In Defense of Narratives

Lately there’s been something of an assault on the narrative as a viable tool in the analyst’s toolbelt.  My friend Barry Ritholtz dissed the narrative in a presentation he gave at his Big Picture Conference in early October (his presentation was titled, “To Win on Wall Street Don’t Tell Stories”), and last week blogger Felix Salmon used the Greg Smith/Goldman Sachs saga to tar the act of narrative construction as little more than an unconscious defense mechanism against the harsh realities of the objective world.

Barry and Mr. Salmon are both on to something — narratives can lead an analyst astray and they can be used as a shield against a difficult-to-accept reality — but neither blogger points out that narratives are indispensable when it comes to making sense of data (in Barry’s case) or reality (in Mr. Salmon’s).  I suspect that both of them would hasten to agree with me on this point, and that both would protest that they were merely arguing against the misuse of narratives.  Nevertheless, I think it’s important to set the record straight, and the record shows that the ability to take facts or data or events and from them construct a narrative is indispensable whether you’re trying to succeed in the market or in life.

Because I’ve got the CFA on the brain, I’ll start there.  On page 572 of Volume 1 of the Level 1 curriculum, authors Fusco, McLeavey, Pinto and Runkle state:

“The absence of an explicit economic rationale for a variable or trading strategy is the “no story” warning sign of a data-mining  problem.  Without a plausible economic rationale or story for why a variable should work, the variable is unlikely to have predictive power.”

The authors go on to cite Leinweber’s 1997 discovery that butter production in Bangladesh, cheese production in the U.S., and the sheep population in both Bangladesh and the U.S. “explain” (statistically) 99% of the movement of the S&P 500 Index.  With a simple narrative such as, “The S&P 500 Index moves in response to myriad monetary, financial, economic, geopolitical, behavioral and idiosyncratic (“company-specific”) factors both domestic and global,” one can easily dispel the notion that butter, cheese and sheep production in Bangladesh is the sole determinant of the S&P 500′s direction.  Without this most basic of narratives, “rational” investors would be flocking to Bangladesh and investing heavily in low latency connections to NYSE’s and Nasdaq’s data servers in New Jersey — and they’d be losing their shirts.  (At least I think they would be.  Truth be told I haven’t bothered to look at the Bangladeshi dairy production figures lately, let alone compare them to the SPX.  Another analyst falls prey to an over-reliance on the narrative!)

I submit that an accurate narrative should in fact be one of the main goals of any analyst — or of any human being who wishes to live in harmony with the rest of the universe — and that attention to facts and data are essential precisely because they allow us to construct a more accurate narrative.  Another of the analyst’s main goals should be to find or construct a viable analytical framework with which he or she may extrapolate future facts or data points.  The narrative is simply the story from the past to the present; the framework sketches out how the story is likely to end based on logic and observation of how similar stories have ended.

More to the point, objective reality is not a dynamic matrix of facts and figures that manifest spontaneously such that the best an analyst can do is assiduously record them and proclaim their existence — as if this constitutes analysis!  Objective reality is an in-progress work of art that simultaneously exists — unequivocally, indisputably, and in fact-and-figure form — and provokes human-artists into acting in ways that alter its existence.  Because human beings naturally construct narratives from their observations of reality, they act according to those narratives.  Thus, objective reality itself takes the form of a narrative, or, more likely, an interplay of myriad narratives.  Facts and figures — “the data” — are there to clue us in to how these narratives are unfolding.

To Barry and Mr. Salmon, I say let’s make a deal:  I’ll concede that too many of us spend too little time observing and gathering data if you’ll concede that proper analysis (and proper living) requires a narrative.

4 thoughts on “In Defense of Narratives

  1. Pingback: 10 Tuesday PM Reads | The Big Picture

  2. First, narrative is inevitable as the associative process of our more primitive System 1 can’t help but put things together. You use narrative as a causal context for prediction without reference to source, perspective, or scale. The Bangladesh example tries to address the first two humorously. The “accurate narrative” fallacy was explained by Nassim Taleb. I am paraphrasing as I cannot find the exact quote. “… the more they can explain the past, the more they think they can predict the future.” Barry abhors predict, though he certainly enjoys a good story.

    As for your assertion that, “… human beings naturally construct narratives from their observations of reality, they act according to those narratives,” you have already bid good-bye to any “objective reality.” As you note, “objective reality itself takes the form of a narrative, or, more likely, an interplay of myriad narratives.” Which is to say, it isn’t objective or a reality. George Soros’ idea of reflexivity maybe useful in this regard. In this quote, a narrative is a form of “mental construct.”

    “… every mental construct distorts to some extent what it represents and every distortion adds something to world that we need to understand. The more we think, the more we have to think about. This is because reality is not given. It is formed in the same process as the participants thinking: the more complex the thinking, the more complicated reality becomes.”

    In other words, if narrative is inevitable (and it is), then the stories we tell ourselves need to be “as simple as possible, but no simpler.”

    But not in the way that you think.

  3. Hmmm, could it be that you have mistaken the word ‘hypothesis’ for the word ‘narrative’? If so, then what you have described is a decent approximation of the scientific method.

    The danger that Felix and Ritholtz describe refers to the forecast narrative, which facilitates all manner of biases and, given that the future is empirically unknowable, is rarely constructive for investing.

    Without any narrative whatsoever, and in fact without any forecast for returns at all, it is possible to construct an efficient portfolio to harvest potential risk premia using creative mathematics. It is these purely unbiased portfolios that offer investors the only true free lunch in investing. A strong narrative only gets in the way.

    • Thoughtful comment, Adam, thanks for that. Your assertion about the efficient, narrative-less portfolio reminds me of Billy Beane’s Moneyball: he built a team that made no sense to “real baseball men” but on the field he got results. And, there might be something to your suggestion that I’m confusing “narrative” for “hypothesis.” You’ve definitely given me something to think about. Thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>