From Data to Decisions: Role of Machine Learning in Predicting Cash Holdings of Manufacturing Firms in Pakistan

Abstract

Effective cash management is essential to maintain a firm's financial health and sustainability. Hence, this study evaluates the prediction performance of various machine learning (ML) algorithms in identifying firm-specific determinants of cash holdings among the manufacturing firms of an emerging market, namely Pakistan. Using secondary data, the analysis employs ML techniques such as multiple linear regression, LASSO regression, ridge regression, elastic net regression, as well as random forest, gradient boosting, support vector regression, and decision tree models. The findings reveal that random forest and gradient boosting models outperformed others in predicting cash holdings, while the decision tree model exhibited the poorest performance. These insights are valuable for managers and decision-makers in optimizing cash retention, capital allocation, and investment planning. Additionally, policymakers can leverage these findings to develop policies that enhance financial resilience and foster growth in the manufacturing sector of Pakistan.