ARIMAX 모형을 적용한 건강보험지출 장기전망

초록

This study is aimed at applying the ARIMAX model—a time-series forecasting approach that takes into account relevant independent variables—to overcome some of the limitations of such traditional approaches as component method and driver-based simulation in forecasting National Health Insurance expenditure. The versions of the ARIMAX approach that we used in this study include first-order autoregressive model, multiple regression model, log ARIMAX model, and autoreg model. The results were that the projected expenditures differed widely depending on the forecasting versions used. For example, the National Health Insurance expenditure in 2050 was projected to be anywhere between KRW430.42 trillion and KWR561.37 trillion. Such differences are traceable in large part to the differences in the assumptions each version incorporates. This study presents the possibility of the future application of an extended multivariate ARIMAX model for long-term projection of National Health Insurance expenditure.

키워드