BNP ECB QE Revisited

www.GlobalMarkets.bnpparibas.com Please refer to important information at the end of the report.
Market Economics | Credit | IR & FX Strategy

ECB QE revisited
17 July 2014

Page

2-3
We analyse the market implications of the reduced probability of ECB QE.

Economics
QE: Failure to launch 5-14
The probability of a broad-based asset purchase programme has fallen following Junes package of measures, which
the ECB is confident will be sufficient to ward off the threat of deflation. We are less convinced and still see good
reasons for doing more. But this is now a risk rather than a central scenario and we have changed our policy forecast
accordingly. We look at the reasons the ECB is reluctant to launch such a programme.

Implementation of TLTROs
The four phases 15-17
We have split the eight available TLTROs into four distinct phases. Banks whose recent de-leveraging has outpaced
their 12-month average might want to use their maximum allowance in Phase I, while banks which have recently
experienced steep de-leveraging could find it easier to borrow during Phase II than Phase III.

Market implications
FX: No QE implies slower EUR depreciation, but stay short
19-20

Money market and ccy basis: Still ample excess liquidity
21-22

Core/semi-core rates and inflation: ZIRP to keep rates and breakeven inflation low
23-25

Peripheral bonds: Worse risk/reward
26-30

Credit: No QE, no cry
31-33

Contacts
Paul Mortimer-Lee Global Head of Market Economics +44 20 7595 8551
paul.mortimer-lee@uk.bnpparibas.com
Ken Wattret Co-Head of European & CEEMEA Market Economics +44 20 7595 8657
kenneth.wattret@uk.bnpparibas.com
Laurence Mutkin Global Head of G10 Rates Strategy +44 20 7595 1307
laurence.mutkin@uk.bnpparibas.coml
Patrick Jacq Europe Strategist +33 1 4316 9718
patrick.jacq@bnpparibas.com
Shahid Ladha Head of UK Strategy & European Inflation +44 20 7 595 8573
shahid.ladha@uk.bnpparibas.com
Ioannis Sokos Europe Strategist +44 20 7595 8671
ioannis.sokos@uk.bnpparibas.com
Camille de Courcel Europe Strategist +44 20 7595 8295
camille.decourcel@uk.bnpparibas.com
Olivia Frieser Global Head of Credit Research and Sector Specialists +44 20 7595 8591
olivia.frieser@uk.bnpparibas.com
Greg Venizelos Senior Credit Strategist Europe +44 20 7595 8296
greg.venizelos@uk.bnpparibas.com
Daniel Katzive Head of FX Strategy North America +1 212 841 2408
daniel.katzive@us.bnpparibas.com

Non Independent research - Marketing Communication
Overview

17 July 2014
ECB QE revisited 2 www.GlobalMarkets.bnpparibas.com

The probability of the ECB implementing QE in the form of a broad-based asset purchase
programme (BAPP) has fallen, following the package of measures announced in June, and this
is no longer our central case.
The TLTRO programme, without a BAPP, implies a somewhat different outlook for financial
markets from our previous central case of ECB buying of sovereign bonds.
The still-ample provision of liquidity via TLTROs means that we still expect the money-
market curve to stay very flat, and the cross-currency basis to fall and the euro to weaken, albeit
by less than previously expected.
But without ECB buying of sovereign bonds, we are now considerably less optimistic
regarding the outlook for inflation expectations and sovereign spreads. We now expect that
inflation expectations will continue to languish, keeping nominal core rates low, and we no
longer expect a further tightening of peripheral spreads this year.
These points are discussed in more detail below.
Note: For the sake of clarity, in this publication we use the term broad-based asset purchase
programme (BAPP) to describe large-scale ECB buying of assets, including sovereign bonds,
for monetary policy purposes, as well as the less precise term quantitative easing (QE).
Economics
QE: Failure to launch
The chances of the ECB launching a broad-based asset-purchase programme have fallen,
given Junes package of measures. With inflation to remain uncomfortably low, the ECB will
keep QE as a contingency option but unless forecasts change, it will probably not deliver.
While we had long expected persistently low inflation to prompt the ECB to deliver a range of
policy initiatives, key aspects of Junes package went beyond our expectations. The
exceptionally long maturity of the TLTROs and their limited conditionality were the main
surprises and could generate similar effects to QE in some areas.
Recent comments from a range of ECB officials highlight the central banks high level of
confidence that the measures already announced will have substantial effects. We are less
convinced. We continue to see downside risks to the ECBs inflation projections and believe
there is a case for doing more.
There are various potential shocks that could yet force the ECB into a broad-based asset-
purchase programme. Inflation, growth and the exchange rate are key swing factors.
We look in detail at the ECBs underlying objections to a QE-style programme.
The four phases
- Phase I: Sep-14 and Dec-14 based on stock of qualifying loans
- Phase II: Mar-15 and Jun-15 based on existing trajectory of lending
- Phase III: Sep-15 to Jun-16 based on trajectory of lending vs April 2015
- Phase IV: Sep-16 mandatory repayment test based on cumulative lending
Banks whose recent de-leveraging has outpaced their 12m average might want to use their
maximum allowance in Phase I, as they may not be able to beat the benchmark in Phases II
and III.
Banks which have recently experienced steep de-leveraging (eg Spanish) could find it easier to
borrow during Phase II than in Phase III.

Overview

17 July 2014

Market implications
We raised our EURUSD forecast targets modestly, to reflect reduced prospects for ECB QE.
We continue to expect the EUR to trade broadly weaker in H2. We favour EURUSD and
EURGBP shorts and recommend funding risk and carry positions in EUR.
TLTROs will lead to a smaller extension of excess liquidity with a slightly shorter duration than
would be the case with a BAPP, but ample excess will remain.
The money-market curve therefore looks too flat to us.
EURUSD cross-currency basis will not widen as much as with QE, but should still reach -12bp
in 5y by end-2014.
Core/semi-core rates and inflation: ZIRP to keep rates and breakeven inflation low
ZIRP implies low volatility and low core rates: we expect 5- and 10-year to further outperform on
the curve. Without QE, the market will remain sceptical that inflation will revive, keeping break-
even inflation and core nominal rates under downward pressure.

Semi-core and SSA spreads should benefit from a zero interest rate policy: we maintain our
forecasts for tighter SSA & semi-core spreads.
TLTROs differ from both 3y LTROs and QE in two important ways:
1. The size and use of the borrowed funds from TLTROs will remain uncertain, providing a
less transparent exit point for investors who have been purchasing peripherals in
anticipation of QE.
2. TLTROs reinforce the link between sovereigns and banks whereas QE could have
weakened them, setting the ground for a non-zero risk weighting on sovereign bonds.
Absence of a direct ECB bid for peripherals should lead to higher spread volatility and so to a
higher risk premium in peripheral spreads
We now anticipate that 10y Spain and Italy spreads will end 2014 at +150bp over Bunds (vs our
prior expectation of +100bp). Further, as we expect higher volatility given lower probability of
QE, we do not see peripheral spreads as cheap enough to buy below 180bp.
We expect 3y peripheral spreads to be better-supported than the 10s, due to the nature of
TLTROs. We now target +60bp at end-2014 (vs +40bp prior expectation)
The implications of a no ECB QE scenario for the European credit markets are a reduced
degree of spread tightening in H2 2014 and less spread compression of high beta vs low beta,
Periphery vs Core, Financial vs Non-Financial and European vs US credit spreads.
Furthermore, we may see some steepening in credit curves, particularly in high beta and
periphery risk.
We expect a relatively high take-up for the TLTRO, although the EUR 1trn maximum amount
mentioned by ECB President Draghi is higher than our estimates. There is no stigma in using
the facility in our view, given that it is conditional to lending to the economy. Although, the
Central Banks may put pressure on the banks to use the TLTRO.
That said, we are not convinced that this will lead to a significant increase in lending, especially
to SMEs. The TLTRO is not that targeted after all. Banks can continue to embark on the
sovereign debt carry trade and the net lending benchmark can be met/exceeded via an
improvement in net lending to large corporates and/or consumers rather than SMEs.

17 July 2014

Economics

17 July 2014
The probability of the ECB launching a broad-based asset-purchase programme has
fallen following the package of measures announced in June.
With inflation to remain uncomfortably low, the ECB will keep QE as a contingency
option. But there is a reluctance to deliver and we have revised our forecast accordingly.
We no longer forecast such a purchase programme in our central scenario.
While we had long expected persistently low inflation to prompt the ECB to deliver a
range of policy initiatives, key aspects of Junes package went beyond our expectations.
The exceptionally long maturity of the TLTROs and their limited conditionality were
the main surprises and could generate similar effects to QE in some areas.
Recent comments from a range of ECB officials highlight the central banks high
level of confidence that the measures already announced will have substantial effects.
We are less convinced. We continue to see downside risks to the ECBs inflation
projections and believe there is a case for doing more.
There are various potential shocks that could yet force the ECB into QE. Inflation,
growth and the exchange rate are key swing factors.
We look in detail at the reasons why the ECB is reluctant to embark on QE.

ECB forced to act by persistently low inflation
Since late last year, a key theme of our eurozone analysis has been the likelihood of persistent
downward surprises in inflation prompting the ECB to deliver additional policy easing, in both
conventional and unconventional form, including asset purchases.
It took a long time for the ECB to react to persistent sub-1% inflation. On 5 June, an unusually
broad range of policy measures was announced, comprising five of the six policy options
included in the contingency framework put forward by ECB President Mario Draghi on 24 April.
The introduction of a negative deposit rate and an asset-purchase programme for asset-backed
securities (ABS) did not come as a surprise to us, nor did the extension of the fixed-rate, full-
allotment procedure for all refinancing operations (though the latter went beyond our
expectations, going out to December 2016 at least). Ending the sterilisation of asset purchases
made under the Securities Markets Programme (SMP) had also been flagged to markets by the
ECB some months beforehand.
Junes package reduces the probability of further action
The main surprises were in the area of very long-term liquidity provision for banks. The ECBs
decision to introduce targeted long-term refinancing operations (TLTROs) with a fixed rate was not
surprising; it was part of Aprils contingency framework. The surprises were in the exceptionally
long maturity of the loans (up to four years) and the limited conditions attached to them.
In tandem with the other measures announced, the TLTRO initiative and its potential impact on
asset markets and bank lending have reduced the probability of the ECB implementing the only
policy option cited in Aprils framework that was not included in early Junes package, namely, a
broad-based asset purchase programme (BAPP).
The chances of the ECB launching a BAPP have diminished and we have revised our forecast
accordingly. We no longer forecast such a programme in our central scenario. The ECB will,
however, continue to dangle the prospect in front of markets and, in the event of a significant
shock (see below), we would expect a QE-style programme to be the cornerstone of the ECBs
response. Below, we look in detail at a range of issues related to our change of forecast.
Unusually broad range of
measures in June
Length of TLTROs was the
big surprise
Chances of a BAPP have
radically diminished

