A strong October TIC data release, but the BEA’s data only deepens the mystery of how the US financed its current account deficit in the third quarter

NOTE: THIS POST WAS EXTENSIVELY REVISED ABOUT 5 HOURS AFTER IT WAS INITIALLY POSTED.  I DIDN’T HAVE TIME TO DO A FULL ANALYSIS IN THE MORNING.    

The big story of the October TIC data is, obviously, the rebound in net capital flows to the US.   That in some sense had to happen: a large deficit cannot be financed without net capital outflows.   

Foreign demand for US securities doubled, more or less, from September – rising to $118b (more like $105b after adjusting for ABS payments).   US demand for foreign securities somewhat surprisingly fell sharply, from $41b in September to $5b in October.    Central bank demand didn’t appear to dominate long-term flows -- though a bit of caution is in order as the TIC data tends to understate official flows.  But central banks were quite active on the short-end: the increased their deposits and short-term securities holdings by almost $20b, offsetting a fall in private short-term claims.  Total official flows ($41.6b) were only a bit smaller than total private flows ($56.2b).   And remember the TIC data tends to understate official flows.  My strong suspicion is that the official sector financed the bulk of the q4 current account deficit.  China and the oil exporters are piling up cash --

Foreign demand for Treasuries was strong -- $45.9b from private investors and $4.0b from the official sector (if you believe the TIC breakdown) – as was foreign demand for equities.    Agencies and corporate bonds are now out of favor.   In effect, there is currently demand for the safest US assets, and what might be considered the riskiest – but not demand for much in between.

Almost all the demand for Treasuries came from Europe – not Asia.   Europe bought $38b net – though the $30b in UK purchases almost certainly were then sold to investors around the world.    Latin America bought $5.6b and the Caribbean another $7.5b.    Asia was basically flat – Japan was a net buyer, but the emerging Asian economies were large net sellers.   Or perhaps some existing bonds just matured and they bought new bonds through London.     It does though seem like most emerging Asian central banks – not just the PBoC – are reducing the Treasury’s share of their reserves.  There has been a big shift toward agencies generally (though not in October).     Korea and Singapore bought US corporate bonds in October; if that demand came from the Bank of Korea or Singapore’s government investment corporation, they acted as stabilizing speculators, selling the Treasuries the market wanted and buying the corporate bonds the market didn’t.

Private demand for equities also rose to $30b; official demand is still negligible.   The TIC data doesn’t provide much support for the widespread thesis that sovereign wealth funds have lent a lot of support to US equities, but then again, the TIC data probably isn’t capturing most official equity purchases (it did pick up $1.8b in net purchases from the Gulf in October, but there is no way to know if that came from an official account).  Americans sold $5b in foreign equities -- generating $35b in net equity financing.  

Most of that demand seems to be coming from offshore financial centers – London ($6.7b), the Caymans and similar islands ($6.5b) and Hong Kong ($4.8b).   Hong Kong is kind of interesting … it could well represent demand from the mainland.  The other large buyer of US equities, strangely enough, was France. 

Those looking for the inflows from China should look in the short-term data, and specifically the banking data, not the long-term data.  Chinese bank deposits were up $23.8b.  Total short-term claims were up $16b, as China reduced its holdings of short-term securities.    Interestingly, the rise in Chinese deposits wasn’t matched by a rise in total official deposits – suggesting either that another central bank reduced its deposits, offsetting the PBoC, or that the rise in deposits came from state banks and state firms. 

Certainly the formation of the CIC and what seems like a policy decision to allow the banks to manage more of China’s reserves (the rise in swaps contracts) has played havoc with the US data.    Identified Chinese flows have generally come at the short-end recently, though there also has been a shift from long-term treasuries to long-term agencies and corporate bonds.   We just don’t know if these shifts are representative of all Chinese purchases.

Who provided the most financing to the US in October?   Simple: the Kremlin.  Russian short-term claims rose by $18.5b, mostly from a rise in short-term Agency holdings.  Russia also bought $4.4b in long-term claims (mostly Agencies).

The Gulf also chipped in, though not on as large a scale.    The Gulf added $4.3b to its short-term holdings, about $3b of long-term securities (mostly equities).

Press accounts have emphasized the ongoing fall in the quarterly US current account deficit.  It came in under $180b for q3, “only” a bit over 5% of US GDP.   

