The US trade deficit is falling, but not as fast as the world's demand for US debt.

The August TIC data was really bad. Even Fox Business News would have trouble putting a happy face on it.

The net outflow in August – from a combination of foreign investors reducing their claims on the US and Americans adding to their claims on the world – was around $160b. Most of that -- $140b – came from the private sector, but the official sector also reduced its claims on the US. The total monthly outflow works out to a bit more than 1% of US GDP. Annualized, that is a 12% of GDP outflow. To put a 12% of GDP outflow in context, it is roughly the magnitude of the private outflow from Argentina in 2001, at the peak of its crisis.

Throw in the United States roughly $70b a month current account deficit and there is a $200b – or 1.5% of GDP monthly, and more like 18% of GDP annualized – gap between the net flows in the August data and the flows needed to sustain the current equilibrium. There is no way to spin that kind of outflow as a positive.

Obviously, there is a big difference between that kind of outflow for a month and that kind of outflow over the course of the year. The rolling 12 month sum of high quality inflows – foreign purchases of long-term securities (see line 19) – is still $785b. That though is below the totals for 2005 and 200. And if the repayment of principal on various Agency guaranteed MBS and other ABS are factored in, net long-term acquisitions of long-term securities (line 21) came in at $594b over the last 12 months – well below what the US needs to sustain the current account deficit.

The August data certainly doesn’t provide any support for the then-popular argument that the US remained a safe haven – despite all the subprime turmoil – in times of stress. Foreign demand for US bonds – and particularly corporate bonds -- disappeared. American demand for foreign bonds and equities didn’t.

I was expecting a fall in foreign demand for US corporate bonds. Net purchases of US corporate bonds by private investors abroad – including London SIVS sponsored by US banks – averaged about $40b a month in 2006. It was close to zero in July, and August saw net sales.

But I was also expecting to see Americans selling some of their overseas holdings. And there is no sign of any repatriation of US funds invested abroad in the data.

Indeed, the flows in the August data make the dollar’s mini-rally in August all the more puzzling. My guess is that some European banks with troubled conduits borrowed in euros from the ECB – or borrowed from other banks who in turn borrowed from the ECB -- in order to obtain the cash they needed to repay their maturing commercial paper. And since they were borrowing in euros to cover dollar liabilities, they needed to hedge.

Alas, I would also have expected to see more signs of the deleveraging process in the US data. That deleveraging could have come from the sale of US securities abroad. Or from a fall in US lending to the rest of the world. Neither happened. Some of the deleveraging may have taken place “off-balance sheet” so to speak.

A few additional points, all derived from data that can be found on the TIC home page:

China reduced its holdings of long-term securities, especially Treasuries. Its logn-term holdings fell by $10b, with a $14.2b fall in its Treasury holdings. However, it added $14.1b to its short-term portfolio (mostly by buying short-term securities other than T-bills – whether Agencies or something else). Net flows from China were still positive – though at $4b, they were rather low.

The same basic story holdings for Russia. Its long-term holdings fell by $5.7b, but its short-term holdings rose by $10b (mostly short-term Agencies).

Brazil’s purchases of Treasuries fell significantly, consistent with the fall in the pace of Brazil’s reserve growth.

Norway bought a ton of Treasuries (after selling a ton earlier in the year). I hope that the complicated trading strategy the Government Fund is using is working for them … these sales and purchases are generally thought to be related to Norway’s need to hedge other positions.

Japan sold a ton of Treasuries ($20.1b). So did Taiwan ($5.6b). Some of Taiwan’s sales are tied to the fall in its reserves in August, but in general Asian central banks seem to have taken advantage of the surge in private demand for Treasuries to pare back some of their own holdings. Good for them. August was precisely the time when central banks could sell without destabilizing the market. Indeed, by selling an asset that private investors had previously shunned but now desperately wanted, the central banks acted as a force for stability (as they should).

Normally I would argue that the $33.2b in UK purchases of Treasuries are likely to be disguised central bank flows – and use them to offset the $29.7b recorded fall in central bank holdings. That story still has some legs – lots of central banks do buy in London rather than New York. But august was a strange month. The Caymans, for example, was a big net buyer of Treasuries as well (to the tune of $26.6b), and central banks aren’t known to spend as much time in the Caribbean as another well-known class of investors. Some UK based hedge funds could also have been buying Treasuries.

Europeans stopped buying US corporate bonds. The eurozone reduced its holdings of US bonds by $20.2b, with a $6.3b fall in holdings of US corporate bonds. France and Luxembourg both scaled back their holdings of US corporate debt. The UK was still a net buyer, but only to the tune of $5.9b. That is way, way down.

One last point: the fall-off in official demand for US assets in August was tied to a large fall in private capital inflows to a set of emerging markets in August, together with relative low (recorded) purchases from China and other emerging economies with large current account surpluses. Saudi Monetary Agency foreign assets actually fell in August, for reasons I still don’t fully understand.

With oil high and with private capital once again flowing towards emerging economies (though perhaps not today), I would expect official inflows to bounce back. I am not as confident that foreign demand for US corporate debt will reemerge.

UPDATE: The FT's US site is currently leading with an article on the TIC -- surely a first. They put a lot of emphasis on the -$53.4b net equity outflow, while I emphasized the fall in net bond flows. That partially reflects my biases -- I am a more of a bond market/ currency guy than an equity guy. I should though have placed a bit more emphasis on the equity outflow. But the fact that net bond flows were flat (private buying, especially of Treasuries, offset official selling) isn't evidence that bond flows don't matter. The US deficit has overwhelmingly been financed by the sale of bonds. Indeed, in 2006, the US sold over a trillion in bonds, or over $80b a month, on average. That number needs to be adjusted for principal payments, but it still comes out to be close to $850b for 2006, or around a $70b net inflow every month.

That is what financed the US deficit. Short-term flows generally offset -- outflows match inflows. Net equity flows were flat or slightly negative. Obviously, it would matter a lot if net equity flows turned massively negative. But the disappearance of net bond flows also would have a huge impact.