17 July 2014
We had forecast that the ECB would announce a QE-style programme in September, in tandem
with another round of downward revisions to the inflation projections. The recent package of
policy measures has radically reduced the likelihood of this happening.
Some of the measures announced in June (the TLTROs and the ABS purchases) will only take
effect with a lag and it is clear that the ECB wants to take some time to assess their effects. In
the absence of a significant shock, therefore, the ECB is likely to focus on its existing initiatives
until at least Decembers second TLTRO and perhaps the third window in March 2015 (the first
of the operations in which the take-up will be determined by banks net lending). That far down
the line, the window for a BAPP is more likely to have closed, as we expect inflation and growth
to be somewhat higher by then, while the EUR is also forecast to depreciate.
as reflected in recent communication
The ECBs confidence in the impact of Junes measures and its desire to take time to assess
their effects were prominent at Julys press conference. ECB President Mario Draghi stated on
more than one occasion that the impact of Junes package would be substantial and that has
been echoed in other speeches since by a range of ECB officials.
Executive Board member Benoit Coeur gave little encouragement to the prospect of a BAPP in
an interview with the Financial Times on 7 July. "I see the odds as being low ... I am very
convinced that what we decided already will work and will prop up inflation," he said.
Having put so much faith in Junes package, it is going to take a long time (or a major shock) for
the ECB to accept that it is has not done enough.
Arguments against QE
It is now 2014. The Fed started to conduct QE in 2008. The ECB is clearly not that enamoured
of QE or it would have undertaken it by now. What are the underlying objections to QE and the
more immediate reasons for not wanting to implement it?
The deepest reason for not wanting to quantitatively ease is the experience in the 1930s, in
particular in Germany, where central bank financing of the government resulted in hyper-
inflation. This makes the Bundesbank and others very cautious about embarking on such a
strategy. It also means that the acceptability of the policy among the German public is low.
Once the Rubicon is crossed, there is a danger that central bank finance for government could
become easier on subsequent occasions. Second is the fear of moral hazard in that central
bank financing of governments reduces the pressure on governments to get their fiscal house in
order. This does indeed seem to be a risk given that the OMT seems to have reduced the
pressure on governments to do the right thing.
Against this backdrop what are the more immediate reasons for wishing to avoid QE, except in
extremis?
It wont work because of the instability of the money multiplier.
QE works, so the argument runs, by increasing the monetary base, because the money
multiplier moves in a way to offset variations in the monetary base almost completely (see
Chart 1). We think this is too narrow an interpretation of how QE works and in any case surely
means TLTROs will not work either.
Bond yields are already very low and QE will not lower them further.
We have some sympathy with this, though we would point out that qualitative and quantitative
monetary easing (QQME) in Japan, where yields were even lower, did lower bond yields (see
Chart 2). This ignores the fact that there is more than one bond yield in the eurozone. The
German bund yield may not fall much further, but QE would bring down the much higher real
bond yields in the periphery (see Chart 3).
The 5 June package has done all that needs to be done.
We do not buy this. TLTROs continue to try to work through bank liquidity measures. They
may work better when bank capital funding issues have been addressed though the asset
quality review (AQR). However, the demand side of bank lending remains weak, even if the
supply side is being supported, and banks attitude to risk (NPLs) remains cautious. QE would
ECB confident in effects of
June measures
Deep fears of central banks
financing governments
It wont work?
Bond yields already low
June 5 was enough

17 July 2014
add liquidity and increase money supply ahead of a recovery in loan demand and would
symbolize a determination for the ECB to be active not passive, thereby boosting demand.
Inflation in Germany is stable around 1% (true) and the adjustments in the periphery are
merely a relative price change. When this is complete, inflation will pick up.
German inflation has surprised on the downside on a number of occasions. It looks like
German inflation is not only partly a function of the German output gap but also partly a
function of the eurozone output gap. Anyway, it is not just peripheral countries that have low
inflation. Inflation in the Netherlands is only 0.3% and in France it is 0.5% in spite of a VAT
hike. There is much more than a relative price change going on here since if it were just that,
the average eurozone inflation rate would not have fallen in the way it has.
Long-term inflation expectations are stable.
The 5y5y implied inflation from the swaps market is drifting slowly down. The 5y measure from
the survey of professional forecasters has edged down to 1.8% (Chart 4). While these
measures have not plunged, the fact that they are declining shows that faith in the ECB to
achieve price stability is fading, due either to perceived unwillingness or an inability to do the
job. In Japan inflation expectations became unanchored after deflation had already set in,
suggesting inflation expectations were at least partially backward looking.
The ECB does not need to conduct QE until deflation is an imminent threat.
This means it could already be too late for QE, as it will be likely to occur only in the most
adverse circumstances.
Monetary policy has reached the limit of its capabilities.
This is a counsel of despair with which we disagree strongly. For a start, the ECB clearly thinks
it could work in extreme circumstances, which is when they have said they would deploy it. If it
will work in such circumstances, why not in less challenging times? Even if we accept that
monetary policy has limited ability to narrow the output gap significantly, the really important
transmission channels in our view that the ECB seems to want to turn a blind eye to are the
currency and inflation expectations.
German inflation is just
fine
Inflation expectations well
anchored
Monetary policy has
reached its limits
Chart 1: Eurozone money multiplier

Chart 2: JGB yields
3
4
5
6
7
8
9
10
11
12
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Multiplier
(M3 / base money)
Base money
(right axis)
EUR bn

Source: Macrobond, BNP Paribas Source: Reuters EcoWin Pro, BNP Paribas
Chart 3: Real interest rate differential

Chart 4: Eurozone inflation expectations

Source: Macrobond, BNP Paribas Source: Macrobond, BNP Paribas

17 July 2014
To us, the short-term arguments look more like excuses than anything and stem from the
atavistic fear in parts of the ECB that QE will bail out fiscally weak governments and lead to very
high inflation. We would suggest this is not the experience in Japan, the US and the UK. There
are no good excuses for a bad policy.
Is Junes package enough?
We can already see beneficial effects of Junes measures on monetary conditions. Short-term
interest rates have fallen and the ECB has been successful in its attempts to dislocate policy-
rate expectations in the eurozone from those in the US and UK (Chart 5). The four-year maturity
of the TLTROs and lengthy extension of the fixed-rate, full-allotment procedure have been
pivotal in this regard.
Interest-rate divergence has not, however, done much to weaken the exchange rate. At the time
of writing, the EUR had fallen by around 2% on a trade-weighted basis (based on the ECBs
daily measure) from its high point prior to Mays dovish press conference. This is important, as
EUR appreciation has been a key reason for the tightening of monetary and financial conditions
in the eurozone, as measured by our own proprietary indicator, the Financial and Monetary
Conditions Index (FMCI, Chart 6). The FMCI for the eurozone has now started to loosen,
though it remains much less accommodative than the US and UK equivalents.
The ECB would argue, too, that key elements of Junes package of measures have yet to take
effect. The impact of a very large TLTRO take-up on the ECB balance sheet, for example, could
lead to more pronounced depreciation of the EUR.
TLTROs matter most
We see the TLTROs as crucial to the success, or otherwise, of Junes package of measures. In
the ECBs view, at least, the TLTROs will be very effective. In a speech on 3 July, Executive
Board member and ECB Chief Economist Peter Praet outlined his views on how the effects
would be seen through several channels, including the following:
Reduced term-funding costs for banks
A substitution effect on term funding in the market for bank bonds, creating a scarcity of
investible assets and reducing yields (easing funding conditions for banks, including those that
did not participate in the operations)
Knock-on effects on other segments of the corporate credit markets, lowering yields on a
range of bonds, in a similar way to the portfolio balance effect cited by other central banks
The conditionality of liquidity provision on the evolution of net lending will lead to an outward
shift in the credit supply curve, reducing the cost of borrowing while expanding new loans
Higher excess liquidity, reducing short-term interest rates.

Beneficial effects are
already visible
TLTROs key to success of
ECBs easing measures
Chart 5: Divergent policy-rate expectations

Chart 6: BNPP Financial and monetary conditions indices
09 10 11 12 13 14 15 16 17
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
Implied by Futures Strip
BoE
Fed Funds Rat e
BoE Bank Rate
Fed
ECB
ECB Refi Rate

00 01 02 03 04 05 06 07 08 09 10 11 12 13 14
L
o
o
s
e
r

T
ig
h
t
e
r
-4
-3
-2
-1
0
1
2
3
4
US
UK
Eurozone

Source: Reuters EcoWin Pro, BNP Paribas, ECB, BoE, Federal Reserve Source: Reuters EcoWin Pro, BNP Paribas

17 July 2014
We agree that the TLTROs will loosen monetary conditions in the eurozone. But we do not see
them as an effective substitute for a broad-based asset-purchase programme (see Box 1).
There is a lack of provisions in the framework to channel bank lending to the small and medium-
sized enterprise (SME) sector. Banks, therefore, may well favour expanding net lending to those
less risky, large non-financial corporates that are less in need of the assistance.
There are other potential drawbacks to the ECBs opting for TLTROs rather than outright large-
scale asset purchases. One is uncertainty over the take-up. While Mr Draghi talked about the
potential for a take-up of as much as EUR 1trn at Julys press conference, this will depend on a
number of factors, including, in part, banks willingness to expand net lending.
With outright purchases, there would more clarity about the scale of the intervention. One could
also argue that TLTROs do not have the same announcement effect (on inflation expectations,
in particular) of large-scale asset purchases. Breakeven inflation rates were unmoved by Junes
announcements.

Box 1: TLTROs and BAPP similarities and differences
Similarities
Both increase bank liquidity (though whether this would increase broad money is doubtful)
Both should lower yields in the bond market
Both should increase demand for government debt
Both should soften the exchange rate
Both should stimulate growth and reduce the inflationary threat

Differences
TLTROs work through the banking system, BAPP goes around it
TLTROs should boost private credit more, BAPP government credit
TLTROs should benefit corporate credit spreads more
TLTROs have less moral hazard with respect to government behaviour
TLTROs are self extinguishing because of their four-year term, bond purchases may mature much later
BAPP removes quantum of government debt in the market
Therefore has more powerful portfolio balance effect
Therefore reduces market risk by more
BAPP is generally at longer maturities than the four-year tenor of the TLTROs
Therefore lowers the longer end by more
TLTRO effect should be more at the short end
BAPP likely to have bigger effect on inflation expectations
BAPP likely to have bigger announcement effect on the exchange rate
BAPP would have been more immediate
Source: BNP Paribas

TLTROs not a substitute
for a BAPP
SME funding, TLTRO
takeup remain uncertain

17 July 2014
Is this QE by stealth?
Nonetheless, there are grounds for taking a more positive view of their broader effects. While
TLTROs are unlikely to be a game-changer for bank lending to SMEs, the light conditionality could
be a boost for asset markets. The first two operations in September and December will allow
banks to borrow up to EUR 400bn collectively for up to four years, with no restriction on what the
funds are used for, including carry trades into sovereign debt markets, even if their net lending
picks up only marginally relative to benchmarks (which have been set rather generously).
The perceived stigma previously associated with borrowing via the three-year LTROs should
also be less of an issue this time, as the ECB, as supervisor as well as lender, will presumably
be encouraging the banks to make full use of their allowances. The introduction of TLTRO
groups, whereby borrowing can be pooled between banks with close links, should also help
to alleviate eurozone financial system fragmentation.
One interpretation of the TLTROs could, therefore, be QE by stealth, replicating the effects of
the prior three-year LTROs, but with the added bonus of potentially increasing bank lending to
the corporate sector. Without the backstop of ECB purchases, however, there is a risk that
foreign buyers will become less keen to expand their holdings of sovereign debt, leaving yield
spreads vulnerable to widening pressures from current levels. This, in turn, could increase the
likelihood of the ECB having to launch a BAPP a scenario we discuss below.
Comprehensive assessment also important
The timing and impact of the ECBs comprehensive assessment of eurozone banks is also
important when considering the TLTRO framework. The ECB wants to ensure more effective
transmission of its monetary policy through the bank lending channel and will be more confident
that this will be achieved once the comprehensive assessment is completed in the autumn. (The
results of the asset quality review (AQR) and European Banking Authority (EBA) stress tests will
be published in October.)
A better-capitalised, more robust banking system should be more willing to extend credit to the
private sector, again reducing the need for the ECB to provide additional stimulus. The timing of
the comprehensive assessment was one of the reasons we thought that if the ECB were going
to be bounced into asset purchases by uncomfortably low inflation, it was more likely to happen
before the process was complete.
Being critical, however, one could contend that the ECB was focusing too much on the
impairment of its policy stance (which is the second aspect of Aprils contingency framework)
when its priority ought to be the third, namely, delivering a meaningful increase in monetary
accommodation via large-scale asset purchases in order to strongly counteract the risk of
deflation. The risk that the ECB may have left it too late remains a significant concern, in our
opinion.
ABS purchase programme has longer-term effects
A key feature of our call for more policy easing from the ECB this year was that it would include
asset purchases. While this forecast generated considerable scepticism when first presented, it
ultimately proved correct. The asset-purchase programme announced, however, is centred on
ABS and dovetails with the second aspect of the 24 April contingency framework, ie, impairment
of policy transmission.
While we still await the details of the ABS programme, in all likelihood, it will have little near-
term effect on monetary conditions. While Mr Draghi suggested at Julys press conference that
the pool of eligible assets was large he cited a figure of EUR 1.4trn for the whole of the EU a
chunk of that is in the form of residential mortgage-backed securities. With the ECB having
excluded mortgage lending from its TLTRO framework, it would be odd to for it to then intervene
in large scale in the RMBS market. Earlier housing bubbles, including in Spain and Ireland, are
a cautionary tale.
We continue to assume, therefore, that the size of the ABS buying programme will be similar to
that for covered bonds (ie, in the tens of billions) initially. The magnitude of the programme
could increase eventually as the market becomes more developed, but this will take time and
will be linked to regulatory, reporting and legal issues. While we believe the development of
alternative sources of finance for SMEs makes sense, this is really a long-term project with
Light conditionality should
boost asset markets
Lack of an ECB buying
backstop could be a worry
ECBs AQR and stress
tests also important
Too much ECB focus on
policy impairment?
ABS purchases will have
little near-term effect
ABS buying only in the
tens of billions initially

