Flows & Liquidity Investors are not overweight bonds:: JPMorgan

We estimate that central banks and commercial banks currently
hold $24tr of bonds or 55% of the global bond universe.
 Non-bank investors, which invest in both bonds and equities, hold
the remaining $20tr of bonds globally, 30% of their combined
holdings of bonds and equities.
 This 30% bond allocation by non-bank investors is marginally
below the historical average since 2002. This casts doubt on the idea
that non-bank investors are overweight bonds globally.
 Global share buybacks are up 50% from a year ago. They are on
track to reach $590bn this year. The previous high was in 2011 at
$640bn and before then in 2007 at $900bn.
 A negative ECB deposit rate will further hurt the Euro area money
market fund industry, which has already contracted by €70bn since
last August.
 Peripheral banks could benefit from a negative deposit rate at the
expense of core banks. Core banks are currently “long cash” by
around €430bn, while peripheral banks are “short cash” by €650bn.
 Italian and Spanish trading volumes continue to normalize.

�� -->

 The BoJ’s asset purchase program is re-igniting questions regarding the
share of the bond market owned by central banks. Indeed, G4 central
banks and FX reserve managers currently own $14tr or 32% of the global
bond universe, which we proxy by the $44tr market value of Barcap’s
Multiverse Bond Index. The central bank share was 15% in 2002 (Figure
1). This year’s planned bond purchases by the Fed and the BoJ will raise
this ratio further towards 35%.
 Of this $14tr of central bank bond holdings, $5.2tr is owned by G4
central banks and $8.7tr is owned by FX reserve managers. We exclude
short term debt from these calculations by removing Tbills from G4
central banks holdings. For FX reserve managers we use information
from the IMF’s annual Coordinated Portfolio Investment Survey.
According to this survey 80% of the $11tr of global FX reserves was
invested in long-term debt in 2011 and this long-term debt share has
stayed relatively stable since 2001.

 But it is not only central banks that are large holders of bonds. Commercial
banks are also large holders. G4 commercial banks held $12.6tr of debt
securities as of last February. Admittedly some of this $12.6tr of debt holdings
is short term debt. Unfortunately we do not have the share of short-term debt
held by commercial banks. For simplicity we assume that 80% of their debt
securities are bonds, i.e. we apply a similar long-term debt share as the FX
reserve managers.
 Under this assumption, banks, i.e. G4 central banks plus G4 commercial banks
plus FX reserve managers, likely hold $24tr of bonds or 55% of the global
bond universe. This share was 40% in 2002 (Figure 1). This is not to say that
FX reserve managers only buy bonds but their allocation to equities is very
small, below 2% according to the IMF’s Coordinated Portfolio Investment
Survey (see section below on the equity buying by central banks). Similarly
commercial banks sometimes hold equities, but these equity holdings typically
represent stakes in other banks. It is rather unattractive for commercial banks
to hold equities because of their high risk weighting vs. zero risk weighting for
government related bonds.
 By excluding banks, i.e. entities that typically invest in bonds rather than
equities, the amount of bonds held by the rest of the world, i.e. non-bank
entities, is around $20tr. This compares to a global equity universe of $47tr,
based on DataStream’s global equity index. That is, non-bank entities, which
invest in both bonds and equities, have an allocation to bonds that is 30% of
their combined holdings of bonds and equities. Figure 2 shows the evolution of
the bond allocation of non-bank entities since 2002. The current allocation of
30% is marginally below the historical average since 2002. While this analysis
provides no conclusive evidence of how non-bank entities are positioned
globally, Figure 2 casts doubt on the idea that non-bank investors are
overweight bonds.
Central banks buy a small amount of equities
 Record low bond yields are pushing even conservative central banks into
higher yielding riskier asset classes. Recent central bank announcements on
the composition of their reserves have shown some central banks have even
increased their allocation to equities. For example, as part of the BoJ’s efforts
to reflate the Japanese economy they will more than double their holdings of
ETFs to ¥3.5tr ($36bn) by 2014. The Bank of Israel bought equities for the
first time ever last year and they now aim to double their equity holdings by
the end of this year to around $4.5bn or 6% of their total reserves. The SNB
also increased their equity holdings last quarter, from 12% to 15% adjusted for
valuation effects (see The SNB puts more faith in equities and non-G4
currencies, Paul Meggyesi, 30 Apr). Other central banks from South Korea to
the Czech Republic have also boosted their equity holdings over the past few
years.
