Deutsche Bank screenshow template - QWAFAFEW New...
Transcript of Deutsche Bank screenshow template - QWAFAFEW New...
Deutsche Bank
QWAFAFEW Presentation
January 2015
Deutsche Bank does and seeks to do business with companies covered in its research reports. Thus, investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider this report as only a single factor in making their investment decision. DISCLOSURES AND ANALYST CERTIFICATIONS ARE LOCATED IN APPENDIX 1. MICA(P) 072/04/2012
Global Quantitative Strategy
Yin Luo, CFA ▪ 212 250 8983 ▪ [email protected]
Managing Director, Global Head of Quantitative Strategy
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
#1 Ranked Global Quant Strategy Team
Source: gettyimages.com, Deutsche Bank Quantitative Strategy
research surveys: America #1; Europe #1; Asia #1 FX quant research #2
All our research can be accessed at: http://eqindex.db.com/gqs
1
New York
— Miguel Alvarez
— Javed Jussa
— Sheng Wang
— Allen Wang
— Gaurav Rohal, CFA
— David Elledge
— Zheyin Zhao
Quant IT
— Sergei Khomiouk
Chile Offshore Support
— Dagoberto Mendez
— Nicolas Magunacelaya
London
— Spyros Mesomeris, PhD
European Head of Quantitative Strategy
— Christian Davies
— Jacopo Capra
— Shan Jiang
— Alison (Shuo) Qu, PhD
— Paul Ward
Quant FX/Commodities
— Caio Natividade
— Vivek Anand
Hong Kong
— Khoi LeBinh
Asian Head of Quantitative Strategy
— Vincent Zoonekynd
— Ada Lau
Mumbai
— Hemant Sambatur
— Yin Luo, CFA
Global Head of Quantitative Strategy
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected] 2
Introduction to quantitative equity investing Quants look at factors
— A factor is simply a systematic way of ranking (and selecting) stocks. It could be as simple as value (e.g.,
P/E) or momentum (e.g., past 12-month returns).
Source: Bloomberg Finance LP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
How do we know it’s a good factor?
And you don’t need to trade it every day The turnover is not too bad
Cheap stocks are (almost) always good
0.6
0.8
1.0
1.2
1.4
1.6
1 2 3 4 5 6 7 8 9 10
Earnings yield, forecast FY1 mean, Quantile average return (%)
(%)
Quantile
20
30
40
50
60
88 90 92 94 96 98 00 02 04 06 08 10 12
Factor turnover, tw o-w ay, (%)
12-month moving average
Earnings yield, forecast FY1 mean
-0.4
0.0
0.4
0.8
1.2
1 2 3 4 5 6 7 8 9 10 11 12
Earnings yield, forecast FY1 mean, Long/short quantile portfolio return decay
(%)
Period
-20
-10
0
10
20
88 90 92 94 96 98 00 02 04 06
Long/short quantile portfolio return (%), Ascending order
12-month moving average
Earnings yield, forecast FY1 mean
(%)
Avg = 1.11%
Std. Dev. = 7.12%
Min = -29.6%
Avg/Std. Dev.= 0.16
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
-15
-10
-5
0
5
10
01 02 03 04 05 06 07 08 09 10 11 12 13
Long/short quantile portfolio return (%), Ascending order
12-month moving average
DB composite options factor
(%)
Avg = 0.82%
Std. Dev. = 2.51%
Min = -10.89%
Avg/Std. Dev.= 0.33
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Jan-00 Jan-01 Jan-02 Jan-03 Jan-04 Jan-05 Jan-06 Jan-07 Jan-08 Jan-09 Jan-10 Jan-11 Jan-12
Rel
ativ
e O
pp
ort
unit
y
Stock-specific Country Style Industry Currency
3
But then the 2008 financial crisis changed everything (maybe forever)
Source: Bloomberg Finance LP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
The profit from a simple value strategy has fallen by 2/3
And learn to live in a macro dominated environment That’s why we need new factors
Momentum has been even more challenging
-8
-4
0
4
8
12
2008 2009 2010 2011 2012 2013
Long/short quantile portfolio return (%), Ascending order
12-month moving average
Earnings yield, forecast FY1 mean
(%)
Avg = 0.36%
Std. Dev. = 3.46%
Min = -6.94%
Avg/Std. Dev.= 0.1 -60
-40
-20
0
20
40
88 90 92 94 96 98 00 02 04 06 08 10 12
Long/short quantile portfolio return (%), Ascending order
12-month moving average
12M-1M total return
(%)
Avg = 1.23%
Std. Dev. = 7.3%
Min = -43.51%
Avg/Std. Dev.= 0.17
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
26/01/2015 08:18:08 2010 DB Blue template
DB Quant Handbook, Part II — The rapid rise of computing power and wide availability of
off-the-shelf backtesting software provided by many data
vendors have given the impression that quant investing is
easy, or is it?