17 July 2014
limited immediate results. In that context, we view the ABS programme as a complement rather
than a substitute for QE.
This too is a reason for thinking that a BAPP could yet come into play. But as with the TLTROs,
the implementation of the ABS buying occurs with a lag, so it will be some time before the ECB
is able to assess its impact and conclude that more needs to be done.
Dangling the carrot
While we have concluded, on the basis of recent developments, that the ECB is now unlikely to
launch a BAPP in the absence of a shock, it will continue to dangle the prospect carrot-like in
front of the markets. Speeches have, and will continue to, make this clear. The idea presumably
is to emulate the effects of OMT by threatening to take action but not delivering. So, what could
force the ECBs hand?
Various shocks could put QE back in play
Various shocks could bounce the ECB into adopting a BAPP. Three are particularly important,
in our view: (1) lower inflation and inflation expectations, (2) a worsening of the economic
outlook, widening an already large, persistent output gap, and (3) tighter monetary and financial
conditions, either via a stronger exchange rate or due to renewed stress in other asset markets.
including downward surprises in inflation and inflation expectations
The ECB has expressed increasing concern about the potential risks associated with
persistently low inflation, citing three issues, in particular: (1) the diminishing safety margin, or
cushion, against a deflationary episode, (2) the more difficult competitiveness adjustment in the
periphery and the potential implications for debt sustainability, and (3) the adverse effect on real
interest rates and the monetary policy stance. Its rhetoric has changed markedly to reflect these
concerns, with the prospect of a BAPP explicitly linked to the risk of too prolonged a period of
low inflation.
How low is too low? A move close to zero in the headline HICP inflation rate from the already
low 0.5% currently would be a big deal in our view, though some Governing Council members
appear to set the threshold for a further policy response rather higher. In a recent speech, for
example, German Executive Board member Sabine Lautenschlaeger suggested that a BAPP
would only be considered in a true emergency such as imminent deflation.
In our forecast, we expect that headline inflation will go broadly sideways in the coming months,
before picking up a little in Q4 (albeit due to base effects). The ECB expects inflation, over time,
to rise back towards its below, but close to 2% definition of price stability. If this diminishes in
probability, the likelihood of the ECB having to deliver more policy easing will increase.
However, in various speeches recently, ECB officials have stated that there is a low probability
of the kind of persistent low inflation which would prompt a BAPP. We are sceptical about the
timing and the speed of the pickup in underlying inflation in Junes staff projections. The output
gap is still very large and we expect core inflation to remain lower for longer than the ECB
expects. Risks to price stability, contrary to the ECBs view, are still to the downside, in our
opinion.
Expectations are important
The extent to which inflation expectations remain unanchored will also be a pivotal influence on
ECB policy. Junes policy announcements will have reinforced its belief in the firm anchoring of
eurozone inflation expectations, though, again, we are less convinced. The five-year
expectation in the ECBs Survey of Professional Forecasters, while still below, but close to
2%, is fairly inert and conditional on the implementation of appropriate policy. In contrast, the
two-year survey has fallen to record lows and suggests that the ECB is losing credibility.
Market-derived measures of inflation expectations have fallen according to Mr Praet and market
participants seem to factor in a longer period of low inflation. We would argue, given this
backdrop, that the ECB should be more willing to take the more radical action, in the form of large-
scale asset purchases, required to raise inflation expectations.

ECB to continue to dangle
the carrot of a BAPP
Various shocks could lead
the ECB to adopt a BAPP
When is inflation too low?
Untethered expectations
to play a key policy role

17 July 2014
Large, persistent output gap
A key reason we are sceptical about the pickup in core inflation expected by the ECB staff is the
large, persistent output gap in the eurozone. Historically, such circumstances have not signalled
a turn in the inflation trajectory until much later.
Should the economic situation in the eurozone deteriorate significantly, this would increase the
likelihood of those disinflationary forces intensifying. This would undermine the credibility of the
projected pickup in inflation and, potentially, unanchor inflation expectations. An ECB response
in the form of a BAPP would, therefore, be very likely, in our view.
The weakening of some leading indicators of economic activity in the eurozone recently merits
close attention. However, the deterioration has been confined to manufacturing sector data to
date, with domestically driven indicators less affected. Moreover, in level terms, most of the
reliable gauges of growth (such as the composite PMI) suggest that while momentum has
stalled, the economy will continue to expand.
The pickup in eurozone growth we expect to see from late this year and through 2015 as
monetary conditions ease, bank lending picks up and global growth accelerates would, again,
be consistent with the view that if the ECB does not implement a BAPP fairly soon (which is
looking less likely), the chances of it happening will diminish over time.
Exchange-rate effects
While the ECB does not target the exchange rate, its impact on monetary conditions and
inflation make it a key swing factor when it comes to assessing the likelihood of a BAPP. The
ECB is probably disappointed that Junes measures have not had a bigger effect on the EUR,
though the effect of the TLTROs on the ECBs balance sheet will take time to filter through.
Moreover, the current EURUSD rate is below the ECBs assumption of a stable 1.38 rate in
Junes projections, which is very different to Marchs projections. Then, the EURUSD rate rose
well above the assumption of a constant rate of 1.36 (peaking just shy of 1.40 ahead of Mays
dovish press conference).
A 10% rise in the effective EUR rate would, other things equal, boost inflation by around 0.4pp.
Using the assumption of 1.38 in Junes ECB projections, a move up to 1.45 would probably be
required for the staff inflation projections to change markedly.
How EURUSD evolves will be sensitive to US developments and the ECB seems keen to wait
for a pickup in US economic conditions. This would improve eurozone growth prospects, boost
confidence and lift asset markets, while normalisation of Fed policy should help soften the EUR.
If the ECB does not embark on a BAPP in the next couple of quarters, therefore, as it wants to
take time to gauge the impact of Junes measures, the window is more likely to close on such
programme as, by early 2015, US economic and policy conditions could reinforce the ECBs
belief that there is no need for further action.
Market turbulence
The exchange rate is, of course, one potential route to tighter monetary and financial conditions
in the eurozone. Renewed turbulence in other asset markets could also increase the downside
risks to growth and inflation. For example, renewed tensions in sovereign debt markets related
to a deteriorating outlook for growth and fiscal dynamics.
This would raise the likelihood that the ECB ends up launching an asset purchase programme,
though there is a question as to whether OMT or a BAPP would be appropriate.
Not out of the woods yet
Our view is that the safety margin against deflation in the eurozone has worn perilously thin and
from a risk-management perspective, therefore, we believe there is a compelling case for the
ECB to do more. The eurozone is one economic downturn away from deflation, with potentially
serious consequences for debt sustainability. A large-scale purchase programme, with the clear
objective of raising inflation expectations, would be a sensible insurance policy at this point.
Some on the Governing Council probably favour taking out the insurance but they have been
silent so far.
The eurozones big output
gap makes us sceptical
The longer the ECB stalls,
the less likely a BAPP
Effects on EUR will take
time to filter through
EURUSD will depend on
US developments
New market tensions
could prompt a BAPP
Safety margin to deflation
is now perilously thin

17 July 2014
We think they should but expect that they will not and so there is certainly still a risk, in our view,
that the ECB will commit a policy error and end up doing too little, too late.
What if?
We have had little guidance from the ECB about the potential modalities of a BAPP, aside from
a few references here and there in some speeches. If the ECB was forced to launch this kind of
programme, the timing of it and the reasons for it would be pivotal to its size and composition.
We assumed in our scenario of a September announcement that the ECB would get away with
a relatively small shot of EUR 400bn, equivalent to around 5% of GDP (small, that is, compared
with the cumulative purchases of other central banks, rather than its first instalment). But this
was based on a scenario in which inflation was low and growth was improving and expected to
benefit from a pickup in global economic conditions thereafter. In other words, the ECB would
act somewhat pre-emptively in order to ward off the threat of deflation.
The bigger the shock and the longer the ECB stalls on delivering what we would consider to be
the most effective response of the options still available (QE), the greater the likelihood that the
asset purchase programme would have to be much larger in scale.
We calculate that on the basis of the purchases programmes of other central banks, it takes QE
of around 1% of GDP to achieve a 0.6% depreciation in the effective exchange rate. As a 10%
fall in the EUR effective exchange rate would raise inflation by around 0.4-0.5pp, this would
imply that the ECB needed a purchase programme of around 12% of GDP, or EUR 1.1trn.
The size of any programme influences significantly the type of assets bought. If small in scale
EUR 400bn, say the ECB could avoid buying sovereign debt if it really wanted to. The
corporate and supranational markets are sufficient to allow purchases of that scale without the
ECB dominating the markets. But if it were in the trillions, the programme would almost certainly
have to include sovereign debt. We continue to believe that capital key-weighted purchases
would be the cleanest way of avoiding legal challenges, if this proved necessary.
Conclusion
The likelihood of a broad-based asset-purchase programme has declined following the package
of measures the ECB announced in June. We have revised our call accordingly and no longer
include QE in our central scenario. Still, as inflation is going to stay uncomfortably low, it will
continue to be talked about as an option. While we had forecast that the ECB would deliver a
range of policy initiatives in the wake of persistent downward surprises on inflation, key aspects
of Junes announcement, in particular on TLTROs, went further than we had been expecting.
Moreover, recent communication from the ECB highlights its high level of confidence that the
measures will have substantial effects.
We are not so convinced. We continue to see downside risks to the staff inflation projections
and believe there is a compelling case for doing more. The ECBs willingness to launch QE in
the absence of a significant shock, however, appears to have diminished significantly.

Paul Mortimer-Lee & Ken Wattret

Time and reason would be
key to size and make-up
The longer the delay, the
bigger QE would need to be
The bottom line

17 July 2014

The four phases

This publication is classified as non-objective research

17 July 2014
The four phases
There are four distinct TLTRO phases:
- Phase I: Sep-14 and Dec-14 based on stock of qualifying loans.
- Phase II: Mar-15 and Jun-15 based on existing trajectory of lending.
- Phase III: Sep-15 to Jun-16 based on trajectory of lending vs April 2015.
- Phase IV: Sep-16 mandatory repayment test based on cumulative lending.
Banks whose recent de-leveraging has outpaced their 12m average might want to use
their maximum allowance in Phase I, as they may not be able to beat the benchmark in
Phases II and III.
Banks which have recently experienced steep de-leveraging (eg Spanish) could find it
easier to borrow during Phase II than in Phase III.

We have split the eight available TLTROs that will take place from Sep-2014 to Jun-2016 into
four distinct phases (Chart 1), with the first two TLTROs that will take place in Sep-2014 and
Dec-2014 constituting Phase 1. For these two TLTROs we already know the maximum take-up
is roughly EUR 400bn (Table 1 shows amount per country), which is 7% of the outstanding pool
of eligible loans at the end of April 2014 (static snapshot).

