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.

Read more from the author/contributor here.

Comments

Comments are closed.