10.1016 J.jebo.2014.06.004 Facebook S Daily Sentiment and International Stock Markets

G Model
ARTICLE IN PRESS
JEBO-3383; No. of Pages 14
Journal of Economic Behavior & Organization xxx (2014) xxxxxx
Contents lists available at ScienceDirect
Journal of Economic Behavior & Organization

journal homepage: www.elsevier.com/locate/jebo
Facebooks daily sentiment and international stock markets

Antonios Siganos a , Evangelos Vagenas-Nanos a , Patrick Verwijmeren a,b,c,
a
b
c
University of Glasgow, Adam Smith Business School, Glasgow, G12 8QQ, Scotland, United Kingdom
Erasmus University Rotterdam, Burgemeester Oudlaan 50, 3000DR Rotterdam, Netherlands
University of Melbourne, 198 Berkeley Street, Victoria 3010, Australia
a r t i c l e
i n f o
Article history:
Received 25 February 2013
Received in revised form 30 April 2014
Accepted 10 June 2014
Available online xxx
Keywords:
Behavioral nance
Sentiment
Facebooks gross national happiness index
a b s t r a c t
We examine the relation between daily sentiment and trading behavior within 20 international markets by exploiting Facebooks Gross National Happiness Index. We nd that
sentiment has a positive contemporaneous relation to stock returns. Moreover, sentiment
on Sunday affects stock returns on Monday, suggesting causality from sentiment to stock
markets. We observe that the relation between sentiment and returns reverses the following weeks. We further show that negative sentiments are related to increases in trading
volume and return volatility. These results highlight the importance of behavioral factors
in stock investing.
2014 Elsevier B.V. All rights reserved.
1. Introduction
An important part of behavioral nance concerns the relation between investor sentiment and stock market returns.
Measuring sentiment is, however, not a trivial exercise. The conventional method to obtain measurements of sentiment
is to take surveys of households. In this type of study, researchers typically select a random number of households and
ask a small number of questions to identify the level of optimism or pessimism per household. The responses are then
aggregated to construct an average sentiment level.1 Although these studies have provided important insights, the survey
method has some important weaknesses. One weakness is that sample sizes and participation rates are typically low. For
example, the Michigan Consumer Sentiment survey is sent to only 500 households, and the Consumer Condence Index
to 5000 households. Another weakness is that the surveys are typically conducted on a monthly frequency. The resultant
studies then typically rely on the assumption that sentiment remains stable from day to day over the survey period.2
We propose to use an alternative measure of sentiment, based on status updates on Facebook, which is the worlds
largest social network site. Facebooks Gross National Happiness Index (FGNHI) has been developed by Facebooks data
team and offers daily sentiment for twenty international markets. The website investorwords.com denes sentiment as a
measurement of the mood of a given investor or the overall investing public, either bullish or bearish. Facebook measures
peoples mood by examining the positive and negative terms used by Facebook participants. The assumption is that happy
Corresponding author at: Erasmus University Rotterdam, Burgemeester Oudlaan 50, 3000DR Rotterdam, Netherlands. Tel.: +31 104081392.
E-mail addresses: antonios.siganos@glasgow.ac.uk (A. Siganos), evangelos.vagenas-nanos@glasgow.ac.uk (E. Vagenas-Nanos), verwijmeren@ese.eur.nl
(P. Verwijmeren).
1
Sentiment indexes based on surveys include the University of Michigan Consumer Sentiment Index, and the Consumer Condence Index (see for
example Brown and Cliff, 2004; Lemmon and Portniaguina, 2006; Qiu and Welch, 2006).
2
Several other studies have used indirect measures of sentiment. Indirect sentiment measures represent economic and nancial variables that are
believed to capture investors state of mind. Examples of indirect proxies are fund ows, trading volume, IPO volume-rst day return, and closed-end fund
discounts (see also Lee et al., 1991; Baker and Wurgler, 2007; Brown et al., 2008).
http://dx.doi.org/10.1016/j.jebo.2014.06.004
0167-2681/ 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Siganos, A., et al., Facebooks daily sentiment and international stock markets. J. Econ.
Behav. Organ. (2014), http://dx.doi.org/10.1016/j.jebo.2014.06.004
G Model
ARTICLE IN PRESS
A. Siganos et al. / Journal of Economic Behavior & Organization xxx (2014) xxxxxx
participants on average use more positive terms when updating their status and unhappy participants on average use more
negative terms.
Although many participants on Facebook are young, Facebook is no longer the exclusive domain of young people, and a
substantial amount of Facebook users are likely to invest. Appendix A reports the percentage of a nations population that
has a Facebook account. Typically, this percentage is close to 50%, which highlights the high participation rates of Facebook.
The average age of Facebook users in recent times is about 31 years (Kramer and Chung, 2011), with more than a quarter of
Facebook users being older than 45. It is also important to note that even when investors are underrepresented on Facebook,
it is still the case that the underlying factors that make Facebook users optimistic, like their nations win in the World Cup,
are also likely to make the investors in that country more optimistic.
The data from Facebook provide some important advantages. First, the index has been constructed based on text analyses
of the status updates of millions of participants, which stands in contrast to the limited sample sizes of household surveys.3
Second, FGNHI represents sentiment on a daily level, which allows us to test contemporaneous relations between sentiment
and stock market returns. Third, status updates on Facebook are undirected by any particular question that may be asked
in surveys, but are self-descriptive messages.4 A fourth benet of our data is the international coverage. Other sentiment
indexes are typically only available for the United States (like the University of Michigan Consumer Sentiment Index) or
for a small number of developed markets (the UBS/Gallup Index of Investor Optimism offers monthly sentiment levels for
France, Germany, Italy, Spain, and the United Kingdom). We obtain a direct measure of sentiment for twenty countries.5
We explore whether Facebooks Gross National Happiness Index is related to stock market returns for the period
September 2007March 2012. Our main hypothesis is that positive sentiment leads to positive biases in returns. This hypothesis follows from the behavioral nance theory of De Long et al. (1990), who predict that noise trader sentiment affects
nancial markets when noise traders are plentiful and there are limits to arbitrage. Other studies have mostly focused on
a related prediction following from De Long et al. (1990), which is that prices will revert to fundamental values in the long
term. Most notably, Schmeling (2009) and Baker et al. (2012) show that their measures of investor sentiment are related
to negative returns in the future, when any overly optimistic or pessimistic expectation is corrected. Our daily sentiment
measure from Facebook allows us to also test behavioral nances predictions on the contemporaneous relation between
sentiment and stock returns.
We nd a signicant positive relation between sentiment and contemporaneous stock market returns, showing that
optimistic (pessimistic) sentiment is related to gains (losses) in the market index. These results hold for different regions,
languages, and religions. Moreover, these results are not solely driven by particular days on which Facebooks measure of
sentiment reaches extremely high or low levels. In the cross-section, we expect that optimism is especially related to stock
