The U.S. Income Balance Puzzle

One of the great puzzles of international economics has been the fact that the U.S. has tended to earn more money on its international investments (and lending) than it pays on its external debts. After all, a country that persistently runs trade deficits (as the U.S. does) should accumulate external debt (as the U.S. has) and pay a bit of interest to the world.

But the U.S. long seemed to defy gravity.

The U.S. balance on international investment income (investment income counts both dividend income on equity and interest income form lending) has been positive—and in fact at times significantly positive—despite the persistent U.S. trade deficit. This seems to confirm the United States’ exorbitant privilege—and special position in the international economic system. Ask the French.

Ricardo Hausmann went so far as to argue that there was “dark matter” on the U.S. balance sheet. A debtor should pay interest or dividends to the world, not collect interest and dividends. Hence, he argued that the U.S. must have a “dark” asset that offset its visible external debts in order to keep generating a surplus in investment income.

Others argued that the U.S. was simply a successful private equity fund, borrowing from the rest of the world to invest in higher yielding assets abroad, collecting large risk and liquidity premiums along the way.

This argument is a bit too smug.

The U.S. is at best a very profligate private equity fund; it borrows a lot to fund current spending, and it has a lot more debt than equity assets.

A detailed examination of the U.S. balance of payments acturally suggests that the U.S. doesn’t borrow to invest in foreign equities; equity inflows and equity outflows are generally balanced over time. Nor does it earn especially high returns on its investment in most of Europe, Japan, China, or North America.

The boring reality is that the income differential is now mostly a function of U.S. corporate tax avoidance—it goes away if the profits of U.S. firms that report earnings in the seven most significant low tax jurisdictions are taken out of the income balance.

Foreign firms are also likely shifting profits out of the United States; the reported yield on foreign direct investment (FDI) in the U.S. is suspiciously low.

Perhaps more significantly, even with the exorbitant gains from tax avoidance, the U.S. income balance is about to move into a deficit. Exorbitant privilege will soon disappear.

How to Assess the Income Balance

There are two ways to look at the U.S. income balance.

One approach, used by many distinguished economists, notes that a higher share of U.S. international investment abroad is in equity than in debt, and the average return on U.S. equity investment abroad is higher than the average return on foreign equity investment in the U.S. The U.S. thus looks like it is harvesting risk premium—providing safe assets to the world and buying riskier assets.

Another approach, which I prefer, accounts for the fact that the overall stock of U.S. external liabilities is higher than the stock of U.S. external assets. Thus, while the U.S. has a higher share of its external investments in equities, total foreign equity investment in the U.S. tops total U.S. foreign equity investment abroad.

A slightly more sophisticated version of this calculation looks not just at the current stock of investment, but the flows over time, which removes the impact of valuation changes (it also generates a longer quarterly time series).**

FDI Flows Match

It turns out that, over time, U.S. FDI flows into the world basically match FDI in the U.S.

Thus, the U.S. isn’t really borrowing from the rest of the world to buy up the rest of the world (the argument Giscard d’Estaing made in the 1960s), but rather swapping investment abroad for foreign investment in the U.S. GM invests in China, Toyota invests in the U.S.; Apple invests in Ireland (or sets up a cost share in Ireland that pays 60% of Apple's R&D budget for the rights to profit from Apple's IP outside the Americas), LVMH buys Tiffany, and so on.

Debt Inflows Finance the Current Account Deficit, Not Equity Investment

If like flows are matched against one another, it quickly becomes apparent that the net inflow into U.S. debt securities far exceeds the (modest) net outflow from U.S. portfolio equity investment.

In fact, net debt inflows basically track the cumulative U.S. current account balance.

A bit more digging shows that the U.S. does pay a little less, on net, than expected on its external debt—but that it does in fact make net interest payments to the rest of the world.

A simple model (the average interest rate over the last five years on five-year Treasuries times the net debt stock) does a pretty good job of forecasting net debt payments. There are some intricacies (to be discussed later) but there isn’t any big mystery.

Portfolio equity dividend payments pretty much map as well (and since U.S. firms do return money to foreign shareholders through buybacks, the income line may understate the actual flow of profits to foreign investors here).

The big gap is the much higher return that U.S. FDI earns abroad compared to the return on FDI in the U.S.

To many, this seems logical, since the S&P has outperformed foreign equities. It is, however, worth digging a bit deeper.

Remember that the most important fact about U.S. FDI abroad is that all the reported earnings, more or less, come from small centers of tax avoidance not from big and populous markets. Net of the return reported in the big tax centers, U.S. FDI abroad earns the same low return as foreign investment in the U.S.

That itself is interesting; FDI in the U.S. seems not to be anymore worth the effort of foreign firms than U.S. FDI in France and the like.

But the bottom line here is that the return puzzle—and U.S. exorbitant privilege—cannot really be divorced from the geography of U.S. foreign direct investment. As shown above, the most important driver of the geography of U.S. foreign direct investment is tax, not economic gravity. Half of the value of U.S. FDI and 60 percent of the profits on U.S. FDI abroad come from the seven easily identified low-tax jurisdictions (Zucman has more in his comprehensive data set).

The Debt Side of the Ledger

The U.S. does currently enjoy a slightly higher interest rate on its external lending than it pays on its external debt.

But there are important differences in the composition of the (small) U.S. stock of external lending and the (much bigger) U.S. stock of external borrowing.

