Data Matters - HSBC PDF

2 May 2019
Multi-Asset
Data Matters Quant
Risk-on, Goldilocks-off
 The current environment looks like a classic risk-on move… Mark McDonald
Head of Data Science and Analytics
 …rather than a Goldilocks scenario HSBC Bank plc
mark.mcdonald@hsbcib.com
+44 20 7991 5966
 April saw a pattern of positive returns from risk-on assets and
Max Kettner
negative returns from risk-off assets Multi-Asset Strategist
HSBC Bank plc
maximilian.l.kettner@hsbc.com
+44 20 7991 5045
The consensus is increasingly looking for a period of stable global growth but low
inflation. This, coupled with the sharp rebound in all asset classes at the beginning of Subhrajit Banerjee, CFA
Fixed Income Strategist
2019, has cemented the belief of many market participants in a so-called Goldilocks HSBC Bank plc
subhrajit.banerjee@hsbcib.com
scenario. +44 20 7991 6851
However, when we look at the pattern of asset returns this year, it appears that the Dominic Kini
Analyst, Quantitative Credit Strategy
true Goldilocks period was confined to January. In contrast, the returns seen over HSBC Bank plc
April were more consistent with a classic risk-on move with positive returns from risk- dominic.kini@hsbcib.com
+44 20 7991 5599
on asset classes such as equities, and negative returns from typical safe havens
Alastair Pinder, CFA
such as Developed Market rates. Equity Strategist
HSBC Securities (USA) Inc.
alastair.pinder@us.hsbc.com
+1 212 525 4131
Correlation
Peter Barnshaw
Analyst, Credit Strategy
This risk-on pattern of returns has been accompanied by an increase in the HSBC HSBC Bank plc
peter.barnshaw@hsbc.com
RORO Indicator, which measures the strength of cross-asset correlations. If we drill +44 20 7991 5022
down further, the correlations within risk-on and risk-off asset groups have risen.
Despite this, the overall level of cross asset correlations remains quite modest. This
suggests that there is not a single dominant factor driving asset classes at the
moment. The HSBC dendrogram analysis identifies three clear clusters of assets:
(i) risk-on, (ii) USD-bloc, and (iii) rates. Interestingly, the rates cluster contains two
sub-clusters with IG credit falling into one cluster and 10y sovereign bonds falling into
the other.
Volatility
Throughout 2019 we have seen an intense period of volatility compression across asset
classes. This has been discussed at great length by many in the market. The usual
focus of these discussions is a search for an answer to the question of why volatility is so
low. We focus instead on the increase in gap risk since the end of the financial crisis.
This is something we discuss in detail in The Great Gapsby, 1 May 2019.
Disclosures & Disclaimer Issuer of report: HSBC Bank plc

This report must be read with the disclosures and the analyst certifications in
View HSBC Global Research at:
the Disclosure appendix, and with the Disclaimer, which forms part of it. https://www.research.hsbc.com
Multi-Asset ● Quant
2 May 2019
Contents
Taking a dip in the data 3
Goldilocks waning; correlations rising 3
Much ado about low vol 8
The Great Gapsby 9
Deep dive 11
Related reports 12
Asset returns 13
Monthly returns (sorted by performance) 14
Monthly returns (sorted by asset) 15
Annual returns (sorted by performance) 16
Annual returns (sorted by asset) 17
Risk indicators 18
Cross asset volatility indicator (CRAVIN) 19
HSBC risk monitor 20
Multi asset volatility indicator suite (MAVIS) 21
Asset class realised volatility histories 22
Asset class volatility monitors 26
Bond/equity portfolio volatility 40
Volatility DNA profile 41
Volatility waves 41
Correlations 42
Cross-asset correlations 43
Equity country correlations 46
Global equity sector correlations 49
US equity sector correlations 52
Europe equity sector correlations 55
Japan equity sector correlations 58
EM equity sector correlations 61
FX correlations 64
2-year rates correlations 67
10-year rates correlations 70
Commodities correlations 73
Bond/equity correlations 76
Methodology 78
Disclosure appendix 83
Disclaimer 87
2
2 May 2019
Taking a dip in the data
Goldilocks waning; correlations rising
Multi-asset performance during 2019 has been a pleasant change from a very challenging
2018. All major asset classes have performed well so far this year (Chart 1), and this has been
accompanied by a widespread perception that we are now in a so-called Goldilocks period for
financial markets.
1. Monthly asset-class total returns
Source: MSCI, Refinitiv Datastream, Bloomberg, HSBC

Note: DM equity = MSCI World Index; EM equity = MSCI EM Index; DM rates = Global Treasury Index; DM IL = Global Aggregate Inflation-Linked Index; USD IG = US
Investment Grade Corporate Credit Index; USD HY = US High Yield Corporate Index; EMD USD = EM USD debt; EM Local = EM local-currency debt; Comdty = Composite
Commodity Index; REITs = Real Estate Investment Trusts Index. All returns are total returns in USD.
However, when we look more closely at the monthly breakdown in Chart 1 we can see that the
true Goldilocks period was concentrated in January, during which all major asset classes
performed well. As 2019 has progressed, the pattern of asset class returns has looked less like
Goldilocks and more like a traditional risk-on move. During April we have seen positive returns
in risk-on asset classes such as equities and negative returns from more classic safe havens
3
2 May 2019
such as Developed Market rates. Nevertheless, April has been a much more challenging
environment than the beginning of the year.
This risk-on move has been accompanied by a tick up in the HSBC RORO Indicator (Chart 2),
which measures the strength of cross-asset correlations. This suggests to us that we might be
seeing a waning of the Goldilocks story and are now simply living through a classic risk-on surge.
2. Strength of correlations across a wide range of asset classes
Source: Bloomberg, Refinitiv Datastream, HSBC
What does this mean for diversification?

When the RORO Indicator is at high levels it typically means that a single factor is dominating
the returns of all asset classes. This, clearly, makes diversification challenging to achieve.
However, whilst the indicator has recently ticked up, the overall level is still low – particularly
relative to the levels seen since the financial crisis. In such a situation, interpreting what the
recent increase in correlation strength means for diversification opportunities is a bit more
subtle.
We first look at what has happened to correlations for risk-on and risk-off asset groups
separately. Since we are looking at a very recent increase in the RORO indicator, we look at
short-term correlations (Chart 3). These correlations for risk-off assets appear to still be rising,
whereas there are signs that the correlations amongst risk-on assets might have peaked.
3. Correlations across risk-on and risk-off asset classes
1.0 0.9
0.8
0.8 0.7
0.6
0.6
0.5
0.4
0.4
0.3
0.2 0.2
0.1
0.0 0.0
2001 2004 2007 2010 2013 2016 2019 2001 2004 2007 2010 2013 2016 2019
Risk-on correlation (rolling 13-week) Risk-off correlation (rolling 13-week)
Source: Bloomberg, HSBC Source: Bloomberg, HSBC

Note: Average correlation between DM equities, EM equities, USD HY, EM debt, Note: Average correlation between gold, US Treasuries, Bunds and JPY-USD
cyclical commodities, DM REITs
4
2 May 2019
These embryonic signs of diversification amongst risk-on assets are bolstered by the evidence
from the more traditional channel of multi-asset diversification – namely between risk-on and
risk-off asset classes. The correlation between equity and bond returns in major developed
markets has been reliably negative since at least the start of the year (Charts 4 and 5) and has
recently decreased further.
4. Bond-equity correlation (US)
Source: MSCI, Refinitiv Datastream, HSBC

Note: Correlations calculated using weekly total returns over 6-month window
5. Bond-equity correlations (Germany)

Note: Correlations calculated using weekly total returns over 6-month window
Drilling down into cross-asset correlations

With the RORO Indicator being at relatively modest levels it suggests that there is not just one
dominant factor driving asset class returns at the moment. At times like this we find it useful to
turn to our dendrogram analysis. These dendrograms are very information-rich charts which
illustrate how different asset classes are moving together. They are produced by an
unsupervised machine learning algorithm1 which finds natural clusters within a dataset.
We show the dendrogram for cross-asset correlations over the last six months in Chart 6. There
are some natural clusters in here.
______________________________________
1 Agglomerative hierarchical clustering, using the complete linkage algorithm.
5
2 May 2019
 Towards the right of the chart we can see (in red) a classic risk-on grouping of assets such
as equities.
 Towards the middle of the dendrogram (in blue) we can see what appears to be a USD-
dominated cluster of USD-FX rates and commodities. To us, it seems likely that the driving
force behind this USD-dominated cluster is the dovish pivot from the Fed seen early this
year.
 Towards the left of the chart (in purple) we can see a DM rates cluster. There are two
interesting points to note about this DM rates cluster, which we outline in more detail below.
6. Multi-asset dendrogram
First, when we look within the cluster we can see that the three IG credit series have formed a
cluster. This is unusual – we more typically see the USD IG corporate credit returns cluster most
strongly with US 10y government bond returns, and similarly for the other currencies. This
suggests that credit spreads have been much more closely correlated lately than usual. Indeed,
Chart 7 shows that correlations among the three IG credit series have risen notably lately.
7. IG credit spreads much more closely correlated recently

