Global investors have been in thrall to the central banks ever since
quantitative easing (QE) started in 2009 and, of course, all eyes are on
the Federal Reserve this week. The Fed has now frozen its QE programme,
and may raise rates sometime this year, though perhaps not as early as
next Thursday. Nevertheless, global investors have been comforted by the
extremely large increases in balance sheets proposed by the Bank of
Japan (BoJ) and the ECB, and the overall scale of worldwide QE has
seemed likely to remain sizeable for the foreseeable future.
However, in recent months, an ominous new factor has arisen. Capital
outflows from the emerging market economies (EMs) have surged, and have
resulted in large declines in foreign exchange reserves as EM central
banks have intervened to support their exchange rates.
Since these reserves are typically held in government bonds in the
developed market economies (DMs), this process has resulted in bond
sales by EM central banks. In August, this new factor has more than
offset the entire QE undertaken by the ECB and the BoJ, leaving global
QE substantially in negative territory.
Some commentators have become concerned that this new form of
“quantitative tightening” will result in a significant reversal of total
central bank support for global asset prices, especially if the EM
crisis gets worse. This blog examines the quantities involved, and
discusses the analytical debate about whether any of this matters at all
for asset prices.
The conclusion is that the EM factor is likely to offset part, but
perhaps not quite all, of the QE planned by the ECB and the BoJ in the
next year. Overall, global QE will provide much less stimulus than it
has since 2006.
The major sources of central bank balance sheet expansion at present are
of course the bond purchase programmes announced by the BoJ and the ECB.
Together these programmes are running at an average of about $130
billion a month. The average maturity of the bonds purchased is probably
around 7 years, so an enormous amount of “bond duration” is still being
removed from private sector hands in the developed economies.
The question is how much of this stimulus is likely to be offset by the
sale of EM bond holdings as a result of the foreign exchange
intervention by EM central banks. Fulcrum estimates that total EM
central bank balance sheets may have declined by about $450 billion in
the 3 months since the crisis worsened in the summer, of which about
$170 billion has come from China alone.
Consequently, global QE, measured by this metric, has probably turned
substantially negative . Nomura (and others) estimate that foreign
exchange intervention by the EMs was probably around $160 billion in
August alone, and this would have directly triggered bond sales in the
US and Europe.
Of course, no-one can prove that this drain of central bank liquidity
caused the rise in global bond yields and the drop in risk assets last
month; market interpretations of Fed policy have probably been just as
important. Nevertheless, it is an interesting fact that has grabbed the
attention of macro investors. The release of China’s foreign exchange
reserve figures has suddenly become one of the most watched global data
releases each month.
What is the outlook for this measure of global liquidity over the next
year or so? The BoJ and the ECB are, if anything, considering further
extensions of their bond purchase programmes. Consequently, global QE
will return to positive territory unless the large drain on EM foreign
exchange reserves continues.
But this drain is likely to be maintained for a while. A recent detailed
analysis of global reserve holdings by Deutsche Bank economists 
suggests that total global reserves could fall by $1,500 billion during
the current drawdown, about a third of which has already happened. This
might take EM central bank balance sheets roughly back to where they
were just before the 2008 financial crash as a share of EM GDP – a
pessimistic but not extreme outcome.
On this and other assumptions, Fulcrum estimates that the total increase
in global central bank balance sheets as a percentage of world GDP,
which is one indicator of the stimulus from global QE, would be fairly
close to zero next year, compared to an average injection of about 2 per
cent of global GDP in recent years. Apart from a short period at the end
of 2009, this would be the lowest rate of expansion since 2006:
There is huge uncertainty here. If China and other EMs stop intervening
in the foreign exchange markets, then the drain on reserves and on
global liquidity would soon end as EM exchange rates fall towards their
equilibrium levels. Alternatively, private sector EM capital outflows
might end spontaneously, as they did after the “taper tantrum” in 2013.
But the deterioration in economic fundamentals in the EMs looks more
serious than in 2013, so the current shock could be long lasting.
What then? Some economists, like Matthew Klein at FT Alphaville, and
Paul Krugman, argue that sales of bond holdings by foreign central banks
do not matter anyway. Krugman argues that they hold bonds of very short
maturity, which are close to cash, and he points out that sales of these
bonds can always be easily offset by Fed action to hold short rates
down. But the average maturity of US bond holdings by foreign central
banks, at 3.95 years, is not negligible. Furthermore, since the Fed is
thinking about raising rates, it may not want to offset the impact of
foreign bond sales on US medium dated bond yields.
In my opinion, one of the few analytical lapses made by the Keynesian
camp after 2010 has been a reluctance to believe that QE – or bond
buying by foreign central banks – could impact asset prices and economic
activity, except through a signalling effect about the future path of
Fed short rates. Yet studies by the Fed  and the ECB  suggest that
these bond purchasing programmes have had important effects on yields
through “portfolio balance” effects, as private investors are induced to
extend bond duration and hold riskier assets.