Back to Portfolio for the Future™

Cause and Effect: Or, Shooting the Messenger

messageOne of the oldest sorts of “media criticism” is the argument that reporting on something bad causes that bad thing. If only there were fewer news reports about violence … or embezzlement …or Ponzi schemes… it wouldn’t occur to anyone to act violently, embezzle money, or pay out earlier investors from the cash received from later investors.

Put that baldly, such assertions seem absurd. It isn’t difficult, though, to dress such a notion up in sensible sounding, even scholarly, garb. In my own humble opinion, however dressed up, this notion that reporting X causes X is still and always fallacious.

Yes, I suppose, there may be times when observing X causes X. According to a venerable thought experiment still used to explain quantum mechanics to undergraduates, an observer who looks into a box under certain hypothetical circumstances will not only see a dead cat, he will by observing the deceased feline make himself a cause of its demise. Still, not even Schrodinger blamed the reporter. Saying out loud, “hey, this cat is dead,” doesn’t kill the cat.

Swiss Finance Institute

The old fallacy takes a new – and a quite nicely dressed-up—incarnation in the work of Rajna Gibson and colleagues associated with the Universities of Geneva and St. Galen. In a new publication [Swiss Finance Institute Paper Mo. 15-25] they suggest that “news suggestive of irrationality within financial markets” may “have an impact on stock returns.” Specifically, if media outlets use words that may be taken as synonyms for “market irrationality,” they increase the volatility of such returns.

I have to ask the obvious: if these cause and effect connections exist, doesn’t that mean that buyers and sellers are using the presence of such vocabulary in such reports as a signal for their marketplace actions? And isn’t that … well … irrational? We seem to have here, only thinly masked, the claim that talk about irrationality causes irrationality.

These scholars derived their “lexicon” of words indicating market irrationality from the Harvard IV 4 Psychosocial Dictionary and General Inquirer, and they brought “three experts from the fields of neuroscience and psychology” into the project to “validate which words would comprise the final lexicon.”

Which words do indicate that the press is pushing a market-irrationality thesis? I won’t reproduce here the full list, but twenty such words correspond to the first two letters of the alphabet alone: Aberrant; absurd; absurdity; accursed; alarming; anarchic; anxious; baffle; baffled; baffling; barbarous; belligerent; berserk; bizarre; bonkers; brainless; bubble; bubbles; and burst.

And on the third day…

Yet there is a twist here. The last three words in the just-provided sample from that full list (bubble, bubbles, and burst) were treated as indicative of irrationality by these scholars for some purposes but not for others. They were not counted when considering which news articles belong in the “market irrationality sentiment measure” but they were used when scoring the articles that were included.

Setting aside their methodological subtleties, the authors’ bottom line is this: “a one standard deviation increase of the market irrationality sentiment measure is associated with a 0.060 increase in the VIX volatility index (significant at the 5% level) on the third day.” The third day? Yes: one of the incidental claims of the study is that the pro-irrationality sentiment showing up in the press on a given day “takes more time to impact prices than other news-based sentiment measures with the most significant impact taking place 3 days after its publication.”

With respect: that claim is just bizarre (oops, used a word on the list!). The need to fiddle around with the timing of the alleged cause-effect nexus is a very good indicator that there is no cause-effect nexus, and that something else is going on here. It is a bit like the famed nexus between the rise of the number of Baptist preachers working in a given town and the amount of public drunkenness.

There may well be such a connection for a number of reasons. But it would be unwise to conclude either that Baptist sermons are driving congregants to the saloon or that drunks are induced to look for salvation in Baptist services.