returns for stocks that are disproportionally held by noise traders (Lee et al., 1991). Because small rms might have relatively
more noise traders as compared to institutional traders, Lemmon and Portniaguina (2006) and Baker and Wurgler (2007)
argue that behavioral biases are expected to be mostly present in the stock returns of small rms. We exploit MSCI indexes
and conrm that our results are strongest for small rms.
Potentially, the relation between sentiment and stock returns is subject to reverse causality, as good market performance
could create positive feelings (Brown and Cliff, 2004). Our data provide substantial research leverage in this regard. As people
also update their status in the evening (after the markets close), we expect to nd that sentiment on day t affects returns
on day t + 1. In line with this expectation, we observe that sentiment is related to the next days market returns. In addition,
we exploit the availability of sentiment data on Sundays. Any sentiment observed on Sunday is not likely to be the direct
result of market returns on Friday, reducing worries of reversed causality when returns are auto-correlated. We nd that
sentiment on Sunday is related to market returns on Monday.
To examine causality further, we use models that adjust for lead-lag effects. The results of our analysis with these models
again suggest that sentiment affects market returns. Although these results provide new insights into the relation between
sentiment and stock returns, it is important to stress that our results on causality have to be interpreted with appropriate
caution, as several events might affect both sentiment and stock returns. For example, NASAs successful Mars landing could
at the same time increase peoples sentiment and increase expected future spending on space programs. Still, we consider
it unlikely that these types of events happen frequently enough to drive our results across international markets and in the
cross-section.6 In addition, we nd that controlling for macroeconomic conditions by using the Policy Uncertainty Index (as
developed by Baker et al., 2013) does not change our conclusions. Our results are further strengthened as we show that the
relation between sentiment and stock returns reverses over the following weeks, indicating a correction to fundamental
values.
Kramer (2010) reports that, on average, over 40 million status updates are posted on Facebook per day.
Facebook users write their status updates in a box that contains an open question, which is typically: How are you feeling?, How are you doing?,
Whats on your mind?, or How is it going?
5
Using indirect measures of sentiment also allows for an international study. In particular, Baker et al. (2012) construct sentiment indexes within six
developed countries, using indicators like volatility premiums, IPO underpricing, and number of IPOs. Schmeling (2009) uses consumer condence levels
within 18 countries as a measure of sentiment.
6
We have checked all the status updates of our Facebook friends over January 2013 and observed that less than one percent of the updates relate to an
event with potentially important effects on the economy.
4
G Model
ARTICLE IN PRESS
We further explore whether sentiment on Facebook is related to trading volume and stock price volatility. We nd
that sentiment has a signicant negative relation to trading volume across international markets, indicating that negative
sentiment is associated with higher transaction volume. This result is in line with evidence from psychology, where negative
sentiment causes investors to trade more, as they look to overcome their negative sentiment with a positive outcome from
an alternative activity (Erber and Tesser, 1992). Similarly, we nd that sentiment on Facebook is negatively related to stock
price volatility, suggesting that negative sentiment is associated with a higher propensity of investors to speculate.
Our study is related to other contemporary papers that use data from social media to examine nancial markets. For
example, Bollen et al. (2011), Zhang et al. (2011) and Yang et al. (2013) examine mood on Twitter.7 In another related study,
Da et al. (2013) exploit the volume of queries related to household concerns in Google, and conclude that this volume predicts
nancial markets in the United States. Most notably, a higher volume of concerns corresponds to lower S&P 500 returns.
Most closely related to our study is Karabulut (2013), who also uses sentiment on Facebook. His study focuses on the U.S.
market, and corroborates our ndings that sentiment is positively related to stock returns. The main contribution of our
study compared to other papers on social media is to exploit Facebooks daily sentiment proxy across twenty international
markets, and to provide insights into potential causality by exploiting sentiment information on non-trading days. The
strength of the Facebook measure in representing sentiment for a specic country on a specic day originates mostly from
the sheer size of Facebook. Facebook has over a billion users, and in 2010 has passed Google to become the most visited
website in the United States, accounting for more than 7% of U.S. web trafc. Importantly, over 80% of Facebook users reside
outside of the United States, which makes Facebook data perfectly suitable for an international study (Wilson et al., 2012).8
The remainder of the paper is structured as follows. Section 2 describes our data and explores the validity of Facebooks
index. Section 3 discusses the empirical results on the relation between sentiment and stock returns. We examine issues
related to causality in Section 4 and consider stock price reversals and additional tests exploiting the international aspect of
our data in Section 5. Section 6 examines the relation between sentiment and volume and volatility, and Section 7 contains
our conclusions.
2. Data and validity of FGNHI
We obtain daily sentiment data from Facebook (http://www.facebook.com/gnh/). Facebook refers to its sentiment index
as the Gross National Happiness index, inspired by the former king of Bhutan, Jigme Singye Wangchuck, who in 1972 began
to construct an index that attempted to capture his nations level of happiness more accurately than the Gross National
Product. Bhutans index measures happiness within a multidimensional framework by using 33 (in the latest 2010 index)
indicators based on the following nine domains: psychological wellbeing, health, education, time use, cultural diversity
and resilience, good governance, community vitality, ecological diversity and resilience, and living standards. The Gross
National Happiness Index developed by Facebook measures happiness based on peoples status updates, which relates to
the dimension of valence. Facebooks index was rst published in 2009. We collect data in March 2012, when sentiment data
are available for the following twenty countries: Argentina, Australia, Austria, Belgium, Canada, Chile, Colombia, Germany,
India, Ireland, Italy, Mexico, the Netherlands, New Zealand, Singapore, South Africa, Spain, the United Kingdom, the United
States, and Venezuela.
FGNHI is estimated by Facebooks data team based on the status updates of millions of Facebook participants. The procedure is explained and validated in Kramer (2010). Based on Text Analysis and Word Count (TAWC) programs, the Facebook
data team analyzes the percentage of positive and negative terms that are used across all participants. They follow the
Linguistic Inquiry and Word Count (LIWC) dictionary to categorize terms as positive, neutral, or negative. For example, a
status update of What a nice day contains one positive term (nice), and all remaining terms are neutral. More specically,
FGNHI is estimated by the Facebook data team as follows:
FGNHIi,j =
xp,i xp,all
p,all
xn,i xn,all
n,all
(1)
where FGNHIi,j is the sentiment index of country j at day i, xp,i and xn,i show the average positive (p) and negative (n)
words used respectively at day i for the country, and xp,all , xn,all p,all , n,all are the average (x) positive and negative words
