Forecasting GDP Per Capita in the USA: Integrating Econometric and Machine Learning Approaches for Policy Insights

This study integrates traditional econometric and advanced machine learning techniques to forecast GDP per capita. GDP, a critical indicator of economic health, reflects the monetary value of goods and services produced within a nation. Using data from 1960–2020, this study examines key macroeconomic variables such as Foreign Direct Investment (FDI) inflows, trade ratios, inflation, and Gross National Product (GNP). Ordinary Least Squares (OLS) regression was employed to quantify the relationships between these variables and GDP per capita. ARIMA and Long Short-Term Memory (LSTM) models were utilized for time-series forecasting, with an ensemble approach combining their outputs to enhance prediction accuracy. Results reveal FDI inflows and trade ratios as key drivers of GDP growth, while inflation negatively impacts economic output. The ensemble model demonstrated superior accuracy compared to individual models. This study offers actionable insights for policymakers to design strategies promoting trade, investment, and inflation control, fostering sustainable economic growth.