At the end of 2020, Americans owed $14 trillion more to the rest of the world than the rest of the world owed to America, according to the latest reading on the nation’s Net International Investment Position (NIIP) from the Bureau of Economic Analysis (BEA). That’s 67 percent of U.S. GDP, up from about 50 percent at the end of 2019 and well above the 16 percent of GDP recorded a decade ago.
This blog discusses two questions: First, how did we get here? After all, U.S. current account deficits have been moderate during the past decade. So why has the U.S. net debtor position increased so much? Second, is this increase in net foreign liabilities a cause for concern? (Spoiler: Not much.)
How did we get here?
The NIIP is the difference between U.S. foreign assets and U.S. foreign liabilities. U.S. assets abroad include the value of overseas affiliates of U.S. multinational corporations (foreign direct investment, or FDI), shares of foreign companies held by U.S. residents (portfolio equity), bonds issued by nonresidents (for instance, Mexican government bonds bought by a U.S. pension fund), U.S. banks’ deposits in foreign banks, etc. Symmetrically, U.S. liabilities include the value of U.S.-based affiliates of foreign corporations, shares of U.S. companies held by nonresidents, U.S. Treasury and corporate bonds held by foreign central banks and other foreign investors, as well as foreign deposits in and foreign loans to U.S.-based banks.
When spending by U.S. residents exceeds domestic production of goods and services (the GDP), the U.S. runs a current account deficit, which is financed by net borrowing from abroad. This borrowing can take the form of foreign purchases of U.S. Treasury securities or shares of U.S. firms, foreign direct investment in the U.S., or sales by U.S. residents of assets held abroad. In the same way in which the U.S. budget deficit adds to U.S. public debt, current account deficits add to the U.S. net debtor position. Figure 1 shows the U.S. NIIP as a share of GDP, together with the cumulative value of U.S. net borrowing from overseas during the period (broken line). If external assets and liabilities did not fluctuate in value and everything was measured perfectly, the two lines would coincide. Instead, there are substantial divergences.
During the “global imbalances” period of 1998 to 2008, the U.S. ran very large current account deficits and did massive net borrowing from overseas. By 2010, it had borrowed the equivalent of about 45 percent of GDP from foreigners. However, its net debtor position in that year was much smaller—only 17 percent of GDP. The primary reason for this gap involves changes in the value of U.S. assets and liabilities driven by exchange rates and stock prices. A substantial share of U.S. assets abroad is denominated in foreign currency (think of a Mexican government bond in pesos, or the value of a U.S. affiliate based in Ireland) while almost all its liabilities are denominated in U.S. dollars. Hence when the dollar weakens (which is what happened between 2000 and 2010), the U.S.-dollar value of foreign assets rises correspondingly. Furthermore, during this period stock prices outside the U.S. rose much more rapidly than U.S. stock prices, thus raising the market value of U.S. equity and FDI holdings overseas relative to foreign holdings in the United States.
Since 2010, external borrowing has waned, and, with the exception of 2020, cumulative net borrowing from abroad has roughly stabilized in relation to GDP. But the net international investment position has plunged by some 50 percent of GDP. This time the valuation effects have worked in reverse: the U.S. dollar has strengthened notably since 2010, and U.S. equity prices have risen much more than foreign equity prices. In other words, the value of foreigners’ investments in the U.S. has risen a lot relative to the value of Americans’ investments abroad. And with growing financial integration (a higher share of external assets and liabilities in relation to U.S. GDP), the valuation effects induced by these asset price changes apply to a larger “base” and are hence bigger.
These effects can be better understood looking at Figure 2, which splits the U.S. NIIP into two components: net debt instruments (debt securities plus loans, deposits, etc.) and net equity instruments (the sum of the net position in portfolio equity and FDI). As the issuer of the preeminent reserve, the U.S. has traditionally received substantial net debt inflows (for instance, purchases of Treasury securities by foreign central banks) and has hence accumulated a large net liability position in debt instruments (solid orange line). However, it has been a net “creditor” in terms of equity instruments, for example through investment overseas by U.S. multinational corporations. However, the run-up in U.S. equity prices and the dollar during the past decade has substantially raised the value of foreign equity investment in the U.S. and generated a massive shift in the net equity position (solid blue line), despite broadly balanced equity inflows and outflows (dotted blue line).
Should we worry?
To address this question, it is first useful to compare U.S. net external liabilities in relation to the size of its economy with those of other countries. If we consider the more than 50 countries with GDP above $150 billion in 2020, only four have a net liability position as a share of their GDP larger than the U.S. Those are Ireland (where the NIIP data is heavily biased by the choice of multinationals to domicile their intellectual property there), Greece, Portugal, and Spain. All of them faced a severe debt crisis a decade ago.
Is that a big risk for the U.S.? For several reasons, the answer is no.
The first is that the driving force of the worsening in the U.S. external position in the past decade has been an increase in the market value of U.S.-based wealth. Stronger prospects for U.S. firms are reflected in higher stock prices. So U.S. wealth has increased substantially during this period, benefiting both residents and nonresidents. Should U.S. prospects wane, a correction in the U.S. stock market and a depreciation of the dollar would correspondingly reduce the value of U.S. net external liabilities (and weaker U.S. demand and a weaker dollar would tend to reduce the U.S. current account deficit).
