Development of a Funding Allocation Formula for Dropout-Related Educational Expenditures within Local Education Finance Grants: Utilizing Dropout Prediction Data

Persistent concerns surrounding student dropout and the allocation of local education finance grants have highlighted the need for more equitable, data-informed funding mechanisms. Although dropout prediction models have been developed, prior studies have primarily focused on identifying risk factors or improving predictive accuracy, with limited attention to connecting prediction outputs to fiscal decision-making. Current education welfare budgets in Korea are also heavily influenced by administrative discretion rather than data-driven needs. In contrast, the state of Nevada in the United States allocates student education funding using dropout prediction data. Against this backdrop, this study constructs a hypothetical formula that integrates predicted dropout data into the calculation and distribution of “dropout-related education funds” within the local education finance grant system. The proposed formula incorporates both the predicted number and proportion of dropout students; proportions are transformed into weights using a sigmoid function to mitigate the cliff effects observed in threshold-based allocations. By adjusting dropout scale weights and the quantile-based center of the sigmoid function, horizontal equity was evaluated using the Gini coefficient and the inverse McLoone Index, allowing for the identification of suitable parameter pairs. Simulation results showed that, at the provincial office level, the largest allocations would be made to Gyeonggi, Seoul, and Busan, while at the school level, Sejong, Jeju, and Daejeon exhibited higher average allocations. A permutation test comparing the proposed dropout-linked allocations with existing education welfare budgets revealed no statistically significant difference between the two distributions. School-level allocations were strongly correlated with the predicted number of dropout students but weakly correlated with predicted dropout rates. Overall, the proposed formula offers a novel approach to enhancing equity and transparency in dropout-related funding by directly linking educational data to fiscal policy and by reducing discontinuities in allocation. However, further work is required to determine total budget size, design implementation programs, and address sample-related limitations, underscoring the need for comprehensive administrative datasets and institutional refinement in future research.