According to standard government statistics, the U.S. has accumulated a massive debt to the rest of the world during the last three decades – over 5 trillion dollars. No statistical tricks, however clever, are going to erase that debt. Or so I would have thought.

But today the brilliant Ilia Rainer plugged me into the most shocking macroeconomic debating point I’ve heard in years. Decades of trade deficits notwithstanding, Americans’ net return on foreign assets (the income we earn from our foreign assets minus the income foreigners earn from U.S. assets) has been roughly constant at positive $30 B per year! Harvard’s Ricardo Hausmann and Federico Sturzenegger’s have a fascinating research agenda which uses this raw fact to topple the conventional wisdom on trade deficits. Here’s a very readable intro:

The Bureau of Economic Analysis (BEA) indicates that in 1980 the US had about 365 billion dollars of net foreign assets (that is the difference between the foreign assets owned abroad and the local assets owned by foreigners). These assets rendered a net return of about 30 billion dollars. Between 1980 and 2004, the US accumulated a current account deficit of 4.5 trillion dollars. You would expect the net foreign assets of the US to fall by that amount, to say, minus 4.1 trillion. If it paid 5 percent on that debt, the net return on its financial position should have moved from a surplus of 30 billion in 1982 to minus 210 billion dollars a year in 2004. Right? After all, debtors need to service their debt.

So let’s look at how much is the actual return on the US net financial position. The number for 2004 is, yes, you’ve guessed it, still a positive 30 billion, just like in 1982! The US has spent 4.5 trillion dollars more than it has earned (which is what the cumulative current account deficit implies) for free!

The upshot: The dollar is not going to collapse. In fact, it looks like we should expect the dollar to slightly appreciate. As Neo says, “Woh.”

How is it possible for the U.S. to sell assets like crazy but keep the net income flow positive? In a nutshell:

Part of the answer is that the US benefited from about 1.6 trillion dollars of net capital gains so that instead of owing 4.1 trillion, it owes “only” 2.5 trillion (which, at best, cuts the puzzle in half, leaving a whole other half to be explained). The other part of the official answer is that the US earns a higher return on its holdings of foreign assets than it pays to foreigners on its liabilities.

It’s easy to miss the latter effect in a simple macro model that has just one “rate of interest.” In reality, of course, there are many interest rates. The basic pattern in world investment seems to be that Americans invest abroad in risky high-return assets, and foreigners invest here in low-risk, low-return assets.

Hausmann and Sturzenegger go so far as to propose a new accounting system to help make these economic realities more transparent:

We start by assuming that if an asset consistently pays more than another asset, then it is worth more, even if they both have the same historical cost or “book value”. We choose to value the assets on the basis of their returns. This is just like valuing a company by calculating its earnings and multiplying by some price-earnings ratio, or valuing a property based on its rental value.

If we take their approach, you will probably initially feel like you’ve entered Bizarro World:

First and foremost, the US does not appear as a net debtor but as a net creditor and, as mentioned above, its net foreign asset position has remained stable over the last 20 years. Japan, consistent with official data, is a growing creditor, while the European Union and the rest of world are net debtors.

But on reflection, it’s the conventional story – that the United States’ three decades of apparent financial tranquility are just about to end in disaster – that’s bizarre.