The next two TLTROs that will take place in Mar-2015 and Jun-2015 constitute Phase 2. We
have put these two TLTROs in a group of their own for two reasons. They are distinct from the
first two TLTROs because they depend on the evolution of net lending from May 2014 to the
reference month of each TLTRO (Jan-15 for the Mar-15 TLTRO and Apr-15 for the Jun-15
TLTRO) versus a pre-defined benchmark, while the first two TLTROs depend on the
outstanding amount of loans at a specific point (April 2014). They also differ from the last four
TLTROs (Sep-2015 to Jun-2016) in the way their benchmark is estimated (Charts 2 and 3 show
the projected benchmarks for Spanish and Italian banks).
Chart 1: Four TLTRO phases Table 1: Take up of first and second TLTROs
Phase 1
Phase 2
Phase 3
Phase 4

DE 23.7%
FR 19.3%
IT 18.9%
ES 13.5%
NL 7.3%
AT 3.8%
BE 2.5%
GR 2.4%
PT 2.1%
IE 1.9%
FI 1.7%
Others 2.8%
EuroArea 100%
TLTROsEligibleLoans(Apr14) TLTROsTakeup
97
5692
158
414
219
139
771
1099
1075
Eligible
LoansPool
140
1350
110
120
House
holds
128
1468
107
114
765
1123
NonFin
Corp.
116
904
87
96
122
605
863
836
373
163
98
117
5227
599
107
429
69
4331
Share
(%)
71
144
127
71
89
3866
TLTRO
Eligibility
(7%)
9.8
94.5
7.7
9.7
54.0
76.9
388
88
105
15.3
8.4
598
887
360
Loansfor
Housing
104
1022
83
398.4
75.3
29.0
11.0
6.8

Source: BNP Paribas, ECB Source: BNP Paribas, ECB
Chart 2: Implied lending benchmarks for Spanish banks Chart 3: Implied lending benchmarks for Italian banks
8th 7th 6th 5th

4th - 2015Apr
benchmark used for
TLTRO Jun-15
3rd - 2015Jan
benchmark used for
TLTRO Mar-15
2014Apr
7% of which is the
allowance for 1st &
2nd TLTROs
600,000
650,000
700,000
750,000
800,000
850,000
900,000
phase 1
phase 2
phase 3
Constant benchmarks
for TLTROs 5th to 8th
phase 4

8th 7th 6th 5th
4th - 2015Apr
benchmark used
for TLTRO Jun-15
3rd - 2015Jan
benchmark used
for TLTRO Mar-15
2014Apr
7% of which is the
allowance for 1st
& 2nd TLTROs
1,020,000
1,040,000
1,060,000
1,080,000
1,100,000
phase 1
phase 2
phase 3
Constant benchmarks
for TLTROs 5th to 8th
phase 4
Source: BNP Paribas, ECB Source: BNP Paribas, ECB
Phase 1: Sep-14 & Dec-14,
initial take-up is already
known (up to EUR 400bn)
Phase 2: Mar-15 & Jun-15,
depending on net lending
since May 2014 but versus
a dynamic benchmark


17 July 2014
Phase 2 TLTROs are different because, in estimating the benchmark, the ECB takes into
account the recent (May 2013 - April 2014) evolution of each banks net lending. If a bank has
been deleveraging aggressively during the 12 months leading up to April 2014, the average
monthly pace of this deleveraging will be used to extrapolate a projected net lending path from
May 2014 until April 2015. Therefore, a bank can continue deleveraging and still borrow more
from the Mar-2015 and Jun-2015 TLTROs if its deleveraging pace has not been as steep as
that between May 2013 and April 2014. For banks that have not deleveraged between
May-2013 and Apr-2014 (unchanged or increased net lending), the benchmark is just the April
2014 outstanding amount of eligible loans (constant benchmark).
However, this extrapolated trajectory of net lending is not taken into account for the Sep-2015 to
Jun-2016 TLTROs for any bank, whether it deleverages or not. The rule is that, to borrow in the
last four TLTROs, net lending must be above the April 2015 projected data point (equal to the
actual April 2014 data point for non-deleveraging banks). In other words, the benchmark based
on the outstanding amount of eligible loans remains constant.
The direct implication of this methodology is that banks which have recently experienced a
steep deleveraging trend, such as the Spanish, could find it easier to borrow additional amounts
(beyond the first and second TLTROs) in the third and fourth TLTROs than in the fifth to the
eighth

(green line in Chart 1). This is especially true if the pace of deleveraging has slowed
compared with the last 12 month average. Moreover, banks whose deleveraging has outpaced
its previous 12 month average might want to use the maximum allowance in the first and
second TLTROs, as they could find it difficult to beat the benchmark in subsequent TLTROs.
Chart 4 shows the year-on-year percentage change in the TLTRO-eligible pool of loans across
different countries since January 2012. In some countries, such as Spain and France, there has
been a recent improvement in net lending while in others, such as Ireland and the Netherlands,
there has been a deterioration.

Mandatory early repayment in September 2016
In this section we analyse eligibility for borrowing additional amounts in the third to the eighth
TLTROs, and the circumstances under which a mandatory early repayment in September
2016 two years before original expiry can be forced (Phase 4). For simplicity we assume
there is only one top-up TLTRO facility in June 2015 (the fourth TLTRO), which will based on
reference net lending data of April 2015 as shown in Chart 5. We also assume that the
hypothetical bank under examination will top up the full amount it is entitled to (assuming that it
is) ie, 3 x (actual lending in April 2015 minus benchmark). We have constructed scenarios for
four hypothetical banks (defined by different colours in Chart 5) and for each scenario we
consider firstly whether there is eligibility for a top-up in June 2015 and, secondly, whether there
will be a mandatory early repayment in September 2016 and for what amount.

Chart 4: Annual % change in LTRO-eligible pool of loans Chart 5: Four lending scenarios
-20%
-15%
-10%
-5%
0%
5%
10%
GER
FRA
FIN
AUS
NET
BEL
ITA
SPA
POR
IRE
GRE

Source: BNP Paribas, ECB Source: BNP Paribas
Phase 3: Sep-15 to Jun-16,
depending on net lending
but versus a constant
benchmark (April 2015)
Deleveraging banks will
have to deleverage less in
order to be able to borrow
additional amounts
Implications of the
benchmark estimation for
participating in TLTROs
Phase 4: Sep-16, banks
could face a mandatory
early repayment if they are
below the benchmark


17 July 2014
Red scenario:
- Top-up eligibility in June 2015? NO, because net lending in April-2015
(reference data point) is below the benchmark (red line below dotted red line).
- Mandatory early repayment in September-2016? YES, because net lending in
April 2016 is below the benchmark (red line still below dotted red line).
Blue scenario:
- Top-up eligibility in June 2015? YES because net lending in April-2015 is
above the benchmark (blue line above dotted red line).
- Mandatory early repayment in September 2016? YES for both the initial take-
up and the top-up amount. This is because net lending in
April 2016 is below the benchmark (blue line below dotted red line). As a
result, the total amount from all TLTROs up to that point, including the first
two in Sep- and Dec- 2014 has to be repaid early.
Black scenario:
- Top-up eligibility in June 2015? YES because net lending in April 2015 is
above the benchmark (black line above dotted red line).
- Mandatory early repayment in September-2016? YES but not for the full
amount, only for the part of the top up amount borrowed in June 2015 and
which exceeds the multiple of 3 x the difference between the black line (actual
lending in this scenario) and the dotted red line (benchmark) in the April 2016
reference point (ie, the difference in the distance of the black line versus the
dotted red line in two different points in time, Apr-15 versus Apr-16 in this
example). The initial take-up in the first and second TLTROs can be
maintained because, in contrast with the RED and BLUE scenarios, actual
lending is still above the benchmark (black line still above the dotted red line).
This appears to be a mandatory repayment rule aiming to prevent banks
temporarily increasing net lending to borrow significant amounts from the
additional top-up TLTROs (manipulation) and then resuming deleveraging.
Green scenario:
- Top-up eligibility in June 2015? YES because net lending in April 2015 is
above the benchmark (green line above dotted red line).
- Mandatory early repayment in September 2016? NO, both for the initial take-
up and the top-up amount. This is because, firstly, the total amount of actual
lending is much higher than the benchmark (so this bank can maintain its take
up in the first and second TLTROs) and, secondly, because the green line in
April-2016 is higher than where it was in April 2015 (in relation to the
benchmark dotted red line) when the bank borrowed from the additional top-
up TLTRO of June-2015.

Ioannis Sokos

17 July 2014

Market implications
Core rates and inflation: ZIRP to keep core rates and breakeven
inflation low


17 July 2014

No QE implies slower EUR depreciation, but stay short
We have raised our EURUSD forecast targets modestly, to reflect the reduced
prospect of QE.
We continue to expect the EUR to trade broadly weaker in H2.
We favour EURUSD and EURGBP shorts and recommend funding risk and carry
positions in EUR.

The reduced likelihood of the ECB announcing a QE programme implies a somewhat slower
path of EUR depreciation than previously anticipated. However, we continue to expect the
currency to trade with a weaker bias in the second half of the 2014 relative to the rest of the
G10 and most EM currencies, as the effects of the June rate cut continue to weigh. We are
making a modest adjustment to our EURUSD forecasts, raising the Q3 2014 target from 1.32 to
1.34 and the year-end 2014 target to 1.32 from 1.30. Our Q1 2015 target has been raised to
1.30 from 1.28 but we continue to look for 1.26 by the end of Q3 2015.
While EURUSD is little changed from levels prevailing prior to the ECBs June rate cut, the EUR
has weakened on a broader basis, losing about 1% on an effective exchange rate basis and
weakening sharply versus the GBP and the commodity bloc currencies. Moreover, even
EURUSD has declined about 2.5% from levels prevailing before the May ECB meeting, when
President Draghi sent a strong signal that rates would be cut in June. EURUSDs lack of
downward progress since May is consistent with a generally weaker dollar prevailing over this
period, with USDJPY also trading lower.
Unlike the rate cuts delivered by the ECB in May and November of 2013, the policy action
announced in June succeeded in substantially lowering EUR money market rates. The
combination of a lower policy corridor, end to SMP sterilisation and plans for the TLTROs
announced in June finally managed to end the downward trend in excess reserves which had
been in place since late 2013, bringing eonia fixings close to zero and knocking 15bp off the
3-month Euribor fixing. As a result, EUR money market rates have fallen to near the bottom of
the G10 and the EUR has emerged as the most attractive funding currency for G10 and EM
carry trades, at a time when the low volatility environment is encouraging a scramble for yield.
Lower peripheral bond yields may also be discouraging inflows into eurozone bonds now,
allowing the broad basic balance to decline to zero in April on a 3-month basis.
The USDs failure to gain ground versus the EUR (and other G10 currencies) over the past six
weeks is consistent with a loss of real yield support for the US currency. As Chart 4 below
We have raised our year-
end EURUSD target to 1.32
from 1.30
The EUR has weakened
since June easing, though
not vs USD
EUR is now the most
attractive funding
currency for carry trades
Chart 1: EUR now the most attractive funding
currency for G10 carry trades

Chart 2: Eurozone broad balance no longer in surplus
thanks to bond outflows

Source: Macrobond, Bloomberg, BNP Paribas Source: Macrobond, Bloomberg, BNP Paribas
EUR has gained real yield
advantage vs USD


17 July 2014
shows, US short-term inflation expectations (as implied by inflation swaps) have jumped since
the June FOMC meeting, outpacing the move higher in US front-end rates over this period. In
contrast, the June ECB easing has not succeeded in reviving confidence in the ECBs capacity
to achieve its 2% inflation target over a two-year horizon.
Going forward, we see scope for the USD to regain the loss in real yield support experienced
over the past few weeks. The approach of Fed rate hikes in the second half of 2015 should
guide US nominal front-end rates steadily higher on a trend basis even as the Feds 2% inflation
target provides an anchor for US inflation expectations. And, while we dont see scope for
eurozone inflation expectations to recover quickly, our rates strategy team notes that the
continued possibility of ECB QE operations should help keep a floor on eurozone inflation
expectations going forward, even as nominal rates remain low and anchored.
Trading implications
We see two sets of trading implications for the EUR in the second half of 2014 from the above
analysis:
Trade policy divergence: We continue to favour staying short EUR vs the GBP and the
USD as ECB policy diverges from that of the BoE and the Fed. The ECBs indication
that QE is more likely if the EUR strengthens has created a perception of asymmetric
risk for the currency, and should mean continued interest in selling the EUR on rallies
vs these currencies.
Fund in EUR: The EUR should increasingly be viewed as an attractive funding
currency for carry plays into higher yield currencies. We expect the EUR to trend lower
vs the commodity bloc and higher yielding EM currencies during periods of stable and
healthy risk appetite. In the options market, the EUR should increasingly be priced to
correlate more with low-yield currencies like the JPY, and less with high-yielders like
the AUD.
Daniel Katzive
Real yields should move
against the EUR
Chart 3: EUR vs USD real yield differentials have moved
in the EURs favour

Chart 4: because US/eurozone inflation expectations
have diverged

Source: Macrobond, Bloomberg, BNP Paribas Source: Macrobond, Bloomberg, BNP Paribas


17 July 2014
Money market & ccy basis: Still ample excess liquidity
TLTROs will lead to a smaller rise in excess liquidity with a slightly shorter duration
than would be the case with QE, but ample excess liquidity will remain.
The money-market curve therefore looks too steep to us.
The EURUSD cross-currency basis will not widen as much as with QE, but should still
reach -12bp in the 5y by end-2014.

