What to Expect from the COFER data

The IMF provides the most authoritative data on the currency composition of reserves. (Think of it as the Ron Swanson of currency composition data) It is reported at the end of the quarter for the proceeding quarter. Tomorrow it is expected to report reserve holdings as of the end of Q1 10.

The COFER data is not perfect. Not every one reports their currency allocation. Notably, China does not and Taiwan, is not a member of the IMF (only countries are) so it too does not appear to report its currency allocation. Roughly speaking China and Taiwan account for almost 80% of what the IMF calls unallocated reserves.

Many observers often see the swing in reserve valuations and try to interpolate the activities of central banks. These observers tend not to appreciate that like other portfolios, shifting valuations are very important.

In Q1, the euro fell 5.6% against the dollar. Sterling fell a little more than 6%. The Japanese yen was fairly flat, losing 0.5%. The Swiss franc depreciated by 1.8%. The changes in currency value would suggest the COFER data is likely to report an increase in dollar holdings, even if central banks did were entirely passive, which they weren't. Custody holdings for foreign official accounts rose by about $60 bln in Q1.

The valuation of the securities that central bank's hold also is part of the calculus.

It is more difficult to assess. Assuming an average maturity of around 5 years among central bank reserve holding, one can look at the performance of 5-year yields in Q1. As investors are well aware, currency volatility is greater than bond market volatility. During Q1 5-year yields in the US were barely changed, while 5-year bunds, gilts and French bonds advanced with yields off around 10-15 bp. In addition to the changes in bond prices, investors would have earned interest. If a valuation adjustment of currencies is net dollar positive, the valuation of bonds may be slightly dollar negative.

In addition, there has been more interest recently in the "other " category of currencies. This would include the Australian dollar, Canadian dollar, Norwegian krona, as well as other misc currencies. This "other" category accounted for 3.1% of the allocated currencies according to the COFER data.

It is unlikely to change very much in Q1, but Q2 could be a different story, though we wouldn't learn of it until the end of September. Indeed, Q2 reserves are going to be more difficult to get handle on for several reasons. First, the SNB has been very active in Q2 and in May alone appears to have bought something around 55 bln euros. To put this in perspective, in Q3 and Q4 09 combined euro holdings rose 42.5 bln euros. Second, the European debt crisis is going to have other impacts. There has been much talk in recent weeks that central banks in Asia and the Middle East may have boosted sterling reserve holdings.

There is also some thought that the debt crisis is forcing reserve managers to avoid peripheral euro zone bonds--that used to be a way to augment returns. This in some ways may have also weighed on the euro share of world reserves, though probably more in Q2 than Q1.

On balance, then the COFER data due out tomorrow and covers Q1 10 reserve holdings should show the dollar's share of international reserve increased. The second quarter is not as clear but also will likely see an increase in the dollar's share. Sterling's reserve status may also have been enhanced by the European debt crisis, but this and the "other" category may be more of a Q2 story.