Rethinking Risk Propagation Mechanism in Public–Private Partnership Projects: Network Perspective

Abstract

A large portion of public–private partnership (PPP) infrastructure projects in China have ceased without delivery, which has drawn attention to the need for effective risk management. Previous studies directly and/or independently correlated risks with consequences. This does not allow for an effective investigation of causal chain and propagation effects among risks. To address this research gap, this paper developed a risk network model to analyze risk interactions and their propagation mechanism in PPP infrastructure projects. A case study of a PPP water supply project in China highlighted the key risks and mitigation actions. The study showed that policy change is frequently a source of risk, government default is profoundly affected by other risks, and safety accidents are one of the critical risks leading to project failure, which previous studies have overlooked. Furthermore, PPP stakeholders should focus on decoupling the following risks to reduce the probability of PPP project failure: safety accidents caused by lack of managing/organizational ability, and safety accidents caused by insufficient/delayed financing. The results indicated that the risk network model can be applied in PPP risk analysis to identify unnoticed but pivotal risks with propagation effects. This paper contributes to the research on PPP project risk management in three ways. First, it encourages both theorists and practitioners to consider PPP risk management as a dynamic, nonlinear process. Second, it illuminates critical risks and propagation mechanisms that have been overlooked by previous studies, thereby supplementing the body of knowledge on PPP project governance/management. Additionally, it proposes effective mitigation actions within risk management practice, thus laying the foundation for management and governance strategies.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 38 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.