There is something about SAFE — maybe the fact that its full name, the State Administration of Foreign Exchange, seems to come straight from the Ministry of (Bureaucratic) Magic –  that seems to inspire Harry Potter analogies.  

Macro Man calls China’s reserve managers Voldemort – in part because investment bank research never (or almost never) refers to China by name.    Diplomatic analysts prefer to write about “Asian accounts.” 

Andy Mukherjee got into the act yesterday as well.  He – drawing on Paul Meggyesi’s work for JP Morgan – argues that China’s reserve managers must have an invisibility cloak.   

Something doesn't quite add up in the U.S. Treasury's International Capital System statistics, or TIC data.

Why, for instance, did U.K. investors begin buying U.S. securities just as the People's Bank of China started increasing its pace of reserve accumulation in 2004?

“Is this more than a coincidence? I believe it is,'' Meggyesi says. For the past three years, the reserve buildup in Russia, China and rest of Asia has overshot the recorded official purchases of U.S. securities.

And the growing difference between the two has moved in lockstep with the reported purchases by U.K. investors, “providing very strong evidence that much of this reserve accumulation is being channeled via the U.K., and is in the process being incorrectly recorded in the TIC data as a private rather than official inflow,'' Meggyesi says.

I try not to be too turf-conscious – the diffusion of knowledge is a good thing.   But I confess to feeling a small urge to defend my (intellectual) turf when I read the Mukherjee column.    

After all, one of the main purposes of this blog has been to try to brew a potion that counters SAFE’s invisibility cloak — and now it seems that JP Morgan is starting to do the same thing. 

I have consistently argued that the US TIC data dramatically understates the role of “official actors”, both central banks and oil investment funds, in the financing of the US external deficit (more here).  And I have spent a fair amount of time trying to construct alternative ways of estimating central bank financing of the US – measures that don’t rely entirely on the US data.  Right now those measures suggest that the US is received close to $800b (annualized) in central bank financing in the first half of 2007.

Indeed, I suspect the three most consistent themes of this blog have been:

• The US trade deficit (which incidentally is far larger than Australia or New Zealand’s trade deficit, as both have large income deficits that contribute to their current account deficits) is creating strong underlying pressure for dollar depreciation.
• The US income balance will eventually turn negative.
• The US TIC data/ the BEA’s capital account data understates central bank financing o the US, and specifically understates Chinese and Middle Eastern financing of the US (there is more than one invisibility cloak floating around) 

The dollar is pretty weak.  I am still waiting for the income balance to swing.   The BEA’s recent data revisions didn’t help my case here, though I think the details do suggest that the income balance will soon deteriorate (the BEA revised income payments down because more US debt is short-term – and that is no longer a big help).  But the BEA’s revisions – based on the survey data – do indicate that the unrevised TIC data systematically understates official inflows.

Indeed, the BEA revised official purchase of Treasuries up by $45b for 2005, and official purchases of Agencies up by close to $25b.   The adjustments to the 2006 data were bigger, but they aren’t formally considered revisions.  And the initial estimates for 2006 were done on the basis of good data (from the survey) for only the first half of 2006.   More revisions will come next year.

The 2006 survey data led the Treasury to revise the UK’s holdings of Treasuries down by $145 billion (see here), from $200b to $55b. China’s holdings of Treasuries were revised up by close to $45b at the same time.  Similar revisions were made to the data on Agencies purchases.   Total Chinese holdings of US debt were revised up by $90b in both mid 2005 and mid 2006.  But even the survey didn’t really pierce the invisibility cloak of many actors – Switzerland and Luxembourg’s holdings of Treasuries, for example, were also revised up … 

We will have to wait a long time for the revised data for the second half of 2006 and 2007.   Still, I am quite confident that British investors haven’t increased their holdings of treasuries by $110b since mid-2006.   If past patterns hold, China will turn out to have bought between $30b and $50b of those bonds.  The 2005 revisions also increased Chinese holdings by about $50b, while reducing the UK’s holdings by close to $100b.

The “visible” data shows that China now buys far more Agency bonds than Treasury bonds.  I consequently expect the next survey will lead to a very large – probably a shockingly large – increase in reported Chinese holdings of US mortgage debt. 

China’s invisibility cloak isn’t perfect.   But it does help Voldemort disguise its activities.   Reported Chinese purchases of US assets have been very low recently – only about 20% of the increase in China’s reserves.    That could mean that Voldemort has found a better invisibility cloak.    Or it could mean that Voldemort is diversifying away from the dollar – and thus putting contributing to the strength of the euro and the pound.

Indeed, I would argue that China’s impact on the euro/ dollar is one of the great unsolved mysteries of international finance.    There is little doubt that Chinese intervention has limited the RMB’s appreciation v the dollar.   And it clearly has pushed the RMB down v the euro and the pound.   But the impact of China on the euro/ dollar cross is subject to substantial uncertainty. 

Macro Man argues that in the absence of Chinese demand for euros and pounds, both the euro and the pound would be substantially weaker against the dollar.  China intervenes primarily in the rmb/ dollar – and consequently is a major net seller of dollars for euros.    And with $500b in dollar reserve growth, China has become a big seller of dollars.

Others – including others close to the market – make the opposite argument.   China tends to end up holding more dollars (as it reduces the amount of dollars it sells for euros and pounds) when the dollar is under pressure.   This may reflect China’s desire to hold a constant share of dollars in its portfolio — as keeping the dollar's share in China's portfolio constant would require holding on to more dollars when the dollar is falling than when it is rising.   Or it may reflect China’s desire not to push the RMB too far down against the euro.  

No matter.  If this describes the PBoC's behavior, the net effect is that China reduces its sale of dollars when the dollar is under the pressure, helping to support the dollar.
And the PBoC – along with other central banks – ends up supplying more financing to the US when no one else wants to.  Absent such financing the dollar would be even weaker against a host of currencies. 

Both arguments are plausible.   And until Voldemort gives up his invisibility cloak, we won’t know the answer.