Missing assets: Exploring the source of data gaps in global cross-border holdings of portfolio equity

Shutterstock / NicoElNino

Gian Maria Milesi-Ferretti, a senior fellow in the Hutchins Center on Fiscal & Monetary Policy and a former deputy director in the research department of the International Monetary Fund, has long studied the flow of money across national borders and the relationship between countries that borrow a lot from other countries and those that lend or invest a lot in other countries. His latest work examines the large gap between the total amount of reported cross-border liabilities (essentially sums that residents of one country owe residents of others) and cross-border assets (investments, such as stocks and bonds, that residents of one country have in others). Here is a Q&A with him about this project.

What prompted you to investigate this difference between total reported cross-border liabilities and cross-border assets?

I have been interested in this for almost two decades, when Philip Lane (now chief economist at the European Central Bank) and I first noticed that investor country reports on the location of their portfolio equity holdings abroad (stocks and shares of investment funds) left a large fraction of holdings of Irish and Luxembourg investment fund shares unaccounted for. Most recently I focused on the role of booming U.S. stock prices in increasing the value of foreign investor claims on the United States, and realized that some partner countries’ reported holdings of U.S. shares fall well short of what the U.S. reports is being held by such countries. I therefore took a more comprehensive look at this mismatch between reported equity assets and liabilities, and found that at the end of 2021 it was close to $4 trillion.

What difference does it make that, when we add up all the available numbers, there is a large and persistent discrepancy between the world’s total financial assets and liabilities (which, by definition, have to match)? Would policy be different if we had complete data? Would imposing sanctions be easier?

The discrepancy I highlight reflects an under-reporting of the shareholdings that investors have in other countries. This will lead to an understatement of wealth held by some countries’ residents. While it is unlikely to affect the conduct of macroeconomic policy, this understatement matters for measuring private sector wealth, for its taxation, and for measuring financial linkages across countries and markets. For instance, how are residents of one country (corporations and households) affected by changes in asset prices in a different country? Addressing it would require progress in exchange of information between different countries—particularly coming from those that are financial centers acting as custodians for financial instruments held by nonresidents. Among other benefits, having a better sense of who are the holders of financial instruments held through intermediaries in financial centers would definitely help target sanctions.

How did you go about filling the gaps in the data? Tell me about the detective work.

Once the global discrepancy was identified, my first step was to compare what countries investing in equity instruments abroad report to be holding in specific destination countries with the total equity liabilities reported by such destination countries. This step allowed me to identify which countries report higher liabilities than other countries claim to be investing there. Once these destination countries were identified, I relied on their own statistics on foreign holders of their shares to establish which partner countries reported lower holdings compared to those reported by the destination countries.

What countries account for most of the gap between actual liabilities and those derived from investor country reports?

The U.S., Ireland, and Luxembourg all report more of their shares are owned by foreigners than the total sum that foreign investors report holding. The U.S. is the largest economy in the world, the U.S. dollar is the preeminent world reserve currency, and the U.S. stock market accounts for a sizable fraction of world stock market capitalization. Hence, to the extent that there are “missing portfolio equity assets,” it is likely that a share will be in the U.S. Ireland and Luxembourg are the largest centers outside the U.S. for the investment fund industry. International investors hold fund shares issued in these centers (some $10 trillion at the end of 2023) that are then invested around the world by these funds. What is remarkable is that for the three of them, the lion’s share of the gap is explained by liabilities vis-à-vis the United Kingdom, with a smaller share related to Switzerland. For example, Irish investment fund statistics indicate that as of end-2022, $1.7 trillion of their shares had immediate counterparties in the United Kingdom, but U.K. statistics report only $400 billion of holdings of Irish shares. Both the United Kingdom and Switzerland are large international financial centers—they manage substantial assets on behalf of international investors. It is therefore likely that most of the unrecorded assets reflect holdings by international investors through these financial centers. These are difficult to capture for statisticians of the investor home country since they are managed and held in custody overseas.

Did anything in this research surprise you?

The size of the discrepancy vis-à-vis the United Kingdom for the U.S., Ireland, and Luxembourg is really striking. It is very unfortunate that in a globalized financial industry, asset holdings held in custody overseas (especially by individual investors) are so difficult to attribute to their country of origin. Financial centers report claims and liabilities of their own residents vis-à-vis foreign investors, but if foreign investors hold financial instruments issued by other foreign jurisdictions, even if managed in the financial centers, these are not accurately reported. If the foreign investor is a foreign institution, their holdings may be reported to its national statistical authorities by the parent company, but if the investor is a wealthy individual, the holdings will be missed by statisticians, even if the individual in question reports it to her/his tax authorities.

Read the full paper here.

  • Acknowledgements and disclosures

    The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.