We apply panel data analysis, starting from pooled ordinary least squares (OLS) as a principle estimation. This method mixes time series and cross-section together. We apply the fixed-effect model to help mitigate the issue of heterogeneity, and instrumental variables estimators to overcome the causality between saving ratio and domestic investments.
We control for macroeconomic variables in order to make the estimations more robust. For example, the inflation rate is measured by the annual percentage change on consumer price index (CPI) to capture the impact of macroeconomic stability. In addition, we include trade openness as a measure of international trade, and growth rate measured by the change of gross domestic product (GDP) to capture the impact of economy size. Moreover, we compromise the saving ratio as a determinant of domestic investment, following in the model of Feldstien- Horioka.
Aggregate data show that FDI inflows have positive effects on public domestic investments, but generate a negative effect on private domestic investments. In addition, FDI outflows improve private domestic economic activities, but negatively affect public domestic investments. Disaggregate data illustrate that greenfield FDI into the oil sector yields a significant and positive effect on public domestic investment: a one-dollar increase in oil FDI leads to an approximate 0.6 cent improvement in public domestic investments. Non-Oil FDI has an ambiguous effect on domestic investments. All the results are robust.
Keywords: FDI, Public domestic investments, Private Domestic Investments, Panel data, Endogeneity