Most U.S. external lending (and holdings of foreign debt securities) is in dollars, so that isn’t the difference (the Treasury survey data). But a relatively large share of U.S. external lending is short-term. And the bulk of U.S. holdings of foreign debt securities are corporate bonds (among other things, a U.S. investor buying a collateralized loan obligation is technically generally buying a dollar-denominated debt security issued by the Cayman Islands; that is why U.S. investors hold around $790 billion in bonds from the Caribbean).

Foreign investors in U.S. debt, by contrast, primarily hold U.S. Treasuries (the fact that foreign central banks hold a lot of reserves in U.S. dollars and the U.S. just doesn’t really hold any non-gold reserves matters here).

A bit of netting shows that the U.S. now actually has a bit more short-term borrowing than short-term lending. Net “other” is negative, so on net, the U.S. is borrowing both short- and long-term from the world.  And it so happens that on net, foreign holdings of U.S. corporate bonds are about equal to U.S. holdings of foreign (mostly corporate) bonds. What is left after this is netting is foreign holdings of Treasuries and Agencies, which more or less add up to the U.S. net debt stock and more or less equal the cumulative current account balance.

So, in a simple model, U.S. equity investment abroad is in a sense financed by foreign equity investment in the U.S, and the U.S. external deficit is financed by debt.

The income surplus thus depends on the return differential on equity exceeding the net interest paid on external debt.

And that’s the rub.

The external interest bill has increased because the low-for-long era is over. Net external interest payments are now 1.3 percent of GDP and are likely to rise a bit more (as short-term rates fall, the interest on U.S. external lending will fall a bit relative to the interest rate on U.S. external borrowing; this is very predictable and mechanical).  Best case, the average external interest should converge to 3 percent or so on a stock of around fifty percent of GDP.  Worst case, that number is much closer to 4 percent on a rising stock of debt.

The surplus on FDI income has started to slip.

That surplus is the hardest to forecast, as it depends a bit on the path of the dollar as a strong dollar reduces the value of foreign profits relative to U.S. GDP.

But it also depends on the trajectory of U.S. tax policy.

If the U.S.—as some Republicans have proposed—tries to win the game of tax competition by matching Ireland’s low rate (the U.S. would need a rate below 15 percent until all the depreciation allowances on existing IP expire) the income surplus would shrink as U.S. firms lose the incentive to book their profit abroad. The $100 billion in profit that Apple and Microsoft together now book in Ireland could become $100 billion in royalty exports if both firms check a box and repatriate their offshored IP. Another $70-100 billion could return if the top six U.S. pharmaceutical companies started booking their profits in the U.S. rather than abroad. The big offshore profit in low-tax jurisdictions is very lumpy, and it depends on the decisions of just a few firms.

Of course, the U.S. offshore profit could also fall if the U.S. raises the tax on offshore profits (and makes it hard for U.S. firms to both invert and shift intellectual property abroad)—as some prominent Democrats have suggested.

If there is no tax advantage to being in Ireland, firms aren’t likely to stay in Ireland.

But despite this complexity, the U.S. income balance is ultimately fairly predictable so long as tax policy is stable.

The U.S. net international investment position tends to converge with the sum of U.S. current account deficits over time (as valuation gains and losses even out)—right now it is at a low because the dollar is strong, and U.S. equities are valued highly relative to foreign equities.

The interest rate on that debt tends to map to the rate on Treasuries for deep structural reasons—

And then it is just a matter of whether U.S. firms will continue to generate more income in offshore tax centers than the U.S. pays on its external debt.

There is no dark matter involved, and no dark magic -- though some would describe corporate tax avoidance as something of a dark art.

The low-for-long era coincided with an era of rampart corporate tax avoidance. The low-for-long has (mostly) ended, and even if the era of tax avoidance hasn't come to an end, the end of low-for-long should be enough to end the income surplus, and structurally pull the overall current account deficit to over 4 percent of GDP.

With a rising trade deficit and a structural shift in the income balance underway, there is every reason to expect the U.S. current account deficit to continue to rise absent a large change in the dollar -- or a shock to U.S. demand.

* The current account balance is the sum of the trade balance, the income balance, and the transfers balance. The trade balance can be split into the service and goods balance and the petrol and non-petrol balance (the big swing in U.S. external accounts has come from the petrol balance, not the services balance, but the reduction in the petrol deficit has been offset by the rise in the non-petrol deficit after the dollar’s rise in 2014-2015. The income balance includes both investment income and labor income (Americans working abroad) and the transfers balance includes both U.S. foreign aid and remittances (it generally shows a small deficit).

The NIIP normally follows the balance of payments convention of showing the FDI balance, the portfolio investment balance and the “other” balance (“other” should be renamed bank and other flows to be a bit more intuitive). I prefer to split portfolio investment into portfolio equity (“stocks”) and portfolio debt (“bonds”) and focus on the net equity and the net debt position. Investment income is receipts on equities net of payments, and interest receipts net of interest payments.

** There was once an argument that the U.S. current account deficit was just the U.S. borrowing off the capital gains on its external equity investments. But with the market value of foreign external investment in the U.S. far more than the market value of U.S. equity investments abroad, this argument no longer works.

*** The U.S. external data is generally very good (China take note; the U.S. is winning this race), but Gian Maria Milesi-Ferretti has highlighted one significant flaw in the U.S. data: the BEA values FDI abroad using the performance of the local stock market in the destination country. So, U.S. investment in, say, Ireland is valued using the Irish stock market. In reality, Apple (Ireland) should be valued based on Apple (U.S.), and thus the market value of U.S. FDI abroad should be a bit higher than is reported (reducing the deficit in the NIIP).