PsyFi Search


Saturday, 16 October 2010

The Language of Lucre

Talk the Talk

Once we get past blood-lust, misogyny and an insatiable desire for saturated fat presented in peculiar forms, language is probably the defining feature of humanity. Thus it’s unsurprising that some psychologists think that human behavior is defined by language, rather than through some cognitive super-controller that oversees our every action.

Discursive psychology, the study of how language impacts human behaviour, would expect to see investors swayed by use of language in a way that more rationalist theories would not. Language deployed by the financial media might, for instance, be employed to promote buying or selling of stocks regardless of their fundamental attractions or otherwise. Which, of course, is a stupid idea and therefore, inevitably, looks like it’s probably correct.

Recursive Discourse

Discourse analysis is as much about what we mean by what we say than what we actually say. Rather than trying to explain this let’s take an example:
“The main headline reads: "Baby Peter's father tries to cash in with demand for £200,000 compensation". Let's try to deconstruct this headline using "shared patterns of talking" - this means, simply, let's use our knowledge of how English is used to unpack the meaning of the text.

Firstly, let's begin with "Baby Peter". The newspaper writers and editors are relying on the fact that this is a well-known case, which is often just referred to in the media as "baby Peter". However, the text refers to the father of baby Peter. Next, the term "tries to cash in" is used. In common language this refers to the act of trying to take advantage of a situation (either for money or in a metaphoric sense, as in trying to take praise for someone else's endeavours). However, the next few words "with demand for £200,000" lets us know that this is not a metaphor but a literal claim for money and in an opportunistic way.

The use of the term "cash in" implies that the writer is unsympathetic with the father and there is the hint that the writer believes that the father is an opportunist rather than a grieving parent. The sentence ends with the word "compensation". This might seem to contradict the notion of "cashing-in" as compensation is associated with genuine loss. However, the word is most likely being used to refer to the legal process required to receive such a payment. Finally, the word "demand" is quite a forceful choice of word. To "demand compensation" implies that the father believes he has this right but that he may recieve some resistance."
Got that?

Speaking With Forked Tongues

The point, of course, is that whenever we use language we’re using it for a purpose – there’s no such thing as a neutral statement. Not even a statement about being neutral. This approach can widened from newspaper headlines to any kind of text and into verbal discourse where the hesitations and slips and what’s not said is regarded as just as important as what is said. This stands in contrast to traditional language analysis which is generally all about syntax and semantics.

Now frankly it’s almost impossible to read anything written by any researcher in this area without getting a headache, so impenetrable is the text as the writers apparently analyse the meaning of their own words as they go along. This book about discourse analysis, for instance, starts off by analysing the meaning of the phrase "discourse analysis".

However, pretty much everyone agrees that discursive psychology is not a scientific approach to analysing human behaviour, because it doesn’t offer the repeatability of traditional laboratory based experimentation. What they don’t agree about – often quite violently – is whether this means it has any validity as a method of analysis for social scientists. Regardless, it’s likely that a discourse analysis of economic behaviour would yield some interesting results. Unfortunately I doubt any of us would be able to understand it.

Financial Media Discourse

Still, there is some relevant research around the impact of newspaper articles on investor behaviour. In particular Joseph Engleberg and Christopher Parsons have managed to extract data from historical records that show that business reporting has a definite impact on stock buying and selling patterns dependent on two things – the reporting has to be relevant to the investors and negative reporting has a stronger effect than positive. This effect appears to be independent of the underlying fundamentals of the stocks in question.

This kind of experiment, which relies on data derived from natural situations outside of the laboratory, is much less likely to suffer from bizarre behaviours caused by the experimental situation, but typically produces results which are far more difficult to interpret. The beauty of The Causal Impact of Media in Financial Markets is that it seems to overcome this problem quite neatly, but leaves us with a hanging question about how to use the information in our today's world of rapidly fragmenting media.

Reporting Chickens and Financial Eggs

First, though, let’s look at the hypothesis which was that media stories about quoted businesses influence investor behaviour to an appreciable degree. Although we might think, as so often in these experiments, that this is a statement of the bleeding obvious that’s hardly worthy of wasting research dollars, it does rather hide a quite complex issue. News stories make it to the media because the news companies reckon that the stories will interest their readers. There’s no guarantee, therefore, that it’s the event which the story is reporting which drives investor behaviour – it could easily be caused just by the publication of the story.

So, if a company reports a major new contract you might well reckon that it’s the fact that it’s won the business rather than the decision of a local newspaper editor that this event is worth reporting that will cause a surge in investors’ buying the stock. However, as the researchers show quite neatly, you’d be wrong: in general most investors were happily oblivious to the wondrously good and thunderously bad stuff happening. All that mattered was the reporting.

Elegant Natural Research

The study covers local newspaper reporting of S&P500 earnings announcements for local businesses in the US during 1991 to 1996. It showed that:
“All else equal, local press coverage increases the daily trading volume of local retail investors, from 10% to nearly 50% depending on the specification”.
The beauty of the study is that the researchers were able to separate the underlying event and the actual reporting. The elegance of this is delightful, as they used naturally occurring interruptions to generate their data. Sometimes this was simply due to publication deadlines, sometimes to major weather events: it snows lots in Minnesota in the winter time and paper deliveries get delayed. In fact severe weather caused the relationship between local reporting of news about local businesses and trading to break down.


Now the familiarity effect, where people have a strong preference to trading in stocks in their locality, is well established – see Puke: Don’t Invest in the Familiar. People prefer to trade what they’re familiar with, which often leads to portfolios which are wildly skewed in terms of diversification, particularly in respect of favouring stocks from the investor’s own backyard. The research here suggests a spin-off effect in terms of news: newspapers will publish stories about local companies because their local readers are investors and their readers will react to this news.

Now if this news directs investors to the relevant earnings announcement and causes them to re-analyse the fundamentals then investor behaviour is fairly rational. Although the researchers are cautious they suggest that this isn’t the case:
“We find extreme earnings surprises are related to the volume of retail trade. However, the media effect we identify is several times larger than this information effect no matter how we define our earnings surprise. Simply put, in our setting, the media is much more likely to drive trade than information”.
Talked into Trading

So basically, it’s not that rational controller sitting in our brains that’s deciding how and when we trade, it’s all about the language. Discourse analysts wouldn’t be surprised and neither should we.

How this result generalises to national and international media and to the vast proliferation of news outlets engendered by the internet is a topic of future research. However, as an discourse analyst would tell you, there’s no such thing as unbiased reporting: study the numbers, not just the words.

Related Articles: The Media, Fear and Stockmarket Manias, Fairy Tales for Investors, Puke: Don’t Invest in the Familiar

No comments:

Post a Comment