Download PPT - OptiRisk Systems

Forum on News Analytics
applied to Trading, Fund Management
and Risk Control
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
GAUTAM MITRA
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
• Background and overview of news analytics in finance
• Sentiment classification
• News and abnormal returns
• News and volatility
• Industry insights, technology, products and services
• Directory of news analytics service providers
• Bibliography
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
DAN DI BARTOLOMEO
Incorporation of Quantified
News into Portfolio Risk
Assessment
Dan diBartolomeo
Northfield and Brunel University
London, 2009
Motivation for the Short Term
Risk Forecasts
• Risk models for asset management (as distinct from
trading operations) have traditionally focused on
estimating portfolio risk from security covariance over
time horizons of a year or more
– Suitable for long term investors such as pension funds
• Investment performance of asset managers is often
evaluated over shorter horizons so they are interested in
shorter term risk assessment. Hedge funds and other
portfolios with high portfolio turnover are even stronger
in this preference
• The proliferation of high frequency trading and
algorithmic execution methods have created demand for
very short horizon risk assessment
A Short Chronology
• A call from Blair Hull in 1996
• diBartolomeo and Warrick (2005) in Linear Factor Models
in Finance, edited by Satchell and Knight
• “Short Term Risk from Long Term Models”, Northfield
research series, Anish Shah, 2007-2009
• “Equity Portfolio Risk Using Market Information and
Sentiment” by diBartolomeo, Mitra and Mitra, 2009
Simple Approach to Short Term
Modeling
• The usual answer:
– Increase the frequency of observations (daily or shorter)
– Use a shorter sample period
– Generally need different factors
• There are serious problems with this approach at the
individual security level
–
–
–
–
High degree of kurtosis in return distributions (well maybe?)
Negative serial correlation due to short term reversal effects
Positive serial correlation on illiquid instruments
Asynchronous trading across time zones makes correlation
estimation very difficult
• Address “shocks” through a GARCH process
What’s the Problem with High
Frequency Data?
• Financial markets are driven by the arrival of information
in the form of “news” (truly unanticipated) and the form
of “announcements” that are anticipated with respect to
time but not with respect to content.
• The time intervals it takes markets to absorb and adjust
to new information ranges from minutes to days.
Generally much smaller than a month, but up to and
often larger than a day. That’s why US markets were
closed for a week at September 11th.
• GARCH models don’t work well on announcements
– Market participants anticipate announcements
– Volume and volatility dry up as investors wait for outcomes
– Reduce volatility into the announcement and boost it after the
announcement, so they are wrong twice
Some Surprising Things Appear
Anticipated
• Lets look at a precipitous decline in the implied volatility
of options on LUV
– All days in 2001 prior to September 7, average of .45 with a s.d.
of .13
– September 7, LUV implied = .22
– September 10, LUV implied = .15
– All days subsequent to September 17, average of .54 with a s.d.
of .18
– September 10 is in bottom 1% of the universe in implied
volatility, September 17 is 91st percentile
• Could this be driven by fundamentals?
Investor Response
to Information Flow
• Several papers have examined the relative market
response to “news” and “announcements”
– Ederington and Lee (1996)
– Kwag Shrieves and Wansley(2000)
– Abraham and Taylor (1993)
• Jones, Lamont and Lumsdaine (1998) show a
remarkable result for the US bond market
– Total returns for long bonds and Treasury bills are not different
if announcement days are removed from the data set
• Brown, Harlow and Tinic (1988) provide a framework for
asymmetrical response to “good” and “bad” news
– Good news increases projected cash flows, bad news decreases
– All new information is a “surprise”, decreasing investor
confidence and increasing discount rates
– Upward price movements are muted, while downward
movements are accentuated
Our Approach is Different
• Continue to use the existing risk models that are
estimated from low frequency return observations
• Use new information that is not part of the risk model to
adjust various components of the risk forecast to shortterm conditions
– Just ask yourself “How are conditions different now than they
were on average during the sample period used for estimation?”
• This approach has multiple benefits
– We sidestep almost all of the statistical complexities that arise
with use of high frequency data
– We get to keep the existing factor structure of the model so risk
reporting remains familiar and intuitive
– Since our long term and short term forecasts are based on the
same factor structure, we can quickly estimate new forecasts for
any length time horizon that falls between the two horizons
– Can be applied to any of our existing models
One Form of Working with “External
Information”
• Risk estimates in our short term model of US equities