Money market
The biggest difference between TLTROs and QE is the size and, potentially, the maturity of the
net cash injection (injection repayments + expiring tenders): TLTROs alone are likely to lead
to a smaller rise in liquidity with a slightly shorter maturity than in the case of QE.
Part of the reason for the smaller net injection of liquidity is that a significant part of the take up
(gross injection) of TLTROs will be accompanied by repayments on the LTROs. The ECB is
targeting a maximum EUR 1trn overall take-up at all eight TLTROs. If we make the extreme
assumption that this EUR 1trn is a substitute for all other open market operations, this would
lead to a net rise in liquidity of EUR 455bn in the current environment. With current excess
liquidity at around EUR 135bn, this would boost excess liquidity to EUR 590bn, the same level
as in early 2013, ahead of the first repayments on the 3y LTRO. At that time, the eonia/deposit
facility spread was 0.08%.
When it comes to the maturity of the injected liquidity, the ECBs securities markets programme
(SMP) portfolio had a higher average maturity than the 3y LTRO. A year after the ECB stopped
purchasing assets under the SMP, the maturity of this portfolio was 4.3y, dropping to 3.9y after
two years. This suggests an average maturity longer than 4.5y and shorter than 5y when the
ECB stopped purchasing. Compared with the 4y TLTRO, QE would therefore seem to offer
longer term liquidity, although the difference is relatively small.
However, whether the ECB implements QE or not, there will be little impact on the money
market curve. The high level of excess liquidity will keep the front end flat and, as the
conditional benchmarks set for the TLTROs are easy to beat, this should last until the end of
2016 and probably longer (the mandatory repayments are likely to be very small). The curve
therefore looks too steep, at least until the end of 2016. The 2y3m OIS is at 0.17%.
Trade idea: Receive 2y3m OIS
However, a QE programme would see tighter OIS/BOR spreads across the board and, as a
result, OIS/BOR spreads could be slightly paid. But, as the ECB is providing ample liquidity at a
cheap price for a long time (full allotment at a fixed rate until the end of 2016), BORs should not
rise significantly. OIS/BOR spreads have the potential to rise slightly above 15bp but we dont
see them moving much further.
EURUSD x-ccy
In the absence of QE, the EURUSD x-ccy basis will probably not widen as much as we
previously thought (our call was for a rewidening of the 5y spot basis towards -15bp and
possibly further in the case of QE). However, we still expect a widening and still like receiving
the basis.
One of the reasons for the less-than-expected widening is the smaller extension of liquidity
expected from the TLTROs compared to a QE programme, which implies a lesser cheapening
of the EUR relative to the USD, reflected by a basis which will also be less negative.
Another reason is that the expected shutdown of USD issuance by European SSAs is less likely
to materialise in the absence of QE. In a QE environment, the lack of USD issuance by
European SSAs, which are heavy users of the cross-currency market, was critical to our -15bp


17 July 2014
and beyond call. Without it, the widening of the cross-currency will be more limited as interest in
paying the basis will be smaller.
However, there are still reasons to expect a further widening of the x-ccy basis from current
levels, even without QE.
First, one of the consequences of the absence of QE is that we are less bullish on peripherals
as uncertainty and volatility pick up on the back of the fading of QE expectations. Episodes of
risk aversion cannot be ruled out in this environment. As a result, the year-to-date decline in the
Italian sovereign CDS spread is likely to continue to slow and the likelihood of a breakthrough of
the four-year low reached on 9 June has diminished, which means the x-ccy basis has less
room to tighten.
Second, even if the expected pickup in excess liquidity from TLTROs is not as large as it would
have been with a QE programme, it will still be significant enough to trigger further widening of
the x-ccy basis from current levels. We expect the take up at the first two TLTROs to be
significant, especially after the ECB released more details of how they will work. For instance, if
excess liquidity rises to EUR 350bn, this would be consistent with a 5y x-ccy basis of -12bp in
our model, all other things remaining equal.
In the meantime however, the volume of excess liquidity is likely to decline by the end of the
summer as repayment of the 3y LTROs continues, and we think it could decline to EUR 90-
100bn in the run up to the first TLTRO in September if the pace of weekly repayments remains
at around EUR 4bn. In this case, the fair value of the 5y basis would fall back towards -7.5bp in
our model, all other things remaining equal, or closer to -6bp should the market move into risk-
on mode at the same time.
To understand the relationship between excess liquidity and the cross-currency basis, keep in
mind that TLTROs could also, to some extent, be used as a cheap USD funding synthetic tool:
a European bank could borrow euros from the ECB through the TLTRO and then swap the
proceeds back to US dollars using the x-ccy basis swap market. The cost of this operation
would be around $L-4bp (ignoring execution costs, haircut at the ECB etc), which in most cases
is much less than through the bond market (Deutsche Bank sold a 3y USD bond in May at
$L+58bp). With this operation European banks receive the basis, thus pushing the basis wider
(ie, into more negative territory).
Furthermore, whether QE takes place or not, after the significant fall in EUR yields there is an
incentive for foreign entities to issue in EUR and swap back to their domestic currency through
the x-ccy basis swap market as the cost of funding is looking increasingly attractive. A
protracted zero interest rate policy in the eurozone would therefore support a widening of the
basis led by a pickup in interest in issuing in EUR and receiving the basis. However, we
acknowledge that under QE, the volume of interest in such operations would probably be more
significant, because in that environment foreign bonds would have more room to tighten due to
the rebalancing effect. That, together with the potential shutdown in USD issuance by European
entities spelled out earlier, was also critical to our -15bp and beyond call.
Overall, we keep receiving the basis even in the absence of QE as the TLTROs are set
to push the basis further into negative territory from September. However, we would not
enter the trade now as it has room to tighten over the summer, and we lower our target
on the 5y spot basis to -12bp from -15bp+ initially. Enter the trade at -6bp.
Patrick Jacq and Camille de Courcel


17 July 2014
Core/semi-core rates and inflation: ZIRP to keep rates
and breakeven inflation low
A zero interest rate policy (ZIRP) implies low volatility and low core rates: we expect
the 5-year and 10-year to outperform further on the curve.
Semi-core and SSA spreads should benefit from a ZIRP; we continue to forecast
tighter SSA and semi-core spreads.
Without QE, the market will remain sceptical that inflation will revive, keeping
breakeven inflation and core nominal rates under downward pressure.

Core rates: Low rates for longer
The ECB's apparent preferred policy stance of strong long-term forward guidance without QE
means low volatility in core rates and uncertainty over the volume of liquidity to be injected into
the system. Both should keep core rates lower than would have been the case under QE.
One of the key differences between QE and the TLTROs is that the volume of liquidity that will
be injected ultimately relies on the banks willingness and ability to expand balance sheet, and
less so on the ECBs. It could therefore end up being less than the ECB envisions.
Another difference is that, in the absence of QE, the so-called announcement effect, consisting
of a bond market sell-off as inflation expectations pick up as soon as it becomes clear that the
ECB is about to embark on QE, will not materialise.
Hence, TLTROs instead of QE mean low rates for longer; the 20bp fall in the 10y Bund yield
since the ECB announced the TLTRO is a good illustration of the difference between the two
policies.
Thus we expect further 2s5s flattening and the 10s to outperform on the curve, as
investors search for yield in an environment of low rate volatility and subdued inflation
expectations.
Semi-cores and SSAs: A protracted ZIRP environment is positive for spreads
For semi-core and SSA spreads, the change in expectations regarding the ECB policy outlook
will have little impact on the spread tightening we previously forecast.
In a low rates environment for longer, the chase for yield is set to continue. Against this
backdrop, SSAs and semi-core EGBs look attractive as they offer yield pickups to benchmark
bonds in exchange for a limited risk extension (semi-core EGBs) and/or in exchange for some
liquidity give-up (SSAs). A protracted ZIRP environment is especially positive for SSAs and
semi-core EGBs when the 10y Bund yield is at such a depressed level: with a 10y Bund yield at
1.20%, any bond with a similar credit and offering 20-30bp pickup looks attractive.
Indeed, since the last ECB meeting, 5-10y EIB, EFSF and KfW have outperformed Germany by
4bp. Crucially, this outperformance has taken place despite the market being back in risk-off
mode, as illustrated by the heavy sell-off seen in peripherals. It has also taken place at a time
when SSAs usually underperform Germany as a result of their strong relationship with the Bund
ASW, which typically widens between mid-June and late-July because of the seasonal collapse
in net supply in the eurozone (note that this year we chose not to enter this seasonal trade in
light of the ECB June announcement because we expected increased appetite for SSAs). So
SSAs performance in this environment demonstrates a particularly strong appetite for them.
The post-ECB performance of semi-core EGBs has been less pronounced but this is
unsurprising after their significant performance achieved in the year-to-date.
Overall, we think QE could have had more power in terms of spread reduction although
looking at recent pricing action it is hard to believe notably because of the rebalancing effect
as far as SSAs are concerned. But a protracted ZIRP environment is still positive for semi-core


17 July 2014
EGBs and for SSAs, for the reasons stated earlier. The main difference between QE and the
TLTROs will therefore probably lie more in terms of the amount of time it will take before our
various targets are reached, rather than the targets themselves, as there is more uncertainty
lying ahead than there would have been with QE: uncertainty regarding the amount of take up
at the TLTROs and their use, and uncertainty with regard to the bond market in general, as the
ECB is a potential big buyer staying on the sidelines, leaving the market more vulnerable to
episodes of risk aversion.
To conclude, we are keeping our target on semi-core unchanged: Our 10y OAT/Bund spread
target is at 25bp.

Inflation: ECB balance sheet expansion should keep 5y5y breakevens in a tight range
5y5y EUR HICPxt breakevens are below 2.10% - close to their lows. Unlike on previous
occasions when the ECB has embarked on unconventional easing, inflation breakevens have
not collapsed (because of a crisis) but rather have been drifting lower since late 2012. This
gradual but protracted decline in eurozone breakevens is more consistent with a structural re-
pricing of inflation expectations. This may be partly responsible for the lack of reaction to the
ECBs June measures and the announcement of TLTROs.
In the US, 5y5y forward breakevens rose by around 30bp on average in response to the Fed
announcement of QE via large scale asset purchases (LSAPs). During QEI (2009) and QEII
(from 2010), the dollar effective exchange rate depreciated by 10%+, helping to stimulate
inflation and at the very least avoiding deflation. Whilst we continue to expect a lower EURUSD
from USD strength as the US undergoes macroeconomic and policy normalisation, the lack of a
protracted decline in the EUR following the ECBs June meeting may cap any upside for
inflation breakevens.
With eurozone HICP Ex-tobacco running at the low level of 0.4% y/y, the inflation market is
waiting for evidence of increasing inflationary pressure. Our economists dont expect a pickup in
EUR inflation above 0.5% until Q4. In the absence of any support from inflation data, 10y
inflation breakevens are likely to stay low and not volatile. Despite not entering QE directly, the
potential for a significant expansion of the ECB balance sheet should be enough to keep 5y5y
EUR HICPxt breakevens above 2.00% - effectively putting a floor on longer dated inflation
expectations.
Chart 1: SSAs compressed further in July Chart 2: Semi-core compressed sharply post June-ECB
20
25
30
35
40
45
50
55
15
20
25
30
35
40
Feb 14 Mar 14 Apr 14 May 14 Jun 14 Jul 14
KfW Jan-20 Inverted Bund ASW (bp) EIB Jan-24 (RHS)
SSA spreads to Germany (bp)
June ECB
July ECB
sharp tightening
in spite of risk-off
sessions

15
20
25
30
35
40
30
35
40
45
50
55
60
65
70
75
Jan 14 Feb 14 Mar 14 Apr 14 May 14 Jun 14 Jul 14
10y BGB 10y OAT 10y RAGB (RHS)
Semi-core spreads to Germany (bp)
June ECB
July ECB
Source: BNP Paribas Source: BNP Paribas


17 July 2014
Without underlying volatility in the European inflation market (or the HICP itself), we keep our
focus on tactical trading opportunities around seasonality and carry. The front end remains
exposed to further repricing of negative carry which, combined with index extensions and the
macro picture should keep the inflation curve steep in the eurozone. We have reversed our key
QE-based recommendation (long EUR vs US breakevens) given the rise in US CPI inflation
above 2% y/y and the strong positive carry from being long US inflation through the summer.
Camille de Courcel and Shahid Ladha