used over the duration of the index and the standard deviation () of those variables. Facebooks data team excludes the
extreme high and low 10% of the days when estimating xp,all , xn,all p,all , n,all to minimize the impact of extreme values
on the estimation of daily sentiment levels. A positive (negative) FGNHI score at day i for a country indicates an optimistic
(pessimistic) sentiment above (below) which is found on a typical day in that country.
7
Our sample period is substantially larger than the sample periods used in these studies. Bollen et al. (2011) examine tweets in 2008 and nd that some
mood dimensions are related to the Dow Jones index. Zhang et al. (2011) use a randomized sample of tweets over six months in 2009 and nd that the
percentage of emotional tweets is negatively related to U.S. stock market returns. Yang et al. (2013) conclude that sentiment in tweet messages is related
to the Dow Jones index for one month in 2013.
8
We further relate to studies on the effects of weather and sports results on stock markets (see for example Saunders, 1993; Hirshleifer and Shumway,
2003; Edmans et al., 2007). In these studies, sentiment cannot be directly observed, but the assumption is that the weather and sport results affect
sentiment, which in turn affects market outcomes. With Facebook data we observe sentiment more directly, which, for example, overcomes the existence
of non-monotonic relations between weather and sentiment, and the fact that different people prefer different types of weather.
G Model
ARTICLE IN PRESS
4
Table 1
Summary statistics.
Argentina
Australia
Austria
Belgium
Canada
Chile
Colombia
Germany
India
Ireland
Italy
Mexico
Netherlands
New Zealand
Singapore
South Africa
Spain
United Kingdom
United States
Venezuela
All countries
America
Europe
Other countries
Number of observations
Average
St. dev.
Max.
Min.
1487
1492
1497
1619
1491
1611
1600
1620
1611
1492
1617
1617
1615
1494
1494
1489
1618
1615
1613
1612
0.012
0.004
0.007
0.005
0.009
0.013
0.006
0.000
0.053
0.013
0.014
0.006
0.015
0.010
0.005
0.008
0.011
0.007
0.012
0.010
0.022
0.017
0.021
0.019
0.018
0.024
0.024
0.020
0.035
0.020
0.032
0.020
0.020
0.020
0.017
0.018
0.021
0.017
0.022
0.027
0.137
0.144
0.146
0.142
0.135
0.142
0.142
0.134
0.144
0.141
0.145
0.143
0.138
0.137
0.146
0.140
0.137
0.136
0.133
0.144
0.089
0.046
0.072
0.059
0.054
0.109
0.072
0.040
0.183
0.072
0.042
0.061
0.070
0.131
0.043
0.051
0.106
0.062
0.058
0.106
31,304
11,031
12,693
7580
0.010
0.010
0.005
0.017
0.025
0.023
0.023
0.030
0.146
0.144
0.146
0.146
0.183
0.109
0.106
0.183
This table reports descriptive statistics for Facebooks Gross National Happiness Index across twenty international markets. Observations during nontrading days are included. In the last four rows of this table we either cluster all countries, all countries in America, all countries in Europe, or all countries
outside of America and Europe.
We exclude daily FGNHI observations above the 99th percentile, as we observe that these typically relate to messages like
Merry Christmas and Happy New Year.9 These messages might not necessarily be informative about peoples sentiment.
Table 1 reports the number of observations and other summary statistics of our Gross National Happiness index across
countries. By construction, the averages are close to zero.
As an untabulated descriptive statistic, we have estimated the correlations of FGNHI across countries. We nd that the
correlations tend to be positive and statistically signicant, with an average correlation coefcient of 0.589. We observe the
highest correlation between FGNHI in the United States and Canada (0.921).
Similar to most other direct sentiment indexes, FGHNI reects sentiment of non-investors. Although Facebook was
initially intended to be used by students (upon its introduction in 2004), the average age has gradually increased throughout
the years. For a sample period from September 2007 to February 2010, Kramer and Chung (2011) report that the average
age is 33, 32, 30, and 31 within the United States, Canada, the United Kingdom, and Australia, respectively. In fact, more
than a quarter of Facebook users are older than 45, and less than 10% of users are younger than 18. Appendix A shows
the high participation rates of a countrys (online) population in Facebook. As such, many investors are expected to be on
Facebook. In addition, the same underlying factors that make a countrys Facebook population happy should also have a
positive inuence on the mood of most of the investing population of that country. In line with this reasoning, studies have
shown that investors respond to sentiments that would also inuence the mood in a country, like the weather and football
results (e.g., Saunders, 1993; Hirshleifer and Shumway, 2003; Edmans et al., 2007; Kaplanski et al., 2013). We therefore argue
that the demographic characteristics of Facebook users do not generate a major concern regarding our studys validity.10
To validate FGNHI empirically, Table 2 tests whether FGNHI is correlated with other recently developed daily sentiment
indexes. In particular, we compare the U.S. FGNHI measure to the Gallup and Google indexes. The Gallup Daily Index is
a sentiment index based on phone call interviews in which U.S. participants are asked about their future expectations.11
The Google sentiment index is developed by Da et al. (2013) and is based on the search activity of U.S. households of
thirty negative terms toward the economy in Google.12 As the Google sentiment index measures pessimistic sentiment, we
multiply the Google sentiment index by minus one.
9
Although stock markets are closed on public holidays, they could be open on, for example, Mothers Day, when similar messages are posted (Kramer,
2010). We also report the results when we exclude values of the FGNHI variable above 0.05 and below 0.05, and when no outliers are excluded from the
sample.
10
Because personal information is deleted by Facebooks data team before they construct the sentiment indexes, the exact demographics of Facebook
users in our sample are not available.
11
See www.gallup.com/poll/122840/gallup-daily-economic-indexes.aspx. We divide Gallups values by 1000 for comparability.
12
These terms include recession, depression, bankruptcy, and unemployment. We manually download the search activity of U.S. households in
the thirty terms through Googles Insight. The sentiment index is the average logarithmic change in search activity.
G Model
ARTICLE IN PRESS
Table 2
Comparison of U.S. daily sentiment indexes.
Panel A: Descriptive statistics
U.S. FGNHI
Gallup Index
Google Index
Average
St. dev
Max.
Min.
Median
Start date
End date
1613
1452
1613
0.012
0.037
0.001
0.022
0.012
0.177
0.133
0.014
0.460
0.058
0.066
0.608
0.013
0.033
0.008
09/09/2007
02/01/2008
09/09/2007
02/03/2012
02/03/2012
02/03/2012
Panel B: Correlation coefcients
U.S. FGNHI
Gallup Index
Google Index
0.434*
(0.000)
0.167*
(0.000)
This table compares U.S. daily sentiment indexes. Panel A shows descriptive statistics for the U.S. FGNHI, Gallup and Google indexes and Panel B shows the
Pearson correlations of the FGNHI index with the alternative daily sentiment indexes. U.S. FGNHI is Facebooks U.S. sentiment index. The Gallup index is
based on phone call interviews regarding households condence in the U.S. economy. The Google index is estimated based on thirty negative terms toward
the economy, as identied by Da et al. (2013), by taking the average logarithmic change in search activity across these terms. We multiply the Google Index
by minus one. Observations during non-trading days are included. P-values are shown in parentheses.
*
Indicates statistical signicance at the 1% level.
Table 3
Sentiment and stock market returns.
Stock market returns
N
Parameter estimate
Standard error
22,361
7888
9063
5410
0.031***
0.029***
0.035***
0.029***
0.008
0.009
0.013
0.011
Panel B: Sample with potential outliers excluded

