You Don’t Have to Be a Contrarian to Disagree
There’s such thing as the wisdom of the crowd, but there’s also a difference between crowds and herds
Takeaways
Alternative data is empirical and not influenced by consensus
While consensus is often an important input to data analysis, consensus is influenced by the herd and data is not
Independent analysis, based on proprietary modeling, informed by appropriate alternative data, should yield the best results
Sometimes, finance analysts need to remember if your picks overlap 100% with those of all your competitors, you’re doing it wrong.
But somehow, the fear of missing out – or the fear of being humiliated if you make a non-consensus call that turns out to be in error – often overrides their confidence in their own quantitative skill.
Is there another way? If you trust your inputs and models, then you need to either a) trust your outputs or b) start polishing up your resume. Consensus is indeed important, but how can it evolve from the individuals who comprise it. When a position’s popularity becomes a proxy for its objective correctness, it can become a cascade of self-fulfilling prophecies.
Bulls, bears, and lemmings
It’s not strictly true that lemmings jump off cliffs en masse just because the lemming in front did, let’s go with that popular myth anyhow. After all, real bears don’t actually short stocks. But herd behavior has become something of a standard.
And not just in finance. This cookie-cutter, follow-the-leader cadence has become part of our culture. Yogi Bear, introduced in 1958, has stood the test of time. There was no need to explain the illustration above. Will the same be said of Uncle Grandpa or Chowder or any of the other Cartoon Network stock keeping units?
“We’re getting boring stuff and not even experimental mistakes(?) because people are afraid of getting cancelled,” tweeted actor/musician Donald Glover, also known as Childish Gambino. “So they feel like they can only experiment w/ aesthetic. (also because some of em know theyre not that good).”
Essentially, Hollywood is producing boring or safe work because content producers are afraid of people cancelling when they make a mistake.
And the same thing is clearly happening where we work, too.
Did you know venture capital as a whole barely broke even as recently as 2016? In 2017, it had a 10.3% internal rate of return; this would otherwise have been considered respectable, if not for the fact that every other alternative investment vehicle – buyouts, private equity, private debt, real estate –surpassed it.
For the 15-year period from 2007 through 2021, according to PitchBook, venture capital’s 15-year IRR has been a middling 11.8% compared to a 10.7% return from the S&P 500 – and that’s without assuming dividend reinvestment.
One major reason for this – not the only one to be sure – is the well-document herd mentality propagating among VC firms. A recent study out of Renmin University of China confirms what has been known for some time: “[H]erd behavior in the venture capital market are driven by positive signals of essential information and a higher degree of information uncertainty.”
A contemporary example is the current rush to fund start-ups in what has come to be known as “climate tech”. PwC’s analysis of Dealroom data suggests that in 2021 $87.5 billion was injected into this industry. As Capital Monitor reports, though, VCs played follow-the-leader into the same industry – then branded as “cleantech” – in 2004 and, by 2014, most of those deals had gone bust. The Massachusetts Institute of Technology published what is essentially a postmortem of that cycle.
As the Chinese study suggests, “a better external information environment would help weaken the herding among venture capitalists, while their reputation concerns might amplify the herding effect.”
Not that groups of people are necessarily less savvy than the individual. Quite the contrary. Economics, finance, cryptography, industrial psychology and political science are constantly coming up with newer and better consensus mechanisms, known collectively as the Wisdom of Crowds. This name comes from the title of business author James Surowiecki’s 2004 book. Surowiecki describes this improvement to decision making as “the collective opinion of a group of individuals rather than that of a single expert.”
It's a problematic book though. This insight was not original with Surowiecki, although he did much to popularize it for a new millennium. In 1908, British social scientist Francis Galton observed at a livestock fair that the weight of an ox could be determined within 1% accuracy by taking the median of the guesses of passersby.
It should also be briefly noted that Surowiecki himself cribbed the title. Actually, he inverted it from an 1841 tome by Scottish journalist Charles Mackay, Extraordinary Popular Delusions and the Madness of Crowds. Mackay and Galton probably wouldn’t have had much to say to each other.
A consensus of one
So is consensus to be embraced or avoided? Neither. Just because you’re right doesn’t mean nobody else is. Still, it doesn’t mean that, just because everyone else agrees, their shared thesis will prove to be correct.
Nobody is suggesting that Alternative Data replace consensus as a decision support tool. Consensus is not going away. Still, deployed correctly and in context, Alternative Data can augment consensus and, in the future, will be given equal weight.
Alternative Data offers a revolutionary proposition: It is 100% empirical and un-influenced by consensus. Yes, this means that might not agree with your hypothesis or the herd, but suddenly beating consensus is quantifiable. You can identify truly surprising opportunities. If you’re following the data rather than the crowd, you can independently corroborate your choice of actions and explain them without resorting to magical thinking or some vague appeal to “the market”. Your competitors have their own data sources and their own models. So it becomes a matter of whether they agree with you, not whether you agree with them.
So in closing: Start incorporating Alternative Data into your homework and worry less about how everybody else answered the question. Stop fearing "missing out" and start fearing "getting it wrong". Take a look at your Alternative Data and make sure your inputs are timely, accurate, and relevant.