 Sovereign wealth funds are much more active in equity markets than FX
reserve managers, and they too appear to have bought more equities in Q1.
The Norwegian petroleum fund invested around $13bn into equity markets in
Q1, which combined with the roughly $33bn gain on its existing equity
investments, brought its total equity allocation up to 62%, inline with the
fund's long run target equity allocation. For comparisons sake, the SNB’s
increased equity holdings amounted to around $9bn.
 Unfortunately, few sovereign wealth funds or reserve managers report their
securities purchases in a very timely fashion, so it is difficult to gauge the full
extent of equity buying by the official sector. However, to give an order of
magnitude for potential buying. The Norwegian petroleum fund’s Q1 equity
purchases were around 2% of their Q4 assets. Total global sovereign wealth
fund assets are around $4tr, so if all sovereign wealth funds bought equities to
the equivalent of 2% of their assets, this would be around $80bn, or around
half the equity flows into global mutual funds and ETFs in Q1.
Share buybacks and Q4 Flow of Funds
 This week’s release of Q4 Flow of Funds data for the Euro area helps us to
shed light on the evolution of corporate and financial sector flows at the end of
last year across the whole of the G4, i.e. the US, Euro area, UK and Japan.
 G4 non-financial corporates appear to have retrenched further in Q4
as shown in Figure 3. Both cash flows and capex declined in nominal terms
across all G4 regions. The non-financial corporate sector, which was the
locomotive of global growth in the 2-3 years following the Lehman crisis,
has been reducing its capital spending since the end of 2011, either due to
high uncertainty or in response to weaker final demand. Since Q4 2011,
cash flows, i.e. profits, are down by €180bn or 5.2%, while capex, which
includes inventory accumulation, declined by $100bn or 3.2%. So
corporate savings are actually down by $80bn since then. This cushions
somewhat the downbeat message from declining cash flows and capex in
the sense that the corporate sector is saving less.
 As mentioned before, despite this rather gloomy backdrop there are some
tentative signs of optimism. Our economists argue that higher frequency
indicators of capex are improving for the most recent months suggesting
that Q1 Flow of Funds data will likely show a rise in capex when the data
become available in three months time. In addition, the above capex
figures represent nominal flows which mix business fixed investment with
inventory accumulation. And the combination of higher business fixed
investment and lower inventories in Q4 created a positive backdrop for Q1.
 Net equity withdrawal was also disappointing in Q4 as it fell sequentially
from Q3 (Figure 4). For 2012 as a whole, net equity withdrawal totaled
$340bn, 35% below the $520bn seen in 2011. This is despite a rise in share
buyback activity quarter on quarter. Net equity issuance is the combined effect
of M&A, LBO and share buyback activity and the more negative it is the more
pronounced a buyer of its own equity the corporate sector is. Net equity
issuance measures net equity withdrawal while share buybacks are reported on
a gross basis and they do not take into account employee stock programs or
other stock issuance.
 Again there are reasons for optimism for this quarter. Share buyback activity
jumped to $168bn in Q1 globally vs. $100bn in Q4 and $74bn in Q3. So net
equity withdrawal should also rise when Q1 Flow of Funds data become
available in three months time. Figure 4 shows that there is decent correlation
between the two. But there are two caveats with expecting a big increase in net
equity withdrawal in Q1. The first caveat is that the share buyback figures
typically reported are announced rather than actual buybacks. This is one
reason why there is no perfect match in the timing between the two series in
Figure 4: reported share buybacks are announced while net equity issuance
reflects actual share purchases. The second caveat is that the current level of
announced share buybacks is not materially different from the level of net
equity withdrawal we saw in Q4 according to Figure 4.
 Finally the share buybacks announced so far this year are high but not
exceptional. Figure 5 shows the share buyback activity both for the first four
months of the year and the full year. The volume of the first four months of the
year is 50% higher than a year ago but it is not unprecedented by historical
standards. It was higher during the first four months of 2011, 2008 and 2007 as
shown in Figure 5. Of these three years, 2007 and 2011 saw very high share
buyback activity for the full year. There is statistically decent correlation
between yoy changes in share buybacks for the full year vs. yoy for the first
four months of each year. Mechanically, the regression model of Figure 6
suggests that the 50% yoy increase in share buybacks for the first four
months of the year translate to a 30% yoy increase for the full year, i.e. it
implies global share buybacks of $590bn for the full year. The previous high
was in 2011 at $640bn and before then in 2007 at $900bn.
