Study Suggests Financial Bloggers May Move Markets

A vote of confidence in financial bloggers – at least those featured on Seeking Alpha - emerges in a paper looking at the effect of their stock predictions.

According to the paper* by Veljko Fotak, a Doctoral candidate at the University of Oklahoma’s Price College of Business, blog recommendations appear to have an impact on stock prices and trading volumes.

The average market reaction is of an increase in stock prices of around 0.4% for long recommendations and a decrease of 0.8% for short recommendations.

The analysis was based on 500 buy and sell recommendations for the year 2006 from SeekingAlpha; of those, 340 are classified as long recommendations and 160 as short. The recommendations originated from 107 different blogs and 122 distinct bloggers.

Fotak found abnormal trading volumes after recommendations, but also that trading volumes appear abnormal in the days preceding the blog posting, which does leave open the possibility that abnormal trading volumes following blog publications are a result of trading momentum. “The stronger impact of short recommendations offers additional evidence in the still open debate as to whether short recommendations do lead to stronger market reactions,” Fotak writes.

Unlike studies of the effects of e-mail spam, Fotak found “no evidence of short-term mean reversion in the days following the event. “This difference in market reaction can be interpreted as an indication of the fact that, while e-mail spam is, often times, an attempt to manipulate markets, blogs are not. E-mail spammers spread false information in order to move the price of a security; eventually, the price reverts to its fundamental value. In the case of blogs, it is plausible to hypothesize that posters are acting out of a genuine desire to spread real information. The lack of mean reversion in prices following publication supports this hypothesis.”

Bloggers do appear, for the most part, to echo information already available in the media, Fotak writes. “Yet, as market reactions to select blogs suggests, some do spread genuine information and analysis, which does lead to a market adjustment.”

*The Impact of Blog Recommendations on Security Prices and Trading Volumes (Veljko Fotak - Michael F. Price College of Business, University of Oklahoma)

Comments

[SA EditorEli Hoffmann]

From what I can tell (I did download the original paper), the author made no effort to separate pure long and short ideas that had no connection to ongoing news from "ideas" that may stem from an analysis of recent company news.

Which leads me to question the validity of this paragraph: "Bloggers do appear, for the most part, to echo information already available in the media, Fotak writes. “Yet, as market reactions to select blogs suggests, some do spread genuine information and analysis, which does lead to a market adjustment.”"

How can he gauge whether the reaction was to the "information already available" or to the ideas of "select blogs"?

I suggest repeating the study, but using only long/short ideas on issues that are not being influenced by actual news (using, say, a Google Finance ticker search for XXX day(s) before/after the idea was published).

[Veljko]

Dear Mr. Hoffmann and Seeking Alpha community,
Thank you very much for your kind feedback.
I try not to comment on my own research, as I think it is for the reader to judge its merits, or lack of.
But I feel compelled to make an exception in this case. I have to apologize in case I failed to make this clear, but the sample used in gauging market reaction was controlled for contemporaneous news. As stated in the “Data Sources” section, I use two news databases, Factiva and Lexis-Nexis, to control for all firm-specific news items appearing on either Dow-Jones Newswires or on The Wall Street Journal on the interval between day -1 (the day preceding blog publication) and day +1 (the day following blog publication). In addition, I excluded all instances in which other blog posts were published during the same time interval. Also, I exclude all instances in which the firm posted earning reports (although those are, in the great majority of cases, reported by Dow-Jones Newswires). As stated and as indicated in the tables, these restrictions reduce my final sample to 208 observations from the original 500. Results are reported for both subsets. The evidence I offer in support of market reactions is based on the reduced sample and so are the quoted results.
The first statement you quote "Bloggers do appear, for the most part, to echo information already available in the media:” is based on observing that only 208 observations report novel information.
While I do appreciate the interest that the blogging community has so far shown, I wish to make clear that this is a preliminary study, which I consider a starting point for a larger project. In particular, I am currently working on dramatically expanding my sample size by using web-crawling algorithms to automate data collection. In addition, the revised version of the paper will contain a discussion on the relationship between market reaction and past performance of bloggers; I hope to be able to establish that markets react more to bloggers who have a better “track record”, defined by the accuracy of their past predictions (I define an accurate recommendation as one which correctly anticipated the sign of the subsequent long-term abnormal return; for example, a long recommendation followed by a subsequent price increase).
Once more, please allow me to thank you for taking the time to review my research and to offer your feedback. Thanks also to Seeking Alpha for the wonderful job done in tracking the evolution of the financial blogging community.