— In this paper, we discuss the seven common mistakes
investors tend to make when they perform backtesting
and build quant models. Some of these may be familiar to
our readers, but nonetheless, you may be surprised to
see the impact of these biases. The other sins are so
commonplace in both academia and practitioner’s
research that we usually take them for granted.
— There are a few unique features in this research that we
have not seen in other places. We deliberate when to and
when not to remove outliers; discuss various data
normalization techniques; address the intricate issues of
signal decay, turnover, and transaction costs; elaborate
on the optimal rebalancing frequency; illustrate the
asymmetric factor payoff patterns and the impact of short
availability on portfolio performance; answer the question
of “how many stocks should be held in the portfolio”; and
review the tradeoffs of various factor weighting/portfolio
construction techniques. Last but not least, we compare
traditional active portfolio management via multi-factor
models, with the new trend of smart beta/factor portfolio
investing.
Seven sins of quantitative investing
4
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
I. Survivorship bias
Ignoring inactive companies — Survivorship bias is one of the common mistakes
investors tend to make. Most people are aware of the
survivorship bias, but few understand its significance.
— Practitioners tend to backtest certain investment
strategies using only those companies that are currently
in business, meaning stocks that have left the
investment universe due to bankruptcy, delisting or
being acquired are not included in the backtesting.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
5
Survivorship bias
Stocks that have survived perform better than average # of stocks in the US and Europe that have survived until today
0
10
20
30
40
50
60
Russell 3000 index (equally weighted)
Survivor universe (equally weighted)
0
100
200
300
400
500
600
MSCI Europe survivor
0
2
4
6
8
10
12
14
MSCI Europe equally weighted
MSCI Europe survivor universe
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Survivorship bias illustrated
Survivorship bias leads to completely opposite conclusions
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
6
Merton distance of default factor on the Russell 3000
universe Factor exposure, the Russell 3000 universe Low volatility factor on the proper S&P 500 universe
Merton distance of default factor on the “survivor
universe” Factor exposure, the “survivor universe”
Low volatility factor performance on the current S&P 500
index constituents
0
5
10
15
20
25
Q1 (worst quality/highest credit risk)
Q5 (best quality/lowest credit risk)
0
20
40
60
80
100
120
140
160
180
200
Q1 (worst quality/highest credit risk)
Q5 (best quality/lowest credit risk)
0
1
2
3
4
5
6
7
8
9Quintile 1 (Low Volatility)
Quintile 5 (High Volatility)
0
20
40
60
80
100
120
140
160
180
Quintile 1 (Low Volatility)
Quintile 5 (High Volatility)
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5Q1 (worst quality/highest credit risk)
Q5 (best quality/lowest credit risk)
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5Q1 (worst quality/highest credit risk)
Q5 (best quality/lowest credit risk)
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
The impact of survivorship bias
1/3 of factors have the opposite signs with the survivorship-biased universe
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
7
Top 20 factors with the largest return differential Top 20 factors with the opposite signs
0.0% 0.5% 1.0% 1.5%
Realized vol, 1Y daily
Float turnover, 1M
Expected dividend yield
Dividend yield, trailing 12M
Normalized abnormal volume
IBES LTG EPS mean
Short interest/float
Payout on trailing operating EPS
Current ratio
Price-to-52 week high
Merton's distance to default
Operating earnings yield, trailing …
Price to 52-week low
Operating cash flow yield (income …
Skewness, 1Y daily
EBITDA to EV
Earnings yield, FY0
YoY change in # of shares …
Cash flow return on equity
Earnings yield, forecast FY1 mean
"Survivor" universe vs. correct universe
-1.5% -1.0% -0.5% 0.0% 0.5% 1.0%
Dividend yield, trailing 12M
IBES LTG EPS mean
Short interest/float
Current ratio
Operating earnings yield, trailing …
Operating cash flow yield (income …
Skewness, 1Y daily
EBITDA to EV
Earnings yield, FY0
YoY change in # of shares …
Earnings yield, forecast FY1 mean
Long-term debt to equity
Return on Equity
IBES FY1 mean EPS growth
# of days to cover short
Altman's z-score
Return on invested capital (ROIC)
IBES 5Y EPS growth
Moving average crossover, 15W-…
IBES FY1 Mean EPS Revision, 3M
Survivor universe Correct universe
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
II. Look-ahead bias
Using data that were unknown — It is the bias created by using information or data that
were unknown or unavailable as of the time when the
backtesting was conducted. It is probably the most
common bias in the backtesting.
— An obvious example of look-ahead bias lies in
companies’ financial statement data.
