The Sound of Failure: Parsing Conference Call Language

Rochester Cahan
212 803-7973
Stock Selection: Research and Results January 2016
January 8, 2016
The Sound of Failure: Parsing Conference Call Language for Red Flags
Words or Deeds?



If there’s one part of our industry that would be nice to do away with, it must surely be the strange ritual of the
quarterly earnings call. In the decade and a half since Regulation FD came into force an inordinate number of hours
have been spent listening to the leaders of companies read meticulously vetted statements over the phone. And yet
the fear of missing out on that one elusive nugget of new information that might be buried deep in the Q&A session
keeps everyone coming back for more.
In our case, we’ve long believed that deeds speak louder than words. However, a growing body of academic research suggests there may yet be something to be gleaned from conference calls, beyond a jammed earnings season
calendar and a splitting headache. Advances in the field of natural language processing mean that vast tranches of
textual information can be systematically analyzed for common patterns by computer. The question is whether
there’s any useful information in those patterns that might foretell future stock returns.
We took a look using a new archive of conference call transcripts that spans over a million calls and contains a
whopping 20 gigabytes of text. A nifty feature of the database is that different parts of the call are tagged, so we can
study the language used in the management discussion section independently from that used by sell-side analysts in
the Q&A session, for example.
Watch out for Naughty Words


We’re particularly interested in whether there’s any difference in the language used by stocks that end up failing
compared to those that go on to win. Over our sample, which begins in 2010, words associated with future failure
had an energy-exposed aftertaste, even after we explicitly omitted the energy and commodities stocks themselves.
Words like “gas,” “projects,” “equipment” and “orders” all featured much more prominently in the calls of eventual
losers than they did in winners’ calls.
In contrast, future winners talked a lot about “innovation,” “data,” “consumers” and “premium brands.” Overall
it’s a similar picture to what we’ve observed using actual financial data: the winners in the post-Crisis era have been
capital-lite, innovation businesses that can tap the top of the income distribution and scale intellectual capital to
produce big free cash flows. Losers have been big capital spenders caught short by the commodities bust and a noinvestment recovery.
A Bunch of Empty Words?



Word counts are a good starting point but they ignore context. A more nuanced approach is to use a so-called topic
model to extract clusters of words that are common across conference calls. For example, the calls of future losers
often feature what we might call a Turnaround Story whereas the calls of future winners most often feature a
Margin Expansion Story.
In Appendix 1 on page 9 we used some of these ideas to screen for potential failure candidates. We started with the
worst decile of our failure model and then sorted the candidates by the linguistic features in their recent conference
calls. We want to be extra-cautious when a stock has all the financial hallmarks of failure plus has been using
language associated with past losers. Under Armour, Lululemon, Yahoo, LinkedIn, NXP Semiconductors, and
Anadarko Petroleum feature, among others.
On the flipside, Appendix 2 on pages 10 and 11 screens for potential winners based on our core stock selection
model and textual analysis. Included are Bank of America, Omnicom, Western Digital, and Lincoln National.
Sungsoo Yang (212) 803-7925 Nicole Price (212) 803-7935 Yi Liu (212) 803-7942 Yu Bai (212) 803-7919 Iwona Scanzillo (212) 803-7915
© 2016, Empirical Research Partners LLC, 565 Fifth Avenue, New York, NY 10017. All rights reserved. The information contained in this report
has been obtained from sources believed to be reliable, and its accuracy and completeness is not guaranteed. No representation or warranty, express or implied, is made as to the fairness, accuracy, completeness or correctness of the information and opinions contained herein. The views
and other information provided are subject to change without notice. This report is issued without regard to the specific investment objectives, financial situation or particular needs of any specific recipient and is not to be construed as a solicitation or an offer to buy or sell any securities or
related financial instruments. Past performance is not necessarily a guide to future results.
Stock Selection: Research and Results January 2016
Conclusions in Brief
 We analyzed the language used in conference calls…
 …Which have come to occupy an inordinate amount our
industry’s time:
Large-Capitalization Stocks
Number of Transcripts Analyzed
2010 Through 2015
20,000
Large-Capitalization Stocks
Transcripts Analyzed by Month
2010 Through 2015
4,000
18,000
3,500
16,000
3,000
14,000
2,500
12,000
10,000
2,000
8,000
1,500
6,000
1,000
4,000
500
2,000
0
0
Conference
Presentation
Sales &
Revenue
Call
Sep
Oct
Nov
Dec
Premium
Inflation
Consumers
Large-Capitalization Stocks (ex-Energy & Materials)
Ratio of Word Use in Future Winners' Q&A Sessions to that in Failures' Q&A
Sessions: Top Ten
2010 Through 2015
Phase
Devices
Enterprise
0.0
Power
0.0
Solutions
0.5
Services
0.5
Orders
1.0
Equipment
1.0
Backlog
Aug
Other
Japan
x
2.5
1.5
Projects
Jul
brands:
1.5
Gas
Jun
 …While future winners touted innovation, data, and premium
2.0
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Source: FactSet Research Systems, Empirical Research Partners Analysis.
 Future losers often talk a big turnaround story (Topic 5)…
Large-Capitalization Stocks (ex-Energy & Materials)
Five Most Common Topics in Future Losers' Q&A Sessions
2010 Through 2014
Topic 3
Good
Quarter
Year
Business
Expect
Margin
Improvement
Product
Price
Sales
May
Earnings Call
2.0
Topic 2
Know
Right
Now
Want
Little
People
Market
Good
Price
Time
Apr
Regulatory
Large-Capitalization Stocks (ex-Energy & Materials)
Ratio of Word Use in Future Losers' Q&A Sessions to that in Winners' Q&A
Sessions: Top Ten
2010 Through 2015
Topic 1
Rate
Long-Term
Expect
Capital
Cash
Cost
Next
Year
Impact
Guess
Mar
Commercial
 Words associated with future failure had an energy-exposed
aftertaste…
x
Feb
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Source: FactSet Research Systems, Empirical Research Partners Analysis.
2.5
Jan
Guidance
Call
Data
Special
Situation &
M&A
Innovation
Analyst,
Investor &
Shareholder
Meeting
Brands
Earnings
Call
Topic 4
Business
Customer
Growth
Product
Revenue
Service
Market
Grow
Margin
Long-Term
 …And try to spin the story with positive language:
%
2.4
Topic 5
Market
Opportunity
Long-Term
Invest
Continue
Work
Time
Focus
New
Plan
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Large-Capitalization Stocks (ex-Energy & Materials)
Net Conference Call Sentiment¹
2010 Through 2015
2.3
2.2
2.1
2.0
1.9
1.8
1.7
Failures
Winners
Q&A Session
Failures
Winners
Management Discussion
Source: FactSet Research Systems, Empirical Research Partners Analysis.