0.8
0.6
0.4
0.2
0.0
Jan-17 Jul-17 Jan-18 Jul-18 Jan-19
Average rolling 40-day correlation USD IG, EUR IG, GBP IG (ASW)
6
2 May 2019
Second, whilst we would typically think of DM rates as being a classic risk-off asset class we
can see that other classic risk-off assets such as the JPY and Precious Metals have been
grouped into a rather strange-looking cluster with MXN and EM debt (local currency). This
grouping is so bizarre that we suspect that it is an anomaly caused by a rapidly changing
correlation structure. We calculate these correlations over a 6-month window: This window
includes the sharp risk-off period towards the end of 2018, the classic Goldilocks moves of early
2019, and the most recent ‘classic’ risk-on move. It is likely that this 6-month window is simply
averaging over too many correlation regimes.
Synchronised again – this time on the downside
What the dendogram in Chart 6 is also showing is a tight cluster of 10-year developed market
(DM) sovereign bonds. This suggests a common, global driver behind the recent moves in long-
term DM rates. From a fundamental point of view the most obvious driving force is the fairly
synchronised slowing of global growth and inflation.
Granted, some of the Q1 GDP prints were not as bad as commonly feared. But this was largely
due to one-offs and, as our economists put it, ‘The wrong type of growth’ (30 Apr 19). This has
been – and indeed is still being – reflected on the long-end of sovereign bond yield curves
almost across the board. As a result, correlations among 10-year sovereign bonds have risen
strongly since the start of the year (Charts 8 and 9). We note Russia, Turkey and South Africa
are simply outliers facing acute idiosyncratic challenges.
8. 10y rates correlations during 2018… 9. … and 2019 year-to-date
Source: HSBC, Bloomberg Source: HSBC, Bloomberg
In Chart 8 we show correlations across 10-year sovereign bond markets for 2018. Chart 9 plots
the equivalent correlations for 2019-YTD. We have deliberately drawn these heat maps small so
we can show them side-by-side. It is the comparison of the colours in the heat map which is
important here, rather than the precise ordering of assets within the heat map. During 2018
there were relatively strong correlations amongst DM 10 rates (the dark red block in the centre
of the heat map), but the other correlations were relatively weak. In contrast, Chart 9 shows a
much larger dark red area, suggesting that the majority of 10-year sovereign bond yields across
the globe have been dominated by a common factor during 2019. We believe this common
factor is that they are increasingly pricing in a synchronised slowdown of global growth and
global inflation.
7
2 May 2019
10. 2y rates correlations during 2018… 11. … and 2019 year-to-date
Source: HSBC, Bloomberg Source: HSBC, Bloomberg
This highly correlated move YTD has not been confined to the 10-year segment though. Charts
10 and 11 show the same correlation matrices as above, but this time for 2-year sovereign bond
yields. Again, the important thing here is to compare the overall shading within the heat map,
rather than the ordering of assets within it. This part of the yield curve reflects central bank
expectations, rather than the outlook for (structural) growth and inflation.
Despite the different natural drivers of 2y and 10y rates, we see a similar story. 2018 saw fairly
weak correlations across the short-end of global sovereign bonds. During 2018 there were
strong policy divergences between different countries. Some markets were pricing rate hikes,
whereas others were still pricing cuts. This correlation profiles has changed notably now. Chart
11 again exhibits much more red, reflecting the widespread and synchronised dovish central
bank pivot in the past four months.
However, it is notable that the shift in correlations at the 2y point is not as stark as seen for the
10y point. This also underlines that the synchronised slowdown theme is well baked into longer
dated rates, while the short-dated yields appear to be still linked with respective Central Bank’s
rhetoric. There are still some rate hikes being priced into the front end of EUR and GBP curves
while for the USD markets are categorically looking for cuts from here on. A subtle element of
policy divergence, at least in the forward rates markets, still remain.
Much ado about low vol
In 2018 we flagged to our readers a growing volatility wave in multi-asset returns (Chart 12).
On the basis of the previous behaviour of volatility waves we expected this wave to continue
growing, with elevated realised volatility spilling over into other asset classes. Clearly we were
wrong – this most recent volatility wave has dissipated.
8
2 May 2019
12. The volatility wave has ebbed
Throughout 2019 we have seen an intense period of volatility compression. This can be clearly
seen in our CRoss-Asset Volatility INdicator (CRAVIN, Chart 13).
Market participants seem to demand precisely one explanation whenever we see periods of
sustained low volatility. However, we rarely believe that low volatility environments can be
explained by any single factor, believing instead that there are a range of possible explanations
with varying degrees of plausibility (see our previous notes on cross-asset volatility: The Bonfire
of the Volatilities, 25 Sep 2017 and Ideas for Jekyll or Hyde, 17 Jul 2014). We believe that the
two important factors behind the recent drop in volatility are the dovish pivot of central banks
and the consensus belief in a Goldilocks scenario.
13. CRAVIN on the back foot again

4.1
1.2 95th percentile
HSBC Cross-Asset
Indicator level
0.6
Volatility Indicator
0.0
-0.6 Periods of
sustained low
5th percentile volatility
-1.2
99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19
For more detail please see

the full piece
The Great Gapsby
In addition to the current low vol environment, there have been several other low vol periods
since the financial crisis. Despite this, the frequency of large daily increases in volatility has
increased since the financial crisis. In essence, the volatility of volatility has increased, leading
to markets being more “gappy” than before.
9
2 May 2019
This gap risk is not just manifested in these spikes in implied volatility. It is sometimes even
more pronounced for intraday price moves. In Charts 14-16 we compare the distribution of daily
ranges for two time periods: (i) 2001-20082 and (ii) 2011-today3. Chart 14 shows the results for
the S&P 500, Chart 15 for EUR-USD, and Chart 16 for 10Y Bund futures.
14. S&P 500 distribution of daily high-low 15. EUR-USD distribution of daily high-low
400 700
350 10 40
600
30
300 500
5 20
250 10
400
200 0 0
4.0% 4.8% 5.6% 6.4% 7.2% 8.0% 300 2% 3% 4%
150
200
100
50 100
0 0
0% 1% 2% 3% 4% 5% 6% 7% 0.0% 0.5% 1.0% 1.5% 2.0% 2.5% 3.0% 3.5% 4.0% 4.5%
2001-2008 2011-today 2001-2008 2011-today
Source: HSBC, Bloomberg Source: HSBC, Bloomberg;
15. 10Y Bund futures distribution of daily high-low

500
5
400
300
0
200 1.90 2.30 2.70 3.10 3.50
100
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4
2001-2008 2012-today
Source: Bloomberg, HSBC
In all three cases there is more weight in the right tail of the distribution. This is particularly
notable for the 10Y Bund future. It is interesting to see that the frequency of right-tailed moves
are higher for Bunds than the S&P and EUR-USD. This could be partly due to the increased
occurrences of politically driven safe haven flows into Bunds. Additionally, Bunds have gone
through a sustained period of collateral squeeze during ECB QE. When the funding market
(Bunds repo) are dysfunctional and squeezed outsized moves on German yields became
common.
The Jekyll and Hyde behaviour of financial markets

Whether this gappy behaviour matters will depend on the time-frame of the investor. For long-
term holders of assets these brief volatility spikes are of little concern. In contrast, for market
makers the combination of generally low volatility with sudden unpredictable bursts of high
volatility is a particularly challenging combination.
______________________________________
2 Jan01-Jun08
3 For the 10Y Bund future we begin during 2012 so as not to include the sovereign debt crisis
10
2 May 2019
Deep dive
Hopefully, like Goldilocks, you have found this first section to be “just right”. If, however, you feel
that there is a gap in your knowledge which can only be filled by a multi-asset quant chart pack,
then we have a wealth of information for you in the following pages.
The deep dive begins with a look at returns from a variety of asset classes (page 13). We follow
this with an in-depth look at volatility (page 18). We analyse volatility not only within individual
asset classes, but also from a multi-asset perspective with our HSBC CRAVIN (CRoss-Asset
Volatility INdicator) indicator. Next we apply our powerful correlation analysis to a variety of
different data sets (page 42). First we investigate how different asset classes have moved
together (page 43), then we focus on the correlation structure within single asset classes. Last,
but not least, we track the correlation between sovereign bonds and equity markets (page 76).
11
2 May 2019
Related reports
Previous Data Matters
Data Matters: Diversification… apart from when you need it, 17 January 2019
Data Matters: RORO Resurgence, 17 October 2018
Data Matters: EM-DM Vol Divergence, 8 August 2018
Data Matters: EM underperforms, US dominates, 23 May 2018
Data Matters: Riddles, mysteries and enigmas, 12 April 2018
Other periodicals
HSBC Multi-Asset Heat Map (interactive webpage)
HSBC Surprise Indices, 1 May 2019
HSBC FX Positioning Indicators, 29 April 2019
FX Volametrics (interactive webpage)
Specials
Machine Learning 102: Tree hugging, 1 May 2019
The Great Gapsby, 1 May 2019
Machine Learning 101: Everybody needs good neighbours, 11 March 2019
Alternative Data – Text analysis of 20,000 earnings calls, 13 September 2018
A question of priorities – Country or sector: what matters more?, 14 February 2018
Comfortably numb, 7 August 2017
FX and the yield curve, 24 July 2017
The seasonality illusion, 3 May 2017
12
2 May 2019
Asset returns
Section contents
Monthly returns (by performance) p. 14

