Forecasting global FDI: a panel data approach

The future patterns of foreign direct investment (FDI) are important inputs for policymakers, even more so during severe economic downturns, such as the one caused by the COVID-19 pandemic. Yet, there is neither empirical consensus nor significant ongoing empirical research on the most appropriate tool for forecasting FDI inflows. This paper aims to fill this gap by proposing an approach to forecasting global FDI inflows based on panel econometric techniques--namely the generalized method of moments--accounting for the heterogeneous nature of FDI across countries and for FDI dependence across time. The empirical comparison with alternative time-series methods confirms the greater predictive power of the proposed approach.