¹ Sentiment based on the Loughran-McDonald Dictionary, available at https://www3.nd.edu/~mcdonald/Word_Lists.html.
2
Stock Selection: Research and Results January 2016
The Sound of Failure: Parsing Conference Call Language for Red Flags
Words or Deeds?
As an industry we collectively spend an extraordinary amount of time listening to what the leaders of companies
have to say for themselves. It seems any time one calls a portfolio manager these days the harried response is always the same: “I’d love to, but let’s wait until after earnings season.” Since the introduction of Regulation FD a
decade and a half ago, the earnings call has become about as original as the Hollywood Summer blockbuster, sans
the explosions and 3D glasses.
First there’s the generic disclaimer about forward looking statements, then some pleasantries from the CEO followed by his or her carefully-scripted assessment of the company’s performance, and finally the Q&A session
where a few analysts ask mostly layup questions lest they get benched for the next coveted corporate access roadshow. Given that, we’ve long emphasized managements’ deeds over their words. Talk is cheap, even if it is being
meticulously vetted by an army of expensive lawyers.
However, a growing body of academic research has suggested there may yet be something to be gleaned from conference calls. In part that’s because technological advances in the field of natural language processing have allowed
a more systematic analysis of the language deployed in calls. For example, a recent paper showed that the aggregate sentiment of conference calls, based on their textual tone, can predict the market’s performance with high sentiment preceding negative returns.1
That study piqued our interest so we took a look at whether we could find any information in conference call transcripts that might help forecast individual stock returns, rather than the overall market. The raw ingredients for our
study come from a new database of conference call transcripts provided by FactSet, a data vendor. The archive begins in 2010 and contains a whopping 20 gigabytes of transcripts spanning more than a million calls. A nifty feature
of the database is that it’s in XML format, which allows parts of a call, like the Q&A session or management discussion for example, to the extracted and analyzed independently.
Exhibit 1 shows the composition of the conference call archive for our large-capitalization universe. The ubiquitous
quarterly earnings call is by far the most prevalent whereas standalone guidance calls are few and far between.
That’s been pretty consistent over the years, with about 80% of all transcripts covering earnings calls (see Exhibit 2).
Exhibit 1: Large-Capitalization Stocks
Number of Transcripts Analyzed
2010 Through 2015
Exhibit 2: Large-Capitalization Stocks
Breakdown of Transcripts Analyzed by Type
2010 Through 2015
100
20,000
%
90
18,000
80
16,000
70
14,000
60
12,000
50
10,000
40
8,000
30
6,000
20
4,000
10
2,000
0
0
Earnings
Call
Analyst,
Investor &
Shareholder
Meeting
Special
Situation &
M&A
Conference
Presentation
Sales &
Revenue
Call
2010
Guidance
Call
Source: FactSet Research Systems, Empirical Research Partners Analysis.
2011
2012
Earnings Call
2013
2014
2015
Other
Source: FactSet Research Systems, Empirical Research Partners Analysis.
When we said portfolio managers are always in the midst of earnings season it was only half tongue in cheek; apart
from the quiet period at quarter-ends there are plenty of calls to dial in to in every other month (see Exhibit 3).
While the data don’t capture how many people listen in to each call, we can get a lower bound by tracking how
many sell-side analysts ask questions. For example, the typical earnings call has eight analysts participating in the
1
Jiang, F., Lee, J., Martin, X., and Guofu Zhou, 2015. “Manager Sentiment and Stock Returns.” Working Paper.
3
Stock Selection: Research and Results January 2016
Q&A session whereas at a conference presentation, like those often hosted by sell-side investment banks, there’s less
scope for a drawn out Q&A exchange (see Exhibit 4).
Exhibit 3: Large-Capitalization Stocks
Transcripts Analyzed by Month
2010 Through 2015
Exhibit 4: Large-Capitalization Stocks
Average Number of Sell-Side Analysts Actively
Participating in Call1
2010 Through 2015
4,000
Analysts
9
3,500
8
3,000
7
6
2,500
5
2,000
4
1,500
3
1,000
2
500
1
0
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Earnings Call
Aug
Sep
Oct
Nov
Earnings
Call
Dec
Guidance
Call
Special
Situation &
M&A
Other
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Analyst,
Investor &
Shareholder
Meeting
Sales &
Revenue
Call
Conference
Presentation
Source: FactSet Research Systems, Empirical Research Partners Analysis.
1
Active participation is defined as asking a question during Q&A.
Watch out for Naughty Words
Is there a significant difference in the language used by firms that fail versus those that go on to win? To study this
we took stocks that performed in the top and bottom 10% in a given year and then analyzed the words used in all
their conference calls in the prior calendar year. Exhibit 5 shows the result for the management discussion section of
the calls. Each dot represents a word and dots above the diagonal line represent words that were used more frequently in future losers’ conference calls than in the calls of future winners.
We plotted the same chart for the Q&A session of the calls too (see Exhibit 6). In both cases it’s obvious that the collapse in commodity prices at the end of the sample period had a big influence on the results. Words associated with
the energy sector, like “production,” “gas,” “wells” and “per” (as in dollars per barrel), were far more common in
the transcripts of future failure stocks compared to winners. On the other hand, companies touting “data,” “products,” “share” (as in share buybacks), “growth” and shiny “new” things, ended up doing well.
Exhibit 5: Large-Capitalization Stocks
Words Used in Conference Call Management
Discussion Sessions by Future Failures and Winners1
2010 Through 2015
Exhibit 6: Large-Capitalization Stocks
Words Used in Conference Call Q&A Sessions
by Future Failures and Winners1
2010 Through 2015
20,000
20,000
18,000
18,000
"Per"
16,000
"Production"
"New"
14,000
12,000
"Revenue"
10,000
"Critical"
"Share"
"Gas"
8,000
6,000
"Products"
4,000
"Know"
Future Failures' Word Count
Future Failures' Word Count
16,000
14,000
"Costs"
12,000
"Growth"
10,000
"Production"
"Wells"
8,000
6,000
"Gas"
4,000
"Products"
2,000
2,000
"Data"
0
0
0
2,000
4,000
6,000
8,000
10,000 12,000 14,000 16,000
18,000 20,000
Future Winners' Word Count
0
5,000
10,000
15,000
20,000
Future Winners' Word Count
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Source: FactSet Research Systems, Empirical Research Partners Analysis.
1
Future failures and winners are stocks that delivered bottom- and top-decile
performance in the following calendar year. Word counts are based on all
conference call transcripts from the prior calendar year.