Monthly returns (by asset) p. 15
Annual returns (by performance) p. 16
Annual returns (by asset) p. 17
13
Monthly returns (sorted by performance)
Source: MSCI, Refinitiv Datastream, Bloomberg, HSBC Note: DM equity = MSCI World Index; EM equity = MSCI EM Index; DM rates = Global Treasury Index; DM IL = Global Aggregate Inflation-Linked Index; USD IG = US Investment Grade Corporate Credit Index; USD HY = US High Yield
Corporate Index; EMD USD = EM USD debt; EM Local = EM local-currency debt; Comdty = Composite Commodity Index; REITs = Real Estate Investment Trusts Index. All returns are total returns in USD.
2 May 2019
Click for: Contents | Asset returns | Risk indicators | Correlations

14
Monthly returns (sorted by asset)
2 May 2019

15
Annual returns (sorted by performance)
3. Annual asset-class total returns over the past 25 years
2 May 2019

16
Annual returns (sorted by asset)
4. Annual asset-class total returns over the past 25 years
Source: MSCI, Refinitiv Datastream, Bloomberg, HSBC Note: DM equity = MSCI World Index; EM equity = MSCI EM Index; DM rates = Global Treasury Index; DM IL = Global Aggregate Inflation-Linked Index; USD IG = US Investment Grade Corporate Credit Index; USD HY =
US High Yield Corporate Index; EMD USD = EM USD debt; EM Local = EM local-currency debt; Comdty = Composite Commodity Index; REITs = Real Estate Investment Trusts Index. All returns are total returns in USD.
2 May 2019

17
2 May 2019
Risk indicators
Section contents
Cross asset volatility indicator (CRAVIN) p. 19

Risk monitor p. 20
Multi asset volatility indicator suite p. 21
Asset class realised volatility histories p. 22
Asset class volatility monitors p. 26
Bond/equity portfolio vol. p. 40
Volatility DNA profile p. 41
Volatility waves p. 41
18 Click for: Contents | Asset returns | Risk indicators | Correlations

2 May 2019
Cross asset volatility indicator (CRAVIN)
5. Composite index of 3-month realised volatility across asset classes
The cross asset volatility indicator is a composite measure of the 3-month realised volatility of
total returns for all of the assets listed in table below. See page 81 for full details of the
construction methodology.
Table 1. Assets included in the Cross Asset Volatility Indicator

Equities Credit Rates FX Commodities
North America USD Investment Grade US 10Y EUR-USD Industrial Metals
Europe USD High Yield German 10Y GBP-USD Precious Metals
Pacific EUR Investment Grade Japan 10Y JPY-USD Energy
EM Asia EUR High Yield Italy 10Y AUD-USD Agriculture
EM Europe & Middle East GBP Investment Grade UK 10Y CAD-USD Livestock
EM Latin America EM USD debt KRW-USD
EM Money Markets MXN-USD
ZAR-USD
HUF-USD
BRL-USD
INR-USD
Source: HSBC
Click for: Contents | Asset returns | Risk indicators | Correlations 19

2 May 2019
HSBC risk monitor
6. Cross-asset risk premia
Source: HSBC, Bloomberg
Table 2. Details of risk measures in the above chart

Risk measure Description
Equity implied vol. VIX Index
Rates implied vol. USD 3M10Y ATM swaption volatility
CDS spreads Markit CDX North America Investment Grade Index
Gold implied vol. Gold 3M ATM implied volatility
FX implied vol. Global Hazard Indicator (composite indicator of EURUSD, USDJPY, EURJPY 3M ATM implied
volatility – see https://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp001.pdf for details)
Source: HSBC

2 May 2019
Multi asset volatility indicator suite (MAVIS)
7. Annualised realised volatility ranges across asset classes
Source: MSCI, Bloomberg, Refinitiv Datastream, HSBC

Note: DM equity = MSCI World Index; EM equity = MSCI EM Index; DM rates = Global Treasury Index; DM IL = Global Aggregate Inflation-Linked Index; USD IG = US
Investment Grade Corporate Credit Index; USD HY = US High Yield Corporate Index; EMD USD = EM USD debt; EM Local = EM local-currency debt; Comdty = Composite
Commodity Index; REITs = Real Estate Investment Trusts Index.; USD = DXY Index
All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data.

2 May 2019
Asset class realised volatility histories

DM equities
8. Developed market equity total return index 3-month realised volatility
EM equities
9. Emerging market equity total return index 3-month realised volatility
DM rates
10. Developed market government bond total return index 3-month realised volatility
Source: Refinitiv Datastream, HSBC

2 May 2019
Inflation-linked bonds
11. Inflation-linked bond total return index 3-month realised volatility
USD IG corporate credit
12. USD investment grade corporate credit total return 3-month realised volatility
USD HY corporate credit
13. USD high yield corporate credit total return index 3-month realised volatility

2 May 2019
EM USD debt
14. Emerging market USD debt total return index 3-month realised volatility
EM local-currency debt
15. EM local-currency debt total return index 3-month realised volatility
Commodities
16. Commodities total return index 3-month realised volatility

2 May 2019
REITS
17. REITs total return index 3-month realised volatility
USD
18. USD index return 3-month realised volatility

Note: The USD index is the Bloomberg DXY index which represents the average value of the USD against major currencies.

2 May 2019
Asset class volatility monitors

Developed market country equity volatility
19. Developed market country equity realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data.

2 May 2019
Emerging market equity volatility
20. Emerging market equity realised volatility ranges


2 May 2019
Global equity sector volatility

21. Global equity realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data. ENE = Energy, MAT = Materials, CAP = Capital Goods, CMS = Commercial
Services, TPT = Transport, AUT = Autos, CDU = Consumer Durables, CSV = Consumer Services, MED = Media, RET = Retailing, FRT = Food Retailing, FBV = Food and
Beverages, HHP = Household Products, HEQ = Healthcare Equipment, PHA = Pharmaceuticals, BNK = Banks, DIV = Diversified Financials, INS = Insurance, RES = Real
Estate, SFW = Software, TCH = Technology Hardware, SEM = Semiconductors, TEL = Telecoms, UTL = Utilities.

2 May 2019
US equity sector volatility

22. US equity realised volatility ranges


2 May 2019
Europe equity sector volatility

23. European equity realised volatility ranges


2 May 2019
Japan equity sector volatility

24. Japanese equity realised volatility ranges


2 May 2019
Emerging market equity sector volatility

25. Emerging market equity realised volatility ranges


2 May 2019
Developed market FX volatility
26. Developed market FX realised volatility ranges


2 May 2019
Emerging market FX volatility
27. Emerging market FX realised volatility ranges


2 May 2019
Developed market 2-year interest rate swap volatility
28. Developed market 2-year interest rate swap realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data and are expressed in basis points.

2 May 2019
Emerging market 2-year interest rate swap volatility
29. Emerging market 2-year interest rate swap realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data and are expressed in basis points. We include all countries for which we
have data since 2006. For Russia and Turkey, the floating rate is fixed against 3m USD Libor.

2 May 2019
Developed market 10-year interest rate swap volatility
30. Developed market 10-year interest rate swap realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data and are expressed in basis points.

2 May 2019
Emerging market 10-year interest rate swap volatility
31. Emerging market 10-year interest rate swap realised volatility ranges

Note: All volatilities are annualised 3-month volatilities calculated on a rolling basis using daily data and are expressed in basis points. We include all countries for which we
have data since 2006. For Russia and Turkey, the floating rate is fixed against 3m USD Libor.

2 May 2019
Commodities volatility
32. Commodities realised volatility ranges


2 May 2019
Bond/equity portfolio volatility

Global portfolios
33. 3-month realised volatility for a range of global equity/bond portfolios
Source: MSCI, Bloomberg, Refinitiv Datastream, HSBC Note: Vol. calculated using a global aggregate bond index and MSCI All Country for equities.
USD portfolios
34. 3-month realised volatility for a range of USD equity/bond portfolios
Source: MSCI, Bloomberg, Refinitiv Datastream, HSBC Note: Vol. calculated using a USD-denominated bond index and MSCI US for equities.
EUR portfolios
35. 3-month realised volatility for a range of EUR equity/bond portfolios
Source: MSCI, Bloomberg, Refinitiv Datastream, HSBC Note: Vol. calculated using a EUR-denominated bond index and MSCI EMU for equities.