— Ideally, we should use point-in-time data for all
backtesting purposes. When PIT data is not available,
we need to make reporting lag assumption.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
8
Look-ahead bias
# of days to file quarterly earnings – international companies # of days to file quarterly earnings – US companies
0
4,000
8,000
12,000
16,000
20,000
24,000
28,000
0 10 20 30 40 50 60 70 80 90 100
Fre
qu
en
cy
Mean = 30 days
Median = 28 days
0
500
1,000
1,500
2,000
2,500
3,000
0 10 20 30 40 50 60 70 80 90 100
Fre
qu
en
cy
Mean = 37 days
Median = 35 days
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
The importance of using PIT data When PIT data is not available, reporting lag assumption is critical
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
9
The performance of the earnings yield factor, using non-PIT data The performance of the earnings yield factor, using PIT data
The impact of reporting lag assumption, ROE in UK The coverage of UK stocks (S&P BMI index)
IC (
%)
1995 2000 2005 2010 2015
-20
0
20
40
60Spearman rank IC (%), Ascending order12-month moving average
Avg = 8.12%
Std. Dev. = 12.43%Min = -26.28%Max = 47.73%
Avg/Std. Dev. = 0.65
IC (
%)
1995 2000 2005 2010 2015
-20
0
20
40
60 Spearman rank IC (%), Ascending order12-month moving average
Avg = 5.11%
Std. Dev. = 12.82%Min = -30.78%Max = 45.78%
Avg/Std. Dev. = 0.4
0
100
200
300
400
500
600
700
800
PIT coverageTraditional database (non-PIT) coverageS&P BMI UK universe
0%
1%
2%
3%
4%
PIT Non-PIT (no reporting lag
Non-PIT (1M lag)
Non-PIT (2M lag)
Non-PIT (3M lag)
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Look-ahead corporate action bias
Look-ahead bias can be tricky — For example, split adjustment factors can potentially
bring look-ahead bias. From time to time, companies
may decide to split their shares (or reverse split), to
improve liquidity or attract certain clientele. For most
modeling purposes, we want everything to be split
adjusted. For example, when we calculate earnings
yield, EPS data typically comes from company financial
statements with low frequency (quarterly, semi-annually,
or annually), while pricing information is from market
data available at least daily. We need to make sure both
EPS and price are split adjusted at the same time.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
10
Performance of the top 25 names with the lowest share price
Sharpe ratio Annualized return
0.1
1
10
100
1000
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014
The portfolio of 25 “low priced” stocks, based on split adjusted price
The portfolio of 25 “low priced” stocks, based on unadjusted price
S&P 500
0%
5%
10%
15%
20%
25%
30%
Adjusted price Unadjusted price S&P 500
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Adjusted price Unadjusted price S&P 500
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
III. The sin of storytelling
How long is long enough?
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
11
Storytelling Earnings yield, 1987-1997, Russell 3000 Earnings yield, 1997-2000, US technology
Earnings yield, 2000-2002, US technology
Earnings yield in US technology sector has never been a
good factor Sharpe ratio
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0
1
2
3
4
5
6
7
0
0.5
1
1.5
2
2.5
-3.00
-2.00
-1.00
0.00
1.00
2.00
3.00
Entire period Before tech bubble burst
After tech bubble burst
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
IV. Data mining and data snooping
Data mining is almost avoidable — Universe: S&P 500
— Two factor weighting algorithms
— In-sample model: select the best factor from each of the
six style buckets (value, growth, momentum/reversal,
sentiment, quality, and exotic) from 2009-2014 and then
backtest the same model over the same period.
— Out-of-sample model: from May 31, 2009, at the end of
each month, we use rolling 60 months of data to
construct our multi-factor model, using data available as
of that time.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
12
Data mining
Factor weighting – Grinold and Kahn MVO algorithm Factor weighting – equally weighting algorithm
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
V. Signal decay, turnover, and transaction cost
Need to balance among signal decay,
transaction cost, and model prediction power
— Different stock-selection factors have different
information decay profile. Faster decay signals require
higher turnover to capture their benefit. Higher turnover,
however, is likely to incur greater transaction costs.
— Adding a turnover constraint at the portfolio construction
process is an easy, but not necessarily ideal solution –
turnover constraint can either help or hurt our portfolio
performance.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
13
Signal decay
Annualized return with different transaction cost
assumptions, Japan dividend-paying stocks
-2%
0%
2%
4%
6%
8%
10%
12%
No cost 10 bps 20 bps 30 bps
One month reversal
Price to book
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00Turnover 40%
Turnover 80%
Turnover 120%
0.00
0.50
1.00
1.50
2.00
2.50
3.00
Turnover 40% Turnover 80% Turnover 120%
Wealth curve of the N-LASR model, with different
turnover constraints Sharpe ratio of the N-LASR model, with different
turnover constraints
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Optimal rebalancing frequency?