1
Future failures and winners are stocks that delivered bottom- and topdecile performance in the following calendar year. Word counts are
based on all conference call transcripts from the prior calendar year.
4
Stock Selection: Research and Results January 2016
In fact, all ten of the most common words that appeared in future losers’ Q&A but not at all in winners’ Q&A were
closely linked to energy (see Exhibit 7). Even for the words that did appear in the Q&A sessions of both future winners and losers, those most skewed towards failure were mainly energy related (see Exhibit 8).
Exhibit 7: Large-Capitalization Stocks
Top Ten Words Appearing in Future Losers' Q&A Sessions
Not Used in Winners' Q&A Sessions
2010 Through 2015
Exhibit 8: Large-Capitalization Stocks
Ratio of Word Use in Future Losers' Q&A Sessions
to that in Winners' Q&A Sessions: Top Ten
2010 Through 2015
Word
Count
8,000
5.0
x
4.5
7,000
4.0
6,000
3.5
5,000
3.0
4,000
2.5
2.0
3,000
1.5
2,000
1.0
1,000
0.5
0
Drilling
Rig
Coal
Acreage
Mine
Tons
Mexico
Basin
Eagle
Ford
Brazil
0.0
Wells
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Oil
Gas
Production Project Equipment
Field
Operations Contract
Capex¹
Source: FactSet Research Systems, Empirical Research Partners Analysis.
1
Also includes variations such as "capital spending," "capital
expenditure," "capital allocation," etc.
Given the decline in the oil price was an exogenous event it’s more instructive to strip energy and materials stocks
out of the analysis. Exhibits 9 and 10 show the same charts with those sectors excluded. It’s noticeable that the
word dots now cluster closer to the diagonal lines, indicating that there’s a greater degree of similarity between the
calls of future failures and winners once we control for these outlier sectors.
Exhibit 9: Large-Capitalization Stocks (ex-Energy & Materials)
Words Used in Conference Call Management
Discussion Sessions by Future Failures and Winners1
2010 Through 2015
Exhibit 10: Large-Capitalization Stocks (ex-Energy & Materials)
Words Used in Conference Call Q&A Sessions
by Future Failures and Winners1
2010 Through 2015
20,000
20,000
"Revenue"
18,000
18,000
16,000
14,000
"Sales"
Future Failures' Word Count
Future Failures' Word Count
16,000
"Customers"
12,000
10,000
"Share"
8,000
"Services"
6,000
4,000
14,000
12,000
"Customers"
10,000
8,000
6,000
4,000
"Guidance"
"Services"
"Software"
2,000
2,000
"Line"
"Revenue"
"Patients"
"Patients"
"Data"
0
0
0
5,000
10,000
15,000
20,000
Future Winners' Word Count
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
20,000
Future Winners' Word Count
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Future failures and winners are stocks that delivered bottom- and top-decile
performance in the following calendar year. Word counts are based on all
conference call transcripts from the prior calendar year.
1
Future failures and winners are stocks that delivered bottom- and topdecile performance in the following calendar year. Word counts are
based on all conference call transcripts from the prior calendar year.
1
Digging deeper, the top ten words that appeared more frequently in losers’ Q&A sessions compared to winners’
calls leaned towards the energy-exposed industrial/capital goods complex, even after stripping out energy and materials explicitly (see Exhibit 11). On the other hand, the top ten words for winners had more of a consumer/technology/health care flavor (see Exhibit 12).
5
Stock Selection: Research and Results January 2016
Words like “phase” (as in the phase of an FDA drug trial), “brands,” “innovation,” “data” and “premium” speak to
the success of the likes of the Five Horsemen (Facebook, Amazon, Google/Alphabet, Netflix, Salesforce) and the
health care industry. Overall it’s a very similar picture to what we’ve observed using actual financial data: in the
post-Crisis era the winners have been capital-lite, innovation businesses that can tap the top of the income distribution and scale intellectual capital to produce big free cash flows. Losers have been big capital spenders caught short
by the commodities bust and a no-investment economic recovery.
Exhibit 11: Large-Capitalization Stocks (ex-Energy & Materials)
Ratio of Word Use in Future Losers' Q&A Sessions
to that in Winners' Q&A Sessions: Top Ten
2010 Through 2015
x
Exhibit 12: Large-Capitalization Stocks (ex-Energy & Materials)
Ratio of Word Use in Future Winners' Q&A Sessions
to that in Failures' Q&A Sessions: Top Ten
2010 Through 2015
2.5
2.5
x
2.0
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0
Gas
Projects
Backlog Equipment Orders
Services Solutions
Power
Enterprise Devices
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Phase
Brands
Innovation
Data
Commercial Regulatory
Japan
Consumers Inflation
Premium
Source: FactSet Research Systems, Empirical Research Partners Analysis.
A Bunch of Empty Words?
This is all interesting, but words alone don’t really give us enough to work with. The English language is full of
fiendish inconsistencies and the meaning of standalone words can be completely altered by context. Rather than going out and shorting all stocks that happen to mention “projects” or “backlogs,” or “orders” in their conference calls
we need to capture more nuance.
One technique we can use is the LDA model, which for the nerds out there stands for Latent Dirichlet Allocation.
The idea is to extract common groups of words, or topics, from a collection of text documents. Sometimes these topics have an obvious real-world interpretation while at other times they’re gibberish. For example, Exhibit 13 shows
the five most common topics in the Q&A section of future losers’ conference calls.
Topic 5 is a good example: it appears to be related to a management team trying to “focus” on turning things
around by “continuing” to “invest” in “long-term opportunities” and coming up with “new plans.” We might call
this topic the Turnaround Story. Unfortunately, history says despite the best intentions more often than not such
turnaround attempts fail.
As a concrete example, take Avon’s 1Q 2014 earnings call. A nice feature of the LDA model is that it estimates the
probability of a given call containing a given topic. In Avon’s case, the Q&A section of that particular call had a
44% probability of containing the Turnaround Story topic, more than double the average probability of about 20%.2
A quick read of the transcript reveals why. Here are some snippets:
2
•
“The same is true as I look at some of the work that the teams are doing across the businesses as it relates to
getting cost out. That takes time, but as I look at the plans and where we are, we're making good progress.”
•
“And that’s about the fields getting active representatives and looking at how we make the earnings opportunity great for them through new products.”
•
“And as we look at some of the areas where we’ve stumbled and we’ve had challenges, and I'll point to
Mexico, what we’re seeing now is the team is back blocking and tackling and focusing on that.”
•
“As we look at the plans that we have in place relative to what the teams are working on to drive costs out
and get the return on investment piece, I feel good about that.”
