The Better Letter: So What?!
Making Behavioral Finance Practical. Part One: The Experiment
Daniel Kahneman is the world’s leading authority on human error. The Kahneman TL;DR? There’s a ton of it and it’s universal.
It is one thing to recognize our foibles, of course, and quite another to do something about them, as Kahneman readily concedes.
Research evidence, for example, suggests that being smarter, more aware, and more educated doesn’t seem to help us deal with our cognitive difficulties more effectively. They may actually make things worse. That’s generally because smart people are clever enough to concoct plausible justifications for their preconceived notions despite powerful disconfirming evidence.
Accordingly, this study found that, in many instances, smarter people are more vulnerable to thinking errors, even basic ones. Moreover, “people who were aware of their own biases were not better able to overcome them.”
That’s not good.
The inherent contradiction of false knowledge is that it feels like it’s true. Our bias blindness is such that, most of the time, only people on the outside can see we’re full of it.
As my friend Jason Zweig insists, “behavioral finance is not the study of how ‘other’ people behave. It is the study of how we all behave. It is not just a window onto the world; it is also a mirror onto ourselves.”
We’re exceedingly poor at reading our own label from inside the jar.
“For desired conclusions,” Thomas Gilovich wrote, “it is as if we ask ourselves ‘Can I believe this?’, but for unpalatable conclusions we ask, ‘Must I believe this?’” With the former, we’re seeking permission to believe. With the latter, we’re looking for an escape route.
It’s the path of least resistance in each direction.
There is something strangely comforting about seeing the failures of others. There’s schadenfreude, of course, plus the relief in recognizing I’m not alone in my sins, and the powerful rush for having figured out why. Behavioral finance offers insight and a sense of power — order amidst what appears impossibly chaotic.
But does it deliver?
I once asked Kahneman what we might do to mitigate our inherent weaknesses in this area. He chuckled and replied, “Not much.” Since I have shared this story before, regular readers might recall the rest of that exchange, but the only-partly-in-jest opening salvo is a most significant one.
My friend, the author and Shaping Wealth founder Brian Portnoy, rightly argued recently that behavioral finance theory has significantly outpaced its application.
“Bottom line: Behavioral finance generates fascinating insights but hasn’t had much practical impact on financial advisors and their clients seeking to make better financial decisions, let alone achieve financial wellbeing.”
It will only become powerful as it becomes practical. We aren’t utility-maximizers acting in our own rational self-interest very often, but we aren’t idiots either. We’re humans whose rationality is bounded. And we need help.
As Bob Dylan sang: “Gonna change my way of thinking, make myself a different set of rules. Gonna put my good foot forward and stop being influenced by fools.”
Over the next few issues of TBL, I’ll be looking at how we might try to do that. We’ll start, this week, with an experiment to set the stage.
If you like The Better Letter, please subscribe, share it, and forward it widely.
Thanks for reading.
The Experiment
As regular readers know, my much-better-half teaches fifth grade at our local elementary school, where our kids went, three doors from our house. For math, she teaches the advanced class — the top students in the school. Her students test-in based upon ability and accomplishment. Further, they and their parents agree to a very fast pace and extra homework. It is a very bright group.
The class had a party recently as a reward for winning a national math competition. They voted – overwhelmingly! – to stay in and play math games at the party instead of more traditionally “fun” activities, inside or out. It is a bright and motivated group.
It shouldn’t be a surprise, then, that the class, in the aggregate, advances well above the norm with many students testing far above grade level by the end of the school year. In my view, it helps a lot that they have a pretty great teacher.
Last week, my darling bride took a quart jar of Cheerios into her math class and asked each student to guess how many Cheerios there were. Each student did it entirely on his or her own and recorded the guess independently. A good prize was offered to the winner.
The experiment was based upon Francis Galton’s famous analysis from more than a century ago, wherein the crowd at a county fair paid a small entry fee to guess the weight of an ox, with the closest guess being awarded a prize. When their 787 individual guesses were tabulated (many from highly knowledgeable farmers and butchers), the median guess was much closer to the ox’s true butchered weight than the estimates of the overwhelming majority of guessers. The median guess (1,208 lbs.) was within one percent of the ox’s actual weight (1,197 lbs.).
The math class experiment results were roughly consistent was Galton’s even though, including only 35 students, the sample size was much smaller than optimal. There were 1,067 Cheerios in the jar – I counted each one. There were two truly exceptional guesses: 1,065 and 1,072 (I asked my family, too – one daughter-in-law guessed 1,046; the other DIL, a math Phi Beta Kappa, guessed 1,144). The median class guess was within 13 percent of the correct answer. The guesses ranged from 70 Cheerios (huh?) to an absolutely wild 364,500. Just ten of the 35 students’ guesses were better than the median, within 13 percent of the right answer.
My DB then ran another round of guesses. She told the class that, this time, they were going to share answers with the class one-by-one. They could share their original answer or change it, with another prize for the winner of this round. The results were lousy.
For this second, “public” data set, the closest guess went from two Cheerios off to 66 Cheerios off. The median went from within 13 percent of the correct total to off by almost half. Roughly half the guesses (17) were closer than the median guess and only five were better than the original median.
Most students took the opportunity to change their guesses. These second-guessers thought they were improving their guesses. Nope. And by a lot.
