It is appropriate, I suspect, for my first post over here – and I want to reiterate my thanks to Nouriel for hosting my blog at RGE for the past 8 months even I after I moved to the Council – is on the latest Treasury capital flows data. The TIC data is both amazing – no other country puts out a comparable data series – and intensely frustrating.
The March data highlights why trying to read the TIC data can a challenge.
The data on long-term flows matches what we more or less know about the world. It shows a lot of official demand for US long-term debt -- with total central bank and sovereign fund purchases of a bit less than $50 billion (mostly Treasuries). It also shows the ongoing absence of demand for US long-term “corporate” debt from either private or official buyers. Foreign investors sold more US corporate debt than they bought in March. That is a change from the pre-August 2007 world. Monthly purchases of $40-50 billion were the norm for a while.
The frustration comes from the data on short-term claims. A $34.1 billion fall in official short-term claims larger offset the $48.1 billion in net official (central bank and sovereign fund) purchases of long-term bonds. The resulting $14.0 billion total was further adjusted (by around $4.7 billiion) for principal repayment on asset-backed securities, producing $9 billion in net official inflows.
That isn’t much for a world where Saudi reserves are growing by $15 billion a month and Chinese foreign assets are growing by $50 billion a month. The total growth of central bank and sovereign fund assets now exceeds $100 billion in an average month.
Net private inflows were negative $48.2 billion in March – and only $35.3 billion for all of q1 -- largely because of a huge fall in short-term bank claims.
The US needs a net inflow of around $175 billion to cover its external deficit, so that leaves a rather substantial gap. I doubt all of it was filled by FDI inflows.
For the first quarter as a whole, official creditors provided $76.5 billion in net financing: A $20.6 billion fall in short-term claims offset $107.6 billion in long-term purchases (and 10.5 billion of principal repayment on asset-backed securities). But private creditors took $41.2 billion out of the US.
There is good reason to think, however, that the TIC data significantly understates official purchases of US assets in the first quarter.
First, private investors in the UK bought $98.5 billion of Treasuries and Agencies. That is suspiciously large. Almost all of the buildup of Treasuries in the UK has – over the past few years – proved to be a mirage. When the survey data comes along, the UK’s holdings get revised down – and the holdings of China, Russia and the Gulf get revised up.
Second, the TIC data and the New York Fed’s custodial data don’t match up.
The Fed’s custodial holdings of Treasuries and Agencies rose by $151 billion in the first quarter ($72.5b in Treasuries, $78.3b in Agencies). The TIC data for the first quarter shows $80.1b in central bank purchases of Treasuries and Agencies, ($59.1b of Treasuries, $21.0b of Agencies). To make the data match up, I added short-term flows to the Treasury and Agency totals. The same holds for March – the $69.8b rise in the Fed’s custodial holdings outpaced the $33.2 billion increase reported in the TIC data.
What is going on?
I suspect central banks decided to put some of the money that they had parked at the front end of the curve – in bills and deposits – to work. The US data captured the fall in their short-term holdings, but not the entire rise in the long-term holdings. Russia and others buy a lot of long-term debt through London. Some of the securities bought in London are then turned over to the Fed for safekeeping.
I also suspect that the unwinding of the “shadow” banking system continues to have a big impact on the data, though in ways that I don’t necessarily fully understand. A lot of the rise in overall cross-border flows over the past few years came from a rise in short-term flows. That changed in August. The scale of capital moving across borders – both inflows and outflows – fell. Something similar seems to have happened in q1. Foreign claims on US banks fell by $200 billion …
It is easier to make sense of the various national data points. Or most of them.
London is buying Treasuries and Agencies but not corporate bonds or equities. But here it is almost certain that the UK is just acting as an intermediary. A European bank buys a Treasury bond that it then sells to the PBoC.
Chinese flows remain smaller than would be expected given the scale of its reserve growth. Reserves were up $120 billion in q1, after adjusting for valuation. Only $30.6 billion of that shows up in the US TIC data, as a $7.4 billion fall in short-term holdings offset $38 billion in long-term purchases. If you think that reported reserve growth understates Chinese foreign asset growth –the state banks had to hold dollars to meet their reserves requirement, some fx was shifted to the CIC – the gap is even bigger.
This fits together with the story about London. I also suspect that some of the (now large) net purchase of US assets from Hong Kong also reflects the activities of China’s State Administration of Foreign Exchange
Russia fits the same pattern, but in a more extreme way. In q1, a $21.2b fall in Russian short-term holdings offset a $14.7 billion rise in its long-term holdings. London again. For some reason, the US TIC data captures Russia’s short-term holdings better than its long-term holdings.
The Asian oil exporters added $12.1 billion to their short-term holdings and $7 billion to their long-term holdings (with almost the entire $7 billion coming from a couple of large equity investments in January … ). Saudi Arabia alone added $40 billion to its foreign assets in q1 – and it is reasonably to think the other Gulf states added a similar amount. So the $19.1 billion increase is actually rather small given the huge sums available for the Gulf to invest abroad.
Norway is also interesting, though in a different way. It bought $20.9 billion of long-term assets (mostly treasuries) and reduced its short-term holdings by $16.9 billion. It had a big impact in March – buying $11.3b of long-term Treasuries while reducing short-term claims by $10.1b. Clearly Norges bank is making some kind of complicated bet on US interest rates – or hedging some other position with large treasury purchases. It often produces large swings in the official flows data.
No more from me. Arpana Pandey of the Council on Foreign Relations and I are working on a series of charts that seek to illustrate the changes in various countries’ holdings graphically. I suspect that our charts will tell they key stories better than more words.