Seeking Alpha: Interesting Sentiment Indicator

We were early supporters of Seeking Alpha. We have had a constructive relationship with the site and their leaders.

We congratulate them on their success, and their growing readership, becoming a top-ranked site in the ratings.

The Seeking Alpha Difference

There are some interesting metrics that one can derive from Seeking Alpha. These stem from their approach, embracing the blogosphere and highlighting the news flow.

This is dramatically different from mainstream media sources, where writers choose their subjects and editors review the overall balance.

Please understand that this is not a criticism of the Seeking Alpha model. We merely point out the difference. The articles appearing there reflect blogger opinion and news flow, and it is done very accurately.

Why This is Helpful

The objective stance of Seeking Alpha helps us to gauge what people are thinking. There are two excellent metrics.

Most Popular Articles

The ranking of most popular articles shows what people are actually reading. Here is the current list as of this writing (skipping the excellent daily briefings):

The Worst Case Scenario (Someone Has to Say It)

A Bull Market That Few Are Buying

Dollar's Purchasing Power Annihilated - The Chart They Don't Want You to See

Book Review: Great Depression Ahead

A Summary of Q1 Bank Earnings: World, You Just Got Hustled

Most Recent News

The site picks up news flow, as it happens. Here is the current list as of this writing:

A Summary of Q1 Bank Earnings: World, You Just Got Hustled

This sure smells like a sucker's rally,

Even GM (GM) can't figure out why anyone's still buying its shares,

will oil prices throttle the economy again?

Our Take

The overwhelming impact of these stories is self-evident.

Once again, we repeat that these stories are not a discretionary choice of the Seeking Alpha editors, who do a fine job. It is an accurate reflection of reader sentiment and news flow. That is what makes it so interesting.

While we have picked the leading articles as of yesterday, it is quite typical of the pattern throughout the recent rally. It shows bloggers and readers, all fighting the rally. It is a strong example of the wall of worry that we described a month ago.

If we looked to the most popular authors and commenters, we would see a similar pattern. It is an accurate reflection of the skepticism of those active in online writing and commentary. This might be a better sentiment indicator than Investors Intelligence or other similar choices.

Many people are wondering whether it is too late to buy stocks. This is one indicator.

We think, like traditional sentiment indicators, it is a contrarian signal. If and when the top stories, authors, and articles are all bullish -- well, that would be the time to think about taking profits from the recent rally.

Comments

[fatcat:]

That contrarian think will get you in trouble....This time is truly different...if you can show me where profits are coming in this time of high unemployment and rampant manipulation,bank lies,hyper creation of money,etc,etc...then I will become bulllish.

Until then,get long at your own peril.

May 13 04:54 AM

[David Jackson:]

V. thought provoking article, Jeff. (And thank you for your kind comments about the work we've put in to grow the site.)

We've recently been discussing the embedded value in contributors' articles, and how we could share that information in a more thorough way.

The first issue is: What is the value of SA contributors' sentiment? There are two possibilities:
(1) Like other sentiment indicators such as Barron's money managers' poll, it's a contrarian indicator (if most writers are bearish, there's still money sitting on the side lines).

(2) Our contributor base represents a small minority of money managers with greater ability to predict correctly -- in which case the sentiment is an accurate (not inverse) indicator.

Jeanne Klempner, one of our editors, makes a good point: once we've extracted the sentiment data, just test it for correlation against the market, and we'll find out whether it's valuable and in what direction.

The second issue is assessing the sentiment of an article. Some articles are explicitly bullish or bearish, but many articles aren't: they discuss individual stocks or some aspect of a company or the market, without coming to a "bullish" or "bearish" conclusion. We need to be careful about assigning sentiment to contributors inaccurately.

I'd be interested to hear more thoughts on this issue.

May 13 06:27 AM

[SA Editor E...:]

Interesting study that suggests fading bloggers may be dangerous to one's financial health:

seekingalpha.com/artic...

May 13 07:22 AM

[John Lounsbury:]

Jeff - - -

Excellent article. A factor that I have been thinking about is whether or not the Seeking Alpha community really reflects all of the broader investment community. I have worried that we may be self reinforcing opinions through "group think". I have noticed that authors expressing more negative outlooks seem to have more representatives near the top of the Top 100 list than do the optimistic. Is that because the negative makes better reading or because a majority of the readers are not optimistic?

I don't read a large enough sampling of SA articles to claim the following observation is statistically significnat, but I'll make it anyway. My own reading of articles with optimistic sounding titles has left me disappointed by the lack of data, the deficiency in analysis and the shortfalls in critical thinking in some of them. I have left comments on such articles but the authors have not engaged in discussion and have not made changes in future articles that would indicate my criticism was considered.

I think the willingness of authors to react to what commenters have to say may have an influence on their popularity, so the popularity of authors with negative themes may also be influenced with how they have interacted with their audience.

All of these thoughts have left me in a quandry about whether or not SA can be used as a sentiment indicator. And once we get by this quandry, if it is an indicator, is it contrary?

Jeff, you may have described a monster research project.

May 13 10:34 AM

[nym:]

It takes me a long time just to read all the titles of the articles each day to decide which I will look at. I would like the editors to be more selective, along the lines that Lounsbury points out above.

I am impressed how much Lounsbury has grown during the last 6 months. An open mind in fun to read.

May 13 12:01 PM

[PainfullyAware:]

As With All Worthwhile Pursuits - Diligence Is Required.

Pessimists tend to give better data - but one can come to any conclusion if the sampling is limited to only what supports the chosen paradigm. However - Optimists tend to "Grasp At Straws" when faced with contrarian indicators and rational tends to morph into rationalization.

Safety Is A Function Of Awareness.

I prefer the "Wild West" of the blogosphere rather than the "Programming" that the other media sources have become. Even if the View of the author is Sunshine Or Doom there is more than likely "Data and Event" to be mined from their prose.

Even Cetin has a "Good Point" once in a while. (I do wonder if he has the highest negative rating? Does he believe his "Reality" or is he an agent provocateur working with the PPT? - Take it as tongue in cheek; all things are filtered through others.)

In The Words Of Black Sabbath -- When You Listen To Fools - The Mob Rules. Choose and Double Check Your Sources.

May 13 01:24 PM

[TinyTim:]

This group is definitly not a normal distribution.
Run a critical article with the word gold in the title & see what happens.

May 13 04:10 PM