Prediction of Cash Flows: Selection of A Sustainable Model

This study addresses the limited evidence on cash flow prediction in Pakistan by examining how aggregate and disaggregated accrual-based models forecast future cash flows of non-financial firms. Using 20 years of firm-level data and panel regression with lag structures, we compare five models under AIC, SIC, SSE, and PIC criteria. The results show that combining aggregated cash flows with disaggregated accruals provides the most accurate three-year forecasts. This finding extends prior work in Pakistan by offering a more sustainable prediction framework. The implications are significant for accounting regulators. It is also suggesting the adoption of enhanced reporting practices under IAS-7 to strengthen transparency and to reduce the earnings management.