20,870
All countries
America
7557
8743
Europe
4570
Other countries
0.048***
0.035**
0.066***
0.037**
0.013
0.014
0.021
0.015
Panel A: Overall sample

All countries
America
Europe
Other countries
This table shows whether sentiment is related to stock market returns. The parameter estimate represents the coefcient of regressing daily stock market
returns on our daily sentiment measure from Facebook. Our sample period is September 2007March 2012. All regressions include day-of-the-week xed
effects and country xed effects. In each panel we estimate the regression four times: once for all countries, once for all countries in America, once for all
countries in Europe, and once for all countries outside of America and Europe. We report standard errors clustered by date. Panel B excludes sentiment
values above 0.05 and below 0.05.
**
***
Panel A of Table 2 offers descriptive statistics of the Gallup and Google indexes. Note that Gallups coverage is shorter
than FGNHIs and Googles coverage. Panel B of Table 2 shows that there are signicantly positive correlation coefcients
between the U.S. sentiment measure from Facebook and the Gallup and Google indexes, with Pearson correlation coefcients
of 0.434 and 0.167 (both signicant at the 1% level), respectively.
3. Empirical results on stock returns
3.1. The relation between sentiment and contemporaneous stock market returns
De Long et al. (1990) predict that optimism (pessimism) of noise traders causes temporary upward (downward) biases
in stock prices. To test the relation between sentiment and stock returns, we start with a relatively simple test. We pool
all countries and focus on contemporaneous relations, i.e. we measure sentiment and stock returns on the same day.13 Our
regression analyses include country xed effects and day-of-the-week xed effects, and we cluster standard errors by date
to account for the correlation in returns across countries. Panel A of Table 3 shows the results of regressing stock market
returns on Facebooks sentiment measure.
13
TOTMK indexes from Datastream are used for countries market returns.
G Model
ARTICLE IN PRESS
Table 4
The relation between sentiment and MSCI return indexes.
All countries
America
Europe
Other countries
Small
Large
Premium
Value
Growth
Premium
0.043***
(0.009)
0.048***
(0.012)
0.042***
(0.012)
0.040***
(0.011)
0.035***
(0.011)
0.032**
(0.014)
0.037**
(0.015)
0.035**
(0.015)
0.009*
(0.005)
0.018***
(0.007)
0.005
(0.007)
0.006
(0.011)
0.041***
(0.012)
0.037***
(0.014)
0.050***
(0.017)
0.028**
(0.013)
0.028***
(0.010)
0.025*
(0.013)
0.028**
(0.013)
0.032**
(0.014)
0.013***
(0.005)
0.012**
(0.006)
0.022**
(0.010)
0.004
(0.005)
This table shows whether sentiment is related to the returns within alternative MSCI indexes. We distinguish between the MSCI indexes for small, large,
value, and growth stocks. The parameter estimate represents the coefcient of regressing daily stock returns on our daily sentiment measure from Facebook.
Our sample period is September 2007March 2012. All regressions include day-of-the-week xed effects and country xed effects. For each style category,
we estimate the regression four times: once for all countries, once for all countries in America, once for all countries in Europe, and once for all countries
outside of America and Europe. The premium is estimated by replacing the dependent variable by the difference in returns of the relevant MSCI indexes.
We report standard errors clustered by date in parentheses.
*
**
***
We nd that FGNHI is positively related to stock returns. This relation is statistically signicant at the 1% level. The
coefcient of 0.031 implies that if sentiment changes from zero to 0.1, then, on average, daily contemporaneous returns are
31 basis points higher. These results suggest that optimistic sentiment is associated with gains in the aggregate market.
Panel A of Table 3 also shows the results for different regions. Most countries in our sample are from either America or
Europe, and we create subsamples based on these regions. Our third subsample pools all remaining countries. The creation
of subsamples is likely to be informative about the robustness of our results. It can be seen that the positive relation between
sentiment and stock market returns is present in all three subsamples.
We further examine whether our results are driven by a few days with very high or low sentiment. Although we have
already excluded observations above the 99th percentile, we extend our exclusion to all FGNHI observations above 0.05
or below -0.05. Panel B of Table 3 shows that excluding these observations does not change our conclusions. In fact, the
coefcient estimates are increased, with the coefcient for sentiment being 0.048 rather than 0.031 for the estimation that
includes all countries.14
3.2. Cross-sectional stock returns

Optimism is expected to be especially related to stock returns for stocks that are disproportionally held by noise traders
(Lee et al., 1991). Baker and Wurgler (2007) and Schmeling (2009) argue that small rms in particular might be associated
with many noise traders and could be more subject to behavioral biases. Indeed, Lemmon and Portniaguina (2006) nd that
investors appear to overvalue small relative to large stocks when consumer condence is high. To test the conjecture that
sentiment is more important for small rms, we download both the MSCI indexes for small and large rms from Datastream.
Moreover, we differentiate between value and growth stocks by downloading the representative MSCI indexes for these
two classications. Kumar and Lee (2006) argue that noise traders overweight value stocks, but Baker and Wurgler (2006)
argue that extreme growth rms are relatively hard to arbitrage, which could also increase the likelihood of behavioral
biases. The latter study nds empirically that the coefcients are similar for both value and growth rms in the United
States. Schmeling (2009) uses an international setting and nds that the relation is stronger for value rms, but observes
that the relation is also present for growth rms.
Table 4 shows our results for small, large, value, and growth stocks. In line with our expectations, we nd that our results
are strongest for small rms. The small size premium is signicant at the 10% level for our overall sample and signicant at
the 1% level for the American sample. Our results further corroborate the ndings of Schmeling (2009) in that the relation
between sentiment and returns is stronger for value rms, but also present in growth rms.15 Overall, we conclude that our
results are relevant for different types of rms.
14
We have also estimated the relation between sentiment and stock returns in a sample where no outliers were deleted. We nd that the parameter
coefcients of the sentiment variable are 0.013 (signicant at the 1% level), 0.011 (signicant at the 1% level), 0.012 (signicant at the 5% level) and 0.016
(signicant at the 1% level) for all countries, America, Europe, and the Other countries, respectively. Hence, the relation between stock returns and
sentiment is present both with and without our sample restrictions.
15
In untabulated results, we focus on the U.S. and corroborate the ndings of Baker and Wurgler (2006) that the coefcients for value and growth rms
are similar in the United States.
G Model
ARTICLE IN PRESS
Table 5
Sentiment and next days stock market returns.
Parameter estimate
Standard error
Panel A: Stock market returns on the next day

All countries
22,397
7890
America
9085
Europe
5422
Other countries
0.021***
0.014*
0.026**
0.024**
0.007
0.008
0.011
0.011
Panel B: Sundays sentiment and Mondays stock market returns

All countries
4488
1573
America
1826
Europe
1089
Other countries
0.042**
0.023
0.050*
0.063***
0.017
0.017
0.027
0.023
This table shows whether sentiment is related to stock market returns on the next day. The parameter estimate represents the coefcient of regressing
daily stock market returns on the lagged value of our daily sentiment measure from Facebook. Our sample period is September 2007March 2012. All
regressions include day-of-the-week xed effects and country xed effects. In each panel we estimate the regression four times: once for all countries,
once for all countries in America, once for all countries in Europe, and once for all countries outside of America and Europe. We report standard errors
clustered by date. Panel B is a sub-set of Panel A and only includes our sentiment measure on Sunday, with market returns on Monday.
*
**
***
4. Examining the causal relation between sentiment and stock returns