have been conditioned for many years based on analysis
of stock option implied volatility
– Every day we look at the implied volatility of options on all US
stocks. We keep a 30 day moving average of the ratio of implied
volatility to historic volatility
– If the implied volatility/historic ratio jumps because of an
information flow to the market (e.g. Bill Gates gets run over by a
bus), the specific risk of that stock is adjusted
– If implied volatility ratio of many related stocks changes, the
implied changes in factor variance are also made. Risk forecasts
change even for stocks on which no options trade
– Requires non-linear optimization process for adjustments
– See Chapter 12, by diBartolomeo and Warrick Linear Factor
Models in Finance, Satchell and Knight, editors (2005)
“Variety” as External Information
• Solnik and Roulet (2000) examine the dispersion of
•
country returns as a way of estimating correlations
between markets
Lilo, Mantegna, Bouchard and Potters use the term
Variety to describe cross-sectional dispersion of stock
returns
– They also define the cross-sectional dispersion of CAPM alpha as
idiosyncratic variety (noted as v(t))
– They find that the average correlation between stocks is
approximately:
C(t) = 1 / [1 + (v2(t)/rm2(t) ]
• diBartolomeo (2000) relates periods of high cross-
sectional dispersion to positive serial correlation in stock
returns (i.e. momentum strategies working)
Other Conditioning Information
• Estimates of volatility based on high/low/open/close
information instead of the dispersion of returns
– Parkinson, Garman-Klass, Satchell-Wang, etc.
• Yield spreads for different classes of fixed income
securities provide an implied default rate and the
potential for large negative skew in stock returns
• Implied distribution of asset returns given the implied
vols of options on market indices across strike prices
• Direct measures of information flow to investors, and
investor attention that can create imbalances between
supply and demand for a given stock
What Makes People Buy or Sell a
Particular Stock?
• They WANT to trade the stock
– They believe the information that supports a valid forecast of
abnormal future return
• They HAVE to trade the stock
– They are trading to implement a change in asset allocation
– They are trading to implement a cash versus futures arbitrage
trade on a stock index
– They are a mutual fund or ETF sponsor responding to investor
cash flows in or out of the portfolio
– They are hedge fund that is forced to transact because of a
margin call
– They are forced to cover a short position by having the stock
called
The Potential for “Have To’s”
• We can fundamentally evaluate the potential for “have
to” trades
– Index arbitrage trades only occur with index constituents and
we know the open interest in futures
– Short interest information is published
– We know what big hedge and mutual funds have big positions in
particular stocks
– We have somewhat out of date information on full mutual fund
holdings and cash flow statistics
– We have fairly up to date information on ETF flows
The Potential for “Want To” Trades
• Investors are responding to information, so just measure
variations in the volume of information about a particular
stock over time
• Judge the magnitude of information flow of news text
coming over services such as Dow-Jones, Reuters and
Bloomberg
– Ravenpack and Thomson Reuters offer real time statistical
summaries of the amount and content of text news distributed
– Lexicons of over 2000 popular phrases are used to score the
content as “good news” or “bad news”
• Judge investor attention directly by measuring the
number of Google and Yahoo searches on ticker
symbols
Incorporating News Flows into Risk
Assessments
• diBartolomeo, Mitra and Mitra (2009) forthcoming in
Quantitative Finance
– Follow the diBartolomeo and Warrick mathematical framework
– Allow the conditioning information set to include both option
implied volatility and variations in text news flows from
Ravenpack (derived from Dow-Jones text feeds)
– Empirical tests on Euro Stoxx 50 during January 17-23, 2008 and
Dow Jones 30 stocks September 18 to 24, 2008
– Evaluate both individual stocks, full index and financial/nonfinancial subset portfolios
• In all cases, inclusion of quantified news flows improved
the rate of adjustment of risk estimates to time variation
in volatility faster than implied volatility alone
Incorporating Investor Attention
• Our next step will be to directly measure the degree of
investor attention to a stock
– Judge investor interest directly by measuring the number of
Google and Yahoo searches on trading symbols
– Avoid company names to eliminate product or service related
searches
– Try it yourself with Google Trends
• Da, Engleberg and Gao (2009) have already documented
•
a strong relationship between abnormal search
frequency and price momentum
Investor attention is not always a good thing
– Bolster and Trahan (2009) document predictable price behavior
in stocks mentioned on the Jim Cramer television show
– Clear strategy: wait two days, then short every stock mentioned
positively or negatively
Crucial Refinement
• diBartolomeo and Warrick (2005), and diBartolomeo,
Mitra and Mitra (2009) both assume that the full impact