2 May 2019
Volatility DNA profile
36. Cross-asset volatility DNA

Note: Volatilities are calculated using total return indices for daily data over rolling six-month time windows.
To construct the DNA profile, we first calculate the six-month realised volatility for each asset
through time. We then map these vol. levels to a value between zero and one based on the
percentile it falls into over the full vol. history for that asset. A deep red region in the DNA profile for
an asset indicates a period of high volatility, while a deep blue highlights a period of low volatility.
Volatility waves
37. Cross-asset volatility waves
The volatility wave chart contains exactly the same information as our DNA charts, but at each
time step we’ve simply sorted by the level of vol. This helps to highlight how broad-based high or
low vol. was across asset classes during different periods of time.

2 May 2019
Correlations
Section contents
Cross-asset correlations p. 43
Equity country correlations p. 46
Equity sector correlations p. 49
FX correlations p. 64
2-year rates correlations p. 67
10-year rates correlations p. 70
Commodities correlations p. 73
Bond/equity correlations p. 76
For each asset class we show:
 Heat map showing correlations over the past 6 months
 Index measuring the overall level of correlations
 Dendrogram highlighting the key correlations
 Major correlations changes over the past 12 months

2 May 2019
Cross-asset correlations
38. Cross-asset correlations over the past 6 months

Note: All correlations are calculated using local-currency total return indices. The assets are ordered such that those with similar correlations to other assets are next to each other.
The heat map shows the correlations between different assets over the past six months.
 Dark red regions indicate strong positive correlations;
 Dark blue regions indicate strong negative correlations;
 Yellow/green indicates weak correlation/uncorrelated assets.
See HSBC Multi-Asset Heat Map for an interactive web version of the above heat map.
Table 3. Assets included in the heat map

EM Europe & Middle East GBP Investment Grade UK 10Y CAD-USD Livestock
EM Latin America EM USD debt KRW-USD
EM Money Markets MXN-USD
ZAR-USD
HUF-USD
BRL-USD
INR-USD
Source: HSBC

2 May 2019
Risk On – Risk Off (RORO) Indicator
39. RORO Indicator measuring the strength of cross-asset correlations
The correlation indicator measures the strength of the correlations between the assets listed in the
table on the previous page. An increase in the indicator implies rising correlations, with a value of
one corresponding to perfect correlations between all assets. See page 78 for more details.
Key cross-asset correlations
40. Dendrogram showing the key cross-asset correlations over the past 6 months
The dendrogram pulls out the key correlations in the heat map on the previous page. As one
moves up the chart, assets are joined based on the strength of their correlation, with the most
correlated assets combined at the smallest distances. See page 78 for more details.

2 May 2019
Major changes in cross-asset correlations
41. Asset pairs where the correlations have changed significantly over the past 12 months
Source: Bloomberg, Refinitiv Datastream, HSBC Source: Bloomberg, Refinitiv Datastream, HSBC
Note: All correlations are calculated over rolling 6-month windows using daily data.

2 May 2019
Equity country correlations

42. Equity country correlations over the past 6 months

Note: All correlations are calculated using local-currency total return indices. The assets are ordered such that those with similar correlations to other assets are next to each other.
Table 4. Countries included in the heat map

Developed markets Emerging markets
Australia Italy Brazil South Africa
Belgium Japan Chile Taiwan
Canada Netherlands China Thailand
Denmark Norway India Turkey
Finland Singapore Indonesia Philippines
France Spain Korea
Germany Sweden Malaysia
Hong Kong Switzerland Mexico
Ireland UK Poland
Israel US Russia
Source: HSBC

2 May 2019
Equity country correlation indicator
43. Indicator measuring the strength of equity country correlations
Key equity country correlations
44. Dendrogram showing the key equity country correlations over the past 6 months

2 May 2019
Major changes in equity country correlations
45. Country pairs for which equity correlations have changed significantly over the past 12 months
Source: MSCI, Refinitiv Datastream, HSBC Source: MSCI, Refinitiv Datastream, HSBC

2 May 2019
Global equity sector correlations

46. Global equity sector correlations over the past 6 months

Note: All correlations are calculated using total return indices. The assets are ordered such that those with similar correlations to other assets are next to each other.
Table 5. Sectors included in the heat map

Code Sector Code Sector Code Sector Code Sector
ENE Energy CDU Consumer Durables HHP Household Products RES Real Estate
MAT Materials CSV Consumer Services HEQ Healthcare Equipment SFW Software
CAP Capital Goods MED Media PHA Pharmaceuticals TCH Technology Hardware
CMS Commercial Services RET Retailing BNK Banks SEM Semiconductors
TPT Transport FRT Food Retailing DIV Diversified Financials TEL Telecoms
AUT Autos FBV Food and Beverages INS Insurance UTL Utilities
Source: HSBC

2 May 2019
Global equity sector correlation indicator
47. Indicator measuring the strength of global equity sector correlations
Key global equity sector correlations
48. Dendrogram showing the key global equity sector correlations over the past 6 months

2 May 2019
Major changes in global equity sector correlations
49. Global equity sectors where the correlations have changed significantly over the past 12 months

2 May 2019
US equity sector correlations

50. US equity sector correlations over the past 6 months


Source: HSBC

2 May 2019
US equity sector correlation indicator
51. Indicator measuring the strength of US equity sector correlations
Key US equity sector correlations
52. Dendrogram showing the key US equity sector correlations over the past 6 months

2 May 2019
Major changes in US equity sector correlations
53. US equity sectors where the correlations have changed significantly over the past 12 months

2 May 2019
Europe equity sector correlations

54. European equity sector correlations over the past 6 months


Source: HSBC

2 May 2019
European equity sector correlation indicator
55. Indicator measuring the strength of European equity sector correlations
Key European equity sector correlations
56. Dendrogram showing the key Europe equity sector correlations over the past 6
months

2 May 2019
Major changes in European equity sector correlations
57. European equity sectors where the correlations have changed significantly over the past 12 months

2 May 2019
Japan equity sector correlations

58. Japan equity sector correlations over the past 6 months


Source: HSBC

2 May 2019
Japan equity sector correlation indicator
59. Indicator measuring the strength of Japan equity sector correlations
Key Japan equity sector correlations
60. Dendrogram showing the key Japan equity sector correlations over the past 6 months

2 May 2019
Major changes in Japan equity sector correlations
61. Japan equity sectors where the correlations have changed significantly over the past 12 months

2 May 2019
EM equity sector correlations

62. Emerging market equity sector correlations over the past 6 months


Source: HSBC

2 May 2019
Emerging market equity sector correlation indicator
63. Indicator measuring the strength of EM equity sector correlations
Key emerging market equity sector correlations
64. Dendrogram showing the key EM equity sector correlations over the past 6 months

2 May 2019
Major changes in EM equity sector correlations
65. EM equity sectors where the correlations have changed significantly over the past 12 months

2 May 2019
FX correlations
66. USD-based FX correlations over the past 6 months

Note: The assets are ordered such that those with similar correlations to other assets are next to each other.
Table 10. USD-based exchange rates included in the heat map

Code Country Code Country Code Country Code Country Code Country
AUD Australia COP Colombia ILS Israel NOK Norway SEK Sweden
BRL Brazil CZK Czech Republic INR India NZD New Zealand SGD Singapore
CAD Canada EUR Eurozone JPY Japan PEN Peru THB Thailand
CHF Switzerland GBP UK KRW Korea PHP Philippines TRY Turkey
CLP Chile HUF Hungary MXN Mexico PLN Poland TWD Taiwan
CNY China IDR Indonesia MYR Malaysia RUB Russia ZAR South Africa
Source: HSBC

2 May 2019
FX correlation indicator
67. Indicator measuring the strength of correlations between USD-based exchange rates
Key FX correlations
68. Dendrogram showing the key USD-based FX correlations over the past 6 months

2 May 2019
Major changes in FX correlations
69. USD-based FX rates for which the correlations have changed significantly over the past 12 months


2 May 2019
2-year rates correlations

70. 2-year interest rate swap correlations over the past 6 months

Table 11. 2-year interest rate swaps included in the heat map
Australia Norway Brazil Israel Singapore
Canada Sweden Chile India South Africa
Denmark Switzerland Colombia Korea Taiwan
Eurozone UK Czech Malaysia Thailand
Japan US Hungary Poland Turkey
New Zealand Hong Kong Russia
Source: HSBC

2 May 2019
2-year rates correlation indicator
71. Indicator measuring the strength of correlations between 2-year interest rate swaps
Key 2-year rates correlations
72. Dendrogram showing the key 2-year swap rate correlations over the past 6 months

2 May 2019
Major changes in 2-year rates correlations
73. 2-year swap rate pairs for which correlations have changed significantly over the past 12 months