4.1. Reversed causality
Brown and Cliff (2004) stress the importance of potential reversed causality. When returns are high, people could become
happier. Moreover, evidence by Heimer and Simon (2012) suggests that traders with good performance are more likely to
communicate about their trading activity on networking sites.
Our daily data provide substantial research leverage in examining causality. Facebook statuses are also updated in the
evening. In fact, a 2012 Oracle white paper reports that Facebook activity is at particularly high levels around 8 pm, although
the overall peak occurs at 3 pm.16 Therefore, as our daily sentiment measure captures some of the sentiment after the close
of the market, we can test whether todays sentiment measures are partially reected in tomorrows stock returns. This
potential relation is unlikely to be explained by reversed causality. Panel A of Table 5 shows the results when we use the
lagged value of our sentiment measure.
The results in Table 5 suggest a positive relation between sentiment on Facebook on day t and stock market returns
on day t + 1. This relation holds for all our different regions. The magnitude of the relation is lower than for sentiment and
contemporaneous returns, as the coefcient estimate for our sentiment measure is reduced to 0.021 in our specication
with all countries included.
A potentially even stronger test to control for reversed causality is to explore whether Sundays sentiment is related
to Mondays market returns. That is, we exploit the availability of our sentiment measures on non-trading days. Fridays
returns could perhaps affect mood on Saturday, but it is unlikely that stock returns on Friday have a strong effect on Facebook
sentiment on Sunday. Therefore, any relation between sentiment on Sunday and stock returns on Monday is unlikely to be
explained by reverse causality, even when the returns on Friday and Monday would be auto-correlated. Panel B of Table 5
shows the results.
We nd that our results remain relatively strong when we focus on the relation between sentiment on Sunday and stock
returns on Monday. As before, sentiment and stock returns are positively related.
4.2. Lead-lag relationships
In this section, we use models representing a more sophisticated method of testing whether sentiment affects stock
returns, as they adjust for multiple lead-lag effects. The goal is to examine interactions between sentiment and stock returns
and establish Granger-type causality. In line with other recent studies in the eld (see for example Schmeling, 2009), we
use ve lags for sentiment and market returns. More specically, we estimate the following model:
Market returnit = a1 + a2 FGNHIit +
5

j=1
16
bij FGNHIitj +
5
cij Market returnitj + uit
(2)
j=1
See http://www.oracle.com/us/products/managing-your-facebook-community-1840523.pdf
G Model
ARTICLE IN PRESS
8
Table 6
Lead-lag effects.
Constant
FGNHIt
FGNHIt 1
FGNHIt2
FGNHIt3
FGNHIt4
FGNHIt5
Returnst1
Returnst2
Returnst3
Returnst4
Returnst5
N
All countries
America
Europe
Other countries
0.001
(0.001)
0.021**
(0.009)
0.001
(0.011)
0.026**
(0.011)
0.001
(0.011)
0.009
(0.011)
0.013
(0.009)
0.028
(0.025)
0.019
(0.032)
0.034
(0.028)
0.025
(0.028)
0.031
(0.033)
22,255
0.001**
(0.000)
0.016*
(0.010)
0.005
(0.011)
0.025
(0.015)
0.005
(0.012)
0.012
(0.012)
0.015
(0.010)
0.017
(0.030)
0.014
(0.035)
0.001
(0.031)
0.028
(0.032)
0.045
(0.036)
7852
0.000
(0.001)
0.018
(0.016)
0.004
(0.020)
0.016
(0.018)
0.007
(0.017)
0.021
(0.020)
0.015
(0.016)
0.026
(0.035)
0.031
(0.040)
0.050
(0.038)
0.028
(0.037)
0.037
(0.041)
9022
0.000
(0.001)
0.033**
(0.015)
0.013
(0.017)
0.041**
(0.017)
0.010
(0.019)
0.013
(0.015)
0.009
(0.015)
0.047
(0.030)
0.002
(0.035)
0.039
(0.030)
0.013
(0.033)
0.003
(0.032)
5381
This table examines the relation between sentiment and stock market returns when allowing multiple lead-lag effects. We estimate the following model:
Market returnit = a1 + a2 FGNHIit +
5
j=1
bij FGHNHIitj +
5
cij Market returnitj + uit
j=1
Five lags are used for both sentiment (FGNHI) and the corresponding market return. Our sample period is September 2007March 2012. All regressions
include day-of-the-week xed effects and country xed effects. We estimate the regression four times: once for all countries, once for all countries in
America, once for all countries in Europe, and once for all countries outside of America and Europe. We report standard errors clustered by date in
parentheses.
*
**
We again include day-of-the-week and country xed effects. Such a xed effects specication allows individual countries
and days of the week to have different regression constants, whereas slope coefcients are restricted to be equal across
countries. Table 6 shows the results of estimating the model.
Table 6 shows that the coefcient estimates for the relation between sentiment (FGNHIit ) and stock returns are again positive. For our estimation with all countries included, the coefcient estimate is 0.021 and the effect is statistically signicant
at the 5% level. As such, our conclusions are unchanged and suggest that sentiment positively affects stock returns.
5. Additional tests
5.1. Stock price reversals
Prior studies (e.g., Schmeling, 2009) have tested the relation between sentiment and stock markets by using monthly
data and examining whether there is a reversed relation between sentiment and stock returns in the subsequent month. The
rationale is that if sentiment results in a contemporaneous increase in stock prices, returns should move back to fundamental
values in the next period. In this subsection, we examine whether patterns of reversals in stock prices are present in our
data.
We estimate a regression model for explaining stock returns that uses up to 30-day lags of Facebooks sentiment measure.
That is, we use 31 main explanatory variables, which are contemporaneous sentiment, sentiment at day 1, sentiment at
day 2, and so on, until sentiment at day 30. We also include day-of-the-week and country xed effects. Table 7 reports the
parameter coefcients. It can be seen that the relation between sentiment and returns tends to weaken for a higher number
of lags. In other words, sentiment does not have a strongly positive relation with stock returns that are measured a few days
later. After nine days, the relation is insignicantly negative. Many of the days do not show signicant effects. The strongest
effect after day 0 is on day 16, when the relation is signicantly negative. When one would sum all the coefcients for the
sentiment measure from day 0 to day t, this sum rstly becomes negative after 20 days. Although we acknowledge that the
G Model
ARTICLE IN PRESS
Table 7
Stock market reversals.
Lags
Parameter
Standard error
Lags
Parameter
Standard error
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0.027**
0.003
0.010
0.003
0.007
0.002
0.012
0.000
0.005
0.012
0.003
0.005
0.007
0.015
0.002
0.010
0.011
0.013
0.010
0.011
0.009
0.010
0.014
0.012
0.010
0.011
0.010
0.011
0.010
0.013
0.013
0.010
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
0.029**
0.018*
0.009
0.027**
0.022
0.001
0.002
0.010
0.008
0.014
0.010
0.005
0.012
0.009
0.014
0.012
0.010
0.011
0.013
0.022
0.011
0.010
0.011
0.011
0.012
0.012
0.013
0.016
0.011
0.010
This table examines the relation between sentiment and stock market returns when adding 30 lagged sentiment variables to our basic regression on the
relation between sentiment and stock market returns. Our sample period is September 2007March 2012 and we include all countries. All regressions
include day-of-the-week xed effects and country xed effects. We report the parameter coefcients for the different lags of sentiment, and also report
standard errors clustered by date.
*
**
analysis is subject to noise, the results are in line with prior studies on price reversals and hint toward the suggestion that
there is a correction to fundamental values.17
5.2. Macroeconomic adjustments