We focus on the Q&A sessions because they are less scripted than the management discussion sections and yield richer variation in language.
6
Stock Selection: Research and Results January 2016
It’s the usual litany of promises: better execution, new products, focus, cost cutting. Unfortunately, it all came to
naught: in the following year, 2015, the stock was down more than 50% versus the market.
Exhibit 14 shows the five most common topics for the future winners. This time Topic 5 is quite different. Instead
of future promises it seems more grounded in what has already been delivered; it’s all about a “good quarter” or
“year” with “price” and “margin increases” that are “expected” to “continue” as the “business grows.” We might
call this topic the Margin Expansion Story. It turns out this topic is by far the most common in the conference calls
of future winners (see Exhibit 15). An analogous Margin Expansion Story also shows up in losers’ calls, see Topic 3
in Exhibit 13, but it’s far less prevalent than it is in the winners’ calls (see Exhibit 16). That suggests one way to
screen for future failures and winners could be to scan their conference call transcripts for exposure to topics like the
Turnaround Story or the Margin Expansion Story. We’ll put this idea into practice at the end of this report.
Exhibit 13: Large-Capitalization Stocks (ex-Energy & Materials)
Five Most Common Topics in Future Losers'
Q&A Sessions
2010 Through 2014
To pic 1
Rate
Long-Term
Expect
Capital
Cash
Cost
Next
Year
Impact
Guess
To pic 2
Know
Right
Now
Want
Little
People
Market
Good
Price
Time
To pic 3
Good
Quarter
Year
Business
Expect
Margin
Improvement
Product
Price
Sales
To pic 4
Business
Customer
Grow th
Product
Revenue
Service
Market
Grow
Margin
Long-Term
Exhibit 14: Large-Capitalization Stocks (ex-Energy & Materials)
Five Most Common Topics in Future Winners'
Q&A Sessions
2010 Through 2014
To pic 1
Rate
Long-Term
Capital
Good
Cash
Revenue
Cost
Right
Balance
Manage
To pic 5
Market
Opportunity
Long-Term
Invest
Continue
Work
Time
Focus
New
Plan
To pic 2
Want
Right
People
Now
Work
Start
Different
Talk
Actual
Number
To pic 3
Market
Business
Product
Customer
Grow th
Long-Term
Continue
Grow
Focus
Invest
To pic 4
Next
Long-Term
Expect
Program
Take
Time
Obvious
Market
Follow
Open
To pic 5
Good
Quarter
Year
Price
Margin
Increase
Expect
Continue
Grow th
Business
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Exhibit 15: Large-Capitalization Stocks (ex-Energy & Materials)
Share of Conference Calls Where Each Topic is
Most Important in Future Winners' Q&A Sessions
2010 Through 2014
Exhibit 16: Large-Capitalization Stocks (ex-Energy & Materials)
Share of Conference Calls Where Each Topic is
Most Important in Future Failures' Q&A Sessions
2010 Through 2014
%
50
%
50
45
45
40
40
35
35
30
30
25
25
20
20
15
15
10
10
5
5
0
0
Topic 1
Topic 2
Topic 3
Topic 4
Topic 5
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Topic 1
Topic 2
Topic 3
Topic 4
Topic 5
Source: FactSet Research Systems, Empirical Research Partners Analysis.
The reason topics are so important is because they come closer to the way humans comprehend language. As an illustration, we looked at how often certain topical phrases appear in the calls of future winners and losers (see Exhibit 17). It should come as no surprise that mentions of capital spending were about 15% less likely to appear in
winners’ calls than in those of losers. Similarly, references to share buybacks were 10% more common in the conferences calls of winners. However, margins, and specifically margin expansion, featured more prominently in the
calls of future losers. That’s somewhat counterintuitive given the importance investors have placed on incremental
margins in the post-Crisis era (see Exhibit 18).
The problem of course is that simply counting the number of mentions completely ignores the context: “we hope to
be able to grow our margins next year if market conditions improve” is very different from “we’ve delivered another quarter of solid margin expansion and expect that to continue next quarter.” Topic models like the LDA model
do a better job of setting words in their actual context.
7
Stock Selection: Research and Results January 2016
Exhibit 17: Large-Capitalization Stocks (ex-Energy & Materials)
Difference in Frequency of Mentions in the
Conference Calls of Future Winners Versus Future Failures1
2010 Through 2015
%
15
Exhibit 18: Large-Capitalization Stocks
Relative Returns to the Best and Worst Quintile
of Incremental Free Cash Flow Margin
Monthly Data Compounded to Annual Periods
Five Years Ending Early-January 2016
%
5
4
10
3
2
5
1
0
0
(1)
(5)
(2)
(10)
(3)
(4)
(15)
Capital Spending
Share Buybacks
Margins
Memo:
Margin
Expansion²
Source: FactSet Research Systems, Empirical Research Partners Analysis.
Past Year
Past Five Years
Best Quintile
Worst Quintile
Source: Empirical Research Partners Analysis.
Includes management discussion and Q&A session.
² Also includes variations such as "margin growth," "higher margins,"
"increased margins," etc.
1
Getting Sentimental
There is one area where word counts can be helpful though: in assessing sentiment. Finance academics have recently created a sentiment dictionary that classifies words as positive, negative, or neutral, based on how they’re used in
financial discussions rather than the English language at large.3 Exhibit 19 gives some examples of positive and
negative words, sorted by frequency of occurrence in typical financial documents likes 10-Ks.4 Using that dictionary it’s possible to compute a simple net sentiment metric for a given piece of text: just count the number of positive
words used minus the number of negative words and divide by the total word count.
We ran both the management discussion and Q&A section of each transcript through the academics’ dictionary (see
Exhibit 20). Two things stand out. First, the scripted management discussion is, on average, much more positive in
tone than the analyst Q&A session. No surprise there. Second, future losers actually tend to have slightly more positive sentiment than future winners. On face value that seems a little counterintuitive, but it’s actually very consistent with what we’ve found in our long history of failure modeling: excessively high expectations, whether
measured through valuation ratios or textual sentiment, tend be a red flag in the long-run. A few exceptional (or
lucky) companies make their near-impossible numbers, but most don’t and flame out instead. It’s also consistent
with the aforementioned academic paper, which found high aggregate sentiment at the market level foretold weaker
future returns.
Putting everything together, Appendix 1 on page 9 sorts the worst decile of our failure model based on information
we can glean from the text of their conference calls in the past year. Stocks at the top of the list warrant extra caution: they have all the financial hallmarks of failure plus the sentiment in their conference calls has been very positive (perhaps excessively so) and they’ve used language that’s historically been associated with disappointment.
