Reserve Diversification

Three somewhat wonky and technical thoughts:

1) Do not underestimate Korea just because it does not hold as many reserves as China and Japan.

$200 billion is $200 billion. Moreover, the combined reserves of Taiwan, Korea, Russia and India are larger than China’s reserves: $685 billion v. $610 billion at the end of 2004. Taiwan, Korea, Russia and India added $152 billion to their reserves in 04. That is a bit behind China’s $200 billion increase, but it not a small sum. If all those reserves went into dollars (unrealistic, i know, particularly since some of the increase reflects valuation gains), it would have been enough to fund about 1/4 of the United States’ 2004 current account deficit.

News about what any of these four central banks intend to do is hardly marginal news in my book. Three of the four -- Russia, India and Korea -- have now indicated a desire to either diversify their reserves or to spend their reserves on "infrastructure." Some say Taiwan has made similar noises as well, though Taiwan’s central bank officially denies any such intent.

Right now, I think it is fair to say that their continued willingness to add to their dollar reserves at the same pace as in 2003 and 2004 is somewhat in doubt -- though there ability to accept the consequences of not adding to their reserves and letting their currencies appreciate remains equally open to question.2) Reserve diversification alone cannot protect a central bank’s balance sheet all that much; the overall size of the central banks stock of reserves probably matters more.

Shifting from dollar reserves to euro reserves back in 2002 certainly would have made a lot of sense: the central bank would have seen the value of its assets rise (in dollar terms) as the euro rose against the dollar, and some central banks might now be sitting on fat capital gains. But the past is the past: central banks cannot buy a euro for less than a dollar anymore. Shifting into euros now only protects the central bank against future falls in the dollar against the euro. It won’t necessarily cut the capital losses many Asian central banks face when their currencies are revalued.

Take the following example. Suppose China lets the value of its currency rise by 20%, from 8.28 renminbi to the dollar to 6.62 renminbi to the dollar. (Incidentally, in local currency terms, the value of the renminbi has gone up by 25%, from 12 cents per renminbi to 15 cents per renminbi -- be careful with percentages with large exchange rate moves). If the euro/ dollar exchange stays constant, the renminbi will also go up be 20% against the euro, rising from 10.76 to the euro to 8.61 to the euro. If the central bank has sterilized its reserve accumulation by selling renminbi debt, and has renminbi liabilities offsetting its dollar and euro assets, the capital loss on its dollar and euro reserves will be identical. Now suppose the Euro/ dollar is not constant. Rather suppose the euro falls against the dollar by 20% (from say 1.30 to 1.04) when China revalues the renminbi. The renminbi then goes from 10.78 renminbi to the euro to 6.88 renminbi to the euro -- the renminbi appreciates v. the dollar and the dollar appreciates v the euro, so the renminbi really appreciates v. the euro. The net result: larger capital losses on euro than on dollar reserves. Now suppose the renminbi appreciates against the dollar and the euro also appreciates against the dollar, so euro rises from 1.3 to 1.56. If I did the math right, one euro is worth 10.78 renminbi before the currency move, and 10.33 renminbi after. The net result: smaller capital losses on euro than on dollar reserves, since the local currency value of the euro stays more or less constant.

So diversification may, or may not, help at this stage: it depends on the future course of the dollar/ euro.

The surest way for Asian central banks to limit their future capital losses is simple. They need to stop adding to their stock of reserves. That at least keeps the central bank’s aggregate currency mismatch (local currency liabilities v. foreign currency assets) from growing. If you are in a hole, stop digging! Of course, ending all intervention means letting your currency appreciate, and taking the capital loss on your existing stock of reserves today, rather than taking a potentially larger loss in the future.

3) Just in case there are any aspiring currency strategists out there, one word of advice: official purchases of Treasuries (as reported in the TIC data) are an increasingly bad proxy for overall central bank purchases of dollar assets. Central banks could be buying Treasuries through intermediaries, or just buying a broader range of dollar-denominated assets.

Life was sort of easy back when Japan was adding to its reserves, since a rise in Japan’s reserves was followed, with a bit of a lag, by a surge in official purchases of Treasuries. However, the central banks of emerging Asia have indicated publicly that they intend to invest in a broader set of securities than just Treasuries. It may be that the rising gap between reported official purchase of Treasuries and the reported increase in central bank reserves indicates some diversification into non-dollar assets. But it also could just indicate that central banks are buying a broader range of dollar-denominated assets, and are doing more of their purchases through harder to trace broker-dealers.

To my knowledge, at this stage, we have no real way of knowing whether Asian central banks are purchasing fewer dollar assets (relative to the increase in their reserves), or just purchasing fewer Treasuries. In one scenario, Asian central banks are buying euros and other currencies. In the other scenario, they are just buying a broader range of (higher-yielding) dollar-denominated assets, and doing their purchases in ways that the US data does not pick up as cleanly.