2 May 2019
10-year rates correlations
74. 10-year interest rate swap correlations over the past 6 months

Table 12. 10-year interest rate swaps included in the heat map
Australia Norway Chile Israel Singapore
Canada Sweden Colombia Korea South Africa
Denmark Switzerland Czech Malaysia Taiwan
Eurozone UK Hungary Mexico Thailand
Japan US Hong Kong Poland Turkey
New Zealand India Russia
Source: HSBC

2 May 2019
10-year rates correlation indicator
75. Indicator measuring the strength of correlations between 10-year interest rate swaps
Key 10-year rates correlations
76. Dendrogram showing the key 10-year swap rate correlations over the past 6 months

2 May 2019
Major changes in 10-year rates correlations
77. 10-year swap rate pairs for which correlations have changed significantly over the past 12 months


2 May 2019
Commodities correlations
78. Commodities correlations over the past 6 months

Table 13. Commodities included in the heat map

Precious Metals Industrial Metals Energy Agriculture Livestock
Gold Aluminium Gasoline Coffee Lean hogs
Silver Copper Heating oil Corn Live cattle
Nickel Natural gas Soybean
Zinc Oil (Brent) Soybean meal
Oil (WTI) Soybean oil
Sugar
Wheat
Kansas wheat
Source: HSBC

2 May 2019
Commodities correlation indicator
79. Indicator measuring the strength of correlations between commodities
Key commodities correlations
80. Dendrogram showing the key commodities correlations over the past 6 months

2 May 2019
Major changes in commodities correlations
81. Commodity pairs for which correlations have changed significantly over the past 12 months


2 May 2019
Bond/equity correlations
US
82. US 10Y government bonds and equities correlation
Source: MSCI, Refinitiv Datastream, HSBC Note: Correlations calculated using weekly total returns for bonds and price returns for equities over 6m windows.
Japan
83. Japan 10Y government bonds and equities correlation
UK
84. UK 10Y government bonds and equities correlation

2 May 2019
Germany
85. Germany 10Y government bonds and equities correlation
Italy
86. Italy 10Y government bonds and equities correlation
Spain
87. Spain 10Y government bonds and equities correlation

2 May 2019
Methodology
Correlation indicators
The various correlation indicators in this report provide aggregate measures of the level of
correlations between broad groups of assets. In each case, the starting point is a set of total
returns4 (or yield differences in the case of rates). We then use principal component analysis
(PCA) to decompose these returns into a series of principal components (PCs), which are a new
set of series that represent the common movements across the original time series.
The first principal component can essentially be thought of as the most important factor driving
markets over a given period of time, the second PC represents the next most important factor, and
so on. The key constraints are that the PCs are all uncorrelated with each other, and they explain
progressively smaller amounts of the variance across the group of asset being considered.
We define each of the correlations indices as the proportion of the variance (across the group of
assets being considered) explained by the first PC. An increase in the indicator implies a general
increase in the strength of correlations, whereas a decrease implies a weakening in correlations.
The correlation indicator in this report include the Risk On – Risk Off (RORO) indicator, which
measures the strength of cross-asset correlations (see page 44), as well as indicators for equities
(page 47 and 50), FX (page 65), rates (page 68 and 71) and commodities (page 74).
Dendrogram construction
Dendrograms sit within the field of cluster analysis, a branch of statistics that aims to combine a
larger number of objects into a smaller number of groups (clusters). There are a wide variety of
different clustering algorithms which can be used, each with their own pros and cons. There is a
commonly-used family of clustering algorithms which are collectively called hierarchical
agglomerative clustering algorithms.
Complicated term, simple concept

This name sounds daunting; however, the idea behind these algorithms is simple.
 Initially the clustering algorithm starts with every object in a separate cluster.
 Then, the two most similar (see ‘defining similarity’ section below) clusters are brought
together into a single cluster (see ‘making a new cluster’ section below). So, if the algorithm
begins with 100 clusters (each containing one object), it now contains 99 clusters (one of
which now contains two object).
 This process continues until there is only one cluster (containing all the objects).
As a result of this process we end up with a hierarchy of different clustering arrangements. From
this hierarchy you can choose the clustering which you believe is most appropriate. For
example, you could specify the number of clusters you want and then simply walk up the
hierarchy until you get to the clustering arrangement which contains exactly that many clusters.
______________________________________
4 For equities the total return is the percentage change in price plus dividends. For bonds it is the percentage change in
price plus coupon and roll down.

2 May 2019
Alternatively, you could specify how dissimilar objects within a cluster can be and then choose
the hierarchical clustering which results at this distance level. Another option would be to
attempt to split the hierarchy according to some systematic rule based on the distances at which
different clustering arrangements occur.
Defining similarity
In order to decide which two objects should be joined together as the first non-unitary cluster,
first you need to define a distance between each pair of objects. There are various ways of
doing this; we use the following, somewhat strange-looking definition:
𝑑𝑖𝑗 = √2(1 − 𝜌𝑖𝑗 ) , where 𝑑𝑖𝑗 is the distance between objects 𝑖 and 𝑗
and 𝜌𝑖𝑗 is the correlation between objects 𝑖 and 𝑗
This definition is not as bizarre as it might initially appear and is strongly linked to the Euclidean
distance between the two time-series5.
Making a new cluster

To create the first cluster, the smallest value of dij is found; objects i and j are then joined into a
cluster and this cluster associated with the value dij. In order to proceed further we need to
define the distances between this new cluster and all other clusters. There are various methods
for doing this: common choices are single linkage, complete linkage, and average linkage. We
use the average linkage method in the following example.
In the average linkage clustering methodology, when cluster a and cluster b are joined to form a
new cluster, the distance at which this happens is the average distance between all pairs of
objects {(𝑥𝑎 , 𝑥𝑏 ), such that 𝑥𝑎 ∈ 𝑎, 𝑥𝑏 ∈ 𝑏}. To make this procedure explicit, we present a simple
example below.
Table 14. Example distance matrix between five objects (A, B, C, D, and E)
A B C D E
A 0
B 0.1 0
C 0.8 0.7 0
D 0.9 1.0 0.2 0
E 0.3 0.35 0.4 0.5 0
Source: HSBC
In Table 14 above we show a distance matrix which specifies the distances between five objects
(A, B, C, D, and E). We have highlighted the smallest distance: 0.1 between A and B. The first
step is that we join A and B into a new cluster (A,B) and calculate the distances between each
other object and this cluster. This gives the new distance matrix shown in Table 15.
Table 15. After combining A and B into (A,B)

(A,B) C D E
(A,B) 0
C 0.75 0
D 0.95 0.2 0
E 0.325 0.4 0.5 0
Source: HSBC
To calculate the distance between cluster (A,B) and cluster C we calculate the average of the
distance between A and C (0.8 from Table 14) and the distance between B and C (0.7). Thus,
______________________________________
5 For those interested in the details, see ‘An investigation into the dynamics of correlation networks in the foreign exchange
market’, Mark McDonald, DPhil thesis, University of Oxford, 2007.

2 May 2019
the distance between cluster (A,B) and cluster C is 0.75, which can be seen in Table 15. The
distances between (A,B) and the remaining clusters are calculated similarly.
In Table 15 we have highlighted the smallest distance: 0.2 between clusters C and D. We thus
join these two into a new cluster (C,D) and calculate the distances between this new cluster and
all others. This gives the new distance matrix shown in Table 16.
Table 16. After combining C and D into (C,D)

(A,B) (C,D) E
(A,B) 0
(C,D) 0.85 0
E 0.325 0.45 0
Source: HSBC
In Table 16 we can see that the distance between cluster (A,B) and cluster (C,D) is 0.85. Where
did this number come from? It is the average of the distances between:
 A and C (0.8)
 B and C (0.7)
 A and D (0.9)
 B and D (1.0)
In Table 16 we have highlighted the smallest distance: 0.325 between cluster (A,B) and cluster
E. We now combine these two clusters into a new cluster (A,B,E) and calculate the distances
between this new cluster and the only remaining cluster: (C,D). We show this in Table 17.
Table 17. After combining (A,B) and E into (A,B,E)

(A,B,E) (C,D)
(A,B,E) 0
(C,D) 0.7167 0
Source: HSBC
We can see from Table 17 that the distance between cluster (A,B,E) and (C,D) is 0.7167. This
number is the average of the distances between:
 A and C (0.8)
 B and C (0.7)
 E and C (0.4)
 A and D (0.9)
 B and D (1.0)
 E and D (0.5)
Recap
Let’s have a quick recap on how we got here:
 Object A and object B were combined into a cluster at a distance of 0.1
 Object C and object D were combined into a cluster at a distance of 0.2
 Object E and cluster (A,B) were combined into a cluster at a distance of 0.325
 Cluster (C,D) and cluster (A,B,E) were combined into a cluster at a distance of 0.7167
This procedure gives us the dendrogram shown in Chart 88.