of the conditioning information should applied to ex-ante
risk estimates
• Shah (2008) introduces formal Bayesian framework for
incorporating conditioning information into models
– Requirement for orthogonal factors is removed
– Non-linear optimization to “fit” the adjustments to correlated
factors is even more complex
– Introduced into Northfield “near-horizon” models in May 2009
– Reduces noise and allows for fitting to different time horizons
Other Differences Between
Long and Short Horizon Risk
• Negative serial correlation
– Daily overreactions & reversals, which cancel out over
time, become significant e.g. under leverage
• Contagion / panic
– Liquidity demands can drive up short-term correlations
• Transient behavior
– A long term model intentionally integrates new
phenomena slowly: Is the future like the past or are we
in and concerned about a present shift?
• Lots more extreme events in the short-term
– 3 std deviations contains less probability mass. 99%
VaR is farther away from the mean
Conclusions
• The key to good short term risk assessment is
•
•
•
understanding how conditions now are different than
they usually are
A broad set of information other than stock
characteristics and past returns are clearly useful in
improving risk estimates
Among the most useful sets of conditioning information
appears to be summaries of textual news flows to
investors
A rigorous Bayesian framework should be employed to
intelligently combine long term and short term
information sets
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
ARMANDO GONZALEZ
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
ARUN SONI
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
JAMES CHENERY
INCOPRORATING NEWS ANALYSIS INTO
TRADING AND INVESTMENT PROCESSES
FORUM ON NEWS ANALYTICS
November 9, 2009
James Chenery
Business Development Manager
EXPLOITING NEWS CONTENT
• News is emerging as differentiated, value generating content set
– Quant strategies – all trading frequencies
– Human decision support – especially with analytic enhancements
• Key uses
– Speed – beat the humans, beat the machines
– Manage scale and scope of events affecting portfolio
– Risk management and loss avoidance
• NewsScope product line - Robust set of capabilities
– Historical data to back test and build algorithms
– Real-time feeds for deployment, including ultra-low latency feed
– News Analytics which convert qualitative text into quantitative data
28
EXPLOITING NEWS CONTENT
• News flow is a good indicator of volume and volatility
• Pricing movements accompanied by news tend to be
momentum in nature; those with a lack of news tend to
reverse to average trends
• The market tends to overreact when there is a lot of news
on something and under-react when there is a small
quantity of news
• For other direction and magnitude signals, find
cause:effect relationships
29
MACHINE READABLE NEWS USE CASES
• Wolf detection / circuit breaker
• News flow algorithms
• Alpha generating signal
• Post trade analysis
• Stock screening tool
• Risk Management
• Compliance / Market abuse
• Fundamental research
• Trader decision support
30
NEWSSCOPE PRODUCT PORTFOLIO
• NewsScope Archive
– Historical database of Reuters and select third-party market moving
sources
• NewsScope Direct
– Ultra-low latency feed of highly structured news and economic data
• NewsScope Analytics (aka NewsScope Sentiment Engine)
– Automated news analysis solution measuring sentiment, relevance,
and novelty of text along with a host of other valuable metadata
• NewsScope Event Indices
– Automated news analysis solution indicating when abnormal
amounts of news occur across various categories
31
HOW TO MAKE MONEY WITH THE
NEWSSCOPE ANALYTICS
S&P1500 stocks in 2008; Daily items >50; Pos vs Neg >50%
Buy on good news
Outperform S&P500 by 5000
basis points over a 60 day
period!
Sell on bad news
32
Questions?
33
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
MARK VREIJLING
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
GANGADHAR DARBHA
Forum on News Analytics applied to Trading,
Fund Management and Risk Control
GURVINDER BRAR
Exploiting news-flow signals
Macquarie Quantitative Research
Gurvinder Brar, Christian Davies, Adam Strudwick, Andy Moniz
[email protected]
Macquarie Capital (Europe) Ltd
Level 2, Moor House, 120 London Wall, London EC2Y 5ET
November 2009
In preparing this research, we did not take into account the investment objectives, financial situation and particular needs of the reader. Before making an investment decision on the
basis of this research, the reader needs to consider, with or without the assistance of an adviser, whether the advice is appropriate in light of their particular investment needs, objectives
and financial circumstances. Please see disclaimer.
The global presence of Macquarie Quant
17 professional staff across the globe
Europe (4)
Gurvinder Brar
Christian Davies
Andy Moniz
Adam Strudwick
Asia (2)
Martin Emery
Viking Kwok
Japan (2)
Custom Products
Patrick Hansen
Ayumu Kuroda
Australia (8)
Research
George Platt
John Conomos
Portfolio
Products
Scott Hamilton
Burke Lau
South Africa (1)
Hannes Uys
Page 38
Quant
Applications
Connah Cutbush
Simon Rigney
George Ferizis
Charles Lowe
Recent publications