Tight turnover constraint ≠ Low rebalancing frequency — Having a tight turnover constraint, however, does not necessarily mean that we should have a very low rebalance frequency. In
many instances, we have heard comments such as “we are long-term value investors; we hold stocks for three to five years;
and therefore, we rebalance once a year”. New information comes in constantly and we should adjust our models and beliefs
accordingly. Even if we have a tight turnover constraint, we may still want to frequently adjust our positions – albeit modestly
each time.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
14
Annual versus monthly rebalance for a low turnover value portfolio (36% one-way turnover per year)
0
5
10
15
20
25
30
35
Monthly rebalance
Annual rebalance
Russell 3000
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Signal decay at the extreme
Example: one-day reversal factor
— A simple backtesting of the one-day reversal
factor (i.e., buying stocks that have fallen the
most on the previous day) seems to suggest
that short-term reversal to be a great strategy.
— The only problem is that the factor itself can only
be computed after the market closes; therefore,
the earliest time we can trade on the signal is at
the next day’s open.
— If we can calculate the one-day reversal factor
and trade on the same day’s closing price, we
can generate a Sharpe ratio of 1.4x – pretty
good for a single factor model. However, in
reality, we can only trade at the second day’s
open, while Sharpe ratio plummets to merely
0.3x (down almost 80%).
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
15
Performance of one day reversal
Annualized return and Sharpe ratio
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5Trading at the same day’s close
Trading at the next day’s open
21%
1.41
4% 0.26
Annualized return Sharpe ratio
Trading at the same day’s close Trading at the next day’s open
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
VI. Outliers – spectacular successes and failures
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
16
Outliers Aggregate earnings yield, using raw data
Aggregate earnings yield, using winsorized data Aggregate earnings yield for the Korean market Aggregate book-to-market for the Hong Kong market
Outlier control and data normalization
— Traditional outlier control techniques
include: winsorization (capping data
at certain percentiles) and truncation
(removing outliers from data sample).
— Data normalization process is closely
related to outlier control.
— Outliers could contain useful
information, but most of the time, they
don’t.
— Data normalization techniques can
have significant impact on model
performance.
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
-50%
-40%
-30%
-20%
-10%
0%
10%
20%
Raw data
Winsorizing 1%
Winsorizing 2%
0.00
0.50
1.00
1.50
2.00
2.50
3.00
Raw data Winsorizing Inter-quartile range
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Data transformation
Four alternative data transformation techniques
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
17
The distribution of Indonesia earnings yield – raw data The distribution of Indonesia earnings yield – z-score transformation
The distribution of Indonesia earnings yield – the ranking
transformation
The distribution of Indonesia earnings yield – our proprietary
transformation
De
nsity
-100 -50 0 50 100
0.0
00
.04
0.0
8
De
nsity
-10 -5 0 5 10
0.0
0.4
0.8
1.2
De
nsity
0.0 0.2 0.4 0.6 0.8 1.0
0.0
0.5
1.0
1.5
De
nsity
-2 -1 0 1 2
0.0
0.1
0.2
0.3
0.4
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
The impact of data normalization techniques
Our proprietary technique can improve model performance by 11% and reduce signal turnover significantly
— Example: an equally weighted four-factor model (earnings yield, 12-1M price momentum, three-month earnings revision, and ROE
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
18
Average model performance (rank IC), using different data
normalization techniques
0% 2% 4% 6% 8% 10%
DenmarkGreece
ThailandFinlandNorway
SpainNetherlands
SingaporeSouth Africa
MalaysiaSweden
BrazilSwitzerland
ItalyHong Kong
GermanyFranceChina
CanadaAustralia
KoreaTaiwan
UKJapan
USA
Normalize based on ranking Normalize based on z-score
Average signal serial correlation, using different normalization
techniques
75% 80% 85% 90% 95%
DenmarkGreece
ThailandFinlandNorway
SpainNetherlands
SingaporeSouth Africa
MalaysiaSweden
BrazilSwitzerland
ItalyHong Kong
GermanyFranceChina
CanadaAustralia
KoreaTaiwan
UKJapan
USA
Normalize based on ranking Normalize based on z-score
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Do outliers contain any useful information?
Maybe… at least for the price momentum factor
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
19
Neutralized momentum factor, without the ranking normalization Neutralized momentum factor, with the ranking normalization
Momentum portfolio performance Aggregate net exposure
-50%
-40%
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%Rank IC
12 month average
Avg = 2.8%
Min=-36%
Max=27%
Avg/ Std. Dev.= 0.31
-50%
-40%
-30%
-20%
-10%
0%
10%
20%
30%
40%
50%Rank IC
12 month average
Avg = 4.3%
Min=-41%
Max= 46%
Avg/ Std. Dev.= 0.29
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
Normalized momentum
Raw momentum
-100%
-80%
-60%
-40%
-20%
0%
20%
40%
60%
80%
100%Normalized momentum
Raw momentum
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Calculating excess returns in event study
Example: dividend announcement
— When we normalize our data, we have to
compute our factors relative to certain universes
or benchmarks. Interestingly, the results can
also be day-and-night depending on which
benchmark we use. We use an event study to
show the impact of benchmark selection bias.