2 May 2019
88. Average linkage clustering dendrogram
Source: HSBC
This dendrogram gives us a clustering arrangement for each distance value. We can specify
how dissimilar objects within a cluster are allowed to be (i.e. the cluster distance) and then this
dendrogram will give us the resulting clusters. Specifically, we can get:
 5 clusters (i.e. all objects in their own cluster) if we decide that we want to split the hierarchy
at a distance less than 0.1
 4 clusters if we split the hierarchy at a level bigger than 0.1 and less than 0.2
 3 clusters if we split the hierarchy at a level bigger than 0.2 but less than 0.35
 2 clusters if we split the hierarchy at a level bigger than 0.35 but less than 0.7167
 1 cluster (i.e. all objects in one big cluster) if we split the hierarchy at a level above 0.7167
Cross asset volatility indicator
The cross asset volatility indicator is an aggregate measure of realised volatility across a broad
set of assets. We start by calculating the 3-month realised volatility through time for the total
return indices of each of the assets listed in Table 18. We then transform the volatility over the
period from the start of the data in 1999 until August 2017 into a z-score. The z-score simply
standardises the volatility for each asset by subtracting the mean volatility over that period and
dividing by the standard deviation. The z-score thus represents the number of standard
deviations that a given volatility is away from the mean. To calculate the composite cross asset
volatility indicator we simply average these z-scores across all of the assets in Table 18 at each
time step.
For dates after July 2017, we repeat the above procedure, but only add the latest data point to
the original series constructed between 1999 and July 2017. The mean and standard deviation
used to calculate the z-scores are based on the full data history, and as new data comes in
these numbers change, and so the z-scores for the full history also change. By fixing the
indicator values over the initial window, we thus ensure that the history for the indicator does not

2 May 2019
change as new data become available and the means and standard deviations are updated. It’s
easy to show though that provided the initial window is reasonably long, the choice of this
window doesn’t have much of an effect on the final indicator.
Table 18. Assets included in the cross asset volatility indicator

EM Europe & Middle East GBP Investment Grade UK 10Y KRW-USD Livestock
EM Latin America EM USD debt MXN-USD
EM Money Markets ZAR-USD
HUF-USD
BRL-USD
INR-USD
Source: HSBC
Identifying major changes in correlations
Identifying pairs of assets that have experienced significant changes in correlations is

challenging because of the large number of possible correlations. For example, there are 33
assets in the table above, which means that there are 528 pairs of assets and thus 528
correlations. If you include all of the other assets that we consider in this document then looking
at a chart of each correlation in turn quickly becomes infeasible. Rather than doing this
manually, we therefore identify asset pairs that have experienced major correlation changes
algorithmically.
We use two key criteria to select the correlations:
1. The correlation has moved over a large range over the past year;
2. The correlation has not demonstrated large day-to-day fluctuations historically; specifically,
we look at the mean absolute deviation in daily correlation changes, which is essentially a
measure of the volatility of the correlations.
The first criterion serves to highlight pairs of assets that have seen a significant change in their
relationship over the past year: this could be a large increase in correlation; a large decrease; or
the correlation could have moved by a large amount in one direction and then back again. The
second criterion acts as a constraint to ensure that we’re not simply identifying those assets
pairs with volatile correlations. If the correlation between a pair of assets moves around a lot
with no clear historical patterns, then changes aren’t particularly informative.
The major correlations that we include in this document satisfy both criteria: they have
experienced significant changes, and these changes are unusual and do not simply reflect a
volatile series. We add a final constraint that each asset can appear at most twice amongst the
12 correlations that we select for each asset class. This ensures some diversity in the assets
included in the correlation charts.

2 May 2019
Disclosure appendix
Analyst Certification
The following analyst(s), economist(s), or strategist(s) who is(are) primarily responsible for this report, including any analyst(s)
whose name(s) appear(s) as author of an individual section or sections of the report and any analyst(s) named as the covering
analyst(s) of a subsidiary company in a sum-of-the-parts valuation certifies(y) that the opinion(s) on the subject security(ies) or
issuer(s), any views or forecasts expressed in the section(s) of which such individual(s) is(are) named as author(s), and any other
views or forecasts expressed herein, including any views expressed on the back page of the research report, accurately reflect
their personal view(s) and that no part of their compensation was, is or will be directly or indirectly related to the specific
recommendation(s) or views contained in this research report: Mark McDonald, Max Kettner, Subhrajit Banerjee, CFA, Dominic
Kini, Alastair Pinder, CFA and Peter Barnshaw
Important disclosures
Equities: Stock ratings and basis for financial analysis
HSBC and its affiliates, including the issuer of this report (“HSBC”) believes an investor's decision to buy or sell a stock should
depend on individual circumstances such as the investor's existing holdings, risk tolerance and other considerations and that
investors utilise various disciplines and investment horizons when making investment decisions. Ratings should not be used or
relied on in isolation as investment advice. Different securities firms use a variety of ratings terms as well as different rating
systems to describe their recommendations and therefore investors should carefully read the definitions of the ratings used in
each research report. Further, investors should carefully read the entire research report and not infer its contents from the rating
because research reports contain more complete information concerning the analysts' views and the basis for the rating.
Fixed income: Basis for financial analysis

This report is designed for, and should only be utilised by, institutional investors. Furthermore, HSBC believes an investor's
decision to make an investment should depend on individual circumstances such as the investor's existing holdings and other
considerations.
HSBC believes that investors utilise various disciplines and investment horizons when making investment decisions, which
depend largely on individual circumstances such as the investor's existing holdings, risk tolerance and other considerations. Given
these differences, HSBC has three principal aims in its fixed income research: 1) to identify long-term investment opportunities
based on particular themes or ideas that may affect the future earnings or cash flows of companies in corporate credit and based
on country-specific ideas or themes that may affect the performance of these bonds in the case of covered bonds, in both cases
on a six-month time horizon; 2) to identify trade ideas on a time horizon of up to three months, relating to specific instruments and
segments of the yield curve, which are predominantly derived from relative value considerations or driven by events and which
may differ from our long-term credit opinion on an issuer. Buy or Sell refer to a trade call to buy or sell that given instrument; 3) to
express views on the likely future performance of sectors, benchmark indices or markets in our fixed income strategy products.
HSBC has assigned a fundamental recommendation structure, as described below, only for its long-term investment opportunities.
HSBC believes an investor's decision to buy or sell a bond should depend on individual circumstances such as the investor's
existing holdings and other considerations. Different securities firms use a variety of terms as well as different systems to describe
their recommendations. Investors should carefully read the definitions of the recommendations used in each research report. In
addition, because research reports contain more complete information concerning the analysts' views, investors should carefully
read the entire research report and should not infer its contents from the recommendation. In any case, recommendations should
not be used or relied on in isolation as investment advice.
HSBC Global Research is not and does not hold itself out to be a Credit Rating Agency as defined under the Hong Kong Securities
and Futures Ordinance.
From 23rd March 2015 HSBC has assigned ratings on the following basis:
The target price is based on the analyst’s assessment of the stock’s actual current value, although we expect it to take six to 12
months for the market price to reflect this. When the target price is more than 20% above the current share price, the stock will
be classified as a Buy; when it is between 5% and 20% above the current share price, the stock may be classified as a Buy or a
83
2 May 2019
Hold; when it is between 5% below and 5% above the current share price, the stock will be classified as a Hold; when it is between
5% and 20% below the current share price, the stock may be classified as a Hold or a Reduce; and when it is more than 20%
below the current share price, the stock will be classified as a Reduce.
Our ratings are re-calibrated against these bands at the time of any 'material change' (initiation or resumption of coverage, change
in target price or estimates).
Upside/Downside is the percentage difference between the target price and the share price.
Prior to this date, HSBC’s rating structure was applied on the following basis:
For each stock we set a required rate of return calculated from the cost of equity for that stock’s domestic or, as appropriate,
regional market established by our strategy team. The target price for a stock represented the value the analyst expected the
stock to reach over our performance horizon. The performance horizon was 12 months. For a stock to be classified as Overweight,
the potential return, which equals the percentage difference between the current share price and the target price, including the
forecast dividend yield when indicated, had to exceed the required return by at least 5 percentage points over the succeeding 12
months (or 10 percentage points for a stock classified as Volatile*). For a stock to be classified as Underweight, the stock was
expected to underperform its required return by at least 5 percentage points over the succeeding 12 months (or 10 percentage
points for a stock classified as Volatile*). Stocks between these bands were classified as Neutral.
*A stock was classified as volatile if its historical volatility had exceeded 40%, if the stock had been listed for less than 12 months
(unless it was in an industry or sector where volatility is low) or if the analyst expected significant volatility. However, stocks which
we did not consider volatile may in fact also have behaved in such a way. Historical volatility was defined as the past month's
average of the daily 365-day moving average volatilities. In order to avoid misleadingly frequent changes in rating, however,
volatility had to move 2.5 percentage points past the 40% benchmark in either direction for a stock's status to change.
Rating distribution for long-term investment opportunities

