China’s Holding of Treasuries

In this edition of MarketView, we decipher why China has amassed one of the largest pools of capital in the form of central bank reserves.

China has amassed one of the largest pools of capital in the form of central bank reserves.  At its peak, it was around $4 trillion.  The composition of these reserves is a closely guarded secret.  However, if the yuan is going to be including the IMF’s SDR, it is anticipated that China will have to report the currency allocation of its reserves.   China would not necessarily publish these though the IMF would include them in its aggregations process that it publishes quarterly.

China’s reserves have fallen for four consecutive quarters through the middle of the year.  Over this period, the reserves have fallen a total of $300 bln.  In Q2 15, reserves fell $40 bln. Stemming from its trade surplus, investment income, and direct investment flows, economists expect reserves to have gone up, rather than fall.

The gap between what the economists’ model says that reserves should be and what the PBOC reports them as become the proxy for capital flight.   As the models differ, the estimate of the capital flight differs.  One large investment bank estimates it at $800 bln.  Another one claims $520 bln.

Many observers take it another step further.  On the assumption that the bulk of China’s reserves is invested in Treasuries, a decline in reserves is a decline in Treasury holdings.   The most authoritative source is from the US Treasury.  This is depicted in graph below.  It shows that as of the end of May, China held $1.27 trillion of  US Treasuries.  It is unchanged ($2 bln more) from when its reserves peaked last June.

The PBOC may hold Agency bonds as part of their reserve holdings, not just Treasuries.  In the first five months of this year, China’s Agency holdings rose by $14.5 bln. By the US reckoning, China’s Treasury holdings rose by $26 bln in the Jan-May period. Some claim that China is disguising its flows.  They argue that this was done on Euroclear in Belgium.  There had been a six-month period in late-2013 and into early-2014 that US data showed a large accumulation of Treasuries in Belgium.  They rose from $173 bln in September 2013 to $381 bln in March 2014.  At first some thought this was Russia shifting its reserves, but it was understood to be related to Euroclear’s exchange function, likely collateral.

Belgium’s holdings of US Treasuries has fallen, especially over in the Feb-May period. This is shown in the second graph below. The US Treasury data shows Belgium’s Treasury holdings fell from $354.5 bln in January to $203bln as of May.   There is no compelling reason to think this was China.   What is its motive?  A big run up and then a big run down in short order does not fit China’s modus operandi.

Perhaps the data is flawed.  US Treasury data only aggregates data of activity through US institutions.  Maybe China has been selling its Treasury holdings on the sly.  If there was such a large seller of US Treasuries as the capital flight story implies surely the markets would show it.  Where is it?  Assuming that China’s reserves are kept invested in the belly of the curve, let’s look at what has happened to 5-year and 7-year US yields. The former has risen by 5 bp since China’s reserves peaked.  The latter has fallen 7 bp.  The dollar-yuan rate is unchanged from the end of last June.

It is not clear why China would sell Treasuries now.  The Fed is preparing the market for a rate hike.  While this may weigh on Treasury prices, as an investor that may hold until maturity, not a trader, the change in prices is of little significance.  The yield is locked.   The key to the total return is the dollar, and Chinese officials are well aware of the divergence of monetary policy.

Maybe the US market is so deep and liquid that it has easily absorbed the Treasury and dollar sales. Could other markets have absorbed the hundreds of billions of dollars that economists’ models suggest have left China?   The European stock and bond markets are obvious places to look.  A surge of Chinese money does not appear to be evident.

Chinese economic data is not often thought to be of high quality.  The methodology it employs does not appear to be very transparent.  It is not clear the independence of its statistical collection and reporting.  This, incidentally, is one of the creditors demands that Greece has accepted.

It does appear that there have been capital outflows from China.   Many are likely exaggerating it. It may stem from incomplete data.  It may also be a function of not understanding the data that does exist.  For example, China’s trade surplus is often treated as a source of capital inflows.  However, what is reported is merchandise trade.  China records a service deficit that offsets part of the merchandise surplus.  China’s reserves are not all sitting idle.  Some have been loaned out to other parts of the government, like the Export-Import Bank, for example.   What is the accounting practice China uses for this?

Since its foreign exchange reserves are so large, valuation, that is changes in asset prices and currencies (since the reserves are reported in US dollars) are an important part of a rigorous analysis. What is the accounting practice for this?   Economists who estimate the composition of China’s reserves then estimate the valuation changes.   For example, using round numbers to illustrate the point, assume China has $4 trillion in reserves, of which 25% or $1 trillion are in euro investments.  Since the end of H1 14, the euro has lost 20% against the dollar.  Holding all else equal, which of course it is not, that alone would account for a $200 bln decline in the dollar value of China’s reserves.