— If we normalize each stock’s return by
subtracting the average return of all dividend-
paying stocks on the same day. On average,
there is no price movement prior to the event
date, i.e., there is probably no leakage of
dividend announcement information.
— However, if we choose the wrong benchmark –
where we use the broad equity market, e.g., the
S&P 500 index, we see stocks actually tend to
go up before the dividend announcement .
— The reason is possibly due to the fact that
dividend-paying stocks tend to earn higher
returns than the broad market. Using the wrong
benchmark makes it impossible to tell whether
the price drift before dividend announcement is
due to the dividend premium or dividend
announcement.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
20
Excess return over the equally weighted average of dividend-paying
stocks
Excess return over the S&P 500 index
-1.0%
-0.5%
0.0%
0.5%
1.0%
cum
ula
tive
exc
ess
re
turn
days
ex-date
announcement date
-1.0%
-0.5%
0.0%
0.5%
1.0%
cum
ula
tive
exc
ess
re
turn
days
ex-date
announcement date
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
VII. The asymmetric payoff pattern and shorting
Alpha from the long and the short — Long portfolio excess return: long the top quartile
stocks (equally weighted) against the average (or
median) return of our investment universe (which is
equivalent to shorting a basket of all stocks in our
universe, equally weighted)
— Short portfolio excess return: short the bottom
quartile stocks (equally weighted) against the average
(or median) return of our investment universe (which is
equivalent to using the proceeds from our short positions
to fund a long portfolio of all stocks in our universe,
equally weighted)
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
21
Asymmetric patterns
Price momentum Earnings yield
0
1
2
3
4
5
6
7
8Long portfolio excess return
Short portfolio excess return
0
1
2
3
4
5
6
7
8Long portfolio excess return
Short portfolio excess return
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Most factors’ payoff patterns are asymmetric
Not all factors are created
equally
— These factors are sorted based
on the spread between “short
excess return” and “long excess
return”.
— The higher up on the list, the
more difficult to capture the
alpha, due to heavier demand for
shorting and likely higher
shorting cost (shorting cost will
be discussed in the next section).
— Value factors generally collect
their premia from the long side,
while price momentum/reversal
and quality factors generate
more alpha from the short side.
Analyst revision factors tend to
show more symmetric payoff
patterns.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
22
The asymmetric payoff pattern
0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0%
EBITDA to EVEarnings yield x IBES 5Y growth
Operating earnings yield, trailing 12MEarnings yield, forecast FY1 meanEarnings yield, forecast FY2 mean
Free cash flow yieldTotal return, 1260D (60M)
Price to BookOperating cash flow yield
IBES FY1 Mean EPS Revision, 3MMean recommendation revision, 3M
Cash flow return on investmentsMoving average crossover, 15W-36W
Asset Turnover# of days to cover short
Sales to EVPrice to SalesGross margin
Short interest/floatYear-over-year quarterly EPS growth
Return on EquityCash flow return on capitalCash flow return on equity
Return on capitalReturn on Assets
YoY change in # of shares outstandingRealized vol, 1Y daily12M-1M total return
Short portfolio excess return Long portfolio excess return
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Accounting for short availability
Using DataExplorer’s global stock lending database
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
23
Coverage for the cost of borrow score (DCBS) from DataExplorer Cost-of-borrow score composition
% of hard-to-borrow names vs. short portfolio performance Performance with and without short constraints, N-LASR model
0
500
1000
1500
2000
2500
3000
3500
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
DCBS=10
DCBS=9
DCBS=8
DCBS=7
DCBS=6
DCBS=5
DCBS=4
DCBS=3
DCBS=2
DCBS=1
Asset Turnover
Sales to EV Cash flow return on equity
# of days to cover short
Short interest/floatPrice to Sales
Target price implied return
Payout on EPS
Abnormal volumeRealized vol, 1Y daily
MomentumPE
Correlation = 20.2%
0%
2%
4%
6%
8%
10%
12%
0% 10% 20% 30% 40% 50% 60%
Sho
rt p
ort
folio
exc
ess
ret
urn
Percentage of hard to borrow names
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Assuming we can short any stock
Assuming we can only short easy-to-borrow stocks
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
High conviction or diversification
How many stocks do we need to hold
— One popular view in the investment world,
especially a view shared by many fundamental
investors, is that we should fully take advantage
of our “high conviction” ideas; therefore, a more
concentrated portfolio is more desirable than a
portfolio holding hundreds of stocks. On the
other hand, some managers (more likely to be
quant) believe in diversification and typically
hold fairly diversified portfolios.