Page 39
Monthly ‘Quantamentals’ report

‘Global Dynamics’ report

Unwrapping value, Oct 08

Focusing on earnings revisions, Oct 08

Quality Control, Nov 08

Beyond Minimum Variance, Jan 09

When the tide turns, Dec 08


Style Outlook 2009, Jan 09
Asset Allocation: Spoilt for choice?, Apr
09

Exploiting dividend uncertainty, Feb 09

Do Technicals Add Value, Mar 09

Asymmetric Style, Apr 09


‘Risky Business’ report

Reality bites: Report on risk and
implementation, Nov 08
Have I got News for You? May 09

Stopping losses, taking profits, Feb 09

Positioning for Recovery, June 09

Portfolio Turnover: Friend or Foe, June 09

Preparing for regime changes, July 09

Spotting Growth, Sept 09

Arming Models with industry-specific
data, Oct 09
QUANTAMENTALS: HAVE I GOT NEWS FOR YOU? MAY ‘09
News-flow research in vogue

Academic literature








The impact of public information on the stock market, 1994
The market impact of corporate news stories, 2004
More than words: Quantifying language to measure firms’ fundamentals, 2007
Equity portfolio risk (volatility) estimation using market information and sentiment, 2008
Investor inattention and Friday earnings announcements, 2008
In search of attention, 2009
Impact of news sentiment on abnormal stock returns, 2009
Data challenges





Timeliness of news – Key newswires, stock exchange statements, press releases from company
websites, national newspapers
Relevance of news – Company names mentioned in headlines/1st paragraph of text
Classification of news – Accounting-related versus strategic news
Independence of news – Mixed versus standalone events
Informational content of news – Identifying good versus bad news (computational linguistics/market
based)
Page
40Macquarie Research, November 2009
Source:
QUANTAMENTALS: HAVE I GOT NEWS FOR YOU? MAY ‘09
Data vendors

Bloomberg - “Black Box Newsfeed” and “Black Box ECO Stats”. Low-latency delivery with 10,000+ daily
corporate and economic headlines and text, 1500 category codes, 18month history

CapitalIQ – Website based search engine with categorized headlines, though without full text of article.

Dow Jones Elementized News feed - Low latency, tagged data feed (from Jan 2004). News categorized
by corporate event

Dow Jones News and Archives - Text feed, 20+ years archive with identifiers, headlines & full stories.
News items are not tagged into categories

Ravenpack - Sentiment scoring for traditional news wires (DJ News), internet sources (CNN Money) and
blogs

Factiva.com – Website based search engine with categorized headlines and text
Source:
Page
41Macquarie Research, November 2009
QUANTAMENTALS: HAVE I GOT NEWS FOR YOU? MAY ‘09
Exploiting news flow strategies

How to define the event? Companies may announce several news items over a month, should we react to
all?

Informational content? Once we see an event, how do we systematically decide on its significance?

