Integrating Predictive Credit Risk and Liquidity Analytics into Strategic Portfolio Project Management for Volatile Financial Markets

• Abbas Ahsun

  Texas University

  Author

The increasing frequency of market disruptions and economic volatility has exposed critical weaknesses in traditional project portfolio risk management approaches, particularly their inability to anticipate credit deterioration and liquidity constraints before they materialize. This study addresses the research gap in predictive financial analytics for portfolio project management by developing and validating an integrated framework that combines machine learning-based credit risk prediction with dynamic liquidity forecasting. Using a retrospective dataset of 121,856 loan records and 40 financial indicators, we implemented a hybrid modeling pipeline employing LightGBM and CatBoost classifiers, achieving a predictive accuracy of 89.4% for credit default identification with an AUC-ROC of 0.786. The framework demonstrated a 22% reduction in forecasted budget overruns through lead-time advantages of 45–60 days over conventional static methods. Results indicate that integrating leakage-aware machine learning pipelines with portfolio project management significantly enhances early warning capabilities and enables proactive resource reallocation. This research contributes a replicable analytical framework for financial institutions seeking to embed predictive intelligence into strategic portfolio decision-making processes, with practical implications for risk officers, portfolio managers, and regulatory compliance functions.

• PDF

Copyright (c) 2026 Abbas Ahsun (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.