This subsection examines the effect of macroeconomic news releases on our results. This follows for example Kumar
and Lee (2006), who control for macroeconomic variables such as ination and GDP to show that the relation between
sentiment and returns is robust. We require a macroeconomic variable that is available on a daily basis. Da et al. (2013)
use the Policy Uncertainty Index as developed by Baker et al. (2013) as one of the variables to control for changes in daily
U.S. macroeconomic conditions.18 We examine the FGNHI coefcients after controlling for the Policy Uncertainty Index and
nd that our results are robust. More specically, we nd that U.S. FGNHI is positively related to U.S. returns, for a U.S.
sample that consists of 1120 observations with all required information: The FGNHI coefcient changes from 0.049 (with a
p-value of 0.096) without controlling for macroeconomic conditions to 0.050 (with a p-value of 0.080) with controlling for
macroeconomic conditions. The Policy Uncertainty Index obtains a coefcient of 0.003, with a p-value of 0.817. A study of
Karabulut (2013) focuses on the United States and corroborates the positive relation between stock returns and sentiment
on Facebook when controlling for an alternative measure of macroeconomic conditions, which is the Aruoba-Diebold-Scotti
Business Conditions index.
As an alternative test of the robustness of our ndings to fundamental news, we exploit variation in the correlations
among countries. Weeks in which important fundamental news on the state of the world economy is released are likely to
be associated with relatively high correlations in the stock returns among the countries in our sample. On the other hand,
weeks in which the correlation among returns is relatively low are less likely to be associated with the release of important
macroeconomic news. We therefore split our sample into weeks in which the average correlation between stock markets
is below the median, and weeks in which the average correlation between stock markets exceeds the median. We then
re-estimate the relation between sentiment and stock returns for each subsample with data availability. Panel A of Table 8
shows the results.
We nd that the relation between sentiment and returns is statistically signicant within both subsamples, indicating
that global macroeconomic news releases are unlikely to drive the observed relation between sentiment and stock returns.
In Panel B of Table 8 we examine subsamples based on the average correlation in sentiment levels around the world, again
distinguishing between weeks with above-median and weeks with below-median correlations. Again, the positive relation
between sentiment and stock returns is present in both subsamples.
17
We nd that our results are similar if we distinguish between the different geographical regions.
The Policy Uncertainty Index captures uncertainty in economic policy through a news-based measure that counts terms like uncertain and decit
in newspaper articles (see http://www.policyuncertainty.com/us daily.html).
18
G Model
10
ARTICLE IN PRESS
Table 8
Results based on subsamples.
N
Panel A: Correlation in returns among countries
Above-median correlations
Below-median correlations
Panel B: Correlation in sentiment among countries
Above-median correlations
Below-median correlations
Parameter
Standard error
10,801
10,482
0.032*
0.032**
0.015
0.010
10,570
10,713
0.044***
0.027*
0.012
0.013
1065
2309
7626
2223
1154
1159
6825
0.029
0.049*
0.046***
0.048**
0.024*
0.001
0.028***
0.023
0.025
0.015
0.020
0.014
0.013
0.008
13,488
6654
1154
1065
0.030***
0.041***
0.024*
0.029
0.009
0.012
0.014
0.023
Panel C: Language
Chinese
Dutch
English
German
Hindi
Italian
Spanish
Panel D: Religion
Catholic
Protestants
Hindu
Buddhist
This table explores subsamples. In Panel A we distinguish between subsamples with above-median and below-median weekly correlations in daily stock
returns among countries. In Panel B we distinguish between subsamples with above-median and below-median weekly correlations in daily sentiment
levels among countries. In Panel C we distinguish between subsamples based on language, and in Panel D we distinguish between subsamples based on
religion. In Panels A and B, we rst split into above- and below-median values for the complete dataset, and then estimate regressions for each group with
all available data. The parameter estimate represents the coefcient of regressing daily stock returns on our daily sentiment measure from Facebook. Our
sample period is September 2007March 2012. All regressions include day-of-the-week xed effects and country xed effects. We report standard errors
clustered by date.
*
**
***
5.3. Language and religion

We further exploit the availability of international data by examining cultural dimensions. Mihalcea et al. (2007) show
that the measurement of sentiment within non-English languages can be challenging due to differing attributions in the
meaning of terms. As Facebooks data team has to use dictionaries in several languages to distinguish positive from negative
terms, we explore the robustness of the observed relation between sentiment and stock returns for alternative languages.
Following Stulz and Williamson (2003), we classify languages based on the language of the majority of households within
a country.
Panel C of Table 8 shows the results. We nd that the positive relation between sentiment and returns is observable
for different languages. The results are not statistically signicant for the Italian and Chinese languages, but it should be
noted that for these languages the number of observations is relatively low. We do observe statistically signicant positive
relations for the languages Dutch, English, German, Hindi and Spanish.
We further examine subsamples based on religion, as prior studies have indicated that individuals adapt their linguistic
behavior due to their religious network (Baker and Bowie, 2010). We follow Stulz and Williamson (2003) in identifying the
primary religion within a country. Panel D of Table 8 shows that the positive relation between sentiment and returns is
statistically signicant for the Catholic, Protestant, and Hindu subsamples, which again highlights the broad relevance of
our results.
6. Empirical results on volume and volatility

In this section we examine whether FGNHI is related to trading volume and volatility. In line with evidence from psychology (Erber and Tesser, 1992), negative sentiment could cause investors to trade more in an attempt to overcome their
negative sentiment with a positive outcome from an alternative activity. Indeed, Chang et al. (2008) nd that cloudy weather
is related to higher transaction volumes. Sentiment could also affect investors propensity to trade. Brown (1999) nds that
unusual levels of sentiment are related to higher volatility in closed-end fund returns. Lee et al. (2002) use the Investors
G Model
ARTICLE IN PRESS
11
Table 9
Sentiment and trading volume.
Trading volume
N
Parameter estimate
Standard error
21,514
7488
8830
5196
5.097***
4.293***
5.783***
5.263***
0.742
0.793
0.865
1.244
Panel B: Trading volume on the next days

All countries
21,503
7484
America
8824
Europe
5195
Other countries
4.306***
4.732***
4.205***
3.787***
0.561
0.696
0.701
1.099
Panel C: Sundays sentiment and Mondays trading volume