Holding period? Dealing with conflicting signals, excessive turnover and breadth of strategy
Information ratios (2001-2009)
2.0
Strategy Performances
600
1.9
1.8
1.8
500
1.6
1.5
1.4
400
1.2
300
0.8
0.7
200
0.4
100
0.0
Earnings
Momentum
(filtered for
news)
News
Momentum (all)
News
Momentum
(Accounting
Related)
Source:
Page
42Macquarie Research, November 2009
Combined
Earnings and
News
Momentum
News
Momentum
(Strategic)
Earnings
Momentum
Strategy
0
Jan-01
Jan-02
Jan-03
Jan-04
Eq-weighted universe
Revisions adjusted for good news
Revision downgrades
Jan-05
Jan-06
Jan-07
Jan-08
Revision upgrades
Revisions adjusted for bad news
Jan-09
Important disclosures:
Recommendation definitions
Volatility index definition*
Macquarie - Australia/New Zealand
This is calculated from the volatility of historic
price movements.
Outperform – return > 5% in excess of benchmark return
Neutral – return within 5% of benchmark return
Underperform – return > 5% below benchmark return
Very high–highest risk – Stock should be
expected to move up or down 60-100% in a year
– investors should be aware this stock is highly
speculative.
Macquarie – Asia/Europe
Outperform – expected return >+10%
Neutral – expected return from -10% to +10%
Underperform – expected <-10%
High – stock should be expected to move up or
down at least 40-60% in a year – investors should
be aware this stock could be speculative.
Macquarie First South - South Africa
Medium – stock should be expected to move up
or down at least 30-40% in a year.
Outperform – return > 10% in excess of benchmark return
Neutral – return within 10% of benchmark return
Underperform – return > 10% below benchmark return
Low–medium – stock should be expected to
move up or down at least 25-30% in a year.
Macquarie - Canada
Low – stock should be expected to move up or
down at least 15-25% in a year.
Outperform – return > 5% in excess of benchmark return
Neutral – return within 5% of benchmark return
Underperform – return > 5% below benchmark return
* Applicable to Australian/NZ stocks only
Macquarie - USA
Outperform – return > 5% in excess of benchmark return
Neutral – return within 5% of benchmark return
Underperform – return > 5% below benchmark return
Recommendation – 12 months
Note: Quant recommendations may differ from Fundamental Analyst
recommendations
Recommendation definitions – For quarter ending 30 September 2009
Outperform
Neutral
Underperform
Page 43
AU/NZ
45.08%
39.77%
15.15%
Asia
54.02%
19.10%
26.88%
RSA
40.00%
45.00%
15.00%
USA
42.31%
43.36%
14.34%
CA
62.86%
31.90%
5.24%
EUR
43.61%
39.85%
16.54%
Financial definitions
All "Adjusted" data items have had the following
adjustments made:
Added back: goodwill amortisation, provision for
catastrophe reserves, IFRS derivatives & hedging, IFRS
impairments & IFRS interest expense
Excluded: non recurring items, asset revals, property
revals, appraisal value uplift, preference dividends &
minority interests
EPS = adjusted net profit /efpowa*
ROA = adjusted ebit / average total assets
ROA Banks/Insurance = adjusted net profit /average
total assets
ROE = adjusted net profit / average shareholders funds
Gross cashflow = adjusted net profit + depreciation
*equivalent fully paid ordinary weighted average number
of shares
All Reported numbers for Australian/NZ listed stocks
are modelled under IFRS (International Financial
Reporting Standards).
Analyst Certification: The views expressed in this research accurately reflect the personal views of the analyst(s) about the subject securities or issuers and no part of the
compensation of the analyst(s) was, is, or will be directly or indirectly related to the inclusion of specific recommendations or views in this research. The analyst principally responsible
for the preparation of this research receives compensation based on overall revenues of Macquarie Group Ltd ABN 94 122 169 279 (AFSL No. 318062 )(MGL) and its related entities
(the Macquarie Group) and has taken reasonable care to achieve and maintain independence and objectivity in making any recommendations.
General Disclaimers: Macquarie Securities (Australia) Ltd; Macquarie Capital (Europe) Ltd; Macquarie Capital Markets Canada Ltd; Macquarie Capital Markets North America Ltd;
Macquarie Capital (USA) Inc; Macquarie Capital Securities Ltd; Macquarie Capital Securities (Singapore) Pte Ltd; Macquarie Securities (NZ) Ltd; and Macquarie First South Securities
(Pty) Limited are not authorized deposit-taking institutions for the purposes of the Banking Act 1959 (Commonwealth of Australia), and their obligations do not represent deposits or
other liabilities of Macquarie Bank Limited ABN 46 008 583 542 (MBL) or MGL. MBL does not guarantee or otherwise provide assurance in respect of the obligations of any of the
above mentioned entities. MGL provides a guarantee to the Monetary Authority of Singapore in respect of the obligations and liabilities of Macquarie Capital Securities (Singapore) Pte
Ltd for up to SGD 35 million. This research has been prepared for the general use of the wholesale clients of the Macquarie Group and must not be copied, either in whole or in part, or
distributed to any other person. If you are not the intended recipient you must not use or disclose the information in this research in any way. If you received it in error, please tell us
immediately by return e-mail and delete the document. We do not guarantee the integrity of any e-mails or attached files and are not responsible for any changes made to them by any
other person. MGL has established and implemented a conflicts policy at group level (which may be revised and updated from time to time) (the "Conflicts Policy") pursuant to regulatory
