Standard theory tells us that a country with a low capital endowment, relative to its labor endowment, should have high capital returns and thus should be attracting foreign capital until capital returns are equal at home and abroad. While there is foreign direct investment from the North to the South, it is by far as high as it should be, and capital returns are far from being equalized. There are proposed answers to this puzzle, from mismeasurement to country-specific risk, but that does not explain why there would be foreign direct investment from the South to the North.
Carlos Pestana Barros, Bruno Damásio and João Ricardo Faria look at the case of Angola investing substantially in its former colonial master, Portugal. They build a model of a open economy subject to corruption practices. It is not quite clear to me how this model maps into the linear equation that is estimated (partly because not all equations display in the paper). But at this points, the interesting results is that this FDI is driven by exports and mostly by corruption. One has to understand that corruption in Angola is among the world's highest. For example, there is an unexplained residual in the country's fiscal account that corresponds to about a quarter of its GDP, which is absolutely mind boggling. This corruption is so big that not only does it dry out the FDI flow from Portugal, it reverses it.