On the flipside, Appendix 2 on pages 10 and 11 looks for potential winners by taking the best quintile of our core
stock selection model and sorting it by the opposite linguistic features, i.e., mellow sentiment and language that’s
been associated with future outperformance.
3
Loughran, T. and Bill McDonald, 2011. “When is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks.” Journal of Finance, Vol. 66, pp. 35-65.
4
The complete dictionary is available at https://www3.nd.edu/~mcdonald/Word_Lists.html.
8
Stock Selection: Research and Results January 2016
Exhibit 19: The Loughran-McDonald Dictionary
Examples of Positive and Negative Words
Po sit ive Wo rds
Effective
Benefit
Gain
Greater
Good
Best
Improvement
Beneficial
Advances
Successful
Negat ive Wo rds
Losses
Against
Claims
Termination
Impairment
Adverse
Failure
Default
Disclosed
Closing
Exhibit 20: Large-Capitalization Stocks (ex-Energy & Materials)
Net Conference Call Sentiment1
2010 Through 2015
%
2.4
2.3
2.2
2.1
2.0
1.9
1.8
1.7
Failures
Winners
Failures
Q&A Session
Source: Loughran, T. and Bill McDonald, 2011. “When is a
Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks.”
Journal of Finance, Vol. 66, pp. 35-65.
Winners
Management Discussion
Source: FactSet Research Systems, Empirical Research Partners Analysis.
1
Sentiment based on the Loughran-McDonald Dictionary, available at
https://www3.nd.edu/~mcdonald/Word_Lists.html.
Appendix 1: Large-Capitalization Stocks
Worst Decile of the Failure Model
Sorted by Average of Conference Call Metrics and Market Capitalization
As of Early-January 2015
Symbol
UA
LULU
YHOO
LNKD
NXPI
APC
TWTR
CXO
HBI
ASH
TMO
MMC
BSX
CERN
DLTR
FNFV
LEN
MDVN
SPLK
GIL
IDXX
QGEN
RLGY
A
NCLH
KMX
XEC
ST
N
SLW
VRTX
BMRN
WDAY
WSH
GRA
CSGP
TOL
SYK
VRX
ILMN
GMCR
TYC
SBAC
HAR
PHM
SCTY
D
PXD
WYNN
FMC
ASML
COO
Company
UNDER ARMOUR INC
LULULEMON ATHLETICA INC
YAHOO INC
LINKEDIN CORP
NXP SEMICONDUCTORS NV
ANADARKO PETROLEUM CORP
TWITTER INC
CONCHO RESOURCES INC
HANESBRANDS INC
ASHLAND INC
THERMO FISHER SCIENTIFIC INC
MARSH & MCLENNAN COS
BOSTON SCIENTIFIC CORP
CERNER CORP
DOLLAR TREE INC
FIDELITY FINL FNFV GROUP
LENNAR CORP
MEDIVATION INC
SPLUNK INC
GILDAN ACTIVEWEAR INC
IDEXX LABS INC
QIAGEN NV
REALOGY HOLDINGS CORP
AGILENT TECHNOLOGIES INC
NORWEGIAN CRUISE LINE HLDGS
CARMAX INC
CIMAREX ENERGY CO
SENSATA TECHNOLOGIES HLDG NV
NETSUITE INC
SILVER WHEATON CORP
VERTEX PHARMACEUTICALS INC
BIOMARIN PHARMACEUTICAL INC
WORKDAY INC
WILLIS GROUP HOLDINGS PLC
GRACE (W R) & CO
COSTAR GROUP INC
TOLL BROTHERS INC
STRYKER CORP
VALEANT PHARMACEUTICALS INTL
ILLUMINA INC
KEURIG GREEN MOUNTAIN INC
TYCO INTERNATIONAL PLC
SBA COMMUNICATIONS CORP
HARMAN INTERNATIONAL INDUSTRIES INC
PULTEGROUP INC
SOLARCITY CORP
DOMINION RESOURCES INC
PIONEER NATURAL RESOURCES CO
WYNN RESORTS LTD
FMC CORP
ASML HOLDING NV
COOPER COMPANIES INC
Price
$79.66
55.86
31.40
225.55
84.44
49.13
22.56
90.29
29.28
102.16
138.75
54.32
17.99
58.16
78.81
10.41
46.71
47.18
57.65
28.09
71.12
26.71
36.36
40.69
57.99
52.37
90.00
45.15
82.38
12.48
122.89
104.92
77.76
47.18
97.25
197.54
32.73
90.02
98.50
181.28
90.08
31.79
105.76
93.11
17.16
52.79
67.47
124.35
68.77
37.98
87.30
131.86
Quintiles (1=Best; 5=Worst)
Conference Call Metrics
Failure
Q&A
Language
Average
Sentiment¹
Used In
Of The
(5=Highest)
Q&A²
Two
5
5
5.0
5
5
5.0
4
5
4.5
4
5
4.5
5
4
4.5
4
5
4.5
4
5
4.5
5
4
4.5
5
4
4.5
5
4
4.5
5
3
4.0
4
4
4.0
4
4
4.0
5
3
4.0
4
4
4.0
5
3
4.0
3
5
4.0
3
5
4.0
3
5
4.0
4
4
4.0
4
4
4.0
5
3
4.0
4
4
4.0
4
3
3.5
3
4
3.5
4
3
3.5
3
4
3.5
3
4
3.5
3
4
3.5
5
2
3.5
1
5
3.0
1
5
3.0
3
3
3.0
3
3
3.0
5
1
3.0
3
3
3.0
4
2
3.0
3
2
2.5
2
3
2.5
2