requirements (including the FSA Rules) which sets out how we must seek to identify and manage all material conflicts of interest. Nothing in this research shall be construed as a
solicitation to buy or sell any security or product, or to engage in or refrain from engaging in any transaction. In preparing this research, we did not take into account your investment
objectives, financial situation or particular needs. Before making an investment decision on the basis of this research, you need to consider, with or without the assistance of an adviser,
whether the advice is appropriate in light of your particular investment needs, objectives and financial circumstances. There are risks involved in securities trading. The price of securities
can and does fluctuate, and an individual security may even become valueless. International investors are reminded of the additional risks inherent in international investments, such as
currency fluctuations and international stock market or economic conditions, which may adversely affect the value of the investment. This research is based on information obtained
from sources believed to be reliable but we do not make any representation or warranty that it is accurate, complete or up to date. We accept no obligation to correct or update the
information or opinions in it. Opinions expressed are subject to change without notice. No member of the Macquarie Group accepts any liability whatsoever for any direct, indirect,
consequential or other loss arising from any use of this research and/or further communication in relation to this research. Clients should contact analysts at, and execute transactions
through, a Macquarie Group entity in their home jurisdiction unless governing law permits otherwise.
Country-Specific Disclaimers: Canada: In Canada, securities research is prepared, approved and distributed by Macquarie Capital Markets Canada Ltd, a participating organisation
of the Toronto Stock Exchange, TSX Venture Exchange & Montréal Exchange. Macquarie Capital Markets North America Ltd., which is a registered broker-dealer and member of
FINRA, accepts responsibility for the contents of reports issued by Macquarie Capital Markets Canada Ltd in the United States and to US persons and any person wishing to effect
transactions in the securities described in the reports issued by Macquarie Capital Markets Canada Ltd should do so with Macquarie Capital Markets North America Ltd. Securities
research is issued and distributed by Macquarie Securities (Australia) Ltd (AFSL No. 238947) in Australia, a participating organisation of the Australian Securities Exchange; Macquarie
Securities (NZ) Ltd in New Zealand, a licensed sharebroker and New Zealand Exchange Firm. The Research Distribution Policy of Macquarie Capital Markets Canada Ltd is to allow all
clients that are entitled to have equal access to our research. United Kingdom: In the United Kingdom, securities research is issued and distributed by Macquarie Capital (Europe) Ltd,
which is authorised and regulated by the Financial Services Authority (No. 193905). Hong Kong: In Hong Kong, securities research is issued and distributed by Macquarie Capital
Securities Ltd in Hong Kong, which is licensed and regulated by the Securities and Futures Commission. Japan: In Japan, securities research is issued and distributed by Macquarie
Capital Securities (Japan) Limited in Japan, a member of the Tokyo Stock Exchange, Inc., Osaka Securities Exchange Co. Ltd, and Jasdaq Securities Exchange, Inc. (Financial
Instruments Firm, Kanto Financial Bureau (kin-sho) No. 231, a member of Japan securities Dealers Association and Financial Futures Association of Japan). South Africa: In South
Africa, securities research is issued and distributed by Macquarie First South Securities (Pty) Limited in South Africa, a member of the JSE Limited. Singapore: In Singapore, securities
research is issued and distributed by Macquarie Capital Securities (Singapore) Pte Ltd (Company Registration Number: 198702912C), a Capital Markets Services license holder under
the Securities and Futures Act to deal in securities and provide custodial services in Singapore. Pursuant to the Financial Advisers (Amendment) Regulations 2005, Macquarie Capital
Securities (Singapore) Pte Ltd is exempt from complying with sections 25, 27 and 36 of the Financial Advisers Act. United States: In the United States, securities research is issued and
distributed by Macquarie Capital (USA) Inc., which is a registered broker-dealer and member of FINRA. Macquarie Capital (USA) Inc, accepts responsibility for the content of each
research report prepared by one of its non-US affiliates when the research report is distributed in the United States by Macquarie Capital (USA) Inc. Macquarie Capital (USA) Inc.’s
affiliate’s analysts are not registered as research analysts with FINRA, may not be associated persons of Macquarie Capital (USA) Inc., and therefore may not be subject to FINRA rule
restrictions on communications with a subject company, public appearances, and trading securities held by a research analyst account. Any persons receiving this report directly from
Macquarie Capital (USA) Inc. and wishing to effect a transaction in any security described herein should do so with Macquarie Capital (USA) Inc. Important disclosure information
regarding the subject companies covered in this report is available at www.macquarie.com/research/disclosures, or contact your registered representative at 1-888-MAC-STOCK, or
write to the Supervisory Analysts, Research Department, Macquarie Securities, 125 W.55th Street, New York, NY 10019. © Macquarie Group
Page 44