As of 02 May 2019, the distribution of all independent ratings published by HSBC is as follows:
Buy 53% ( 30% of these provided with Investment Banking Services )
Hold 37% ( 28% of these provided with Investment Banking Services )
Sell 10% ( 21% of these provided with Investment Banking Services )
For the purposes of the distribution above the following mapping structure is used during the transition from the previous to current
rating models: under our previous model, Overweight = Buy, Neutral = Hold and Underweight = Sell; under our current model Buy
= Buy, Hold = Hold and Reduce = Sell. For rating definitions under both models, please see “Stock ratings and basis for financial
analysis” above.
Definitions for fundamental credit and covered bond recommendations from 22 April 2016
Overweight: For corporate credit, the issuer’s fundamental credit profile is expected to improve over the next six months. For
covered bonds, the bonds issued in this country are expected to outperform those of the other countries in our coverage over the
next six months.
Neutral: For corporate credit, the issuer’s fundamental credit profile is expected to remain stable over the next six months. For
covered bonds, the bonds issued in this country are expected to perform in line with those of the other countries in our coverage
over the next six months.
Underweight: For corporate credit, the issuer’s fundamental credit profile is expected to deteriorate over the next six months.
For covered bonds, the bonds issued in this country are expected to underperform those of other countries in our coverage over
the next six months.
Prior to this date, fundamental recommendations for corporate credit were applied on the following basis:
Overweight: The credits of the issuer were expected to outperform those of other issuers in the sector over the next six months.
Neutral: The credits of the issuer were expected to perform in line with those of other issuers in the sector over the next six
months.
Underweight: The credits of the issuer were expected to underperform those of other issuers in the sector over the next six
months.
84
2 May 2019
Distribution of fundamental credit and covered bond recommendations

As of 01 May 2019, the distribution of all independent fundamental credit recommendations published by HSBC is as follows:
All Covered issuers Issuers to whom HSBC has provided Investment Banking in the past 12 months
Count Percentage Count Percentage

Overweight 93 23 48 52
Neutral 227 56 104 46
Underweight 88 21 15 17
Source: HSBC
For the distribution of non-independent ratings published by HSBC, please see the disclosure page available at
http://www.hsbcnet.com/gbm/financial-regulation/investment-recommendations-disclosures.
Recommendation changes for long-term investment opportunities

To view a list of all the independent fundamental ratings disseminated by HSBC during the preceding 12-month period, please
use the following links to access the disclosure page:
Clients of Global Research and Global Banking and Markets: www.research.hsbc.com/A/Disclosures
Clients of HSBC Private Banking: www.research.privatebank.hsbc.com/Disclosures
HSBC and its affiliates will from time to time sell to and buy from customers the securities/instruments, both equity and debt
(including derivatives) of companies covered in HSBC Research on a principal or agency basis or act as a market maker or
liquidity provider in the securities/instruments mentioned in this report.
Analysts, economists, and strategists are paid in part by reference to the profitability of HSBC which includes investment banking,
sales & trading, and principal trading revenues.
Whether, or in what time frame, an update of this analysis will be published is not determined in advance.
Non-U.S. analysts may not be associated persons of HSBC Securities (USA) Inc, and therefore may not be subject to FINRA
Rule 2241 or FINRA Rule 2242 restrictions on communications with the subject company, public appearances and trading
securities held by the analysts.
Economic sanctions imposed by the EU and OFAC prohibit transacting or dealing in new debt or equity of Russian SSI entities.
This report does not constitute advice in relation to any securities issued by Russian SSI entities on or after July 16 2014 and as
such, this report should not be construed as an inducement to transact in any sanctioned securities.
For disclosures in respect of any company mentioned in this report, please see the most recently published report on that company
available at www.hsbcnet.com/research. HSBC Private Banking clients should contact their Relationship Manager for queries
regarding other research reports. In order to find out more about the proprietary models used to produce this report, please contact
the authoring analyst.
Additional disclosures
1 This report is dated as at 02 May 2019.
2 All market data included in this report are dated as at close 01 May 2019, unless a different date and/or a specific time of
day is indicated in the report.
3 HSBC has procedures in place to identify and manage any potential conflicts of interest that arise in connection with its
Research business. HSBC's analysts and its other staff who are involved in the preparation and dissemination of
Research operate and have a management reporting line independent of HSBC's Investment Banking business.
Information Barrier procedures are in place between the Investment Banking, Principal Trading, and Research businesses
to ensure that any confidential and/or price sensitive information is handled in an appropriate manner.
4 You are not permitted to use, for reference, any data in this document for the purpose of (i) determining the interest
payable, or other sums due, under loan agreements or under other financial contracts or instruments, (ii) determining the
price at which a financial instrument may be bought or sold or traded or redeemed, or the value of a financial instrument,
and/or (iii) measuring the performance of a financial instrument.
85
2 May 2019
MSCI Disclaimer
The MSCI information included in this report is for your internal use only, may not be reproduced or redisseminated in any form
and may not be used as a basis for or a component of any financial instruments or products or indices. None of the MSCI
information is intended to constitute investment advice or a recommendation to make (or refrain from making) any kind of
investment decision and may not be relied on as such. Historical data and analysis should not be taken as an indication or
guarantee of any future performance analysis, forecast or prediction. The MSCI information is provided on an "as is" basis and
the user of this information assumes the entire risk of any use made of this information. MSCI, each of its affiliates and each other
person involved in or related to compiling, computing or creating any MSCI information (collectively, the "MSCI Parties") expressly
disclaims all warranties (including, without limitation, any warranties of originality, accuracy, completeness, timeliness, non-
infringement, merchantability and fitness for a particular purpose) with respect to this information. Without limiting any of the
foregoing, in no event shall any MSCI Party have any liability for any direct, indirect, special, incidental, punitive, consequential
(including, without limitation, lost profits) or any other damages. (www.msci.com)
Production & distribution disclosures

