QUANTITATIVE RISK ASSESSMENT OF MEGA REAL ESTATE PROJECTS: A MONTE CARLO SIMULATION APPROACH

Abstract

This systematic literature review synthesizes how Monte Carlo simulation is designed, implemented, and interpreted for quantitative risk assessment in mega real estate projects. Following a PRISMA-aligned protocol, we searched Scopus, Web of Science, ASCE Library, ScienceDirect, IEEE Xplore, and Emerald Insight, complemented by Taylor & Francis Online and snowballing, screened records in two stages, and extracted standardized methodological variables. The final sample comprises 115 studies, analyzed across four decision-shaping themes: input distributions, dependence modeling, sampling discipline, and integration of cost, schedule, and finance. Most studies adopt triangular or PERT-style inputs, with lognormal alternatives common, while heavy-tailed forms remain underused despite escalation and long-lead exposures. Explicit dependence is unevenly treated, many models assume independence, whereas rank correlation, common drivers, or copulas reveal wider tails. Latin hypercube sampling dominates, quasi Monte Carlo appears in a minority, and only a subset justifies run size against precision targets. Where models link schedule to cost and propagate into cash flow and financing metrics such as debt service coverage ratio, credible bands widen and breach probabilities rise, improving alignment with lender and board thresholds. Sensitivity analysis is frequent yet mixed in rigor, with global and tail-focused measures offering clearer mitigation leverage than tornado charts, and validation by hindcasting or exceedance tests remains rare but decisive for credibility. We conclude that governance-ready practice requires tail-aware input fitting, co-movement, variance-efficient designs with accuracy goals, integrated cost, schedule, and finance modeling, and transparent sensitivity and validation protocols that yield auditable percentile contingencies and covenant risk statements for stage-gate decisions.