Tokyo-Sao Paulo-Rio-New York-Washington …

Plenty of signs suggest that a fair amount of money is leaving Japan to look for higher yields elsewhere.    More and more of those outflows seem to be coming from leveraged investors who borrow yen to buy other currencies – look at the excellent analysis by Hali Edison and Chris Walker in the IMF’s Asian regional outlook (box 1.4).

But it doesn’t seem like much of it going to the US.  At least not directly. 

There are plenty of places that offer higher yields than Japan that don’t even have current account deficits – the eurozone for one (its deficit is very small).   Brazil for another.   Russia too.

There are places that offer higher yields than the US but run smaller deficits.   The UK comes to mind.   India too.  Australia perhaps fits here – its overall deficit is about the same size as the US deficit (though its composition is different; Australia has a smaller trade deficit but a large deficit in investment income) 

And then there are places with bigger deficits than the US.    New Zealand.   Iceland.   Turkey.   All offer higher yields than the US.

So how then has the US managed to finance its deficit?   After all, current ten year Treasury yields don’t suggest “foreign flight”? 

Let’s start in Sao Paulo (or Brasilia).    Brazil’s reserves rose by $4.6b last week.    The workers in Brazil’s central bank were on strike last week; foreign investors were not.    The central bank was back in the market again today, to no one's surprise.

$4.6b in a week works out to an average of $0.9b a day -- just under $1 billion.  $1 billion a day is a China-like pace of reserve growth.  Or at least it used to be a China-like pace – I suspect China has raised the bar this year.  Its reserve growth has likely accelerated from about $20b a month over the past couple of years to around $30b a month, if not more. 

Brazil still runs a current account surplus, but its current account surplus maybe explains $1-2b a month of reserve growth.  Not $1b a day.   Brazil is attracting large amount of foreign capital.

Let’s assume that that the money flowing into Brazil comes from a hedge fund that borrowed yen from a bank in London that itself borrowed yen in the interbank market in Tokyo.    That is an assumption, but not a totally implausible assumption.  In practice, I suspect a lot would be done through derivatives, with banks hedging their derivative position generating a lot of the “actual” flows.   No matter.  What counts is that Japan’s spare savings is effectively channeled, with a bit of help from the world's banks and a few hedge funds, to Brazil.  

The logic of this trade is easy to understand.  Real-denominated debt pays 12%.  Borrowing at 1-2% in a depreciating currency and getting 12.5% in an appreciating currency is fun. 

But Brazil doesn’t actually need the money.  It has a current account surplus.  It saves more than it invests.    Its central bank doesn’t want to cut rates (which might encourage more investment) and it doesn’t want the real to appreciate (Brazil has a manufacturing sector that competes with China as well as a commodity sector that sells to China).   So it buys the dollars and sterilizes by selling central bank bills at high rates.  

The central bank effectively ends up paying the 12% interest rate in real to foreign investors. 

And what does it do with the dollars is buys with the real-denominated debt it issues to sterilize the inflow?  

It sends them to the US, and probably has the New York Fed invest them in Treasury securities, financing the US government in DC.    We know that Brazil’s holdings of Treasuries are rising rapidly, and I would assume that the BCB accounts for some of the rapid growth in the Fed’s custodial holdings.   

What does it get on its Treasuries?   A bit less than 5%.   Brazil's central bank effectively borrows in real at 12.5% to build up reserves that earn less than 5%. 

I can see why Brazil needs around $100b in reserves (10% of its GDP is a reasonable benchmark for reserve adequacy).  But not $200b.  Yet unless its pace of reserve growth slows, it will hit $200b rather quickly.  Well before the end of the year.  If last week’s pace continues, it could hit $200b in about four months (think $1b a day *83 working days) …

That is how a portion of the US deficit is now financed. 

Tokyo-Rio-New York though isn’t the only route that brings Japan’s spare savings to the US.   Tokyo-Mumbai-New York also works.  The RBI though likes euro and pounds, so it sends almost as much to London and Frankfurt as to New York.   It isn’t the most efficient routing.  Tokyo-Moscow-New York has the same problem.    It really is Tokyo-Moscow-London-New York, and some funds never make it all the way to New York.

Beijing-New York-DC is more efficient.

Brazil reports its reserve growth with a very tiny lag, so we know more or less what its central bank did last week.       India and Russia report with a week’s lag, so we only know what they did in the second week of April. 

India’s reserves increased by $2.8b.   The euro rose by about 1% that week and the pound by a bit less than 1%, so the rising dollar value of India’s existing euros and pounds explains maybe $1b of that increase.  The rest came from intervention, as the central bank resisted pressure for further rupee appreciation.

Russia’s reserves increased by a whooping $10.3b in the second week of April (after increasing by $7.6b in the first week of April).   Russia has a lot of euros and pounds, but the rise in their value explains at most $2b of the increase.    Russia’s current account surplus explains at most $2b – and probably far less – of the increase.   That implies at least $6b, if not a bit more, in net capital inflows to Russia …

Suppose that after adjusting for valuation, India added $2b to its reserves in week, Russia $8b and Brazil $4-5b.   China is now adding at least $5b a week.  Since the PBoC had to drive the RMB down against the euro and pound in the second week of April, I wouldn’t be surprised if it bought even more.    Say $10b.     The BRICs then are currently adding $20-25b a week to their reserves.  $20b a week is $1 trillion a year.   $25b a week is $100b a month, $300b a quarter, $1.2 trillion a year …

That truly is unimaginable large.

If you think that scale of reserve growth can continue, there is little reason to worry.

But in my view there is an interesting race going on now.    Will private investors (re)discover their appetite for US assets before the world’s central banks loose their appetite for US assets?  Or will the central banks lose their appetite for dollars before the private market regains their appetite?   

On the surface, you can say little has changed since the 2005.     The US is still running a large deficit.   It still finances that deficit by selling bonds.  Treasury bonds still yield between 4.5 and 5%. 

Under the surface though something has changed.  

In the later part of 2005, the world’s central banks were adding about $175b a quarter to their reserves, and putting a pretty high fraction of those reserves into euros and pounds.  Their dollar reserves rose by only about $100b a quarter.  

Now they are adding almost twice as much to their reserves.  If they are trying to keep the dollar share of their reserves constant – or just finding it hard to sell dollars without moving the market – the pace of their dollar reserve growth has increased by even more.  

I am truly stunned by the pace of reserve growth that I am seeing in the emerging world right now.   Emerging market central banks are potentially adding to their dollar reserves at a pace sufficient to finance the entire US current account deficit.