Crowdsourced investor sentiment – News & Commentary
US Companies with good social media sentiment have seen their shares outperform the rest of the market by 56% in the 23 months leading up to last November
The inverse holds true for companies which are subject to negative sentiment
As with anything in the social media world, the universe of shares subject to positive and negative sentiment varies rapidly
In highlighting which US stocks rank highly by applying the new social media factor, we build on the Markit research paper: Extracting Market Sentiment From 140 Character which is available here.
Methodology
This analysis of the social media driven market sentiment mines the 500 million plus Tweets per day made by the 241 million active Twitter users over the last two years and which were filtered for investment relevance by SMA.
The filtered tweets are then analysed using a sentiment dictionary to gauge the investor mood. Just to give an indication of how mainstream Twitter has become within the investment community, 1200 of the shares within the 3000+ shares in the Markit US Total Cap universe are covered within the MSA database of financially significant Tweets.
Markit’s new social media factor suite spans 22 metrics which gauge such things as Tweet volume and dispersion within the Twitter community. Perhaps most significant within this suite of factors is the composite S-Score™ which gauges both relative and absolute sentiment of a stock within the twitter community.
Positive score outperform
Using the S-Score™ to create long and short portfolios which select only the shares with the best (worst) twitter sentiment, we see that shares with positive twitter sentiment have consistently outperformed over the last two years.
To give an idea of the how relevant this sentiment is, the compound average returns for shares in the buy portfolio in the 24 months leading up to November of last year is 76%, over 50% more than the returns seen in the S&P 500.
To give an idea, shares within the positive sentiment universe on January 16th are Sprint, Morgan Stanley and Callaway.
Negative sentiment underperforms
On the other side of the coin, shares which exhibit negative sentiment as per a low S-Score™ have fell over the same observation period with an average daily fall of 3 basis points, 14 basis points lower than the positive universe. Note that this sentiment degrades over time, unlike shares in the positive sentiment universe which continue to outperform the market over the following five trading days.
To give an idea of the how relevant this sentiment is, the compound average returns for shares in the buy portfolio in the 24 months leading up to November of last year is 76%, over 50% more than the returns seen in the S&P 500.
To give an idea, shares within the positive sentiment universe on January 16th are Sprint, Morgan Stanley and Callaway.
Negative sentiment underperforms
On the other side of the coin, shares which exhibit negative sentiment as per a low S-Score™ have fell over the same observation period with an average daily fall of 3 basis points, 14 basis points lower than the positive universe. Note that this sentiment degrades over time, unlike shares in the positive sentiment universe which continue to outperform the market over the following five trading days.
Shares with a low S-Score™ in January include Metro Bancorp, Revlon and Sunpower.
Beware of turnover
While this cursory paraphrase good work undertaken by Markit fails to give justice the nuances of the SMA factor suite, it does outline the fact that social media based investing has explanatory powers over stock returns.
A realistic investment strategy has to account for the high turnover nature of the factor suite in the vastness that is the Twittershpere as the shares experiencing the worst and best investor sentiment turn over by over 90% over any given day.
Read more –
Crowdsourced investor sentiment – News & Commentary