4168
All countries
1394
America
1751
Europe
Other countries
1023
2.426**
2.447
3.451**
0.816
1.138
1.522
1.419
1.802

All countries
America
Europe
Other countries
Panel D: Lead-lag effects

Constant
FGNHIt
FGNHIt1
FGNHIt2
FGNHIt3
FGNHIt4
FGNHIt5
Volumet1
Volumet2
Volumet3
Volumet4
Volumet5
N
All countries
America
Europe
Other countries
0.031
(0.036)
5.043***
(1.339)
0.204
(0.709)
1.262**
(0.636)
0.220
(0.556)
0.188
(0.592)
2.052***
(0.629)
0.421***
(0.016)
0.143***
(0.013)
0.058***
(0.012)
0.064***
(0.012)
0.105***
(0.011)
18,609
0.079**
(0.033)
2.894**
(1.437)
1.434
(0.980)
0.961
(0.885)
0.061
(0.837)
0.728
(0.877)
1.594*
(0.864)
0.347***
(0.022)
0.147***
(0.021)
0.071***
(0.017)
0.061***
(0.015)
0.104***
(0.014)
5960
0.106**
(0.044)
6.882***
(1.728)
2.354**
(1.026)
0.763
(0.900)
0.046
(0.794)
0.607
(0.862)
2.383***
(0.833)
0.498***
(0.023)
0.126***
(0.021)
0.047**
(0.020)
0.062***
(0.021)
0.098***
(0.019)
8194
0.058
(0.039)
5.952***
(1.423)
0.002
(1.166)
1.809
(1.144)
1.059
(1.114)
1.641
(1.067)
1.308
(0.948)
0.415***
(0.023)
0.151***
(0.021)
0.050**
(0.019)
0.071***
(0.020)
0.101***
(0.019)
4455
This table shows whether sentiment is related to trading volume. The parameter estimate represents the coefcient of regressing daily standardized trading
volume on our daily sentiment measure from Facebook. Our sample period is September 2007March 2012. All regressions include day-of-the-week xed
effects, country xed effects, and ve day lagged returns. In each panel we estimate the regression four times: once for all countries, once for all countries
in America, once for all countries in Europe, and once for all countries outside of America and Europe. We report standard errors clustered by date. Panel
A examines contemporaneous relations, whereas Panel B examines the relation between sentiment on day t and trading volume on day t + 1. Panel C is
a sub-set of Panel B and only includes our sentiment measure on Sunday and trading volume on Monday. In Panel D, we estimate the following model:
trading volumeit = a1 + a2 FGNHIit +
5
j=1
bij FGNHIitj +
5
cij trading volumeitj + uit
j=1
Five lags are used for both sentiment (FGNHI) and the corresponding trading volume.
*
**
***
Intelligence sentiment index19 in the United States and report that bearish shifts in sentiment lead to upward revisions in
the volatility of returns. To our knowledge, ours is the rst study to examine the relation between sentiment, stock price
19
The Investors Intelligence index is based on classications of advisory services into bullish and bearish. The Investor Intelligence Sentiment Index
score is the number of investment advisory services that are bullish in relation to the total bullish and bearish advisory services.
G Model
ARTICLE IN PRESS
12
Table 10
Sentiment and volatility.
Volatility
N
All countries
America
Europe
Other countries
Panel B: Volatility on the next day
All countries
America
Europe
Other countries
Parameter estimate
Standard error
21,113
6705
9023
5385
0.108***
0.080***
0.108***
0.133***
0.011
0.014
0.014
0.018
21,102
6701
9017
5384
0.115***
0.083***
0.119***
0.136***
0.012
0.014
0.015
0.018
0.004
0.002
0.006
0.001
0.003
0.004
0.005
0.001
Panel C: Sundays sentiment and Mondays volatility

