Tuesday, September 25, 2007

When The Words Of Financial Writers Are Indistinguishable From Those of Psychiatric Patients

Over the last week or so--really ever since the Fed decision--I've had a vague sense of dis-ease with respect to many of the online financial columns and blog posts I've been reading. Much of them have seemed oddly bearish given Asian markets that are moving to new highs and S&P 500 and NASDAQ 100 indexes that are not far from bull peaks. But that hasn't been the whole story. Something else has made me uncomfortable with what I've been reading.

When I reflected on how I was reacting emotionally, I suddenly figured it out. Much of what I'm reading is identical to the speech of psychiatric patients. I don't mean this hyperbolically at all. I mean it in a very literal sense. If these writers spoke the words to me that they are writing, I would conclude that they're both emotionally conflicted and constricted.

Allow me to illustrate what I mean:

A man comes into my office, having just lost tens of thousands of dollars in trading. From his facial features, he is clearly upset; it's money his family can't afford to lose. He says to me:

"The market moved away from its average and that made the sellers jump in. It was a large loss and now it's just a matter of looking to the future. Things like this happen. You just have to figure it out. The economy can't be all that bad. Stocks have to come back. Just look at interest rates. That has to give things a boost."

What would I conclude from this speech sample? I would see that our patient is upset, but I would also note that he doesn't own any of it. Indeed, he barely speaks about himself. He doesn't talk about what happened in any detail, and he certainly avoids any kind of talk about his feelings.

In other words, he's conflicted (upset about his loss) and he's constricted (not able to talk about his experience). That's a common combination among psychiatric patients. The role of the shrink is to see behind the words to the experience underneath. That's where the real information lies.

Suppose we treat the writings of financial pundits the way we would treat the language samples of patients in therapy. Fortunately, there is software available to accomplish this for us, developed by James Pennebaker and his research team at the University of Texas. This software, called Lingustic Inquiry and Word Count (LIWC), will parse any text into various categories and components.

So what I did was take 7 financial columns from very respected sources, all with distinctly bearish themes. I then used the LIWC program to analyze the 15 pages of single-spaced text from these columns.

According to LIWC, exactly .55% of the text overall consisted of the word "I". The writers barely acknowledged themselves in their writings. They were offering personal views on the markets, but not presenting those as such.

Despite the profoundly bearish themes of the articles and the evident concern of the writers, only about 4% of the content consisted of emotion-related words. Oddly, those were evenly divided between positive emotion and negative emotion terms. In other words, the actual text of what was said did not match the emotional tone of the underlying message.

So what we have is a group of writers offering their personal perspectives and expressing their concerns about the Fed decision, inflation, the weak dollar, etc. Their texts, however, are written in such a way as to present the material as factual and neutral: not coming from a personal source and not connected to any emotional response.

It's just like the hypothetical patient above: worried people who talk in impersonal, non-emotional ways.

In a patient, this is known as "defense"--avoiding one's true sources of concern. In the world of online journalism, it passes for objective reporting.

Don't get me wrong: I don't think online columns and blog posts are worthless. On the contrary: linguistic distortions mask real, underlying information. Any good shrink knows that.


The Psychology of Taking Losses


John Forman said...

Hi Brett,

Very interesting post. I've been giving a lot of thought to the emotional content of market commentators over the recent weeks.

The question I have is how the statistics on LIWC from the recent articles you pulled compared to figures from other, less intense market periods.

I ask because it occurs to me that these folks might be like that all the time due to the nature of what they are attempting to accomplish. Could they just simply be folks with views attempting to present them in a non-personal way as a professional mode of communication? Certainly, anyone with a personal bias is likely to express that somehow in her/his writing, even when trying to be balanced. I'm just wondering if the figures back up the idea that it's happening more now than at other times.


Mat said...


Great commentary as usually. You stated "linguistic distortions mask real, underlying information. Any good shrink knows that." Now I could probably make a guess as to what "information" is behind the commentary you examined, but what are your thoughts on the underlying what is really being conveyed.

Second, what implications- if any- could an investor/trader expect when we see this mental change occurring.


The Financial Philosopher said...

I've noticed the same behavior but lack the education to express it scientifically. I've observed that the current "financial crisis" has been the topic of many financial blogs for months prior to its "arrival" in August when those bloggers seemed to take ownership in forecasting its certain occurance. Now, the crisis has arrived, mainstream media is covering it, and most bloggers stop short of their "forecasting" behavior as to the ultimate impact of the crisis.

I find it especially interesting that most bloggers are still giving so much energy to something that the mainstream media is simultaneously reporting, given the fact that they (we) typically behave as the "anti-media."

I would draw the analogy of a car "wreck" that is easily foreseen by the cars behind it. Once the wreck occurs, only the fools continue to look at it after they have already passed but will tend to look at it more intently and for longer periods of time, especially if the wreck is gruesome and even more so if the media is there...

Stop the "rubber-necking." It's time to move forward...

bzak said...

Have you compared the word count from bearish articles to bullish ones? It's hard to make any conclusions without putting the findings in context.

bruce said...

Don't really get it. The notion that a financial writer should be saying "my market went down", "this is awful for me", "my indicators are killing me", etc, seems rather odd to me.

bryan said...

Really enjoyed this post, Doc. I thought the first part that drew on your practical psychoanalytic skills was fascinating. Thank you.
Bryan Wendon

Brett Steenbarger, Ph.D. said...

Hi Everyone,

Thanks for the excellent comments on this provocative post. I agree that an ongoing investigation of how language is used across posts and websites would be most interesting.

My basic point is relevant to my own postings, not just those of others. Most of what we read online is not objective analysis. Rather, we're reading the writer's *interpretations* of news and markets.

That means that everything, my own posts included, must be read critically.

Online financial blogging is not financial journalism. It's the difference between and editorial and reading a news report of an event. When a commentator appears on TV (Ann Coulter, Al Franken, Rush Limbaugh, Andy Rooney, etc), we know that we're getting opinion. It's stated emotionally, and it's stated in the first person.

It's when editorial content is presented as news or objective analysis that I find my antennae rising. It doesn't matter if the content is bullish or bearish: if it's expressing the sentiment of the writer, but masquerading as impersonal, objective truth, the real information lies in what is not being acknowledged: the emotional state of the writer.


Bill aka NO DooDahs! said...

Care to list the bloggers and posts that were analyzed?

Brett Steenbarger, Ph.D. said...

Hi Bill,

No, as I note in my comment to the post, my aim is to help readers be more critical of what they read. I'm not aiming to criticize any particular writer or blog.