 The other important information in the Flow of Funds is regarding the
investment behavior of pension funds and insurance companies. During Q4,
G4 pension funds and insurance companies bought $116bn of bonds and
sold $28bn of equities (see Chart A22). That is, there is no evidence of a
“Great Rotation” in Q4. Chart A23 also shows that the allocation to bonds and
equities was practically unchanged in Q4. The allocation by pension funds and
insurance companies to bonds and equities experienced a structural break post
Lehman. Bond allocations went up to 50% after Lehman vs. 40% before the
Lehman crisis. The inverse happened to equity allocations. Equity allocations
went down from around 40% to 30%. These new allocations, 50% in bonds
and 30% in equities, have been remarkably stable since 2009 according to
Chart A23.
 The stability in their bond allocations makes pension funds and insurance
companies contrarian to bond selloffs, effectively limiting the extent of
these bond selloffs. The reasoning is that if yields rise and bond prices fall,
bond allocations will decline sharply, especially if equity prices go up at the
same time. Under this scenario, if pension funds and insurance companies
stick to their 50% target bond allocation, they will have to buy bonds to
maintain their target bond allocation. And this appears to have happened in Q4
as pension funds and insurance companies bought bonds and sold equities.
 Finally Q4 Flow of Funds confirmed the high ability of non- financial
corporates to service their debt, a signal which is supportive for corporate
credit. The interest expense over profits ratio for non-financial corporate
remained at a low level of 15% for the G4 and 20% for US non-financial
corporates. These are levels seen before during mid phases of credit cycles, in
mid 1990s and mid 2000s (Figure 7).
A negative ECB deposit rate to drain liquidity from the
Euro area banking system
 This week’s ECB press conference reignited expectation of a cut in the ECB’s
deposit rate to negative. But a hurdle for a deposit rate cut to negative is high
not least because of the unintended consequences that negative rates entail.
 An obvious negative consequence is the potential damage to the money market
fund industry and repo markets. We highlighted these unintended
consequences in the past in F&L, July 6th 2012. Last summer’s move by the
ECB to cut the deposit rate to zero had led to the closure to new money of a
number of European money market funds. This is because with rates at zero,
these funds would have a negative yield after fees, which presents a significant
obstacle to investors. Since August last year, the outstanding amount of the
Euro area money fund industry had shrunk by €70bn or 7% to €900bn as of
last February. That process will certainly accelerate if the ECB were to cut the
deposit rate to negative. In the US, the low-for-long policy introduced
following the crisis has already hurt the money market fund industry. Since
January 2009, US money market funds have lost $1.3tr or 34% of assets. The
big beneficiaries of this have been bond funds. Since 2009, flows into bond
funds have been the mirror image of money market funds.
 The other unintended consequence is in repo markets. A shrinkage of the
money market fund industry hampers liquidity in money markets as the former
are important participants. We have mentioned in the past three impacts on
repo markets: 1) Narrowing spreads in the repo market are likely to induce
some lenders of collateral to draw back from the market, on the grounds that
returns from securities lending are no longer adequate. That would impair
bond market liquidity, by making it more difficult to cover short positions in
repo. 2) Incentives to cover shorts are reduced at zero rates which can result in
higher fails volumes than seen before, hampering liquidity and volumes even
further. 3) Some counterparties (mainly real money) may not be able to trade
repo at negative levels (operational, legal, economic reasons etc) which will
again likely reduce repo market volume/activity. Indeed the European repo
market contracted in H2 2012 according to the latest ICMA repo market
survey.
 Another unintended consequence is the damage to the profitability of banks
and the risk that banks increase lending rates. Assuming that banks will be
reluctant to pass negative rates to retail or corporate depositors, they might
increase lending rates to offset the decline in their profitability. This is
especially true for core banks that hold lots of deposits with the ECB such as
German, Dutch and French banks which held €223bn, €152bn and €98bn of
deposits with the ECB, respectively, as of last March (Figure 8).
 In fact a deposit rate cut will likely drain liquidity from the Euro area banking
system and accelerate the early repayment of LTRO funds. Negative interest
rates on deposits incentivize banks to “get rid” of their excess deposits rather
than suffer an erosion of capital. But the amount of reserves in the system can
only be changed if Euro area banks decide to collectively reduce their reliance
on the ECB. If a commercial bank makes a loan or buys a bond to avoid
negative rates, they simply pass reserves on to another bank, which ultimately
end up back at the ECB. As such, excess reserves would become something of
a ‘hot potato’, with no bank wanting them at the end of the day. Core banks are
more susceptible to negative deposit rates. Core banks have €620bn of deposits
and have borrowed €190bn (gross) from the ECB. In other words they are
collectively “long cash” by around €430bn (€620bn - €190bn). In contrast
peripheral banks are “short cash” by €650bn.