3
2.5
1
4
2.5
4
1
2.5
2
3
2.5
3
2
2.5
2
3
2.5
2
3
2.5
2
2
2.0
2
2
2.0
1
3
2.0
1
2
1.5
1
1
1.0
1
1
1.0
Deciles (1=Best; 10=Worst)
Select Failure Model Factors
Capital
NineArbitrage
Gross
Spending
Month
Risk
Cash Flow
-toPrice
(1=Low est, Dow nside
Yield
Depreciation Trend 10=Highest) Risk
10
10
3
8
4
10
9
9
10
10
10
4
9
8
6
10
7
5
8
10
8
7
7
10
4
10
3
10
9
10
10
6
10
9
9
4
10
8
10
6
10
4
7
9
4
10
3
8
1
4
9
5
5
3
9
na
na
5
1
2
9
2
4
7
1
9
7
9
6
7
10
4
7
8
8
na
na
8
7
2
10
8
6
5
9
9
2
10
10
9
10
9
6
9
9
9
10
4
7
6
10
7
7
8
9
8
6
4
7
1
na
na
8
7
8
9
8
6
3
6
6
5
4
8
6
10
9
9
6
2
4
7
7
9
10
7
9
9
5
2
10
1
7
7
4
5
10
9
10
9
10
1
4
9
9
10
10
9
9
7
10
3
6
6
9
na
na
6
4
4
10
6
6
3
1
10
9
4
5
7
10
1
7
6
9
10
5
5
3
3
6
10
10
10
10
10
9
6
9
9
8
8
10
10
10
9
3
9
1
7
8
2
8
6
7
7
3
10
4
4
10
6
8
7
7
10
10
9
10
10
4
10
7
2
3
6
9
8
9
9
6
10
10
10
10
10
5
9
8
8
9
8
7
2
6
7
7
10
8
7
Core
Model
Rank
10
10
10
10
10
10
10
10
8
9
5
10
10
10
9
10
10
6
10
10
8
6
10
5
8
9
10
10
10
10
7
10
9
10
6
9
10
8
9
7
9
9
8
10
6
10
10
9
10
10
9
10
Failure
Mo del
Rank
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
YTD
Return
(1.2) %
6.5
(5.6)
0.2
0.2
1.1
(2.5)
(2.8)
(0.5)
(0.5)
(2.2)
(2.0)
(2.4)
(3.3)
2.1
(7.3)
(4.5)
(2.4)
(2.0)
(1.2)
(2.5)
(3.4)
(0.8)
(2.7)
(1.0)
(3.0)
0.7
(2.0)
(2.6)
0.5
(2.3)
0.2
(2.4)
(2.9)
(2.3)
(4.4)
(1.7)
(3.1)
(3.1)
(5.6)
0.1
(0.3)
0.7
(1.2)
(3.7)
3.5
(0.3)
(0.8)
(0.6)
(2.9)
(1.7)
(1.7)
Market
Capitalization
($ Million)
$17,192
7,786
29,653
29,574
29,216
24,965
15,407
11,661
11,488
6,845
55,375
28,371
24,200
19,897
18,505
10,111
9,813
7,724
7,477
6,826
6,486
6,219
5,335
13,516
13,306
10,346
8,510
7,684
6,548
5,053
30,196
16,919
15,008
8,502
6,939
6,418
5,723
33,848
33,795
26,539
13,801
13,439
13,342
6,628
5,991
5,152
40,167
18,575
6,985
5,075
37,795
6,366
Source: Empirical Research Partners Analysis, FactSet Research Systems.
¹ Sentiment based on net number of positive words used in conference call Q&A sessions in 2015. Word sentiment based on Loughran-McDonald Dictionary,
available at https://www3.nd.edu/~mcdonald/Word_Lists.html.
² Failure language is based on Q&A exposure to Topic 5 (the Turnaround Story) from analysis of past failures.
9
Stock Selection: Research and Results January 2016
Appendix 2: Large-Capitalization Stocks
Best Quintile of the Core Model
Sorted by Average of Conference Call Metrics and Market Capitalization
As of Early-January 2015
Symbol
BAC
OMC
WDC
LNC
TSS
GT
MT
UGI
GS
GM
MPC
GLW
RCI
EFX
RF
CAM
CTAS
VRSN
WU
JKHY
TDC
JPM
C
IBM
AIG
PSX
DAL
AAL
EXC
DFS
FITB
ETR
XLNX
KLAC
XRX
KSS
GPN
INGR
NUAN
OC
ZION
CPN
JEC
HLF
URBN
HD
GILD
AMGN
BA
DOW
VLO
DISH
ABC
STJ
CTL
TSO
STX
HOLX
WAT
ALK
LEA
XRAY
UTHR
FTI
AR
MAN
STM
HP
G
RIG
NAVI
PBI
ESV
FOSL
Company
BANK OF AMERICA CORP
OMNICOM GROUP
WESTERN DIGITAL CORP
LINCOLN NATIONAL CORP
TOTAL SYSTEM SERVICES INC
GOODYEAR TIRE & RUBBER CO
ARCELORMITTAL SA
UGI CORP
GOLDMAN SACHS GROUP INC
GENERAL MOTORS CO
MARATHON PETROLEUM CORP
CORNING INC
ROGERS COMMUNICATIONS -CL B
EQUIFAX INC
REGIONS FINANCIAL CORP
CAMERON INTERNATIONAL CORP
CINTAS CORP
VERISIGN INC
WESTERN UNION CO
HENRY (JACK) & ASSOCIATES
TERADATA CORP
JPMORGAN CHASE & CO
CITIGROUP INC
INTERNATIONAL BUSINESS MACHINES CORP
AMERICAN INTERNATIONAL GROUP
PHILLIPS 66
DELTA AIR LINES INC
AMERICAN AIRLINES GROUP INC
EXELON CORP
DISCOVER FINANCIAL SVCS INC
FIFTH THIRD BANCORP
ENTERGY CORP
XILINX INC
KLA-TENCOR CORP
XEROX CORP
KOHL'S CORP
GLOBAL PAYMENTS INC
INGREDION INC
NUANCE COMMUNICATIONS INC
OWENS CORNING
ZIONS BANCORPORATION
CALPINE CORP
JACOBS ENGINEERING GROUP INC
HERBALIFE LTD
URBAN OUTFITTERS INC
HOME DEPOT INC
GILEAD SCIENCES INC