The set comprising the first guesses provides a decent representation of what is possible when smart, incentivized individuals, with different bases of knowledge, backgrounds, and experiences, attack a problem independently. A few get an answer close to the truth. But, because the world is messy, complex, and difficult to understand, most do not. And a few get it wildly wrong (364,500!?). However, the aggregate answer is remarkably good.
On the other hand, the set comprising the second guesses provides an excellent example of how things usually are (and how markets are).
Most importantly, the answers got a lot worse. But that’s not all.
Certainty and accuracy did not correlate. Eleven students didn’t alter their guess from round one to round two and only one of those guesses was top quintile. The student who guessed 70 Cheerios (the right answer is over 15 times that amount) held firm to his guess while each of the top quintile guessers but one changed it when given the opportunity and none improved — most ended up way, way off (the mean top quintile first guess was within 29 Cheerios of the correct answer while the mean second guess among those top quintile first guessers was off by almost 400 Cheerios).
The top math students, as a whole, didn’t outperform the rest of the (already advanced) class at guessing Cheerios, either.
Similarly, alleged experts who opine publicly on what various economic measures will be have an ignominious history of failure. Moreover, those errors are compounded because the prediction sets an expectation that, when missed, causes negative reactions in the markets. I suspect the failure of collective predictive intelligence in financial markets is because (a) the guesses aren’t independent; (b) they are subject to incentives other than being accurate; and (c) they set an expectation and the inaccuracy relative to the expectation makes markets respond disproportionately, impacting the initial guess (and subsequent guesses, too — for every 50,000 jobs the previous month’s job growth estimate was off, the current month’s is likely off by about 20,000 in the opposite direction, for example).
In the same way Robert Shiller famously noted that, “Investing in speculative assets is a social activity,” guessing games, markets, and economic predictions are social activities, too. In short, they are much more like the second-guesses above than the first. Aggregation doesn’t get rid of shared errors due to herding, availability bias, social proof, the bandwagon effect, anchoring, recency bias, motivated reasoning, optimism bias, status competitions, and the like.
The bias, noise, and error that mar or even destroy our analysis, our views, and our beliefs rarely seem like a big deal. They neither appear nor feel dangerous in the moment. They’re not explosions or gunshots. They’re paper cuts. They’re pinpricks. And therein lies the danger.
When we don’t perceive a threat, we’re not on guard. The tiny wounds bleed, and the bleed-out can be so slow and gradual that we don’t recognize the threat until it’s too late to stop – if we notice at all.
Ultimately, truth has no skin in the game unless we make it so.
We humans are shockingly prone to bad ideas, ideas that grow into terrible decisions, and then metastasize into actions that undermine, damage, or even destroy our lives. We’d all like to think that we’re a lot smarter than the “average bear,” but vanishingly few of us have a consistently good track record of decision-making and none of us is as good as we think we are. None of us is unhurt, unscathed, or unbroken.
As Koen Smets so pithily put it, “We are bamboozled by biases, fooled by fallacies, entrapped by errors, hoodwinked by heuristics, deluded by illusions.” Worst of all, these weaknesses are largely opaque to us. They leave no cognitive trace.
We are always compromised — often badly compromised — by cognitive and behavioral biases, noise, and the limits of our own abilities. We pay attention to the wrong information, evaluate the available evidence poorly, solve the wrong problems, and listen to the wrong people.
That makes for interesting information. Perhaps it’s even insightful. But, so what?! Can we do anything about the problem?
Behavioral finance is highly limited in its predictive power and proposes seemingly contradictory ideas: When are we risk averse and thus overly cautious and when are we overconfident and excessively risk-seeking? Does it offer an alternative model for human and market behaviors (or even a model at all)? Can we get beyond Kahneman’s “not much”?
Tune in to the next few issues of TBL as I grapple with these questions.
Totally Worth It
Feel free to contact me via rpseawright [at] gmail [dot] com or on Twitter (@rpseawright) and let me know what you like, what you don’t like, what you’d like to see changed, and what you’d add. Praise, condemnation, and feedback are always welcome.
Growing abuse from players and parents has led to a widespread exodus of youth sports umpires and referees, with shortages causing scores of canceled games and tournaments. From 2018 to 2021 an estimated 50,000 high school referees quit, after being followed to their cars, attacked by players on the field, and struck by objects thrown by spectators.
Don’t forget to subscribe and share TBL. Please.
Of course, the easiest way to share TBL is simply to forward it to a few dozen of your closest friends.
This is the best thing I read this week. The loveliest. The sweetest. The stupidest. The saddest, unless it was this. The coolest. The most significant. The most brilliant. The most fun. The most inspiring. The most incredible, unless it was this. The most interesting. The most horrific. The most outrageous. The best thread. The least surprising. The crossover. Huh. China. Oops. Also oops. Uh-oh. Insane. A big mistake. Karma. More karma. No translation required. Smart dad. Well done. 2+2=5. “[N]othing left to ban.” Math for the win. We all need a little help sometimes. Factcheck: True.
Please send me your nominees for this space to rpseawright [at] gmail [dot] com or via Twitter (@rpseawright).
The TBL Spotify playlist now includes more than 200 songs and about 15 hours of great music. I urge you to listen in, sing along, and turn the volume up.
Benediction
This week’s benediction comes from Charity Gayle.
Amen.
Thank you for reading.
Issue 113 (May 13, 2022 — Friday the 13th)