Financial Risk Management Using a Bayesian Network in Public-Private Partnership Projects: A Middle School Construction Case

Abstract

Public-private partnerships (PPPs) can help finance public education facilities, but their long contracts and reliance on future cash flows mean that risks rarely occur in isolation. This paper uses a Bayesian network to examine financing risk in a Beijing middle-school construction case. The model draws on a risk framework from the literature, a project questionnaire, and case information on government, market, and social conditions. The original study distributed 130 questionnaires and reported 100 returns, although it also classified 30 responses as invalid; this inconsistency limits the precision of the survey evidence. Within the reported results, market and social-environment risk was 63.4% and government risk 63.0%. Changing market demand (34%) and lower-than-expected market returns (23%) were the largest sub-factors. The wider case matrix also reported government credit risk at 35%, approval delay and late completion at 20% each, and legal change at 5%. By showing how these factors depend on one another, the network provides a practical basis for allocating and monitoring risks at different project stages. The case suggests that robust demand forecasts, transparent public commitments, meaningful public participation, and regular updating of risk probabilities are more useful than a static risk checklist. The reported percentages should nonetheless be treated as case evidence rather than population estimates until the underlying survey records can be reconciled.