A second argument has to do with measurement. As discussed in a short appendix at the end of this post, the way the market value of FDI claims and liabilities is estimated may exaggerate the deterioration in the U.S. net equity position. An alternative estimate of net FDI, also published by the BEA (the blue line in Figure 3), shows a still-positive net FDI position, some 20 percentage points of GDP stronger than the one implied by the market value estimate (the orange line).
But what about “external sustainability”? In most countries, rising net external liabilities imply a growing debt service burden, which may ultimately require a large exchange rate adjustment and compression of domestic demand to generate the trade surplus needed for such debt service. For countries borrowing in foreign currency, an exchange rate depreciation would increase external debt measured in domestic currency. Indeed, historical evidence suggests that large net external liabilities increase the risk of external crises, especially if they reflect net debt liabilities (Catao and Milesi-Ferretti, 2014).
However, the U.S. borrows exclusively in its own currency so there is no “balance sheet risk” associated with dollar depreciation. On the contrary, a weaker dollar improves the U.S. external position, as discussed earlier. And as the issuer of the pre-eminent reserve currency, there are two favorable factors which temper the need for the U.S. to run a trade surplus:
- The average yield on U.S. external liabilities over the past two decades has been below the growth rate of GDP, reflecting to an important extent the substantial decline in yields on U.S. Treasury and corporate bonds.
- The yield on its assets has consistently exceeded the one on its liabilities—to the point that the U.S. investment income balance remains positive despite its large net debtor position.
While the favorable yield differential between U.S. foreign assets and its liabilities has been the subject of a heated academic debate, the evidence points clearly to the yield differential on FDI between U.S. investment abroad and foreign investment in the U.S. as the most important explanatory factor (Figures 4-5; see also Curcuru, Thomas, and Warnock, 2013). In turn, at least part of this differential reflects strategies of U.S. multinational corporations designed to book profits in low-tax jurisdictions (Guvenen et al, 2017), thus exaggerating U.S. returns on its foreign assets.
However, this additional measurement issue does not affect the assessment of the U.S. external position: reduced investment income would be offset by higher exports. And the U.S. investment income position has been broadly stable during the past decade at about 1 percent of GDP, despite the NIIP deterioration.
Still, there are reasons for some concern as well. Long-term real interest rates could rise, and that has the potential to reduce the favorable yield differential from which the U.S. has benefited. And more importantly, given the speed of the projected U.S. recovery, most forecasters predict a sizable widening of the U.S. current account deficit this year as strong demand fueled by the re-opening of the economy and the very large fiscal stimulus boost imports, and a delayed recovery elsewhere takes some toll on U.S. exports. This widening of the U.S. current account deficit should be temporary—under the assumption that the rest of the world recovers more in 2022, but the starting point is a much larger net debtor position than the one in the late 1990s, which could make “benign neglect” of these developments less likely, including on the political front. The dynamics of the dollar will play an important role here, both through their valuation effects and through their effects on U.S. exports.
Appendix: FDI measurement issues
The BEA estimates FDI positions at market value. However, while portfolio equity investment typically consists of holdings of shares of listed companies, where prices are easily available, FDI generally consists of holdings of unlisted affiliates, whose market value is more difficult to assess. The method applied by the BEA uses the stock market index of the location where the FDI is booked to revalue the equity component of FDI holdings (see users’ guide, particularly pp. 11-9 and 11-10). This implies using the S&P 500 for foreign FDI in the U.S. and foreign stock market indices for U.S. FDI abroad. Given the much more rapid increase of U.S. stock prices in recent years when compared to stock market indices elsewhere (44 percent between end-2017 and end-2020, vs. 6 percent for a stock market index for the ROW), the value of FDI in the U.S. has risen much more rapidly than U.S. FDI abroad.
However, taking an example for 2020, it is not clear why the stock price boom of major U.S. tech companies should substantially raise the value of, say, Deutsche Bank or Toyota affiliates in the U.S. Indeed, U.S. multinationals have a substantial presence in the S&P 500: over 40 percent of their sales represent goods and services produced and sold abroad. This suggests that the valuation of the production facilities of U.S. firms abroad may be reflected in U.S. stock prices as much as in foreign stock prices. Furthermore, the value of U.S. FDI abroad is difficult to assess on the basis of stock prices in the location of affiliates, given that the relevant economic activity often does not take place in the first FDI destination, which is often a low-tax jurisdiction where the affiliate is a holding company.
The BEA also provides alternative estimates of FDI claims and liabilities. The one depicted in Figure 3 (FDI at current cost), attempts to capture the replacement value of capital in foreign affiliates evaluated at today’s prices, without adjusting for stock market developments. It suggests a more modest deterioration in the U.S. net FDI position.
Acknowledgements and disclosures
The author thanks Manuel Alcala Kovalski for help with figures and Lorena Hernandez Barcena for fact-checking.