The fastest growing US export: Debt, without a doubt.
Exports of long-term debt securities have grown from $404.5 billion in 2001 to $890 billion in 2004. That is an impressive 120% increase over four years.
Exports of goods and services rose from $1007 to $1146 billion. That is a much less impressive 14% increase.
And lest anyone think reserve financing from Asian (and other) central banks is a US birthright, it is worth remembering reserve financing on the current scale is a rather recent phenomenon. Back in 2001, the US only sold $3.5 billion in long-term Treasuries to the world’s central banks (and only $18.5 billion Treasuries to foreigners); a far cry from the $203 billion in recorded central bank purchases of Treasuries. A decent chunk of the $357 billion in Treasuries sold abroad were also bought by central banks, only indirectly, and central banks have also been buying some of the $232.5 billion of Agencies the US exported in 2004.
The most recent edition of the Treasury bulletin lets us analyze the external market for US debt securities in a bit more depth, since it provides geographic breakdown of the market for US debt.
What does this tell us?
(Warning: the following material scores high on the wonk quotient. All data comes tables CM-V-1 and CM-V-3 of the Treasury Bulletin. That alone should scare most people off!)
Some analysts, including Joseph Quinlan of Banc of America (if you subscribe, check out his comments in Barron’s), have argued that growing interest by Mexico and Canada in US securities indicates all is well. After all, securities (both debt and equity) purchased by Mexico and Canada surged in 2004, reaching $56 billion (up from $38.5 billion in 03), and indeed exceeded debt and equity securities purchased by China ($47 billion) and the eurozone countries ($54 billion).
That, to my mind, is misreading the data. Canadians purchases about as many US securities in 2004 as they did in 2003 (though the bough more debt and less equity in 04). Mexicans did buy more US assets in 2004, but oil prices were also up. Above all, though, some financing from Canada and Mexico should be expected. Both run current account surpluses, and the global current account only balances is all the countries running current account surpluses lend most of their surplus back to the US. Australia and the UK and Eastern Europe also run current account deficits, but the Eastern European deficit is clearly financed by the rest of Europe (the convergence trade is Europe’s carry trade), and the Australian and British deficits pale in comparison to the US deficit (see this table).
The bigger mistake, though, is arguing that Mexico and Canada are more important to financing of the US deficit than China. Anyone who thinks China only bought $47 billion in US securities in 2004 is smoking something. China’s overall reserves rose by $200 billion. Did they really put $150 billion into the world’s banking system, or into euro denominated securities? Clearly not. If so, the BIS data would show a bigger surge in Chinese deposits in foreign banks (up around $40 billion through the third quarter, and that includes dollar deposits abroad by private Chinese citizens), and the Eurozone would either have to be running a current account deficit (financed by China), or be buying far more US debt that the data indicates. Excluding the UK and Norway, Europe bought $90 billion of US debt in 2004 - that number is not directly comparable to Quilan’s $54 billion, as he includes equities as well and excludes countries like Switzerland that do not belong to the eurozone.
UK purchases of US debt surged in 2004, and reached $231 billion (up $63.5 billion from 2003). It is likely that some "UK" buying reflects indirect buying by China, and others. It certainly cannot reflect net flows of that magnitude from the UK itself. Remember, Britain, like the US, runs a current account deficit, so it is drawing on the rest of the world for financing, not financing the rest of the world.
What else, other than the surge in purchase of US securities through the UK, jumped out at me:
1) While net purchases of US securities ($14 billion debt, $20 billion including stocks) by "oil exporters" in the gulf surged in 2004, their total purchases clearly lagged well behind the growth in their oil revenues. Either these countries are buying US securities through the UK, or they are investing in Asia and Europe (who, in turn, have more funds to lend, in aggregate, back to the US). The same point is broadly true of Russia. It bought $14.7 billion in dollar debt, mostly agencies in 2004; its overall reserves rose by $43.8 billion.