A quick note: the opinions here expressed are mine and do not necessarily reflect those of Price College or of the University of Oklahoma .

Regards,
Veljko Fotak
[e-mail]

[SA EditorEli Hoffmann]

Veljko,

Thank you for your detailed response. I do, indeed stand corrected. It is an honor for us to host academic research on our site.

The question that remains, in my mind, is this: How can we determine whether it was bloggers who "moved" the market by means of their blog posts (and, as your upcoming research may or may not prove, that "better" bloggers move markets more), or whether bloggers, when suggesting long/short ideas, are using the information available to them to make better-than-50/50 investment decisions, the result of which is that long ideas tend to demonstrate positive returns, and short ideas the opposite (and better bloggers put forward better ideas, and the market rewards them accordingly).

In other words, would the results of your research differ significantly if bloggers were posting in a vacuum?

[kkin365]

I think a blogger, when quoting some other sources, should -- must -- disclose his/her source. When done, a reader can easily establish whether a blogger is merely relying some info or is, indeed, creating some info for others to quote.

[Veljko]

Mr. Hoffmann,
The underlying assumption, pinned on market efficiency, is that stock price changes are a reaction to new information. If no new information is offered by the traditional media (that is, no news are published, as I am controlling for those) then the new information that moves the market price must come from the blog itself. In other words, the underlying assumption is that, in the absence of news, bloggers are indeed “posting in a vacuum”.
You are absolutely correct at pointing this out as the difficult “selling point” of the paper. I am conscious of the fact that there is a risk of not capturing all new information. As I mentioned, I am currently revising the study; one of the aspects on which I hope to improve is my control for external news, to make it absolutely certain that no other events are motivating the price change. The emphasis on this issue originates from feedback such as yours and, for that, I am extremely grateful.
Another aspect I am focusing on is the use of more accurate measures of blog readership. My reasoning is, if the market reaction is somehow linked to the size of the audience, then we would have additional support for the hypothesis that the blog itself is moving the market.
Once more, thanks a lot for your feedback. I will make sure to send you the updated version of the paper as soon as it becomes available.
Regards,
Veljko
[e-mail]

[TERN]

Investors, especially retail investors, may be getting more information from their PC screens than from the old print media.
They check their portfolio on-line, read the associated articles and blogs and..form the basis of their decisions.
I very much appreciate Veljko's study and believe the study needs to be complemented by research into the changing reading habits of investors and small traders.

[Donald E.L. Johnson]

The question is what is new information and what are new opinions?

New info may be a scoop on a new product, service or CEO. We seldom see that in blogs. New info also may be a new way of looking at available information, putting it in perspective and connecting the dots. And new info may be the blog itself and the opinions expressed in that blog. If a blogger digs into a conference call and finds important information that is being over looked by most speculators, that is new information for those traders. And if a blogger snags an exclusive interview with an executive of a company or some of its customers, the blogger makes news, adding new information.

So how new information is defined and classified will influence the outcome of the study, I suspect.

As to the impact of a blogger's track record, I'm wondering how many readers pay attention to bylines and recall bloggers' previous posts. In other words how important is a blogger's reputation. There aren't many Jim Cramers blogging. And trying to figure out Cramer's track record is hard enough, much less the record of an individual blogger.