Chart 3: 5y5y US & EUR Inflation swap fwds
vs key unconventional policy

Chart 4: Currency (effective EUR) vs core inflation
and 5y5y EUR HICPxt forwards

15.0
10.0
5.0
0.0
5.0
10.0
15.0
20.0
0.00
0.50
1.00
1.50
2.00
2.50
3.00
Jan06 Jan07 Jan08 Jan09 Jan10 Jan11 Jan12 Jan13 Jan14 Jan15
5y5yEURHICPxtInflationfwd
CoreHICP%y/y
EUREffectiveexchangerate(9mlag),RHS


17 July 2014
The greatest change in our forecasts resulting from the reduced probability of QE is
in peripheral sovereign spreads.
TLTROs differ from both 3y LTROs and QE in three important ways:
- The size and use of the borrowed funds from TLTROs will remain uncertain,
providing a less transparent exit point for investors who have been purchasing
peripherals in anticipation of QE.
- TLTROs reinforce the link between sovereigns and banks whereas QE could
have weakened them, setting the ground for a non-zero risk weighting on
sovereign bonds.
- The absence of a direct ECB bid for peripherals should lead to higher spread
volatility and so to a higher risk premium in peripheral spreads.
We now anticipate that 10y Spanish and Italian spreads to Bunds will end 2014 at
+150bp (vs our prior expectation of +100bp). Furthermore, as we expect higher
volatility given the lower probability of QE, we do not see peripheral spreads as cheap
enough to buy below 180bp.
We expect 3y peripheral spreads to be better-supported than the 10s, due to the
nature of TLTROs. We now target +60bp at end-2014 (vs our +40bp prior expectation)

TLTROs vs QE: A different outlook for sovereign spreads
While the anticipated effect of the TLTRO programme may differ little from the anticipated effect
of QE in terms of its effect on liquidity, it does have substantially different effects on peripheral
sovereign bonds (see Box 2: TLTRO market impact versus 3y LTRO and QE).
For peripheral sovereign spreads, what are the implications?
1. Reduced probability of imminent QE reduces anticipated spread narrowing
We had expected QE to mean a large (eg EUR 30-40bn/month) price-insensitive bid from the
ECB for sovereign bonds, with the ECB purchases distributed across maturities (probably in the
0-10y) and across sovereigns (in line with ECB capital key weights).
Compared to this baseline expectation, the TLTRO programme raises uncertainty for sovereign
assets in all these vectors ie, size and distribution across sovereigns (depends on banks
qualifying loan growth and asset preference), price sensitivity (banks may wait to buy if they
think yields are rising) and maturity (how much curve risk banks would wish to take versus 4-
year ECB money).
Conclusion: This adds up to a significant rise in uncertainty for holders of peripheral sovereign
bonds compared to our previous base case of full-scale QE and implies that our peripheral
sovereign spread forecasts should be revised up considerably.
2. Strong forward guidance that ZIRP will persist is still in place: Still good for spreads
Compared to QE, the TLTRO programme provides similarly strong forward guidance that the
ECBs ZIRP will remain in place for several years. The experience of Japan is that even without
QE, ZIRP does reduce volatility and bring down all credit spreads, notably in shorter maturities.
Conclusion: the ECBs explicit and implicit (because of the term of the TLTROs) ZIRP forward
guidance is supportive of sovereign spreads, but no more so than would have been the case
under our previous base case of full-scale QE.


17 July 2014

3. Higher spread volatility due to greater uncertainty implies wider spreads
Spread volatility in the Spanish and Italian 10y has fallen over the past 12 months (Chart 1) to
its lowest levels since 2011 (by a concomitant fall in spreads). But expected spread volatility
should be higher under the TLTRO programme than under a QE programme for two reasons:
The increased uncertainties regarding the size, distribution, persistence and price
sensitivity of demand under the TLTRO programme compared to QE; and
Individual event risks, as they occur, should be expected to have a larger impact on
market pricing under the TLTRO programme than QE because the implied backstop
bid (and credit backstop) from the ECB is weaker.
But spread volatility is unlikely to return to the extreme levels seen during summer 2011
summer 2012 in our view, because
There are now support mechanisms in place which were not present before (outright
monetary transactions (OMT), TLTROs, strong forward guidance and the possibility
that the ECB will still eventually adopt QE);
Sovereign credits have generally become less vulnerable, with ratings generally
improved, or at least no longer worsening; and
Banks are better capitalised.
Conclusion: Higher expected spread volatility should lead to demand for a higher risk premium
in sovereign spreads than compared to our previous base case of full-scale QE, but not as high
as during the crisis period leading to the OMT announcement. The incremental risk premium
should be higher in longer than shorter maturities, due to the existence of TLTROs. This implies
that our sovereign spread forecasts should be revised up, particularly in longer maturities.
Chart 1: 10y ITA/GER spreads annualised volatility (1m, 3m, 6m)

Source: BNP Paribas
4. Supply and demand
Supply and demand considerations are still broadly positive, although of course not as positive
as they would have been under our previous base case of full-scale QE.
On the supply side: Spain and Italy have already completed more than 70% of their planned
2014 issuance.
On the demand side: Since July 2013, non-bank ownership of Spanish and Italian bonds has
increased significantly, as non-bank buying has greatly outstripped banks purchases.


17 July 2014
Investors other than domestic banks have bought EUR 135bn Italy and EUR 96bn Spain (net)
since June 2013 (Chart 2). Should these non-domestic bank buyers decide to take some profit,
there will be considerable secondary market supply. On the other hand, the TLTROs clearly
offer the banks cheap funding to absorb some of that potential selling. If the Spanish and Italian
banks took all the funding available in the first two TLTROs and invested it all in sovereign debt,
they could buy some EUR 75bn Italy and EUR 55bn Spain.
Chart 2: Cumulative purchases of Italian and Spanish debt by investors other than
domestic banks (EUR bn)

Source: BNP Paribas

Conclusion: supply and demand considerations are less favourable to peripherals than
compared to our previous base case of full-scale QE this year, but unless non-bank owners turn
into sellers, the supply and demand picture remains favourable.
New peripheral spread forecasts
In the absence of QE, the uncertainties for peripheral spreads are greater and spread levels
harder to forecast.
We do expect their path in the coming quarters to be much more event-driven and volatile than
would have been the case under QE. Furthermore, because spread volatility and the level of
spreads are closely correlated, there is a danger that spread widening could be self-reinforcing.
Therefore, we believe that investing in peripherals will depend much more on valuation (spread
vs carry) than was previously the case:
If 10y spread volatility remains around the levels to which it has recently rebounded, we expect
10y Spanish and Italian spreads to end 2014 at around present spread levels of 150bp. But at
present levels of spread volatility we would not consider Spain and Italy cheap unless spreads
rose to around 180bp.
For 3y spreads, as explained above, we see less chance of higher volatility. If 3y spread
volatility remains around current levels, we expect 3y Spanish and Italian spreads to end 2014
at 60bp, marginally tighter than current levels. We see less prospect of a rise in spread volatility
in sub-4y peripherals, so we recommend remaining long 3y Spain and Italy at present spread
levels versus Germany.
Laurence Mutkin & Ioannis Sokos


17 July 2014
Box 2: TLTRO market impact versus 3y LTRO and QE
The size and use of the borrowed funds from TLTROs will remain uncertain, providing a less transparent exit point for investors who have
been purchasing peripherals in anticipation of QE.
TLTROs reinforce the link between sovereigns and banks whereas QE could have weakened them, setting the ground for a non-zero risk
weighting on sovereign bonds.
We anticipate more volatility in peripherals, which could lead to investors demanding a higher premium in order to keep buying peripheral
bonds given the lack of a clear backstop mechanism in the near term.
TLTROs vs 3-year LTROs market impact
There are some key differences and similarities between TLTROs and LTROs. The main difference was supposed to be the fact
that the former target the real economy, but the ECB has attached few conditions to TLTROs, leaving significant room for a
different use of the borrowed amounts. The carry trade might not be as attractive as it was in early 2012, but it is still a profitable
trade given the very cheap funding rate of 25bp. Another key difference is the full allotment, which is not yet known in TLTROs.
There is a maximum allowance, which is already known for the first and second TLTROs and which for subsequent TLTROs will
depend on net lending evolution. Banks will therefore not be able to borrow as much as they want and, on top of this, they will
have to meet certain criteria (beating the benchmark) in order to keep the borrowed amounts until September 2018 and not face
a mandatory repayment order in September 2016.
Table 1 shows that, for the first and second TLTROs, only 32.4% of the total amount will be allocated to Italian and Spanish banks
(assuming full take-up by all), whereas this percentage was 53.5% in the 3-year LTROs. Moreover, the maximum allowance in the
first two TLTROs is roughly EUR 195bn short of the current outstanding of the 3-year LTROs for Italian and Spanish banks, on
aggregate. In terms of market implication, the current environment is so different to early 2012 with Greeces debt restructuring
and subsequent elections, the ECBs outright monetary transactions and the situation for European banks at that time that a
comparison with the 3-year LTROs of early 2012 is meaningless. The only impact we can predict is that the short end of peripheral
curves should be supported more than longer maturities potentially leading to a steepening of the curve. Chart 1 shows the yield
change among different maturity buckets in Italy and Spain around the time of the 3-year LTROs in late 2011early 2012.
TLTROs vs QE market impact
Over the past few months the market has started pricing in the probability of QE, leading to a strong compression in peripheral
spreads across the curve, but this case has weakened since the central banks June and July meetings. If QE were to be
implemented, it would have to wait until an assessment of the impact of the TLTROs has been made, most probably towards
the end of H1 2015. At this point, the important question is whether or not the TLTROs and the other recent ECB
announcements are strong enough to keep peripheral spreads at their recent low levels.
We look firstly at the potential size and country breakdown in the case of QE versus TLTROs. Our initial estimate for a QE
programme was EUR 400bn in its first stage, although there was talk later of EUR 1trn. This is similar to the EUR 400bn maximum
initial take-up in the first and second TLTROs and the EUR 1trn that ECB President Draghi mentioned for all TLTROs up to June
2016. The country breakdown is also roughly similar, as shown in Table 2, since the ECB capital key is not far from the weighting
according to the outstanding amount of TLTRO eligible loans. However, while in the case of QE we would know the total amount of
each countrys debt being bought by the ECB in advance, this is not true for the TLTROs. There is uncertainty both with respect to
the initial take-up (we only know the maximum allowance) and to any allowance in subsequent TLTROs (the third to the eighth).
Table 1: Allocations differ between TLTROs and LTROs Chart 1: Impact of 3-year LTROs in 2011 and 2012
DE 23.7% 7.4% 16.4%
FR 19.3% 13.7% 5.6%
IT 18.9% 25.0% 6.1%
ES 13.5% 28.5% 14.9%
NL 7.3% 0.0% 7.3%
AT 3.8% 1.4% 2.4%
BE 2.5% 3.4% 1.0%
GR 2.4% 12.3% 9.8%
PT 2.1% 2.2% 0.1%
IE 1.9% 1.5% 0.4%
FI 1.7% 0.4% 1.4%
Others 2.8% 4.3% 1.5%
EuroArea 100% 100% 0%
TLTROsTakeup TLTROsvs3yLTROs vsoutst.LTROs
457
0
35
20
1
170
155
0
4
10
Share
(%)
TLTRO
LTRO(%)
LTROs
Outst.
10
50
15.2
3.6
43.9 2
0.0
14.4
35.0
125.0
22.0
Total
LTROs
75.0
140.0
254.7
290.0
1.4
529.5
10.6
1018.7
6.5
65.0
170.0
50.0
139.0
2nd
LTRO
TakeUp
23.0
40.0
10.0
14.0
120.0
90.0
115.7
1st
LTRO
TakeUp
12.0
35.0
5.2
8.0
2.2
489.2
33.2
7.9
60.0
Share
(%)
TLTRO
Eligibility
(7%)
9.8
94.5
7.7
9.7
54.0
76.9
15.3
8.4
398.4
75.3
29.0
11.0
6.8
26.6
12.3
5.8
9.0
58.6
TLTRO
Existing
LTRO
84.5
26.9
94.7
101.0
29.0
11.3
0.2
9.7