1. This report was produced and signed off by the author on 02 May 2019 15:53 GMT.
2. In order to see when this report was first disseminated please see the disclosure page available at
https://www.research.hsbc.com/R/34/KlVxVCQ
86
2 May 2019
Disclaimer
Legal entities as at 30 November 2017 Issuer of report
‘UAE’ HSBC Bank Middle East Limited, Dubai; ‘HK’ The Hongkong and Shanghai Banking Corporation Limited, Hong Kong; HSBC Bank plc
‘TW’ HSBC Securities (Taiwan) Corporation Limited; 'CA' HSBC Securities (Canada) Inc.; HSBC Bank, Paris Branch; HSBC 8 Canada Square, London
France; ‘DE’ HSBC Trinkaus & Burkhardt AG, Düsseldorf; 000 HSBC Bank (RR), Moscow; ‘IN’ HSBC Securities and Capital E14 5HQ, United Kingdom
Markets (India) Private Limited, Mumbai; ‘JP’ HSBC Securities (Japan) Limited, Tokyo; ‘EG’ HSBC Securities Egypt SAE, Telephone: +44 20 7991 8888
Cairo; ‘CN’ HSBC Investment Bank Asia Limited, Beijing Representative Office; The Hongkong and Shanghai Banking Fax: +44 20 7992 4880
Corporation Limited, Singapore Branch; The Hongkong and Shanghai Banking Corporation Limited, Seoul Securities Website: www.research.hsbc.com
Branch; The Hongkong and Shanghai Banking Corporation Limited, Seoul Branch; HSBC Securities (South Africa) (Pty) Ltd,
Johannesburg; HSBC Bank plc, London, Madrid, Milan, Stockholm, Tel Aviv; ‘US’ HSBC Securities (USA) Inc, New York;
HSBC Yatirim Menkul Degerler AS, Istanbul; HSBC México, SA, Institución de Banca Múltiple, Grupo Financiero HSBC;
HSBC Bank Australia Limited; HSBC Bank Argentina SA; HSBC Saudi Arabia Limited; The Hongkong and Shanghai Banking
Corporation Limited, New Zealand Branch incorporated in Hong Kong SAR; The Hongkong and Shanghai Banking
Corporation Limited, Bangkok Branch; PT Bank HSBC Indonesia; HSBC Qianhai Securities Limited
This document is issued and approved in the United Kingdom by HSBC Bank plc for the information of its Clients (as defined in the Rules of FCA) and those of its affiliates only. If this research is
received by a customer of an affiliate of HSBC, its provision to the recipient is subject to the terms of business in place between the recipient and such affiliate. In Australia, this publication has
been distributed by The Hongkong and Shanghai Banking Corporation Limited (ABN 65 117 925 970, AFSL 301737) for the general i nformation of its “wholesale” customers (as defined in the
Corporations Act 2001). Where distributed to retail customers, this research is distributed by HSBC Bank Australia Limited (ABN 48 006 434 162, AFSL No. 232595). These respective entities
make no representations that the products or services mentioned in this document are available to persons in Australia or are necessarily suitable for any particular person or appropriate in
accordance with local law. No consideration has been given to the particular investment objectives, financial situation or particular needs of any recipient.
The document is distributed in Hong Kong and Japan by The Hongkong and Shanghai Banking Corporation Limited and has been prepared for the New York office of HSBC Bank USA, National
Association. In Korea, this publication is distributed by either The Hongkong and Shanghai Banking Corporation Limited, Seoul Securities Branch ("HBAP SLS") or The Hongkong and Shanghai
Banking Corporation Limited, Seoul Branch ("HBAP SEL") for the general information of professional investors specified in Article 9 of the Financial Investment Services and Capital Markets Act
(“FSCMA”). This publication is not a prospectus as defined in the FSCMA. It may not be further distributed in whole or in part for any purpose. Both HBAP SLS and HBAP SEL are regulated by
the Financial Services Commission and the Financial Supervisory Service of Korea.
Each of the companies listed above (the “Participating Companies”) is a member of the HSBC Group of Companies, any member of which may trade for its own account as Principal, may have
underwritten an issue within the last 36 months or, together with its Directors, officers and employees, may have a long or short position in securities or instruments or in any related instrument
mentioned in the document. Brokerage or fees may be earned by the Participating Companies or persons associated with them in respect of any business transacted by them in all or any of the
securities or instruments referred to in this document. This publication is distributed in New Zealand by The Hongkong and Shanghai Banking Corporation Limited, New Zealand Branch incorporated
in Hong Kong SAR.
The information in this document is derived from sources the Participating Companies believe to be reliable but which have not been independently verified. The Participating Companies make no
guarantee of its accuracy and completeness and are not responsible for errors of transmission of factual or analytical data, nor shall the Participating Companies be liable for damages arising out
of any person’s reliance upon this information. All charts and graphs are from publicly available sources or proprietary data. The opinions in this document constitute the present judgement of the
Participating Companies, which is subject to change without notice. From time to time research analysts conduct site visits of covered issuers. HSBC policies prohibit research analysts from
accepting payment or reimbursement for travel expenses from the issuer for such visits.
This document is neither an offer to sell, purchase or subscribe for any investment nor a solicitation of such an offer. HSBC Securities (USA) Inc. accepts responsibility for the content of this
research report prepared by its non-US foreign affiliate. All US persons receiving and/or accessing this report and intending to effect transactions in any security discussed herein should do so
with HSBC Securities (USA) Inc. in the United States and not with its non-US foreign affiliate, the issuer of this report. In Singapore, this publication is distributed by The Hongkong and Shanghai
Banking Corporation Limited, Singapore Branch for the general information of institutional investors or other persons specified in Sections 274 and 304 of the Securities and Futures Act (Chapter
289) (“SFA”) and accredited investors and other persons in accordance with the conditions specified in Sections 275 and 305 of the SFA. Only Economics or Currencies reports are intended for
distribution to a person who is not an Accredited Investor, Expert Investor or Institutional Investor as defined in SFA. The Hongkong and Shanghai Banking Corporation Limited, Singapore Branch
accepts legal responsibility for the contents of reports pursuant to Regulation 32C(1)(d) of the Financial Advisers Regulations. This publication is not a prospectus as defined in the SFA. This
publication is not a prospectus as defined in the SFA. It may not be further distributed in whole or in part for any purpose. The Hongkong and Shanghai Banking Corporation Limited Singapore
Branch is regulated by the Monetary Authority of Singapore. Recipients in Singapore should contact a "Hongkong and Shanghai Banking Corporation Limited, Singapore Branch" representative
in respect of any matters arising from, or in connection with this report. Please refer to The Hongkong and Shanghai Banking Corporation Limited Singapore Branch’s website at
www.business.hsbc.com.sg for contact details. HSBC México, S.A., Institución de Banca Múltiple, Grupo Financiero HSBC is authorized and regulated by Secretaría de Hacienda y Crédito Público
and Comisión Nacional Bancaria y de Valores (CNBV).
The document is intended to be distributed in its entirety. Unless governing law permits otherwise, you must contact a HSBC Group member in your home jurisdiction if you wish to use HSBC
Group services in effecting a transaction in any investment mentioned in this document. HSBC Bank plc is registered in England No 14259, is authorised by the Prudential Regulation Authority
and regulated by the Financial Conduct Authority and the Prudential Regulation Authority and is a member of the London Stock Exchange. (070905)
In Canada, this document has been distributed by HSBC Securities (Canada) Inc. (member IIROC), and/or its affiliates. The information contained herein is under no circumstances to be construed
as investment advice in any province or territory of Canada and is not tailored to the needs of the recipient. No securities commission or similar regulatory authority in Canada has reviewed or in
any way passed judgment upon these materials, the information contained herein or the merits of the securities described herein, and any representation to the contrary is an offense.
If you are an HSBC Private Banking (“PB”) customer with approval for receipt of relevant research publications by an applicable HSBC legal entity, you are eligible to receive this publication. To
be eligible to receive such publications, you must have agreed to the applicable HSBC entity’s terms and conditions (“KRC Terms”) for access to the KRC, and the terms and conditions of any
other internet banking service offered by that HSBC entity through which you will access research publications using the KRC. Distribution of this publication is the sole responsibility of the HSBC
entity with whom you have agreed the KRC Terms.
If you do not meet the aforementioned eligibility requirements please disregard this publication and, if you are a customer of PB, please notify your Relationship Manager. Receipt of research
publications is strictly subject to the KRC Terms, which can be found at https://research.privatebank.hsbc.com/ – we draw your attention also to the provisions contained in the Important Notes
section therein.
© Copyright 2019, HSBC Bank plc, ALL RIGHTS RESERVED. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, on any form or by any means, electronic,
mechanical, photocopying, recording, or otherwise, without the prior written permission of HSBC Bank plc. MCI (P) 065/01/2019, MCI (P) 008/02/2019
[1119770]
87

Data Matters - HSBC PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Data Matters - HSBC PDF

Hochgeladen von

Copyright:

Verfügbare Formate

2 May 2019

Disclosures & Disclaimer Issuer of report: HSBC Bank plc

Taking a dip in the data

Goldilocks waning; correlations rising

1. Monthly asset-class total returns

Source: MSCI, Refinitiv Datastream, Bloomberg, HSBC

2. Strength of correlations across a wide range of asset classes

Source: Bloomberg, Refinitiv Datastream, HSBC

What does this mean for diversification?

3. Correlations across risk-on and risk-off asset classes

Risk-on correlation (rolling 13-week) Risk-off correlation (rolling 13-week)

Source: Bloomberg, HSBC Source: Bloomberg, HSBC

4. Bond-equity correlation (US)

Source: MSCI, Refinitiv Datastream, HSBC

5. Bond-equity correlations (Germany)

Source: MSCI, Refinitiv Datastream, HSBC

Drilling down into cross-asset correlations

Source: Bloomberg, Refinitiv Datastream, HSBC

7. IG credit spreads much more closely correlated recently

Synchronised again – this time on the downside

8. 10y rates correlations during 2018… 9. … and 2019 year-to-date

Source: HSBC, Bloomberg Source: HSBC, Bloomberg

10. 2y rates correlations during 2018… 11. … and 2019 year-to-date

Source: HSBC, Bloomberg Source: HSBC, Bloomberg

Much ado about low vol

12. The volatility wave has ebbed

Source: Bloomberg, Refinitiv Datastream, HSBC

13. CRAVIN on the back foot again

1.2 95th percentile

For more detail please see

15. 10Y Bund futures distribution of daily high-low

Source: Bloomberg, HSBC

The Jekyll and Hyde behaviour of financial markets

Previous Data Matters

Data Matters: RORO Resurgence, 17 October 2018

Data Matters: EM-DM Vol Divergence, 8 August 2018

Data Matters: EM underperforms, US dominates, 23 May 2018

Data Matters: Riddles, mysteries and enigmas, 12 April 2018

HSBC Multi-Asset Heat Map (interactive webpage)

HSBC Surprise Indices, 1 May 2019

HSBC FX Positioning Indicators, 29 April 2019

FX Volametrics (interactive webpage)

Machine Learning 102: Tree hugging, 1 May 2019

The Great Gapsby, 1 May 2019

Machine Learning 101: Everybody needs good neighbours, 11 March 2019

Alternative Data – Text analysis of 20,000 earnings calls, 13 September 2018

A question of priorities – Country or sector: what matters more?, 14 February 2018

Comfortably numb, 7 August 2017

FX and the yield curve, 24 July 2017

The seasonality illusion, 3 May 2017

Monthly returns (by performance) p. 14

1. Monthly asset-class total returns

2. Monthly asset-class total returns

3. Annual asset-class total returns over the past 25 years

4. Annual asset-class total returns over the past 25 years

Cross asset volatility indicator (CRAVIN) p. 19

18 Click for: Contents | Asset returns | Risk indicators | Correlations

Cross asset volatility indicator (CRAVIN)

5. Composite index of 3-month realised volatility across asset classes

Source: Bloomberg, Refinitiv Datastream, HSBC

Table 1. Assets included in the Cross Asset Volatility Indicator

Click for: Contents | Asset returns | Risk indicators | Correlations 19

HSBC risk monitor