— Let’s use our N-LASR global stock selection
model (which has shown great live performance)
as an example.
— Without short constraint, as we hold more and
more diversified portfolios, alpha (i.e., active
return) goes down.
— With short constraint, as our portfolio becomes
more diversified, Sharpe ratio also goes up
significantly.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
24
Annualized return, long/short N-LASR portfolios
Sharpe Ratio, long/short N-LASR portfolios
0%
10%
20%
30%
40%
50%
25 names 100 names 400 names
Assuming we can short any stock
Assuming we can only short easy-to-borrow stocks
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
25 names 100 names 400 names
Assuming we can short any stock
Assuming we can only short easy-to-borrow stocks
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
A hands-on tutorial
How to build a realistic model
— Five common factors:
— Value (trailing 12 month earnings yield)
— Growth (year-over-year quarterly EPS growth)
— Quality (ROE)
— Momentum (12-1M total return), and
— Sentiment (IBES three-month earnings revision)
— How to avoid the seven sins
— Survivorship bias. We perform our backtesting on the Russell 3000 index universe, using those companies in the index
as of a given point in time.
— Look-ahead bias. We use point-in-time data to calculate all of our factors. Company fundamental data is sourced from
Compustat point-in-time database, which reflects whatever was available at each month end.
— Story telling and data history. We follow the convention for the direction of each factor: buying stocks that are cheaper,
that enjoy higher growth, that are more profitable, that have stronger price momentum, and that have more positive
analyst sentiment. Our backtesting is conducted over the past 20 years, from 1994 to 2014, covering multiple economic
cycles.
— Data mining and data snooping bias. The four factor weighting algorithms are extensively tested across multiple
countries/regions and asset classes.
— Signal decay and turnover. We avoid fast decay factors in this exercise. Portfolio performance is computed after
transaction costs.
— Outlier control. We use our proprietary data normalization technique to transform each factor to a standard normal
distribution, before we combine them together into multi-factor models.
— The asymmetric payoff pattern and shorting cost. We study the impact of short availability in detail in this section.
25
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Comparing factor weighting algorithms
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Deutsche Bank Quantitative Strategy
26
Average rank IC Risk adjusted rank IC
Grinold & Kahn factor weights Alpha risk parity factor weights Minimum tail dependence factor weights
— Equally weighting
— Grinold & Kahn (i.e.,
mean-variance
optimization)
— Alpha risk parity
— Minimum tail dependence
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Quality Growth Value Momentum Sentiment
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Quality Growth Value Momentum Sentiment
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Quality Growth Value Momentum Sentiment
… and the winner is – our proprietary minimum tail dependence model
0%
1%
2%
3%
4%
5%
6%
0.00
0.10
0.20
0.30
0.40
0.50
0.60
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Country and sector neutralization
Risk control at the model building stage
— Traditionally, analysts focus on model building,
while portfolio managers are responsible for risk
control at the portfolio level. In this section, we
show the benefit of adding some risk control at
the alpha model construction stage.
— Company characteristics (e.g., valuation, growth
profile, and profitability) vary greatly from
country to country, and from industry to industry.
A model that ranks stocks regardless of their
country/sector essentially engages in not only
stock selection, but also country/sector rotation.
— One way to make our stock selection model
more robust and less volatile is to control for
country/sector difference via a technique called
neutralization.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
27
Average rank IC
Risk adjusted rank IC
0%
1%
2%
3%
4%
5%
6%
Orignal Sector neutral
Equal weight GKW Alpha Risk Parity Min Tail Dependence
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
Orignal Sector neutral
Equal weight GKW Alpha Risk Parity Min Tail Dependence
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Smart beta investing via factor portfolios
Each factor portfolio is constructed on mean-variance optimization with realistic constraints
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
28
Annualized return Annualized volatility
Sharpe ratio Maximum drawdown
0%
1%
2%
3%
4%
5%
6%
7%
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
-30%
-25%
-20%
-15%
-10%
-5%
0%
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Active portfolio management versus smart beta investing
Pros and cons
— Active portfolio management via multi-factor
models tend to have higher realized risk than
smart beta portfolios. The second stage
optimization in multi factor-portfolios further
reduces risk.
— Active portfolio management via multi-factor
models tend to produce higher Sharpe ratios –
especially with more sophisticated portfolio
construction techniques like alpha risk parity and
minimum tail dependence, as these models are
more efficient than multi factor-portfolios.