-450
-400
-350
-300
-250
-200
-150
-100
-50
0
3y
SPA/GER
5y
SPA/GER
10y
SPA/GER
30y
SPA/GER
3y
ITA/GER
5y
ITA/GER
10y
ITA/GER
30y
ITA/GER
Spreads'changefrom8Dec2011to1Mar2012


17 July 2014
This uncertainty makes the exit strategy of investors who bought peripherals in anticipation of QE less transparent. Obviously
we do not know whether investors have bought peripherals in anticipation of QE or because of a broader improvement in the
fundamentals of each country and the need for extra yield in a zero interest rate policy world. However, data have shown that
investors other than domestic monetary financial institutions (MFIs) have increased their participation in peripheral markets
since last summer. Chart 2 shows the cumulative monthly purchases of Italian and Spanish debt by investors other than
domestic MFIs. This group includes both non-resident investors and all other domestic investors apart from domestic banks. It
shows the relatively heavy market positioning which has been long the periphery for a while now. Charts 3 and 4 show the
impact of ECB meetings on peripheral yields and spreads across different maturity buckets.
As shown in the charts above, peripheral spreads and yields have compressed between ECB meetings, except in May when
spreads widened. However, since the July ECB meeting, peripheral yields and spreads have both risen on the back of fading
expectation of imminent QE. Note there was no such reaction to June meeting, because back then it was unclear if TLTROs
might in some way replace QE or add to its impact. However, the ECBs rhetoric at the July meeting pointed in a different
direction; QE is still on the table but only after an assessment of the effect of the TLTROs.
In summary, we believe that the impact of TLTROs on the market differs from that of QE. There is a lot of uncertainty around the
size and the timing of support for peripheral bonds under TLTROs, which would not be the case under QE. The recent market
pricing in of QE has reversed recently, and the anticipation of QE was one of the reasons for the market positioning long in
peripherals. The recent increase in volatility has only just begun to be compensated by higher yields and the lack of imminent QE
could lead investors to demand further protection. The fact that QE is still on the table could cap yields, but we will still revise our
targets for the end of 2014 assuming no QE during this period.
Table 2: Similar allocations
in TLTROs and QE
Chart 2: Net cumulative purchases of peripheral debt by
investors other than domestic MFIs since 2012
DE 23.7% 13.8 38.9 2.9% 0.6% 3.5%
FR 19.3% 4.4 23.9 2.2% 1.3% 3.5%
IT 18.9% 3.8 9.8 0.9% 2.5% 3.4%
ES 13.5% 6.5 69.9 9.1% 11.8% 2.7%
NL 7.3% 6.2 1.9 0.5% 3.0% 3.4%
AT 3.8% 4.2 5.1 2.3% 1.2% 3.5%
BE 2.5% 4.1 4.9 3.5% 0.2% 3.3%
GR 2.4% 1.5 3.4 2.5% 5.8% 3.3%
PT 2.1% 1.6 6.1 5.1% 8.2% 3.1%
IE 1.9% 1.4 6.3 5.7% 8.8% 3.0%
FI 1.7% 0.4 3.4 3.5% 0.5% 3.0%
Others 2.8% 2.0
EuroArea 100% 2.2 0.04% 3.35% 3.39%
Implied
Change
TLTROsTakeup vsQE TopupTLTROs
Implied
Net
Lending
May14
toApr15
Lending
May13
toApr14
10.0
6.4
7.2
9.0
400.0
TLTRO
QE(bn)
EUR600bn
Topup
TLTROs
142.3
115.9
113.3
81.3
43.6
23.1
14.7
14.7
12.6
11.6
10.3
16.6
600.0
QEEUR
400bn
108.3
81.3
71.4
47.5
22.8
11.1
13.9
11.2
Share
(%)
TLTRO
Eligibility
(7%)
9.8
94.5
7.7
9.7
54.0
76.9
15.3
8.4
398.4
75.3
29.0
11.0
6.8

-100
-80
-60
-40
-20
0
20
40
60
80
100
120
01-12 03-12 05-12 07-12 09-12 11-12 01-13 03-13 05-13 07-13 09-13 11-13 01-14 03-14 05-14
NetCumulativepurchasesofperipheraldebtbyinvestorsothersthan
DomesticMFIssinceJan2012(EURbn)
Italian Debt
Spanish Debt
Chart 3: Average SPA/ITA spreads vs Germany change
per ECB meeting
Chart 4: Average SPA/ITA yields change per ECB
meeting
-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
3y 5y 10y 30y
AverageITA/SPASpreads'changessince20Feb
Post July ECB Post June ECB Post May ECB
Post April ECB Post March ECB 20 Feb to March ECB

-110
-100
-90
-80
-70
-60
-50
-40
-30
-20
-10
0
10
20
3y 5y 10y 30y
AverageITA/SPAYields'changessince20Feb
Post July ECB Post June ECB Post May ECB
Post April ECB Post March ECB 20 Feb to March ECB


17 July 2014
The implications of a no ECB QE scenario for the European credit markets are a
reduced degree of spread tightening in H2 2014 and less spread compression of high
beta vs. low beta, Periphery vs. Core, Financial vs. Non-Financial and European vs.
US credit spreads. Furthermore, we may see some steepening in credit curves,
particularly in high beta and periphery risk.
We expect a relatively high take-up for the TLTRO, although the 1tn maximum
amount mentioned by President Draghi is higher than our estimates. There is no
stigma in using the facility in our view, given that it is conditional to lending to the
economy. Although, the Central Banks may put pressure on the banks to use the
TLTRO.
That said, we are not convinced that this will lead to a significant increase in lending,
especially to SMEs. The TLTRO is not that targeted after all. Banks can continue to
embark on the sovereign debt carry trade and the net lending benchmark can be
met/exceeded via an improvement in net lending to large corporates and/or
consumers rather than SMEs.

The pullback in market expectations of large scale ECB asset purchases following recent ECB
rhetoric has been promptly reflected in eurozone risk premia moving wider. Indicatively, our
measure of ECB QE implied probability in Bono/BTP spreads over Bunds or iTraxx Main has
dropped below 50% over the past couple of weeks, see Chart 1, although the idiosyncratic
BES/ESF story has also been a contributing factor to the recent move in risk premia.
A lack of asset purchases/QE does not alter our view for credit spreads to grind tighter into
year end, although it may curb the degree of spread tightening and compression in high beta vs.
low beta, Periphery vs. Core, Financial vs. Non-Financial spreads (given that direct buying of
sovereign debt would benefit banks balance sheets more directly) and by extension, less
compression in European vs. US credit spreads.

For Cash credit, the adjustment to our in-house forecasts for Periphery and Core rates under a
no QE scenario carries implications for our expectations for Excess and Total Returns across
credit for the rest of the year, as well as for the relative performance of vs. $ credit spreads.
In terms of Excess Returns we continue to prefer maturities at the belly of the curve or
shorter, as we expect some steepening pressure in credit curves under a no QE scenario.
Such pressure could arise in higher beta credit and periphery spreads in particular, given
that the no QE is likely to translate into steeper Bono/BTP spreads vs. Bunds.
Chart 1: Bono/BTP spreads to Bunds and iTraxx have
pulled back from expecting QE (probability below 50%)
Chart 2: Total Returns across traded credit indices, the
recent risk off has aided duration and hindered high beta
100%
75%
50%
25%
0%
25%
50%
75%
100%
Dec Jan Mar Apr May Jun Jul
QEprobability,SPITBund10y QEprobability,iTraxxMain
E
C
B
5
J
u
n
e

CDS IG 1.3%
CDS FinSen 1.9%
CDS FinSub2.8%
CDS HY 6.7%
CDS $IG 1.0%
CDS $HY 1.3%
TRS IG 5.0%
TRS HY3.8%
TRS $IG 5.6%
TRS$ HY4.8%
TRS IG 5.0%
97
98
99
100
101
102
103
104
105
106
107
108
Dec Jan Mar Apr May Jun Jul Aug
Source: Markit iBoxx, Bloomberg, BNP Paribas Source: Markit iBoxx, BNP Paribas
TRS for Cash, CDS HY for iTraxx Xover


17 July 2014
In terms of Total Returns the absence of QE should be more beneficial to duration risk
(read longer maturity IG credit) at the expense of credit risk (read shorter maturity HY
credit). The benefit to duration risk would come via a bull flattening pressure on the Bund
curve, given that the lack of QE could push expectations for inflation to revert towards its
2% target further out in time. Two factors that could counteract this are: (i) any potential
bear steepening in real rates (but less likely if growth momentum keeps lagging); and (ii)
the fact that the 2-10y Bund curve is already trading just 2bp wider to its five year low of
116bp.
The Total Return performance across credit indices over the past couple of weeks has already
started to reflect this change in dynamics in favour of duration risk, see Chart 2, although the
flight to safety on the back of BES/ESF has contributed to this. Notably, Xover has given up
some of its recent outperformance vs. IG cash across currencies, with the performance gap
between Xover and IG Cash, for example, narrowing from +c.3% shortly after the ECB
meeting on 5 June to +c.1.5% currently. HY Cash has also given up some of its ytd
outperformance recently and now lags IG Cash by over one point (3.8% vs. 5%). Under a no
ECB QE regime we could likely be faced with the prospect of comparable Total Returns in both
IG and HY credit in 2014, as the carry component over half a year would take both to a full
year return of c.6% (if we are to assume no further tightening in credit spreads and ignoring the
curve rolldown).
In respect to CDS spreads under a no ECB QE scenario, we would also expect a moderate
spread tightening and spread compression. This does not alter our expectation for iTraxx Xover
to outperform iTraxx Main, Sen Fin to outperform Main and Main to outperform CDX IG, but this
outperformance is bound to be more moderate. For example, under an ECB QE scenario we
could envisage Main trading inside 50bp (e.g. 48bp) and Sen Fin well inside Main (e.g. 42bp) to
be consistent with a 10y Bono/BTP spread of 100bp over Bunds. Under a no ECB QE
scenario, however, 10y Bono/BTP spread would be at 150bp over Bunds and the analogous
levels for iTraxx would be mid/high 50s for Main (e.g. 57bp) and mid 50s for Sen Fin (e.g.
56bp).
Banks incentivised to use the TLTRO to the maximum
A key difference between QE and TLTRO in terms of impact on asset prices could be the size
and the timing of each programme. If the ECB were to announce a broad based asset purchase
programme, they would be likely to specify the timing of its deployment and the amounts
involved, thus giving certainty to the market. However, the aggregate amount of the TLTROs
cannot be known at the outset, giving rise to uncertainty given that the second phase
commencing in March 2015 would be highly dependent on net lending. Therefore, while the
TLTRO take-up will be incremental and uncertain, in the case of QE, the market would benefit
more from an announcement effect that comes with more clarity.

Despite these differences, we try here to estimate the potential TLTRO take-up. We believe that
banks will have an incentive to maximise their allowance for the TLTROs for the following
reasons:
- No stigma, given the conditionality of Private Sector lending.
- 4y term funding at 25bp is very attractive.
- Banks can continue with the particularly attractive sovereign carry trade given its 0% RW.
- Banks in the periphery that are still dependent on the 3Y LTROs can refinance.
- The Central Banks will put pressure on the banks to tap the TLTRO.

