Stephen Jen argues that the US-Asian relationship can be understood as a debt for equity swap. Asian - led by China - buys US debt; the US buys Asian equities/ invests in Asian plant and equipment. That idea stems, I think, from Dooley, Garber and Folkerts-Landau. And as Goldstein and Lardy point out, it has a ton of problems.
It also doesn't describe well the current US relationship with Asia (or the Middle East). Most investment in Asia is financed by Asian savings intermediated by Asian financial institutions, not by swap of US savings for Asian savings.
We don't yet know the size of (net) US purchases of foreign equity in 2005 - but judging from what we do know, the US maybe added $45b in net portfolio equity. The US bought $122 billion of foreign equities over the last 12 months (data is November to November), while foreigners bought $78 billion in US equities. And in the first three quarters, the US firms' direct investment abroad totaled about $20 billion, while foreign direct investment in the US totaled around $90 billion. Add it all up, and the US gets a (tiny) net FDI inflow.
US outward FDI is low because, in the third quarter, the US reduced its "direct investment abroad" by $27 billion. The Homeland Investment Act and all. No matter.
The US didn't, in aggregate, issue a lot of debt to buy a lot of foreign equity. Equity flows were pretty balanced.
There was no debt for equity swap either. There was a massive debt for Asian goods swap, and a massive debt for oil swap.
What does that mean? Rather than paying for its current imports with current exports, the US promised to pay for its current imports with future exports. Looking toward 2006, I suspect the US will need to engage in a $1 trillion debt for goods swap -- at least if oil stays above $65 and the US economy grows a bit faster than it did in the fourth quarter.
What of the US net international investment position. Can we find a debt for equity swap there?
At the end of 2004, the US has $10 trillion in foreign assets, and $12.5 trillion in external liabilities. At the end of 2005, I suspect the US will have close to $10.75 trillion in foreign assets, and liabilities of $13.9-14 trillion. And - for the reasons I outlined above - most of the increase in both US assets and US liabilities will come from good old-fashioned debt.
To make things simple and transparent though, I'll use the 2004 numbers.
$5.8 trillion of the United States $10 trillion of foreign assets was "equity" -- $3.3 trillion of FDI and $2.5 trillion of portfolio equity.
And $4.6 trillion of $12.5 trillion in total foreign claims on the US came were "equity" claims -- $2.7 trillion of FDI, and $1.9 trillion of portfolio equity investments.
So in broad terms, the US external balance sheet consists of:
- A $4.6 trillion equity for equity swap. This swap generates most of the United States dark matter - whether because of superior US know-how or tax arbitrage.
- A $1.2 trillion debt for equity swap. This swap also generates dark matter. It can be divided into a $0.6 trillion debt for FDI swap and a $0.6 trillion debt for portfolio equity swap.
- A $4.2 trillion debt for debt swap. That is mostly bank claims for bank claims. And mostly dollar debt for dollar debt. The US only holds $1.2 trillion in foreign long-term debt securities, and of those, only $0.26 trillion ($260 billion) were denominated in foreign currency.
- And $2.5 trillion in net debt. That is debt not offset by any offsetting external asset. Call it the legacy of a series of debt for goods swaps. It is a claim on future US exports.
The debt for equity swap is the smallest item on the US external balance sheet.
We should make one small adjustment though. The US estimates (officially) that foreigners hold $317 billion in US currency abroad. Some say that is more like $500-600 billion. That is a good trade for the US. The US prints pieces of paper that a Colombian drug lords, Russian Mafiosi and others use in a range of transactions. So long as those pieces of paper circulate - or are held under the mattress - outside the US, the US effectively has financed a portion of its trade deficit by printing money.
Or providing "liquidity services" to the world, in Hausmann and Sturzenegger's terms.
So the US balance sheet includes a $300-600 billion "cash for debt" or "cash for goods" swap. That generates a gain for the US. Indeed, Willem Buiter argues that this swap is the only verifiable source of dark matter on the US balance sheet.
At the end of 2005, the equity for equity swap will be a bit bigger, the debt for equity swap will be about the same size, and debt for debt swap will be a bit bigger (offsetting bank claims) and so on. But the biggest change is that the US net debt will rise by about $700 billion, to around $3.2 trillion. It would have been $800 billion but for some valuation gains. Foreign equities did much better than US equities in 2005.
Rising net debt is a consequence of paying for current imports with IOUs. The IOUs eventually do start to pile up.
Looking ahead, the US net external debt should rise - barring valuation gains - by about $1 trillion a year for the next several years. So it will be $4.2 trillion at the end of 2006, $5.2 trillion at the end of 2007 and so on.
That is what happens is you swap debt for goods. Your debt rises.
OK, it rises in the absence of valuation gains, whether from a fall in the dollar or a rise in the value of foriegn equities. But that is a topic for another day.
And unless the US generates a lot of offsetting dark matter, the rising net debt will generate a rising deficit in the investment income line of the balance payments. More debt equals more interest on the debt.
Actually, the income balance will deteriorate even faster, since the interest rate the US has to pay on its existing debt for equity and debt for goods swap is rising. That pesky Fed.
Some of that interest will, in Hausmann and Sturzenegger's terms, be offset by the existing dark matter on the US external balance sheet. But in order to keep the US income balance positive, the US needs to create a lot of new dark matter.
Or, as Higgins, Klitgaard and Tille argue, generate even more dark anti-matter.
One of the core assumptions that Hausmann and Sturzenegger make is that US net interest payments should be valued at a 5% discount rate. So if the US generated $36 billion or so of net investment income in 2004, the US really has external assets of $720 billion, not $2.5 trillion in external debt.
Higgins, Klitgaard and Tille did a small thought experiment. They decided to value all foreign claims in the US on the basis of the income that they generate, using a 5% discount rate. And all US assets abroad at a 5% discount rate.
What did they find? At end the of 2004, foreigners had only $6.8 trillion in US assets ($340 billion in investment income), not the $12.5 trillion reported in the US investment position. What explains the difference? $5.7 trillion in "dark antimatter."
The same procedure values US assets abroad at $7.5 trillion, not $10 trillion. So there is only $2.5 trillion in "dark antimatter" there.
The key points: Reported returns on US assets and US liabilities are both quite low, in aggregate - well below 5%; and positive US investment income comes not from superior returns on US assets abroad, but rather from very low returns on foreign assets in the US. Antimatter destroys value.
We can debate whether those low returns are real, a figment of transfer pricing and other forms of corporate tax arbitrage, or a mix of both.
But there is no getting around the fact that, judging from the profits that foreign firms report on the income line of the balance of payments, the US has been a terrible place to invest over the past few years.