AMGEN INC
BOEING CO
DOW CHEMICAL
VALERO ENERGY CORP
DISH NETWORK CORP
AMERISOURCEBERGEN CORP
ST JUDE MEDICAL INC
CENTURYLINK INC
TESORO CORP
SEAGATE TECHNOLOGY PLC
HOLOGIC INC
WATERS CORP
ALASKA AIR GROUP INC
LEAR CORP
DENTSPLY INTERNATIONAL INC
UNITED THERAPEUTICS CORP
FMC TECHNOLOGIES INC
ANTERO RESOURCES CORP
MANPOWERGROUP
STMICROELECTRONICS NV
HELMERICH & PAYNE
GENPACT LTD
TRANSOCEAN LTD
NAVIENT CORP
PITNEY BOWES INC
ENSCO PLC
FOSSIL GROUP INC
Price
$16.43
73.57
60.40
49.10
48.24
32.01
4.04
33.85
177.14
33.31
51.24
17.91
34.31
109.45
9.44
62.93
88.29
83.95
17.62
76.53
26.25
63.62
51.13
135.95
60.43
80.08
48.66
40.91
27.97
52.71
19.55
68.58
45.80
68.69
10.30
49.55
62.58
92.80
19.66
46.78
26.71
14.45
41.92
54.93
22.75
131.07
98.01
158.34
140.50
49.93
69.94
57.29
101.87
60.40
25.11
107.10
36.27
37.83
130.35
78.40
120.71
58.86
155.54
29.22
22.68
81.82
6.59
54.18
24.20
12.55
11.50
20.38
15.89
34.29
Conference Call Metrics
Winner
Q&A
Language
Average
Sentiment¹
Used In
Of The
(5=Highest)
Q&A²
Two
1
2
1.5
2
1
1.5
2
1
1.5
1
2
1.5
1
2
1.5
2
1
1.5
1
2
1.5
1
2
1.5
1
3
2.0
2
2
2.0
3
1
2.0
3
1
2.0
1
3
2.0
3
1
2.0
2
2
2.0
2
2
2.0
3
1
2.0
2
2
2.0
3
1
2.0
1
3
2.0
2
2
2.0
1
4
2.5
1
4
2.5
2
3
2.5
1
4
2.5
2
3
2.5
2
3
2.5
1
4
2.5
1
4
2.5
2
3
2.5
4
1
2.5
1
4
2.5
1
4
2.5
2
3
2.5
3
2
2.5
4
1
2.5
4
1
2.5
3
2
2.5
2
3
2.5
4
1
2.5
1
4
2.5
2
3
2.5
2
3
2.5
2
3
2.5
4
1
2.5
5
1
3.0
2
4
3.0
1
5
3.0
2
4
3.0
4
2
3.0
4
2
3.0
1
5
3.0
5
1
3.0
2
4
3.0
3
3
3.0
3
3
3.0
4
2
3.0
5
1
3.0
5
1
3.0
3
3
3.0
5
1
3.0
5
1
3.0
1
5
3.0
4
2
3.0
3
3
3.0
5
1
3.0
2
4
3.0
2
4
3.0
2
4
3.0
1
5
3.0
1
5
3.0
4
2
3.0
1
5
3.0
4
2
3.0
Quintiles (1=Best; 5=Worst)
Super Factors
Earnings
Quality
Capital
and
Valuation Deployment
Trend
1
4
na
2
2
1
1
1
3
1
1
na
3
2
1
1
1
2
1
1
2
2
2
2
1
2
na
1
3
3
1
1
1
1
1
4
2
3
2
4
5
1
1
1
na
3
1
1
4
1
1
3
1
1
1
1
2
3
2
1
1
1
2
1
3
na
1
2
na
1
2
3
1
1
na
2
4
1
2
1
2
2
2
1
1
2
3
1
1
na
1
2
na
1
2
5
4
1
4
3
1
1
1
1
2
1
1
5
2
3
1
2
2
3
3
3
1
2
3
1
1
2
na
1
1
2
1
1
4
1
1
2
1
1
1
4
1
2
1
1
1
2
1
1
2
2
1
2
2
1
1
1
1
3
1
1
1
1
3
3
1
1
1
2
2
1
2
2
1
1
2
4
2
1
3
1
1
3
2
1
1
1
1
4
1
2
2
1
1
2
2
1
1
5
2
1
1
3
1
1
4
1
3
1
2
3
1
1
1
3
1
1
na
1
2
2
1
2
2
1
1
2
Market
Reaction
2
3
5
4
1
1
5
3
4
2
1
4
2
1
2
1
1
1
4
2
5
2
3
4
2
1
1
2
3
3
1
4
1
1
4
5
1
1
1
1
2
5
3
2
5
1
2
3
2
1
1
4
4
4
5
1
5
1
2
1
1
1
3
5
5
2
5
5
3
4
5
4
5
5
Co re
M o del
Rank
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
YTD
Return
(2.4) %
(2.8)
0.6
(2.3)
(3.1)
(2.0)
(4.3)
0.3
(1.7)
(2.1)
(1.2)
(2.0)
(0.4)
(1.7)
(1.7)
(0.4)
(3.0)
(3.9)
(1.6)
(2.0)
(0.6)
(3.0)
(1.2)
(1.2)
(2.5)
(2.1)
(4.0)
(3.4)
0.7
(1.7)
(2.7)
0.3
(2.5)
(1.0)
(3.1)
4.0
(3.0)
(3.2)
(1.2)
(0.5)
(2.2)
(0.1)
(0.1)
2.4
(0.9)
(3.1)
(2.5)
(2.8)
(3.0)
(1.1)
0.2
(1.8)
(2.2)
(0.2)
1.6
(1.1)
(2.2)
(3.1)
(2.6)
(1.7)
(3.3)
(0.7)
0.7
4.0
(2.9)
(1.1)
1.2
(3.1)
1.4
0.4
(1.3)
3.2
(6.2)
Market
Capitalization
($ Billion)
$171.3
17.8
14.0
12.2
8.9
8.6
6.7
5.8
79.0
53.3
27.4
21.2
17.7
13.0
12.3
12.0
9.5
9.4
8.9
6.1
3.5
234.2
152.3
131.9
75.3
42.7
38.3
26.2
25.7
22.6
15.6
12.2
11.8
10.7
10.4
9.5
8.1
6.6
6.1
5.5
5.5
5.2
5.2
5.1
2.8
166.2
142.0
119.5
94.3
57.8
33.8
26.5
21.1
17.1
13.9
12.9
10.8
10.7
10.6
9.9
9.1
8.2
7.1
6.7
6.3
6.0
6.0
5.8
5.2
4.6
4.2
4.0
3.7
1.7
Source: Empirical Research Partners Analysis, FactSet Research Systems.
¹ Sentiment based on net number of positive words used in conference call Q&A sessions in 2015. Word sentiment based on Loughran-McDonald Dictionary,
available at https://www3.nd.edu/~mcdonald/Word_Lists.html.
² Winner language is based on Q&A exposure to Topic 5 (the Margin Expansion Story) from analysis of past winners.
