This study examines how perceived power asymmetries in client–contractor relationships contribute to construction payment delay and develops an early-warning model based on project-interaction information assessable by contractors. It addresses the limitation of existing payment-prediction models, which depend on client financial information that payment receivers cannot access.
A model was refined through a four-round Delphi study and tested using questionnaire data from 486 UK construction projects. Partial least-squares structural equation modelling was used to validate the behavioural pathways, and the supported structural and moderation pathways were converted into a Bayesian belief network for predicting payment-delay severity. Predictive performance was evaluated on a held-out test set and benchmarked against alternative classifiers.
Perceived coercive power asymmetry was the most consistent driver of client and contractor opportunistic financial strategies, whereas perceived reward and expert power asymmetries generally reduced opportunism. Perceived forced payment obligation moderated mediated-power pathways, while digital infrastructure establishment moderated selected non-mediated-power pathways. On the held-out test set, the Bayesian network correctly predicted 141 of 146 cases, achieving a log-loss of 0.2432, a Brier score of 0.0898 and a Macro-F1 of 0.945 and produced the strongest performance among the compared models.
The study extends power theory by explaining the differentiated, behaviourally mediated and contextually conditioned effects of power asymmetry, while shifting prediction from inaccessible client financial data to project-interaction information assessable by contractors during project interactions. It applies an established SEM–BBN sequence to retain these explanatory pathways within an interpretable model through which contractors can anticipate and update risk during delivery.
