Early Warning of Development-Project Cost Escalation: Comparing Static, Trend-Based and Hybrid Alerts in Simulation

Project expenditure dashboards frequently display budget deviations without demonstrating whether an alert rule can detect a developing problem early enough for intervention. This simulation creates 240 projects, each with 24 monthly planned and actual expenditure values. Seventy-five projects contain a known four-month cost-escalation process; the remaining 165 do not. Three decision rules are evaluated within a fixed seven-month alert window: a static 15% monthly variance threshold, a three-month increase trend, and a hybrid trigger.The static rule detects 67 of 75 risk projects, with no false alerts among non-risk projects and median warning lead time of zero months. The trend rule detects all 75 risk projects two months ahead on median, but produces eight false project alerts. The hybrid rule detects all 75 risk projects one month ahead on median without false alerts in this particular seeded simulation.The findings illustrate the potential benefits and trade-offs of contextual alerting. They do not establish proven effectiveness in live government projects, where operational validation and institutional safeguards would be required.