— The biggest benefit of smart beta via multi
factor-portfolios is that it empowers asset
owners by providing additional investment
instruments to their asset allocation strategies.
— To add value, active managers need to have
more unique and proprietary factors in their
multi-factor models.
Source: Bloomberg Finance LLP, Compustat, IBES, Russell, S&P, Thomson Reuters, Worldscope, Deutsche Bank Quantitative Strategy
29
Realized portfolio risk
Sharpe ratio
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
Active portfolio management via “multi-factor models”
Smart beta investing via “multi factor-portfolios”
Equal weight GKW Alpha Risk Parity Min Tail Dependence
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Active portfolio management via “multi-factor models”
Smart beta investing via “multi factor-portfolios”
Equal weight GKW Alpha Risk Parity Min Tail Dependence
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
26/01/2015 08:18:11 2010 DB Blue template
Appendix 1 Important Disclosures Additional Information Available upon Request
DOUBLE CLICK IN
For disclosures pertaining to recommendations or estimates made on securities other than the primary subject of this research, please see the
most recently published company report or visit our global disclosure look-up page on our website at
http://gm.db.com/ger/disclosure/DisclosureDirectory.eqsr
30
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Special Disclosures N/A
Analyst Certification
The views expressed in this report accurately reflect the personal views of the undersigned lead analyst(s). In addition, the
undersigned lead analyst(s) has not and will not receive compensation for providing a specific recommendation or view in this report.
[Yin Luo, Miguel-A Alvarez, Javed Jussa, Sheng Wang, Allen Wang]
Hypothetical Disclaimer
Backtested, hypothetical or simulated performance results discussed herein have inherent limitations. Unlike an actual performance
record based on trading actual client portfolios, simulated results are achieved by means of the retroactive application of a backtested
model itself designed with the benefit of hindsight. Taking into account historical events the backtesting of performance also differs
from actual account performance because an actual investment strategy may be adjusted any time, for any reason, including a
response to material, economic or market factors. The backtested performance includes hypothetical results that do not reflect the
reinvestment of dividends and other earnings or the deduction of advisory fees, brokerage or other commissions, and any other
expenses that a client would have paid or actually paid. No representation is made that any trading strategy or account will or is likely
to achieve profits or losses similar to those shown. Alternative modeling techniques or assumptions might produce significantly different
results and prove to be more appropriate. Past hypothetical backtest results are neither an indicator nor guarantee of future returns.
Actual results will vary, perhaps materially, from the analysis.
31
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Regulatory Disclosures 1. Important Additional Conflict Disclosures Aside from within this report, important conflict disclosures can also be found at https://gm.db.com/equities under the “Disc losures Lookup” and “Legal” tabs. Investors are strongly encouraged to review this information before investing.
2. Short-Term Trade Ideas Deutsche Bank equity research analysts sometimes have shorter-term trade ideas (known as SOLAR ideas) that are consistent or inconsistent with Deutsche Bank’s existing longer term ratings. These trade ideas can be found at the SOLAR link at http://gm.db.com.
3. Country-Specific Disclosures Australia & New Zealand: This research, and any access to it, is intended only for "wholesale clients" within the meaning of the Australian Corporations Act and New Zealand Financial Advisors Act respectively. EU countries: Disclosures relating to our obligations under MiFiD can be found at http://www.globalmarkets.db.com/riskdisclosures. Japan: Disclosures under the Financial Instruments and Exchange Law: Company name - Deutsche Securities Inc. Registration number - Registered as a financial instruments dealer by the Head of the Kanto Local Finance Bureau (Kinsho) No. 117. Member of associations: JSDA, Type II Financial Instruments Firms Association, The Financial Futures Association of Japan, Japan Investment Advisers Association. Commissions and risks involved in stock transactions - for stock transactions, we charge stock commissions and consumption tax by multiplying the transaction amount by the commission rate agreed with each customer. Stock transactions can lead to losses as a result of share price fluctuations and other factors. Transactions in foreign stocks can lead to additional losses stemming from foreign exchange fluctuations. "Moody's", "Standard & Poor's", and "Fitch" mentioned in this report are not registered credit rating agencies in Japan unless “Japan” or "Nippon" is specifically designated in the name of the entity. Russia: This information, interpretation and opinions submitted herein are not in the context of, and do not constitute, any appraisal or evaluation activity requiring a license in the Russian Federation.
32
Deutsche Bank
Yin Luo, CFA ▪ 1.212.250.8983 ▪ [email protected]
Global Disclaimer The information and opinions in this report were prepared by Deutsche Bank AG or one of its affiliates (collectively "Deutsche Bank"). The information herein is believed to be reliable and has been obtained from public sources believed to be reliable. Deutsche Bank makes no representation as to the accuracy or completeness of such information.
