Risk-based cost forecasting for construction projects in emerging markets: an empirical study

Thesis Abstract

This study addresses the persistent misalignment between initial cost estimates and actual project expenditures in construction within emerging markets, where volatile macroeconomic conditions, political risk, and underdeveloped data systems undermine reliable forecasting. The aim is to develop a risk-based cost forecasting framework that improves estimation accuracy for construction projects operating under high uncertainty. Specific objectives are (1) to identify and quantify key cost drivers and risk factors affecting forecast accuracy in emerging markets; (2) to examine the relationship between identified risk factors and cost overruns using empirical data; (3) to develop a probabilistic forecasting model integrating qualitative risk assessments with quantitative cost drivers; (4) to validate the model's predictive performance against historical project data; and (5) to formulate actionable recommendations for project stakeholders to mitigate forecast risk throughout the project lifecycle. The study adopts a mixed-methods design anchored in a positivist-interpretivist stance to capture both measurable cost drivers and contextual risk perceptions. The population comprises construction projects undertaken in three emerging economies over the past decade, with a target sample of 200 completed projects and 50 ongoing projects for prospective validation. Data collection employs (i) archival project records, including baseline budgets, change orders, and final costs, (ii) structured surveys of project managers, quantity surveyors, and contractors to capture risk perception and forecasting practices, and (iii) semi-structured interviews to explore nuanced risk factors and decision-making rationales. Instrument validity is ensured through pilot testing, expert panel validation, and triangulation across sources; reliability is assessed via Cronbach’s alpha for survey scales and test–retest procedures. Analytical techniques include regression analysis to identify statistically significant cost drivers, multivariate regression with heteroskedasticity-consistent standard errors to model forecast error, and a Bayesian network approach to integrate qualitative risk data with quantitative indicators. A Monte Carlo simulation will be used to generate probabilistic cost forecasts under different risk scenarios, with model performance evaluated by mean absolute percentage error (MAPE) and calibration plots. The theoretical framework integrates the value of information (VOI) concept with risk management theory and the Flexible Budgeting and Real Options paradigms to justify a probabilistic, information-rich forecasting approach. Expected findings indicate that volatile exchange rates, material price volatility, regulatory delays, and design scope changes are primary predictors of forecast inaccuracy, with interactions between macroeconomic shocks and project-specific factors amplifying cost overruns. The probabilistic model is anticipated to reduce forecast error by 15–25% compared with traditional deterministic budgeting, particularly when incorporating expert-elicited risk priors and scenario analysis. The study contributes to knowledge by (i) operationalizing a practically implementable risk-based forecasting framework suitable for data-constrained environments, (ii) demonstrating how qualitative risk perceptions can be systematically integrated with quantitative cost drivers to improve forecast accuracy, and (iii) providing empirical evidence on the transferability of Bayesian networks and Monte Carlo simulation in emerging-market construction contexts. Policy and practice implications include standardized risk registers linked to cost forecasts, decision-support tools for contingency planning, and guidance for procurement strategies that embed probabilistic budgeting in early-stage project planning. The conclusion is that incorporating structured risk information into cost forecasting yields measurable improvements in accuracy and decision quality, especially in settings characterized by price volatility and regulatory uncertainty. Recommendations emphasize the development of sector-wide data-sharing platforms, capacity-building for risk assessment among project teams, and the adoption of probabilistic budgeting as a standard practice in emerging-market construction projects.