Why doesn’t capital flow from rich to poor countries? That was the puzzle Bob Lucas raised in his famous 1990 paper. A huge literature is devoted to answering this question. Generations of researchers have invested considerable effort into finding ever more sophisticated explanations for the apparent lack of capital flows from developed to developing countries.
Now a new paper suggests there might not be a puzzle after all. Matteo Maggiori, Brent Neiman, and Jesse Schreger (MNS) have a research project recalculating foreign investment positions from the bottom up – with fascinating results.
Conventional statistics on cross-border capital movements rely on surveys of firms asking them about their holdings of foreign assets. The survey responses are then aggregated into measures of stocks and flows of foreign investment in the context of national income accounts and the balance of payments statistics. Importantly, these statistics are based on the residence of a firm, not the nationality of its owners. That means, for instance, the foreign assets of an Irish affiliate of an U.S. firm are recorded as Irish, not U.S., foreign assets.
MNS, on the other hand, construct nationality-based measures of capital stocks. They start from a dataset covering the universe of publicly traded securities (26 million bonds and stocks) identified by a unique number, the CUSIP code. Using the CUSIP codes, and combining a variety of data sources linking subsidiaries to their owners, they match the securities issued by any firm with its ultimate parent. Summing up over securities by country of their ultimate owner, they end up with measures of foreign assets and liabilities for each country based on the nationality of the ultimate issuer.
Furthermore, MNS are able to convert conventional residency-based capital stocks into nationality-based measures. This results in drastic revisions of bilateral investment positions.
The main reason for these revisions is the role global tax havens. In a nutshell, large firms based in the U.S., U.K., E.U. and other developed countries issue large amounts of securities through their foreign subsidiaries in places like the Cayman Islands, Luxembourg, or Panama. This is mostly to avoid taxes, but also to circumvent financial regulations and capital controls.
For example, whereas the national statistics for 2017 list the United States as holding $160 billion in Chinese equities, we find the position to be worth about $700 billion. These positions are largely associated with Variable Interest Entities (VIEs), structures designed to avoid China’s capital controls that restrict foreign ownership in key industries. We report that U.S. investment into Brazilian corporate bonds equal $50 billion, much larger than the $8 billion position listed in TIC [the US Treasury’s data]. EMU holdings of Russian debt triples from $36 billion in CPIS [the IMF’s data] to $107 billion in our restated tables. Similar patterns are found for U.K. investment in emerging market securities. The value of developed country positions in Bermuda, the Cayman Islands, Ireland, Luxembourg, the Netherlands, and Panama plunge. Our work provides a sizable upward revision in investments by developed countries in large emerging economies such as Brazil, China, India, Russia, and South Africa (the “BRICS” countries).
Maggiori et al (2020), p. 3.
The authors’ baseline estimate “raises investment from the United States to the BRICS in corporate bonds from $19 to $122 billion, a 540 percent increase.”
Here are some other astonishing results:
[F]oreign-currency corporate bonds account for a greater share of portfolio investment from large developed countries to large emerging markets. For some emerging markets, nearly all of the corporate sector’s bond financing from developed market investors is intermediated through subsidiaries in tax havens. Emerging market sovereigns, by contrast, issue externally under their own name. As a result, the standard residency- based datasets overstate the importance of sovereign bonds relative to corporate bonds.
Maggiori et al (2020), p. 3. [emphasis added]
Bottom line: the official residence-based statistics massively understate capital flows from rich to poor countries and significantly understate the share of private securities in these flows. Great stuff.