4259
All countries
1344
America
1826
Europe
Other countries
1089
Panel D: Lead-lag effects

Constant
FGNHIt
FGNHIt1
FGNHIt2
FGNHIt3
FGNHIt4
FGNHIt5
Volatilityt1
Volatilityt2
Volatilityt3
Volatilityt4
Volatilityt5
N
All countries
America
Europe
Other countries
0.003*
(0.002)
0.026***
(0.009)
0.022**
(0.011)
0.009
(0.009)
0.016
(0.011)
0.013
(0.010)
0.002
(0.009)
0.662***
(0.073)
0.009
(0.060)
0.078
(0.066)
0.108
(0.069)
0.311***
(0.056)
21,107
0.002
(0.001)
0.012
(0.010)
0.030***
(0.011)
0.000
(0.012)
0.006
(0.017)
0.002
(0.012)
0.001
(0.010)
0.606***
(0.078)
0.050
(0.071)
0.071
(0.068)
0.118
(0.072)
0.365***
(0.068)
6704
0.005*
(0.002)
0.030**
(0.015)
0.034**
(0.017)
0.022
(0.015)
0.036**
(0.017)
0.021
(0.014)
0.009
(0.014)
0.660***
(0.091)
0.005
(0.073)
0.092
(0.085)
0.065
(0.088)
0.298***
(0.068)
9022
0.004**
(0.002)
0.036**
(0.014)
0.000
(0.016)
0.002
(0.016)
0.018
(0.016)
0.016
(0.021)
0.019
(0.014)
0.726***
(0.078)
0.005
(0.068)
0.046
(0.071)
0.212**
(0.100)
0.292***
(0.087)
5381
This table shows whether sentiment is related to stock price volatility. The parameter estimate represents the coefcient of regressing the volatility as
measured by GARCH(1,1) on our daily sentiment measure from Facebook. Our sample period is September 2007March 2012. All regressions include dayof-the-week xed effects, country xed effects, and ve day lagged returns. In each panel we estimate the regression four times: once for all countries, once
for all countries in America, once for all countries in Europe, and once for all countries outside of America and Europe. We report standard errors clustered
by date. Panel A examines contemporaneous relations, whereas Panel B examines the relation between sentiment on day t and stock price volatility on day
t + 1. Panel C is a sub-set of Panel B and only includes our sentiment measure on Sunday and stock price volatility on Monday. In Panel D, we estimate the
following model: volatilityit = a1 + a2 FGNHIit +
5
j=1
bij FGNHIitj +
5
cij volatilityitj + uit
j=1
Five lags are used for both sentiment (FGNHI) and the corresponding volatility.
*
**
***
volatility, and trading volume in an international context. We hypothesize that negative sentiment leads to increased trading
volume and volatility.
We standardize trading volume by subtracting the mean trading volume in a country and dividing by the standard
deviation of trading volume in a country, and use GARCH(1,1) to measure daily volatility (e.g., Bollerslev, 1986). We control
for ve-day lagged returns in all estimations (coefcients are not reported due to space considerations). The empirical results
are shown in Tables 9 and 10. In line with our hypothesis, the results in Panel A of Table 9 indicate that sentiment and trading
G Model
ARTICLE IN PRESS
13
volume have a negative contemporaneous relation. The relation is highly statistically signicant and is observed in different
regions. Similarly, Panel A of Table 10 shows evidence of a negative relation between sentiment and stock price volatility.
Panels B, C, and D of Tables 9 and 10 examine the robustness of these ndings by examining lagged sentiment, the relation
between Sundays sentiment and Mondays market characteristics, and by controlling for multiple lead-lag effects. Overall,
both sentiment and trading volume and sentiment and stock price volatility are negatively related.20
7. Conclusion
We employ Facebooks daily sentiment index and examine its relation to stock returns, trading volume, and stock price
volatility across twenty international markets. Facebook is the worlds largest social network site, and their sentiment index
is based on textual analysis of the status updates of millions of participants.
We observe a positive relation between sentiment on Facebook and stock market returns. We further nd that sentiment
is negatively related to trading volume and volatility. The daily frequency of our data allows for some novel tests on the
causality of these relations. Most notably, we examine the relation between sentiment on Sunday, when stock markets are
closed, and stock market characteristics on Monday. Our ndings suggest that sentiment has a causal effect on stock market
characteristics in different geographical regions, highlighting the importance of behavioral nance for stock markets around
the world.
Acknowledgements
We would like to thank the Guest editors, three anonymous referees, Bruce Grundy, Guy Kaplanski, Meir Statman, Chris
Veld, Vadym Volosovych and seminar participants at the University of Glasgow for valuable suggestions, and Lisa Zhang for
her support during data collection.
Appendix A. Facebook coverage
Data are obtained from Socialbakers.com (last updated on 4th January 2013). The percentage of the online population in
a country that is on Facebook can exceed 100%, as more than one Facebook account is possible per internet connection.
Argentina
Australia
Austria
Belgium
Canada
Chile
Colombia
Germany
India
Ireland
Italy
Mexico
Netherlands
New Zealand
Singapore
South Africa
Spain
United Kingdom
United States
Venezuela
Percentage of population
Percentage of online population
50.00
55.28
36.07
47.45
54.85
57.77
39.69
30.98
5.34
49.05
38.33
35.77
45.52
54.00
62.39
13.19
37.83
53.17
54.37
36.31
142.08
70.27
48.50
61.44
66.55
125.62
103.84
37.56
68.19
72.63
71.16
114.22
50.27
63.34
81.22
104.78
58.03
62.87
73.44
95.63
References
Baker, M., Wurgler, J., 2006. Investor sentiment and the cross-section of stock returns. J. Finance 61, 16451680.
Baker, M., Wurgler, J., 2007. Investor sentiment in the stock market. J. Econ. Perspect. 21, 129151.
Baker, W., Bowie, D., 2010. Religious afliation as a correlate of linguistic behavior. Working paper.
Baker, M., Wurgler, J., Yuan, Y., 2012. Global, local and contagious investor sentiment. J. Financ. Econ. 104, 272287.
Baker, S., Bloom, N., Davis, S., 2013. Measuring economic policy uncertainty. Working paper.
Bollen, J., Mao, H., Zeng, X., 2011. Twitter mood predicts the stock market. J. Comput. Sci. 2, 18.
Bollerslev, T., 1986. Generalized autoregressive conditional heteroskedasticity. J. Econometrics 31, 307327.
20
We nd that our results are robust for including the Policy Uncertainty Index as an additional control variable. Tetlock (2007) shows that pessimism
in the Wall Street Journal predicts increases in trading volume, and that this relation is non-linear. In untabulated analyses, we explore any non-linearity
between sentiment and trading volume. We use the squared term of our sentiment value and nd some evidence for non-linear effects. Importantly, the
non-squared term remains consistently negative in these analyses.
G Model
14
ARTICLE IN PRESS
Brown, G., 1999. Volatility, sentiment and noise traders. Financ. Anal. J. (March/April), 8290.
Brown, G., Cliff, M., 2004. Investor sentiment and the near-term stock market. J. Empir. Finance 11, 127.
Brown, S., Goetzmann, W., Hiraki, T., Shiraishi, N., Watanabe, M., 2008. Investor sentiment in Japanese and U.S. daily mutual fund ows. Manager. Finance
34, 772785.
Chang, S., Chen, S., Chou, R., Lin, Y., 2008. Weather and intraday patterns in stock returns and trading activity. J. Bank. Finance 32, 17541766.
Da, Z., Engelberg, J., Gao, P., 2013. The sum of all fears: investor sentiment and asset prices. Working paper.
De Long, B., Shleifer, A., Summers, L., Waldmann, R., 1990. Noise trader risk in nancial markets. J. Pol. Econ. 98, 703738.
Edmans, A., Garcia, D., Norli, O., 2007. Sport sentiment and stock returns. J. Finance 4, 19671998.
Erber, R., Tesser, A., 1992. Task effort and the regulation of mood: the absorption hypothesis. J. Exp. Soc. Psychol. 28, 339359.
Heimer, R., Simon, D., 2012. Facebook nance: how social interaction propagates active investing. Working paper.
Hirshleifer, D., Shumway, T., 2003. Good day sunshine: stock returns and the weather. J. Finance 58, 10091032.
Kaplanski, G., Levy, H., Veld, C., Veld-Merkoulova, Y., 2013. Do happy people make optimistic investors? J. Financ. Quant. Anal. (forthcoming).
Karabulut, Y., 2013. Can Facebook predict stock market activity? Working paper.
Kramer, A., 2010. An Unobtrusive Behavioral Model of Gross National Happiness, Proceedings CHI. ACM Press, New York, pp. 287290.
Kramer, A., Chung, K., 2011. Dimensions of self-expression in Facebook status updates. In: Proceedings of the 5th international AAAI Conference on Weblogs
and Social Media.
Kumar, A., Lee, C., 2006. Retail investor sentiment and return comovements. J. Finance 61, 24512486.
Lee, C., Shleifer, A., Thaler, R., 1991. Investor sentiment and the closed-end fund puzzle. J. Finance 46, 75109.
Lee, W., Jiang, C., Indro, D., 2002. Stock market volatility, excess returns and the role of investor sentiment. J. Bank. Finance 26, 22772299.
Lemmon, M., Portniaguina, E., 2006. Consumer condence and asset prices: some empirical evidence. Rev. Financ. Stud. 19, 14991529.
Mihalcea, R., Banea, C., Wiebe, J., 2007. Learning multilingual subjective language via cross-lingual projections. Working paper.
Qiu, L., Welch, I., 2006. Investor sentiment measures. Working paper.
Saunders, E., 1993. Stock prices and Wall Street weather. Am. Econ. Rev. 83, 13371345.
Schmeling, M., 2009. Investor sentiment and stock returns: some international evidence. J. Empir. Finance 16, 394408.
Stulz, R., Williamson, R., 2003. Culture, openness, and nance. J. Financ. Econ. 70, 313349.
Tetlock, P., 2007. Giving content to investor sentiment: the role of media in the stock market. J. Finance 62, 11391168.
Wilson, R.E., Goslin, S.D., Graham, L.T., 2012. A review of Facebook research in the social sciences. Perspect. Psychol. Sci. 7, 203220.
Yang, S., Mo, S., Zhu, X., 2013. An empirical study of the nancial community network on Twitter. Working paper.
Zhang, X., Fuehres, H., Gloor, P., 2011. Predicting stock market indicators through Twitter: I hope it is not as bad as I fear. Procedia Soc. Behav. Sci. 26,
5562.

10.1016 J.jebo.2014.06.004 Facebook S Daily Sentiment and International Stock Markets

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10.1016 J.jebo.2014.06.004 Facebook S Daily Sentiment and International Stock Markets

Uploaded by

Copyright:

Available Formats

G Model

JEBO-3383; No. of Pages 14

Journal of Economic Behavior & Organization xxx (2014) xxxxxx

Contents lists available at ScienceDirect

Journal of Economic Behavior & Organization

Facebooks daily sentiment and international stock markets

JEBO-3383; No. of Pages 14

JEBO-3383; No. of Pages 14

JEBO-3383; No. of Pages 14

Panel B: Correlation coefcients

Panel B: Sample with potential outliers excluded

Panel A: Overall sample

JEBO-3383; No. of Pages 14

3.2. Cross-sectional stock returns

JEBO-3383; No. of Pages 14

Panel A: Stock market returns on the next day

Panel B: Sundays sentiment and Mondays stock market returns

4. Examining the causal relation between sentiment and stock returns

cij Market returnitj + uit

JEBO-3383; No. of Pages 14

cij Market returnitj + uit

JEBO-3383; No. of Pages 14

5.2. Macroeconomic adjustments

5.3. Language and religion

6. Empirical results on volume and volatility

JEBO-3383; No. of Pages 14

Panel B: Trading volume on the next days

Panel C: Sundays sentiment and Mondays trading volume

Panel A: Overall sample

Panel D: Lead-lag effects

cij trading volumeitj + uit

JEBO-3383; No. of Pages 14

Panel C: Sundays sentiment and Mondays volatility

Panel D: Lead-lag effects

cij volatilityitj + uit

JEBO-3383; No. of Pages 14

Percentage of online population

You might also like