Estimating round-tripping FDI from firm-level data in China

When capital leaves a country and then flows back as foreign direct investment (FDI), we call it round-tripping FDI. It is widely suspected that China's official FDI statistics contain a substantial amount of round-tripping FDI. However, it is difficult to quantify the round-tripping FDI due to the lack of data. In this paper, we propose two methods to identify round-tripping FDI. The first one tracks capital flows at the firm level. If a firm in China invests in a foreign firm and this foreign firm makes an investment back to China shortly after, then we consider this investment to China as round-tripping FDI. Our second measure of round-tripping FDI adds to the first measure by including investments in China made by Chinese investors registered in tax havens. The first estimate of round-tripping FDI accounts for up to 3% of China's total FDI from 1999 to 2015, whereas the second estimate accounts for up to 70% in the period. Our firm-level analysis shows that industrial firms facing higher tax burdens are more likely to make round-tripping FDI. We also show that at the city level, adjusted FDI statistics by subtracting the estimated round-tripping FDI are better predictors of imports and exports. Finally, we show that provinces receiving higher shares of round-tripping FDI are more likely to be punished for illegal financial activities. Taken together, these findings suggest that our measures of round-tripping FDI, although noisy, are indicative of real transactions.