Japanese reserve management v. Chinese reserve management; transparent v. not.

Billmon has an interesting post up that lays out his take on the current state of the economy. It includes an interesting analysis of how a slowing US will interact with the new Bretton Woods 2 system of reserve financing. His bottom line: it increases the probability that Asian central banks will keep on financing the US, though he expresses that view in considerable more colorful language.

I want to take issue, though, with one specific point Billmon raises in his post, namely that Japanese purchases of US securities during the course of 2004 constitute potential evidence for the reemergence of the yen carry trade.

I’ve tried to get a sense of how important the yen carry trade already is for the U.S. balance of payments, which isn’t easy from the sources I have available. The Commerce Department doesn’t break its international transactions data down enough to tell what’s going on. However, I see from the Treasury’s monthly capital reports that Japanese purchases of Treasuries and agency securities totaled over $200 billion in 2004 -- even though the Bank of Japan itself withdrew from active dollar support operations (and thus stopped adding to its U.S. bond portfolio) in March of last year.

I certainly think it is important to watch for signs of the reemergence of the yen carry trade (yen carry trade =borrowing in yen to invest in dollars since US interest rates exceed Japanese interest rates). But I don’t think the 2004 data provides the needed evidence.

Here is why:

Japan stopped intervening in the foreign exchange market in March. But it took some time for the Bank of Japan to invest all the dollars it bought (on behalf of the Ministry of Finance) into US markets. This shows up clearly in the data that the Japanese authorities make available to the world on the IMF web page.

At the end of March 2004, Japan had $625.8 billion in "securities" and $180.2 billion in various bank accounts. By the end of September, Japan had $688.6 billion of securities and only $122.6 billion in the bank. That is more or less how it ended the year, with $699.4 billion of "securities" and $124.9 billion in the bank. (This link provides the raw data)

In broad terms, Japan started 2004 with $127 billion in the bank (up from $63.7 b at the end of 2003) and $526 b of securities, and it ended the year with $125 b in the bank and $699.5 b in securities. But it got from $127b to $125b via $180b ... It built up its bank account while it was intervening massively, and then slowing converted dollars on deposit into Treasuries.

The most likely explanation for the increase in the value of Japan’s securities in Q4? Simple: changes in the dollar/ euro. Japan holds most of its reserves in dollars, but even 10-15% of $800 b ($700 b in securities) is a big number. $70b of euro securities at the end of September 2004 would be worth around $77.5 billion at the end of December, $100 b of euro securities at the end of September would be worth $109 billion.

An interesting aside, at least for true reserve geeks: Japan’s reported reserves fell by $3.5 billion in January (the euro also fell by about 3.5% against the dollar, suggesting Japan has around $100 billion in non-dollar -- mostly euro -- reserves). US reserves also fell in January, after rising from September to December. Why the correlation: probably because the US holds about 1/2 the amount of euro and euro linked reserves as Japan (note: some US reserves are in SDR’s, the IMF’s internal currency, and the euro is a big part of the SDR).

Note that the government of Japan’s holdings of securities increased by $173.4 billion in 2004. The Japanese don’t publish a currency breakdown of their holdings. But their activity does show up in the US data. However, the US only tells how many securities Japanese investors -- private as well as public -- have bought, not how many US securities the Bank of Japan has bought.

But if you put the two data sources together, is not too hard to figure out what happened. According to the US data, in 2004, Japan bought $166.4 billion of Treasuries. Only $8.7 billion of those were bought in the fourth quarter, when we know the Bank of Japan was (mostly) on the sidelines. That $166.4 b matches up pretty well with the increase in the BOJ’s security holdings.

During the course of 2004, Japanese investors also bought $44.4 billion of agency debt, and $30.4 b of corporate bonds. $22.1 billion of those Agencies and $7.4 b of corporate bonds were bought in the fourth quarter - they certainly were not bought by the BOJ. Indeed, it is likely that all of the corporate bonds and all but a tiny fraction of the agencies were purchased by private Japanese investors.

All told, Japan bought about $242 b of US securities in 2004. Since we know from the Japanese data that the BOJ bought only $173.4 b of securities in 2004, including euro denominated securities, clearly, private Japanese investors also bought US debt. Indeed, they necessarily bought at least $70 b of US debt, and perhaps a bit more, since the BOJ may have bought a few euro denominated securities.

Still, if you add everything up, it is pretty clear that the government of Japan was by far the biggest Japanese buyer of US securities, particularly Treasury securities. Conversely, it is pretty clear that the Bank of Japan probably accounted for well over half, and perhaps almost 3/4s, of the $203 billion in recorded central bank purchases of US Treasury notes and bonds in 2004.

Let’s quickly turn to China.

China’s reserves increased by $207 b in 2004, more than Japan’s.

China doesn’t publish a breakdown between its holdings of securities and its bank deposits, at least to my knowledge. It certainly does not make such information available on the IMF’s web page.

Can the US data help us out? A bit. The US data indicates China bought $47.6 billion in long-term US debt ($22.8 b in the fourth quarter, when China’s reserves soared), and an additional $17.2 billion in short-term bills -- a total of $64.8 billion.

The Chinese bought a relatively diverse portfolio of US debt: $18.9 b of Treasuries, $16.4 b of Agencies, and $12.3 b of corporate debt.

But the US data raises more questions that it answers.

$10-11 billion of the $207 billion overall increase in China’s reserves comes from valuation gains on China’s estimated euro holdings (see this IMF paper). That still leaves roughly $196 billion in new reserves, reserves that needed to be invested somewhere. However, the US recording system only identified $ 65 b in net Chinese purchases.

What did China do with the remaining $131 billion?

I don’t think it all was invested in euros, pounds, yen and Australian dollars, in World Bank bonds, or dollar bonds issued by Asian countries. Among other things, it is hard to make the global current account add up if China doesn’t invest a large fraction of its reserves in the US. The US needs the inflows to finance its current account deficit, the Eurozone doesn’t.

I suspect some of it was invested in the US, but China invested through third parties (private brokers) in London or Singapore or somewhere else. Or China could have just bought US securities already held abroad, say in Europe, from their previous holder. The US data doesn’t track the ultimate owner of US debt, it just tracks the sale of US securities by US residents to foreigners.

Clearly, the big banks in London are doing a lot of buying for someone; the UK (or entities based in the UK) bought $232 billion of long-term US debt in 2004, including $80.5 b of US Treasuries ...

For reserve accumulation geeks, and for those who think all this has an impact on the US fixed income market, the difficulties tracking what China is doing is important. Japan is now out of the reserve accumulation game -- as MTC noted in the comments, the Japanese have convinced the markets that they will keep the yen from rising too much, and consequently, convinced private market participants that it is ok to buy dollars.

China, though, is NOT out of the reserve accumulation game. Indeed, China will probably add more reserves to its stockpile in 2005 than it did in 2004. It is just a lot harder to track what China is doing with its money ... more soon