 The first response by core banks would be to repay most of the extra funds
they borrowed via the 3y LTROs. But for the excess deposits in the Euro area
banking system to decline further, we would need to see peripheral banks
reducing their reliance on the ECB. The resulting search for yield and increase
in the velocity of reserves, i.e. passing on the “hot potato”, within the Euro
area banking system has the potential to improve capital flows back to
peripheral banks and reduce TARGET2 imbalances further, especially now
that the ECB has back-stopped the system with the OMT. From this
perspective, negative deposit rates, could be a useful policy tool to allow the
periphery to reduce its reliance on the ECB and ultimately induce financial reintegration
in the Euro zone, even if they result in a reduction in liquidity. And
the ECB has the tools to cushion the reduction in liquidity i.e. excess reserves
by cutting reserve requirements to zero. This mechanically increases excess
reserves in the Euro area banking system by €100bn ceteris paribus.

Italian and Spanish trading volumes continue to normalize
 April data from MTS showed a small rise in Italian bond trading volumes and a
small fall in Spanish bond trading volumes. Both Italian and Spanish bond
volumes have been steadily rising since around the middle of last year, as
activity in peripheral bond markets started to normalize following the ECB’s
OMT announcement (Figure 9). Spanish bond trading volumes are actually back
to the levels seen before the onset of the Euro area crisis, while Italian volumes
have lagged behind somewhat recently, due to the uncertainty created by the
Italian election result. The latest rise in volumes likely reflects renewed
confidence following the formation of a government, and activity in Italian bond
markets appears back on track to join Spain in returning to pre-crisis levels.
Clarification: Short-sale ban on sovereign CDS
 In a piece last week we showed that since the short-sale ban on sovereign CDS
was imposed, core European countries have seen a dramatic reduction in net
notional CDS outstanding and that the ban was an important factor in this.
Also, that the only major sovereign CDS market that is not contracting is Italy.
 We also looked at whether net notional CDS outstanding on European
financials has changed significantly since the ban came into effect to look for
evidence of a migration from sovereign to financial CDS. It is worth clarifying
that in using DTCC’s net notional data, which is a metric of aggregate market
size, it is not possible to distil the myriad of trading behaviors in CDS markets,
or changes to them since the ban came into effect – net notional can change for
when either long or shorts unwind positions, including trade compression and
clearing, which complicates a simple analysis.
 Our analysis was not to suggest that the short-sale ban has had little or no
effect on CDS markets and/or trading behavior – quite the opposite – and in
the past we have extensively written about the problems, shortfalls and
unintended consequences of short-sale bans. For instance, we have commented
that short-selling bans in European bank stocks appeared to hurt liquidity
rather than support stock prices (see F&L Mar 3, 2012 and July 27, 2012), that
the short-sale ban in sovereign CDS has been a factor in the demise of the
SovX CDS index (F&L Jun 22, 2012), and we also gave a holistic account of
recent short-sale bans and their macro implications, including the sovereign
CDS ban just before it came into effect last year (F&L Oct 26, 2012).
 Our analysis last week was about whether there had been a major migration
from sovereign to financial CDS, as one might expect to observe an increase in
financial net notional data, and that the evidence when viewed from this
perspective is inconclusive. For instance, net notional CDS outstanding on
single-name EU sovereigns, iTraxx senior financials and Intl dev financials
have all fallen since August 2012, but proportionally CDS in EU sovereigns
has fallen significantly more (Figure 10). This suggests that there hasn’t been a
1-for-1 migration from sovereign to financial CDS. While some investors may
be using financial CDS to short sovereigns, this cannot be discerned from
publically available data.
 It is also worth noting that trading volumes in sovereigns have clearly been
more affected by the ban than in financials, as was anticipated before the ban.
For instance, Figure 11 shows that trading activity in EU sovereigns has fallen
since the ban, and continues to do so. By comparison, trading volumes in
iTraxx financials has rebounded since the start of the year to levels seen before
the ban. The extent to which this divergence is due to the short sale ban is
difficult to determine.