2) The US received smaller inflows from the Caribbean than I would have expected. Hedge fund activity in the Caribbean does not provide true external financing -- a security purchased through the Caymans is offset by a US bank loan to the Caymans, so there is no net financing for the US. It is not how the US finances its current account deficit. But I still would have expected more than $22.6 billion in net purchases of Treasuries through the Caribbean (up from $8.7 billion in 2003). Caribbean-based hedge funds actually appear to have been heavy sellers of US Treasuries in the fourth quarter; their holdings fell by $29.8 billion in q4). All told, the Caribbean bough $107.5 billion of US debt in 2004, most of which presumably was financed by loans from US banks and non-banks.
3) If you net out the UK (and oil exporter Norway), European investment in US debt securities remains relatively modest - only $90 billion. $90 billion is nothing to sneeze at. It is broadly consistent with the US bilateral deficit with Europe/ Europe’s global current account surplus. But it is not an enormous sum either. Of course, some flows through the UK may reflect flows that originate elsewhere in Europe ... not just flows that originate in the Middle East or Asia.
4) Asia remains the most important source of financing for the US, by far. Asia purchased $369 billion of US debt in 2004 (up from $295 billion in 2003). $241 billion of that came from Japan, and only $128 billion from China and the Middle East and the rest of Asia. Clearly, someone somewhere in Asia ultimately owns some of the $231 billion purchased through London. Presumably close analysis of the UK’s balance of payments data would reveal who is funding the UK (and UK based financial intermediaries), and thus who is funding the purchase of the $231 billion in US securities bought by "investors" resident in the UK.
Remember, Emerging Asia (East Asia and India) and Japan’s reserves increased by about $535 billion in 2004. The Middle East is sitting on a substantial oil windfall; the increase in the (public and private) assets of the petro-sheiks does not figure in the $530 billion. Even if both East Asia and the Middle East built up their bank accounts, I suspect they bought more than $369 billion of US debt securities.
4) Total Japanese purchases of US debt securities ($241 billion) exceeded the growth in Japanese reserves ($177 billion). The Bank of Japan did not draw down its bank accounts between the end of 2003 and the end of 2004 (its dollar deposits rose sharply in the first quarter, but then fell back), so clearly there is some private Japanese financing as well. In Q4, Japanese investors -- presumably private investors, since the MOF/ bank of Japan were out of the market -- bought $38 billion of US securities, mostly Agencies.
An aside: General Glut has previously has argued that Nouriel Roubini and I underplay the role of Japan in our analysis of Bretton Woods 2. There is a grain of truth there -- I do think that China, far more than Japan, is the anchor of Bretton Woods 2. Why? Because I suspect that Japan will continue to lend its current account surplus back to the US, in one way or another. If private Japanese investors are not willing to finance the US, the MOF will step back in. Japan doesn’t face the same problems sterilizing its growing reserves that other Asian economies face, so that is less of a constraint. Moreover, suppose Japan started investing more in China and less in the US. What would happen? From the US point of view, not very much. Higher capital inflows from Japan would finance more rapid growth of Chinese reserves, and thus more Chinese financing of the US! China, unlike Japan, is doing far more than just lending its current account surplus back to the US. It also takes the money the rest of the world invests in China and lends that back to the US. That is why Nouriel and I talk of the People’s Bank of China’s role as a global financial intermediary, and wonder how long China will be willing to transform the world’s interest in Chinese assets into demand for low-yielding US debt.
Remember that before the 97-98 Asian crisis, Japan still had a current account surplus, which was used, in broad terms, to finance current account deficits in the rest of Asia. The big swing since then does not stem from Japan, but rather from the shift of all of emerging Asia into a substantial current account surplus.
5) And who knew, the biggest purchaser of US Treasuries in Europe, after the UK ($80.5 billion), is Norway ($12.3 billion). Putting their petrokrone to work! The Norwegians bought a fair number of Treasuries in 2003 too. You would think they might be able to find a better place to park their funds. It makes you wonder what other oil exporters are doing with their surplus funds.