Deutsche Bank may engage in securities transactions, on a proprietary basis or otherwise, in a manner inconsistent with the view taken in this research report. In addition, others within Deutsche Bank, including strategists and sales staff, may take a view that is inconsistent with that taken in this research report.
Opinions, estimates and projections in this report constitute the current judgement of the author as of the date of this report. They do not necessarily reflect the opinions of Deutsche Bank and are subject to change without notice. Deutsche Bank has no obligation to update, modify or amend this report or to otherwise notify a recipient thereof in the event that any opinion, forecast or estimate set forth herein, changes or subsequently becomes inaccurate. Prices and availability of financial instruments are subject to change without notice. This report is provided for informational purposes only. It is not an offer or a solicitation of an offer to buy or sell any financial instruments or to participate in any particular trading strategy. Target prices are inherently imprecise and a product of the analyst judgement.
As a result of Deutsche Bank’s March 2010 acquisition of BHF-Bank AG, a security may be covered by more than one analyst within the Deutsche Bank group. Each of these analysts may use differing methodologies to value the security; as a result, the recommendations may differ and the price targets and estimates of each may vary widely.
In August 2009, Deutsche Bank instituted a new policy whereby analysts may choose not to set or maintain a target price of certain issuers under coverage with a Hold rating. In particular, this will typically occur for "Hold" rated stocks having a market cap smaller than most other companies in its sector or region. We believe that such policy will allow us to make best use of our resources. Please visit our website at http://gm.db.com to determine the target price of any stock.
The financial instruments discussed in this report may not be suitable for all investors and investors must make their own informed investment decisions. Stock transactions can lead to losses as a result of price fluctuations and other factors. If a financial instrument is denominated in a currency other than an investor's currency, a change in exchange rates may adversely affect the investment.
All prices are those current at the end of the previous trading session unless otherwise indicated. Prices are sourced from local exchanges via Reuters, Bloomberg and other vendors. Data is sourced from Deutsche Bank and subject companies.
Past performance is not necessarily indicative of future results. Deutsche Bank may with respect to securities covered by this report, sell to or buy from customers on a principal basis, and consider this report in deciding to trade on a proprietary basis.
Derivative transactions involve numerous risks including, among others, market, counterparty default and illiquidity risk. The appropriateness or otherwise of these products for use by investors is dependent on the investors' own circumstances including their tax position, their regulatory environment and the nature of their other assets and liabilities and as such investors should take expert legal and financial advice before entering into any transaction similar to or inspired by the contents of this publication. Trading in options involves risk and is not suitable for all investors. Prior to buying or selling an option investors must review the "Characteristics and Risks of Standardized Options," at http://www.theocc.com/components/docs/riskstoc.pdf If you are unable to access the website please contact Deutsche Bank AG at +1 (212) 250-7994, for a copy of this important document.
The risk of loss in futures trading, foreign or domestic, can be substantial. As a result of the high degree of leverage obtainable in futures trading, losses may be incurred that are greater than the amount of funds initially deposited.
Unless governing law provides otherwise, all transactions should be executed through the Deutsche Bank entity in the investor's home jurisdiction. In the U.S. this report is approved and/or distributed by Deutsche Bank Securities Inc., a member of the NYSE, the NASD, NFA and SIPC. In Germany this report is approved and/or communicated by Deutsche Bank AG Frankfurt authorized by the BaFin. In the United Kingdom this report is approved and/or communicated by Deutsche Bank AG London, a member of the London Stock Exchange and regulated by the Financial Services Authority for the conduct of investment business in the UK and authorized by the BaFin. This report is distributed in Hong Kong by Deutsche Bank AG, Hong Kong Branch, in Korea by Deutsche Securities Korea Co. This report is distributed in Singapore by Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch, and recipients in Singapore of this report are to contact Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch in respect of any matters arising from, or in connection with, this report. Where this report is issued or promulgated in Singapore to a person who is not an accredited investor, expert investor or institutional investor (as defined in the applicable Singapore laws and regulations), Deutsche Bank AG, Singapore Branch or Deutsche Securities Asia Limited, Singapore Branch accepts legal responsibility to such person for the contents of this report. In Japan this report is approved and/or distributed by Deutsche Securities Inc. The information contained in this report does not constitute the provision of investment advice. In Australia, retail clients should obtain a copy of a Product Disclosure Statement (PDS) relating to any financial product referred to in this report and consider the PDS before making any decision about whether to acquire the product. Deutsche Bank AG Johannesburg is incorporated in the Federal Republic of Germany (Branch Register Number in South Africa: 1998/003298/10). Additional information relative to securities, other financial products or issuers discussed in this report is available upon request. This report may not be reproduced, distributed or published by any person for any purpose without Deutsche Bank's prior written consent. Please cite source when quoting.
Copyright © 2013 Deutsche Bank AG
33