Fiscal Forecast Errors in Public Health Expenditure
Fiscal Forecast Errors in Public Health Expenditure: Based on a Spatial-Temporal Approach
This research delved into the forecast errors of budget expenditures and their underlying causes within local governments’ social welfare and health sectors in South Korea, utilizing data from 2008 to 2022 to introduce an approach to enhance estimation and prediction accuracy. This study employed the Spatial Integrated Nested Laplace Approximations model to address the shortcomings of previous studies, which did not account for spatial correlations and temporal dynamics in expenditure forecast errors. Furthermore, it leveraged Spatial-Temporal Graph Convolutional Networks to predict regional expenditure errors. The findings indicate that the influence of various factors on forecast errors in both the social welfare and health domains necessitates a consideration of the spatial interactions and temporal changes unique to each local government. It was found that local government volatility in social, healthcare, and socio-economic aspects significantly impacts expenditure forecasts. Specifically, regions experiencing significant fluctuations in the basic national livelihood security rate and health insurance coverage tended to underestimate social welfare budgets, while areas with a high variance in tuberculosis patient numbers were prone to overestimating expenditures in both social welfare and health. This pattern also emerged in areas with notable population mobility or significant regional gross income fluctuations. However, areas with substantial fluctuations in fiscal autonomy were more likely to underestimate health expenditures. The forecasting results, which consider spatial interactions and temporal changes, underscore the necessity of considering regional dynamics. Notably, the social welfare field in Seoul and the metropolitan area is anticipated to experience significant forecast errors, highlighting the need for targeted improvements in social welfare resource allocation and policy adjustments in these regions.