Forecasting University Admission Rates in China's Gaokao Using ARIMA and ARIMAX Models

Abstract

This study employs ARIMA, ARIMAX, VAR, and GLM models to forecast China's Gaokao admission rates using annual data from 1977 to 2024. The analysis focuses on the admission rate as the dependent variable, with GDP, newborn population, and policy dummies serving as exogenous factors. The results demonstrate that the ARIMAX model, which incorporates economic and demographic variables alongside PCA-based dimensionality reduction, outperforms the other models in terms of both accuracy and interpretability. Predictions indicate an admission rate of 78.47% for 2025 and 83.55% for 2029. The findings emphasize the crucial role of external factors in forecasting and offer valuable methodological and policy insights for higher education planning.