Interesting enough, the improvement in the income balance (the income surplus rose from $12.7b to $20.5b) was a bigger source of improvement in the overall deficit than the improvement in the trade deficit ($173.2b v $178.4b).  "Other" US income (mostly interest income) rose faster than "other" US income payments.  I frankly don't understand this part of the data (at least not yet) – as rising debts should imply rising (net) payments over time -- but it certainly lends support to the "dark matter" thesis.

The balance on FDI income also improved by about $5b.   That isn’t as much of a mystery: The rest of the world is doing better than the US, pushing up the profits of US firms operating abroad.   I would though also note that the vast majority of the overall $55b FDI income surplus– US receipts on investment abroad net of payments on foreign direct investment in the US – comes from reinvested earnings, not actual dividend payments.   The surplus on dividends and interest on intra-company loans was $11.5b.   The surplus on reinvested earnings was more like $44b.

US firms invested $65b in profits abroad, and foreign firms about $20b in the US.    The gap between the stock of US investment abroad and foreign investment in the US isn’t nearly as wide.   I strongly suspect that the United States policy of not taxing the overseas profits of US firms (at least not until they are repatriated) explains much of the difference: both US and European firms like to show profits in low-tax Euroepan jurisdictions.

This difference on reinvested earnings will explain maybe $175b of the $205b FDI income surplus – it is real money.

The real story though is on the capital account side, as financing for the current account deficit fell far faster than the current account deficit.

Gross inflows and gross outflows both fell, but inflows fell faster.   Indeed, identified inflows of $250b simply cannot finance identified outflows of $155b and a $180b deficit.   The error term was huge.

There is no doubt that demand for US securities collapsed in q3.    Private investors abroad only bought $2.5b in long-term US securities, with purchases of Treasuries just offsetting sales of everything else.   Private investors in the US, by contrast, bought a lot -- $79b-- of foreign securities.  Even taking into account official demand for US securities, on net, there was an outflow of around $48b.   That is, to put it mildly, a change.   The US has until now financed its deficit largely by the sale of securities.

I would also note that recorded official inflows ($40b, including a $10b increase in bank deposits) are small relative to global reserve growth.  Christian Menegatti and I estimate global reserve growth was around $240b in q3; if that is right, keeping the dollar share of central bank reserves constant would imply $215b of dollar reserve growth.

My guess is that the US data understates official purchases and overstates private purchases.   If true, that means private investors abroad were net sellers of US securities in q3.

So where did the financing come from?

Foreign direct investment.  Net FDI inflows amount to around $25b.   That though is far to small to finance a $180b (rounded) current account deficit and $50b (rounded) in US purchases of foreign securities (net of foreign purchases of US securities).  $25b in inflows cannot over $230b in outflows.

Net inflows from banks, broker-dealers and other non-bank financial intermediaries provided about $110b of financing.  That too is a big change -- claims from banks and non-banks have not typically been a major source of net financing for the US.   This generally is very short-term financing.

And that leaves a $85b or so gap.  "Errors" – the statistical discrepancy – were very large.

A few final observations:

The US current account deficit with China was $79b in q3 – or only a bit under ½ the total.  Comparing the bilateral deficit to the overall deficit is dangerous, but China’s global surplus is now only a little under ½ the size of the US deficit as well.   There is good reason to focus on China – and good reason why US-Chinese (and Euroepan-Chinese) economic relations have been a major source of friction.

Chinese purchases of US assets have fallen from $81.3b in q1 and $57.5b in q2 to $34.15b in q3 ($34b is obviously well below $80b) -- something is going on, though we still don’t yet know what.   My guess is that the US data fails to pick up purchases from Chinese banks.   But on the surface it seems like the rise in China’s surplus over the course of 2007 has coincided with a fall in Chinese demand for US debt.

One point of caution: the $173b in identified Chinese flows in the first three quarters is still far larger than the $5.3b in identified inflows from the Middle East.

The Fed’s flow of funds data provides a bit of interesting color:  the $8.2b fall in foreign purchases of US corporate debt (a huge fall from an average of around $100b of purchases a quarter) was matched by a $50b fall in US holdings of foreign corporate paper.  Basically, a lot of offshore “entities” where issuing corporate paper to US investors (money market funds, state funds, etc) and buying US asset-backed securities.   That stopped.   The result: the US has fewer liabilities to the world (ABS) and fewer foreign assets  (ABCP issued by foreigners).

Finally, the fall in demand for US corporate bonds came entirely from Europe.   Asia kept buying at basically the same pace as before.    That fits a lot of other anecdotes about the operation of the shadow banking system (also see Naked Capitalism’s critique of Tett and Davies)

I’ll have a lot more to say about the current account deficit – and its financing – in my forthcoming response to Richard Iley.