Checking Your Blind Spots

I’m an expert at predicting things that already happened. The election will be tighter than expected. Quibi will fail. Wouldn’t investing be easy if we could just focus on the known-known and avoid the unknown? This is not another story about Quibi, or the election, but they both remind me of how often we fall prey to hindsight bias when judging decisions based on outcomes. I thought I’d share my own lessons about uncertainty, decision making and finding a balance between conviction and over-confidence.

This chain of thought really crystallized in 2018 when I listened to one of the world’s most successful venture capitalists talking to a packed audience of LPs and aspiring GPs. To paraphrase him, “The longer I do this, the less confident I become.” Here was one of the most successful venture capitalists on the planet, basically an investing immortal, admitting his uncertainty when it came to investing. I was SO relieved. A rare few have earned the right to be as self-deprecating. Many more people in our profession exude a false sense of certainty and confidence. His comments have stuck with me and the Quibi example brought them back to the surface.

How confident are you?

Investing is an exercise in predicting the future. It’s like being a weather person before supercomputers. Experts in all kinds of fields are asked to make predictions. It’s not that investing is hard, it’s that predicting the future is hard! Consider these examples from totally unrelated fields;

· Philip Tetlock, currently a professor at the University of Pennsylvania, is famous for a series of forecasting tournaments funded by the U.S. Intelligence Advanced Research Projects Activity (IARPA). Tetlock recruited thousands of geopolitical “experts” and collected more than 1,000,000 predictions. Topics might include whether a European Union member country would withdraw from the EU, who would win a political contest or predicting economic recessions. He then compared the expert predictions against actual events and graded the predictions for accuracy. The experts performed worse than random guesses would have performed. And even more interesting, the more well-known the expert, the worst they performed! (Scientific American or The Atlantic)

· In the field of hard science, Dr. Michael Lauer, the Deputy Director of the Nation Institute of Health, measured the predictive accuracy of scientific experts during the peer review process for NIH grants. The NIH uses a scoring system which allows a uniquely quantitative baseline to compare with outcomes. They measured outcomes by scoring grant recipients by the # of citations their work received by other scientists (a measure of importance) and $ spent on their research (to measure Return on Investment). There was no association between grant percentile ranking and grant outcome. In other words, the scientific experts asked to judge grants had no predicative power. Based on these results, the NIH continues to evaluate other systems for allocating grant resources. (National Library of Medicine publication)

What does the NIH have to do with Quibi?

Nothing, other than both stories are instructive to those of us that are asked to predict the future. Some decisions seem obvious at the time. More often they seem obvious in hindsight. Without knowing the outcomes, what would you predict about the fate of a few of these VC investments?

· Twist — an app to text friends when you are running late.

· Wakie — originally conceived as a human alarm clock service.

· Yo — allows users to send the word “Yo” to one another with one character.

Yes, these got funded. I don’t mean any disrespect to the investors in these deals and unquestionably there was a larger vision that is not obvious to an outsider. It’s super easy to critique from the balcony. But what about Quibi? It had an all-star management team. It had strategic support from most of the major studios including Disney, Sony, and Time Warner. It utilized an artist friendly agreement to ensure quality content (they even won 2 Emmy Awards), and it was well capitalized. I was a skeptic, but I was a balcony skeptic. My skepticism was no more insightful than anyone else. In this case we skeptics just happened to be right. But there is a plausible universe in which not-investing in Quibi could have turned into a classic Type II error, best exemplified by Bessemer’s famous decisions to pass on companies like Airbnb, Tesla and others.

What’s the lesson for Investors?

For me the lesson is two-fold. First, do everything you can to model better decision making. And second, realize you’ll still be wrong sometimes. Predicting outcomes is hard, even for the best investors on the planet. Humility is a strong predictor of future outcomes.

There are literally hundreds if not thousands of books, seminars, blogs or podcast dedicated to better decision making. I’ll avoid a full review of the literature, but simply share a take-away from The Tetlock study. They identified a group they called “superforecasters”, people who were consistently & statistically better than others at making forecasts. How those superforecasters made decisions is instructive. They tended to think like foxes, open to a wide range of sources, willing to admit uncertainty and untethered to a grand theory. They think in possibilities and probabilities, not certainties. And instead of changing their mind dramatically, they change their mind gradually as new information is presented.

“Have strong convictions, weakly held.” Paul Saffo

So that’s my lesson in better decision making. You’ll be wrong some of the time. That’s the definition of probabilities. Which brings me to an observation that worries me. One of the mega trends with private market Limited Partners has been a trend toward concentration and away from diversification. The first wave of concentration was largely a result of disintermediation of fund of funds as institutions built internal staffs and built their own portfolios. This created a natural tendency to own fewer underlying holdings. The second and third wave have largely co-existed as LPs focus more capital on a relatively smaller number of General Partners and seek to co-invest alongside those GPs creating even more concentration. To be clear, most LPs have not taken this approach to its extreme and I don’t think over-diversified portfolios are beneficial. Obviously balance matters.

But finding that balance is both art and science. On this point, I found an interesting paper from May 2019 by HarbourVest called Rethinking Risk: How Diversification Amplifies Selection Skill. Yes, HarbourVest has a vested interest in selling diversification but if you want to be a fox like a Tetlock superforecaster, it pays to consider the argument. One of the key findings of their paper was that diversification improves outcomes to a greater extent when the distribution of returns is positively skewed. That part of their argument was related to US Venture Capital, where most people subscribe to a power-law distribution of outcomes (the Airbnb & Tesla examples from earlier). The math for venture is clear that it pays to be diversified, which is one reason I changed my mind and decided to largely outsource venture capital to venture fund of funds when I was an investment consultant. Picking one venture fund per year for a client was just way too concentrated and not in their best interest.

There are other sub-categories of private markets with high dispersion of returns, typically they include the word “emerging”… emerging markets and emerging managers. High dispersion is a blessing for skilled investors that embrace uncertainty and diversification. The balance of diversification vs. concentration should not be uniformly applied to all sub-categories of private markets. There are categories where a degree of diversification probably makes more sense than a concentrated approach. As long as the cost of the structure doesn’t overwhelm the positive benefits.