What counts as foreign direct investment

Foreign direct investment is one of the most cited numbers in economic diplomacy and one of the least understood. It is not a measure of factories built, jobs created or technology brought in. It is a financing category in the balance of payments, defined by a single feature: an investor in one country has acquired a lasting interest in an enterprise in another, large enough to give a say in how it is run.

That threshold is a bright line set by international convention, not a judgement about influence in any particular case. Above it, the money is direct investment. Below it, the identical dollars buying the identical shares are portfolio investment and are recorded elsewhere. A pension fund that quietly accumulates a large economic stake without crossing the voting threshold does not appear in the direct investment figure; a small company that buys outright control of a workshop does.

Three things that all count as the same word

Direct investment arrives in three forms, and they behave nothing alike. The first is equity: cash paid to acquire or expand a stake, which includes buying an existing company outright and building something new. The second is reinvested earnings, which is profit an existing foreign-owned affiliate made locally and did not send home. No money crosses a border at all, yet it is recorded as an inflow. The third is intercompany debt, a loan between parent and affiliate, which can reverse next quarter.

This matters because a headline inflow can be dominated by earnings that were already in the country and by loans that are a treasury operation. New buildings and new payrolls are only one of the three doors, and often not the widest.

An announcement is not a flow

When a foreign company announces a plant, the figure it names is a plan, sometimes covering a decade, sometimes conditional on permits, subsidies or offtake contracts that do not exist yet. The balance of payments records money as it moves. The two numbers refer to different objects and are frequently reported as though they were the same one, which is how a single project can be counted several times over: once when announced, once when revised, and again as it is actually spent.

The chart above shows why that distinction bites. Construction put in place counts the work, whoever financed it. A foreign-funded plant is inside that total, indistinguishable from a domestically financed one next to it. If you want to know how much foreign capital is in the ground, an activity series will never tell you; only the financial accounts will, and they are compiled from company reporting rather than from the site.

Whose money it is said to be

Flows are normally recorded against the country of the immediate counterparty. An investment routed through a holding company in a third country carries that country's label, so the map of where capital comes from is partly a map of where corporate structures sit. Statistical agencies publish ultimate-ownership estimates precisely because the immediate-counterparty numbers mislead. This is also why screening regimes look through the chain of ownership rather than at the flag on the wire transfer.

Economists disagree, seriously and in public, about what direct investment does once it lands. One view holds that foreign ownership brings management practice, supplier standards and technology that diffuse into the wider economy. Another finds the spillovers small and concentrated, and argues that screening and conditions on entry can deter capital that would have been useful. The evidence is mixed because the effects depend on the sector, on the capability of local suppliers and on what the investor was buying: a market, a workforce, or a resource.

Next time an investment figure is put in front of you, put three questions to it. Is this an announcement or a recorded flow? Is it gross, or net of earnings repatriated and loans repaid in the same period? And is the country attached to it the immediate counterparty or the ultimate owner? A number that survives all three is worth arguing about. Most do not get past the first.