robust, easy-to-check basis for estimating cost contingencies, thereby enabling a more reliable assessment of final construction costs.