10
Stock Selection: Research and Results January 2016
Appendix 2 (Cont.): Large-Capitalization Stocks
Best Quintile of the Core Model
Sorted by Average of Conference Call Metrics and Market Capitalization
As of Early-January 2015
Symbol
INTC
MO
MCD
PNC
CCL
LYB
ANTM
AMAT
STI
HPQ
UAL
TSN
FE
LRCX
CA
CTXS
CCE
KEY
JNPR
TW
VOYA
RE
FLR
FSLR
FLEX
MUR
NRG
ORCL
CSCO
ABBV
MS
MCK
COF
ATVI
LUV
CNQ
NOV
BBY
IPG
DRI
SNI
FTR
LM
CHK
MSFT
TGT
EBAY
NV DA
FL
Q
DOX
AES
CDNS
CNP
Company
INTEL CORP
ALTRIA GROUP INC
MCDONALD'S CORP
PNC FINANCIAL SERVICES GROUP INC
CARNIV AL CORP/PLC (USA)
LYONDELLBASELL INDUSTRIES NV
ANTHEM INC
APPLIED MATERIALS INC
SUNTRUST BANKS INC
HP INC
UNITED CONTINENTAL HLDGS INC
TYSON FOODS INC -CL A
FIRSTENERGY CORP
LAM RESEARCH CORP
CA INC
CITRIX SYSTEMS INC
COCA-COLA ENTERPRISES INC
KEYCORP
JUNIPER NETWORKS INC
TOWERS WATSON & CO
VOYA FINANCIAL INC
EV EREST REINSURANCE GROUP LTD
FLUOR CORP
FIRST SOLAR INC
FLEXTRONICS INTERNATIONAL
MURPHY OIL CORP
NRG ENERGY INC
ORACLE CORP
CISCO SYSTEMS INC
ABBV IE INC
MORGAN STANLEY
MCKESSON CORP
CAPITAL ONE FINANCIAL CORP
ACTIVISION BLIZZARD INC
SOUTHWEST AIRLINES
CANADIAN NATURAL RESOURCES
NATIONAL OILWELL VARCO INC
BEST BUY CO INC
INTERPUBLIC GROUP OF COS
DARDEN RESTAURANTS INC
SCRIPPS NETWORKS INTERACTIV E
FRONTIER COMMUNICATIONS CORP
LEGG MASON INC
CHESAPEAKE ENERGY CORP
MICROSOFT CORP
TARGET CORP
EBAY INC
NV IDIA CORP
FOOT LOCKER INC
QUINTILES TRANSNATIONAL HLDG
AMDOCS
AES CORP
CADENCE DESIGN SYSTEMS INC
CENTERPOINT ENERGY INC
Price
$33.99
57.39
117.58
93.16
54.20
87.18
139.21
18.47
41.58
11.60
55.61
52.96
31.50
77.73
28.16
74.17
48.65
12.96
27.41
123.00
36.52
181.08
47.71
66.72
11.11
22.95
11.53
35.75
26.41
57.61
31.48
194.67
70.74
37.62
41.96
21.57
34.61
30.65
22.82
62.75
53.84
4.65
39.73
4.95
54.80
73.55
26.43
32.37
65.31
67.04
54.23
9.43
20.53
18.27
Quintiles (1=Best; 5=Worst)
Conference Call Metrics
Super Factors
Winner
Earnings
Q&A
Language
Quality
Average
Sentiment¹
Used In
Capital
and
Of The
(5=Highest)
Q&A²
Valuation Deployment
Trend
Two
4
3
1
1
1
3.5
5
2
4
3
1
3.5
2
5
4
2
1
3.5
4
3
1
2
na
3.5
5
2
4
2
1
3.5
3
4
1
2
2
3.5
5
2
1
2
2
3.5
5
2
3
1
2
3.5
5
2
1
1
na
3.5
3
4
1
1
5
3.5
4
3
1
2
4
3.5
5
2
2
2
1
3.5
2
5
1
4
3
3.5
4
3
2
1
1
3.5
3
4
2
1
4
3.5
2
5
2
3
1
3.5
5
2
2
2
1
3.5
5
2
1
1
na
3.5
3
4
4
1
1
3.5
3
4
4
3
1
3.5
2
5
1
1
na
3.5
3
4
1
1
na
3.5
4
3
1
1
2
3.5
5
2
3
1
4
3.5
4
3
2
1
3
3.5
2
5
1
1
2
3.5
2
5
1
1
3
3.5
3
5
2
1
4
4.0
3
5
3
1
1
4.0
3
5
2
2
4
4.0
4
4
1
2
na
4.0
3
5
1
1
1
4.0
4
4
1
1
na
4.0
5
3
4
2
2
4.0
4
4
3
1
1
4.0
4
4
1
1
2
4.0
4
4
1
1
3
4.0
5
3
1
1
3
4.0
4
4
3
1
1
4.0
4
4
2
3
1
4.0
4
4
1
1
3
4.0
4
4
1
1
2
4.0
5
3
1
1
na
4.0
5
3
1
1
4
4.0
3
1
2
4
5
4.5
5
4
1
1
1
4.5
4
5
1
2
5
4.5
4
5
4
1
2
4.5
5
4
3
2
1
4.5
5
4
3
1
1
4.5
4
5
2
1
2
4.5
4
5
1
1
1
4.5
5
4
2
2
1
4.5
5
5
1
2
2
5.0
Market
Reaction
2
1
1
2
1
2
4
4
2
5
2
1
3
1
4
1
3
3
1
1
4
2
3
1
3
5
5
4
3
2
5
4
4
1
1
5
5
5
2
1
4
5
5
5
1
2
2
1
1
1
3
5
1
4
Co re
M o del
Rank
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
YTD
Return
(1.3) %
(1.4)
(0.5)
(2.3)
(0.5)
0.3
(0.2)
(1.1)
(2.9)
(2.0)
(2.9)
(0.7)
(0.7)
(2.1)
(1.4)
(2.0)
(1.2)
(1.7)
(0.7)
(4.3)
(1.1)
(1.1)
1.0
1.1
(0.9)
2.2
(2.0)
(1.7)
(2.0)
(2.8)
(1.0)
(1.3)
(2.0)
(2.8)
(2.6)
(1.2)
3.3
0.7
(2.0)
(1.4)
(2.5)
(0.4)
1.3
10.0
(1.2)
1.3
(3.8)
(1.8)
0.3
(2.4)
(0.6)
(1.5)
(1.3)
(0.5)
Market
Capitalization
($ Billion)
$160.8
112.5
108.0
47.5
41.8
39.5
36.3
21.4
21.4
20.9
20.8
19.5
13.3
12.3
12.2
11.7
11.0
10.8
10.6
8.5
7.9
7.8
6.8
6.7
6.2
3.9
3.7
150.4
134.2
94.2
61.0
44.8
37.8
27.5
27.3
23.6
13.0
10.6
9.3
8.0
6.9
5.4
4.3
3.3
437.7
45.5
31.8
17.4
9.0
8.3
8.2
6.4
6.1
7.9
Source: Empirical Research Partners Analysis, FactSet Research Systems.
¹ Sentiment based on net number of positive words used in conference call Q&A sessions in 2015. Word sentiment based on Loughran-McDonald Dictionary,
available at https://www3.nd.edu/~mcdonald/Word_Lists.html.
² Winner language is based on Q&A exposure to Topic 5 (the Margin Expansion Story) from analysis of past winners.
11