Art Cashin's Refresher On "Post Hoc" Syndrome

Tyler Durden's picture

Nassim Taleb rants against it all the time: the propensity for the media to frame a narrative, or a plotline, to explain market moves. His contention is that for the human mind it is always far more reasonable to have a cause and effect relationship to what is effectively an engine of chaos at the margin, especially these days when the margin is defined 70% by various algorithms, all of which engage in often times illogical feedback loops (such as the ES is high because of a high EURUSD, which however is high due to stressed French banks liquidating USD-assets and repatriating the funds to shore capital) and/or with levered synthetic products such as ETFs, amplifying the noise. On the other hand, sometimes a narrative fits: what Art Cashin describes today as the "post hoc" syndrome. Is he right, or is the human mind desperately grasping to attribute a pattern, and thus pretend it is in control, when faced with the strange attractor that modern capital markets have become. You decide. Here is Art explaining the basics of "post hoc", aka Monday Morning quarterbacking.

From UBS' Art Cashin:

Tear Gas, Timing And Logical Fallacies - On the Friday before Thanksgiving, I maintained that the afternoon selloff that led to a technical breakdown the day before had accelerated when headlines about unrest in Athens turned uglier. A few readers questioned what they saw as a time gap between the headlines and when the market reacted. That merits a review for a couple of reasons. The most important of these is probably the “post hoc” syndrome.

As you probably recall from your fifth grade classes in epistemology and logic, there are about seven or eight logical fallacies. The three most frequently cited are usually - begging the question; hasty generalization and post hoc. The full title is post hoc, ergo propter hoc. That, as you recall, is “after this, therefore because of this”. Since B happened after A, it was probably caused by A.

The post hoc fallacy is quite common in Wall Street. The main stream media often credits an up market to some piece of economic data which came out hours and hours before the rally that produced the “up day”. You’ll often see headlines like “market rallies on claims data”. The problem is that the claims data hit at 8:30 and the market didn’t even get into plus territory until, maybe, 2:45 in the afternoon. That’s classic “post hoc”.

While data like claims and payrolls tend to be finite triggers, some information can be more of a process, where the impact may be cumulative. That was the case, for example, in the Kennedy assassination.

The first headline was something like: “Shots reported fired at President’s motorcade”. (That’s bad.) Then, minutes, later “Reports that President may have been hit”. (Worse but no detail). Then, more minutes, “Motorcade diverted to Parkland Hospital”. (Even more serious.) That’s when they closed the Exchange. The series of headlines had a cumulative impact.

The sharp selloff on the Thursday, a week before Thanksgiving was a reaction to cumulative headlines. In Friday’s Comments we wrote:

Around 12:30, just as U.S. markets were retesting the morning lows for a second time, things changed.  Headlines hit that anti-austerity demonstrations in Athens had turned ugly - maybe very ugly. Clashes with police were said to be intense. That brought more selling in stocks breaking the morning lows and almostinstantly the key support at the bottom boundary of that universally discussed “triangle”.

That phrasing was intended to indicate that cumulative impact. We were also quite reinforced by the fact that Dennis Gartman and other veteran market observers saw the same cumulative trigger in Athens. But a review of “post hoc” is always in order. See you back in logic class.