One number to watch

My nomination for the single number that best captures the state of total’s world economy -- and a key risk going forward:

The annual increase in central banks dollar reserves.

It is not exactly a number that Wall Street analysts breathlessly track, in part because it is not reported by any national statistical agencies. But, to my mind, it tells us a more about the current global economy almost any other number that comes to mind.

After all, the increase in central bank dollar reserves should be tells us the share of the US current account deficit that is financed by world’s central banks, and thus, the share that is left for private markets.

In 2003, the global increase in dollar reserves was $441 billion (actually $486 billion, including the reserves China used to recapitalize two state banks). The US current account deficit was $530 billion. The US only had to raise something like $50 b privately to finance a current account deficit of over $500 billion. Not bad.

In 2004, I suspect the global increase in dollar reserves will turn out to be above $400 b. The global increase in reserves was roughly $700b; and, after netting out roughly $80 b in valuation gains from the shift in the euro/ dollar, a 2/3 dollars, 1/3 euros and others distribution gives you a dollar reserve increase of 400b. That is a pretty safe bet for another reason: recorded central bank inflows in the US are running ahead of the 2003 pace -- recorded inflows should total at least $320b. We know that the US recording system is imperfect, and total dollar reserve accumulation usually exceeds the recorded in inflow to the US.

$400 b would finance a bit under two-thirds of the 2005 current account deficit, which is probably going to be around $660 b. That leaves only 260b (net) for private investors. And if dollar reserve accumulation turns out to be $450b rather than 400b, even less (on net) had to come from private investors.

In 2005, the current account deficit is -- barring a major change -- likely to be close to $800 billion. A $800 b current account deficit corresponds with a monthly trade deficit of 60b -- the November 2004 number. With oil back close to 50, the November number looks to be typical of early 2005. If strong domestic demand growth in the US continues and that drives continued strong import growth, the current account deficit could well be bigger than 800b.

If dollar reserves go up $200b, 600 b of that has to be financed privately. I personally don’t think that would be possible: the US is unlikely to go from 50b (net) in financing from private investors abroad in 2003 to 600b (net) in financing from private investors abroad in 2005.

If dollar reserves go up by $400b, 400 b needs to be financed by foreign private investors.

If they go up by an unprecedented $600b, only 200b needs to be financed privately. 200 b is about what the US raised privately this year, assuming dollar reserve accumulation in the 400 to 450 b range.

Of course, some of those private flows may come from the tax holiday the US government is offering for the repatriation of profits held by US firms overseas. The FT reported on Monday that Wall Street now estimates that American firms will bind home about $100 billion in assets now held abroad in a FOREIGN currency to take advantage of the special tax holiday the US is offering this year. (Bringing home dollars held abroad is wash -- the US ends up with fewer external assets and fewer external liabilities, but no net financing).

This repatriation of foreign profits because of the tax holiday needs to be kept in perspective. The current account deficit, by my reckoning, is on track to increase from roughly $650 b to around $800b in 2005. The 100b in estimated financing from the tax holiday would not even provide all of the incremental financing needed to cover the expected growth of the current account deficit in 2005. Even with $100 b from US firms bringing profits home, the US would still needs to raise $700b from foreign investors to cover the rest of its current account defict -- more than it raised in 2004.

Still $100 billion is $100 billion. If the US gets 450b in financing from the increase in dollar reserves in 2005, and 100b in financing from the repatriation of corporate profits now held abroad in foreign currency, the US would only need an additional $250 b (net) in private financing. If foreign demand for US equities continues for the entire year, that is a scenario that just might work. It certainly would be a lot easier to fund the US current account deficit if foreign demand for US stocks started to exceed US demand for foreign stocks.

The FT also reported that Greg Anderson, a currency strategist at ABN Amro in Chicago, estimates that this 100 b is will push the broad dollar up 5%:

"All else being equal, $100bn is equivalent to a 5 per cent rise in the dollar’s trade-weighted index."

Interesting.

Just to have some fun, let’s suppose that central bank inflows lend as much support to the dollar as corporate repatriation, and that the $100 b = 5% rule of thumb applies to central banks flows as well.

Central banks are now adding $400 b (or more) per year to their dollar reserves. A more normal number is something like 100b. That suggests that the net $300 b in financing from usually high levels of central bank intervention kept the broad dollar 15% higher than it otherwise would have been in 2003 and 2004 ...

Maybe central bank reserve accumulation is another variable Stephen Jen might want to add to his model.

15%, by the way, is a conservative estimate. Dollar reserve accumulation was higher than 400b in 2003, and may well have been a bit higher in 2004 -- and, as importantly, some private flows only happen because private investors expect that central banks will keep the dollar from falling too far too fast.

One technical note: it is hard to track the increase in dollar reserves in real time. But is not that hard to track the overall increase in central bank reserves in close to real time. Most of the big players release their overall reserve data fairly close to the end of the month. Then all you need is an estimate of the distribution of the global increase between dollars and euros ...