Prediction of US GDP Growth Rate Based on ARIMA and ETS

Abstract

Based on the demand for predicting the quarterly year-on-year growth rate of the US GDP, this paper conducts time series prediction research using the ARIMA model and the ETS model. The background of research stems from the fact that GDP, as a core indicator for measuring a country's economic strength, its accurate prediction is of vital importance to policymaking and market decision-making. This paper captures the characteristics of data autocorrelation and differential stabilization by constructing the ARIMA model and combines the dynamic modeling ability of the ETS model for horizontal, trend and seasonal components to compare and analyze the performance of the two methods in the prediction of GDP growth rate. The research results show that both models can effectively depict the dynamic characteristics of GDP growth. However, the ARIMA model demonstrates higher prediction accuracy on MSE and BIC by effectively identifying the data structure. The significance of this study lies in providing methodological references for macroeconomic forecasting, verifying the complementarity of the two classical models in GDP growth rate prediction, and contributing to enhancing the robustness of economic forecasting and the efficiency of policy response.