President Draghi mentioned that the overall take-up could reach a maximum of 1tn. We
assume that this corresponds to 400bn in the first phase in September and December 2014
(i.e. the full allowance of 7% of existing Private Sector loans) and 600bn in the second phase
starting in March 2015 (this suggests 600bn/3= 200bn of additional lending over and above
the benchmarks, i.e. a strong 3.5% annual growth in total private lending across the eurozone),
well above the initial estimates of 150-300bn by our Banks Sector Specialists. For the
conditional TLTROs, while beating the benchmark of minus 3.3% 12-month growth in private
lending should not be difficult over 24 months, the actual take-up may be lower than the figure
suggested by President Draghi. For more details on our views on TLTROs we refer readers to


17 July 2014
the marketing material published by our Sector Specialists Gildas Surry and Geoffroy de
Pellegars: TLTROs for banks: no more stigma and Tweaked LTROs.
The EBA data from the EU-wide 2013 Transparency exercise indicates that 73% of banks (from
a representative subsample representing 61% of the sector) have reduced their lending from
December 2012 to June 2013. We can extrapolate this trend into April 2014. If we then assume
that these deleveraging banks will just stop reducing their lending by keeping it flat from April
2014 onwards, they would be entitled to 191bn of TLTRO funding in the last of the six
additional allowances. We can also model that the lending banks (27% of this subsample)
would continue lending at the same pace and that would give them 110bn of TLTRO funding.
The total (301bn) for 61% of the sector would suggest an overall 492bn of take-up in the
additional TLTROs for all eurozone banks.
Will the TLTROs lead to a substantial increase in lending?
There will be a lot of scrutiny on how lending evolves as the banks tap the TLTROs. The
average benchmark (net lending for one year until end of April 2014) of minus 3.3% leaves
ample room for improvement, nevertheless, certain factors could hinder a significant increase in
lending:
- The banks argue that they need more demand for borrowing from solvent credits.
- The economy remains vulnerable, especially in peripheral countries and thus banks will be
wary of creating the next wave of NPLs.
- Lending to corporates, especially lower-rated or unrated corporates including SMEs,
remains capital intensive.
- Banks, especially weak banks (a few of those being in the periphery) continue to see
mounting pressure in terms of capital requirements. For example, the AQR/Stress Tests
are leading to more banks cleaning up their balance sheet, further deleveraging and higher
capitalisation. We calculate that 33bn of equity was raised by European banks year-to-
date in 2014, on top of 29bn of AT1 issuance. In a similar vein, the anticipation of a higher
minimum requirement for the leverage ratio (from 3% to potentially 4%) also acts as a
deterrent to lending.
- It will take some time before the reform of the ABS framework can effectively help create
some risk transfer and capital relief for banks. Joint proposals from the ECB and BoE are
heading in the right direction, but this will take time to implement.
Furthermore, the key aim for European authorities was to find a solution for the lack of funding
to SMEs. However, the benchmark could be met by lending to large corporates or consumers,
rather than SMEs. We were surprised that the TLTRO was not more targeted to SME lending.
The reason could be that defining what is an SME is tricky, e.g. an SME in Germany is not the
same size as an SME in Spain. The picture on lending could also differ significantly from one
country to another. This means that we could see lending improve more in the core countries
than in peripheral countries.
That said, SMEs stand as the most profitable lending segment to rebuild net interest margins in
a low rate environment, improving earnings for banks and thus generating capital internally. We
would expect banks to guide towards a gradual impact on SME lending while the demand for it
continues to increase slowly. Also, without banks significantly increasing lending, banks will
fund each other more cheaply in a low rate and cheap TLTRO environment.
So far, the banks are still working on their numbers to assess their 7% allowance. Some banks
have already indicated that the TLTRO could be attractive, such as KBC at its investor update
on 17 June. Erste Bank in Austria said on 7 July, that the TLTRO might be a very good source
of funding and that it will look into it, even though it would be a small contribution. At the country
level, Bank of Italy Governor Ignazio Visco suggested on 10 July that Italian banks could borrow
more than 200bn in TLTRO funds. Note that the 7% allowance for Italian banks in the two
initial TLTROs amount to c. 75bn. At the same time, Unicredits CEO indicated his bank could
draw on 14-15bn, half of which would be in Italy, while Monte dei Paschis CEO said that his
bank could draw up to 6bn of TLTRO.
Olivia Frieser, Greg Venizelos

RESEARCH DISCLAIMERS
This document has been written by our strategy teams. It does not purport to be an exhaustive analysis, and may be subject to conflicts of
interest resulting from their interaction with sales and trading which could affect the objectivity of this report. This document is a marketing
communication and is not independent investment research. It has not been prepared in accordance with legal requirements designed to
provide the independence of investment research, and is not subject to any prohibition on dealing ahead of the dissemination of investment
research.

STEER is a trade mark of BNP Paribas
The information and opinions contained in this report have been obtained from, or are based on, public sources believed to be reliable, but no representation or
warranty, express or implied, is made that such information is accurate, complete or up to date and it should not be relied upon as such. This report does not constitute
an offer or solicitation to buy or sell any securities or other investment. Information and opinions contained in the report are published for the assistance of recipients, but
are not to be relied upon as authoritative or taken in substitution for the exercise of judgement by any recipient, are subject to change without notice and not intended to
provide the sole basis of any evaluation of the instruments discussed herein. Any reference to past performance should not be taken as an indication of future
performance. To the fullest extent permitted by law, no BNP Paribas group company accepts any liability whatsoever (including in negligence) for any direct or
consequential loss arising from any use of or reliance on material contained in this report. All estimates and opinions included in this report are made as of the date of
this report. Unless otherwise indicated in this report there is no intention to update this report. BNP Paribas SA and its affiliates (collectively BNP Paribas) may make a
market in, or may, as principal or agent, buy or sell securities of any issuer or person mentioned in this report or derivatives thereon. Prices, yields and other similar
information included in this report are included for information purposes. Numerous factors will affect market pricing and there is no certainty that transactions could be
executed at these prices.
BNP Paribas may have a financial interest in any issuer or person mentioned in this report, including a long or short position in their securities and/or options, futures or
other derivative instruments based thereon, or vice versa. BNP Paribas, including its officers and employees may serve or have served as an officer, director or in an
advisory capacity for any person mentioned in this report. BNP Paribas may, from time to time, solicit, perform or have performed investment banking, underwriting or
other services (including acting as adviser, manager, underwriter or lender) within the last 12 months for any person referred to in this report. BNP Paribas may be a
party to an agreement with any person relating to the production of this report. BNP Paribas, may to the extent permitted by law, have acted upon or used the
information contained herein, or the research or analysis on which it was based, before its publication. BNP Paribas may receive or intend to seek compensation for
investment banking services in the next three months from or in relation to any person mentioned in this report. Any person mentioned in this report may have been
provided with sections of this report prior to its publication in order to verify its factual accuracy.
This report was produced by a BNP Paribas group company. This report is for the use of intended recipients and may not be reproduced (in whole or in part) or
delivered or transmitted to any other person without the prior written consent of BNP Paribas. By accepting this document you agree to be bound by the foregoing
limitations.
Certain countries within the European Economic Area:
This report is solely prepared for professional clients. It is not intended for retail clients and should not be passed on to any such persons. This report has been approved
for publication in the United Kingdom by BNP Paribas London Branch. BNP Paribas London Branch (registered office: 10 Harewood Avenue, London NW1 6AA; tel: [44
20] 7595 2000; fax: [44 20] 7595 2555) is authorised by the Autorit de Contrle Prudentiel et de Rsolution and the Prudential Regulation Authority and subject to
limited regulation by the Financial Conduct Authority and Prudential Regulation Authority. Details about the extent of our authorisation and regulation by the Prudential
Regulation Authority, and regulation by the Financial Conduct Authority are available from us on request. BNP Paribas London Branch is registered in England and
Wales under no. FC13447. www.bnpparibas.com
This report has been approved for publication in France by BNP Paribas SA, incorporated in France with Limited Liability and is authorised by the Autorit de Contrle
Prudentiel et de Rsolution (ACPR) and regulated by the Autorit des Marchs Financiers (AMF) whose head office is 16, Boulevard des Italiens 75009 Paris, France.
This report is being distributed in Germany either by BNP Paribas London Branch or by BNP Paribas Niederlassung Frankfurt am Main, a branch of BNP Paribas S.A.
whose head office is in Paris, France. BNP Paribas S.A. Niederlassung Frankfurt am Main, Europa Allee 12, 60327 Frankfurt is authorised and supervised by the
Autorit de Contrle Prudentiel et de Rsolutionand it is authorised and subject to limited regulation by the Bundesanstalt fr Finanzdienstleistungsaufsicht (BaFin).
United States: This report is being distributed to US persons by BNP Paribas Securities Corp., or by a subsidiary or affiliate of BNP Paribas that is not registered as a
US broker-dealer to US major institutional investors only. BNP Paribas Securities Corp., a subsidiary of BNP Paribas, is a broker-dealer registered with the U.S.
Securities and Exchange Commission and a member of the Financial Industry Regulatory Authority and other principal exchanges. For the purposes of, and to the
extent subject to, 1.71 of the U.S. Commodity Exchange Act, this report is a general solicitation of derivatives business. BNP Paribas Securities Corp. accepts
responsibility for the content of a report prepared by another non-U.S. affiliate only when distributed to U.S. persons by BNP Paribas Securities Corp.
Japan: This report is being distributed to Japanese based firms by BNP Paribas Securities (Japan) Limited or by a subsidiary or affiliate of BNP Paribas not registered
as a financial instruments firm in Japan, to certain financial institutions defined by article 17-3, item 1 of the Financial Instruments and Exchange Law Enforcement
Order. BNP Paribas Securities (Japan) Limited is a financial instruments firm registered according to the Financial Instruments and Exchange Law of Japan and a
member of the Japan Securities Dealers Association and the Financial Futures Association of Japan. BNP Paribas Securities (Japan) Limited accepts responsibility for
the content of a report prepared by another non-Japan affiliate only when distributed to Japanese based firms by BNP Paribas Securities (Japan) Limited. Some of the
foreign securities stated on this report are not disclosed according to the Financial Instruments and Exchange Law of Japan.
Hong Kong: This report is being distributed in Hong Kong by BNP Paribas Hong Kong Branch, a branch of BNP Paribas whose head office is in Paris, France. BNP
Paribas Hong Kong Branch is registered as a Licensed Bank under the Banking Ordinance and regulated by the Hong Kong Monetary Authority. BNP Paribas Hong
Kong Branch is also a Registered Institution regulated by the Securities and Futures Commission for the conduct of Regulated Activity Types 1, 4 and 6 under the
Securities and Futures Ordinance.
Israel: BNP Paribas does not hold a licence under the Investment Advice and Marketing Law of Israel, to offer investment advice of any type, including, but not limited
to, investment advice relating to any financial products
Singapore: BNP Paribas Singapore Branch is regulated in Singapore by the Monetary Authority of Singapore under the Banking Act, the Securities and Futures Act
and the Financial Advisers Act. This document may not be circulated or distributed, whether directly or indirectly, to any person in Singapore other than (i) to an
institutional investor pursuant to Section 274 of the Securities and Futures Act, Chapter 289 of Singapore ("SFA"), (ii) to an accredited investor or other relevant person,
or any person under Section 275(1A) of the SFA, pursuant to and in accordance with the conditions specified in Section 275 of the SFA or (iii) otherwise pursuant to,
and in accordance with the conditions of, any other applicable provisions of the SFA.
Australia: This material, and any information in related marketing presentations (the Material), is being distributed in Australia by BNP Paribas ABN 23 000 000 117, a
branch of BNP Paribas 662 042 449 R.C.S., a licensed bank whose head office is in Paris, France. BNP Paribas is licensed in Australia as a Foreign Approved Deposit-
taking Institution by the Australian Prudential Regulation Authority (APRA) and delivers financial services to Wholesale clients under its Australian Financial Services
Licence (AFSL) No. 238043 which is regulated by the Australian Securities & Investments Commission (ASIC).The Material is directed to Wholesale clients only and is
not intended for Retail clients (as both terms are defined by the Corporations Act 2001, sections 761G and 761GA). The Material is subject to change without notice and
BNP Paribas is under no obligation to update the information or correct any inaccuracy that may appear at a later date.
Brazil: This document was prepared by Banco BNP Paribas Brasil S.A. or by its subsidiaries, affiliates and controlled companies, together referred to as "BNP Paribas",
for information purposes only and do not represent an offer or request for investment or divestment of assets. Banco BNP Paribas Brasil S.A. is a financial institution
duly incorporated in Brazil and duly authorized by the Central Bank of Brazil and by the Brazilian Securities Commission to manage investment funds. Notwithstanding
the caution to obtain and manage the information herein presented, BNP Paribas shall not be responsible for the accidental publication of incorrect information, nor for
investment decisions taken based on the information contained herein, which can be modified without prior notice. Banco BNP Paribas Brasil S.A. shall not be
responsible to update or revise any information contained herein. Banco BNP Paribas Brasil S.A. shall not be responsible for any loss caused by the use of any
information contained herein.
Some or all the information reported in this document may already have been published on https://globalmarkets.bnpparibas.com
BNP Paribas (2014). All rights reserved.

BNP ECB QE Revisited

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

BNP ECB QE Revisited

Transféré par

Droits d'auteur :

Formats disponibles

www.GlobalMarkets.bnpparibas.com Please refer to important information at the end of the report.

Market Economics | Credit | IR & FX Strategy

8th 7th 6th 5th

Vous aimerez peut-être aussi