They say the devil is in the details, but according to David Epstein’s bestselling book, Range: Why Generalists Triumph in a Specialized World, so are faulty projections. It references a study in which private equity professionals were asked to forecast returns on a deal they were working closely on, detailing the steps required to achieve success. At the same time, they were asked to forecast outside deals which they only evaluated superficially. On average, they projected a return on their own deals that was 50% higher than the outside ones, despite all of the deals sharing strikingly similar characteristics.
Caught up in their own minute scrutiny, the professionals had convinced themselves they were going to crush it. But when reminded of the large gap in their forecasts for deals which seemed so similar, the test subjects backed off, sharply lowering projections of their own deals.
The research Epstein references was conducted by Dan Lovallo, who had previously collaborated with Daniel Kahneman, an academic who shared an economics Nobel for work in behavioral economics. Kahneman, who passed away last March at age 90, loved to investigate how easily humans are fooled (his book Thinking, Fast and Slow is another fascinating read on this topic if you are interested).
What’s crazy is that people are fooled by the presence of more data and the inside view that cultivates. In other words, the quantity of case-specific information taken into account and the optimism of the resulting forecast were directly related. But in a way it’s intuitive: the more detailed exposure we have to one particular scenario, the more likely we are to think it will occur.
Tell someone about the specific physical qualities of a racehorse you like, and they will become more willing to believe it will win, even without knowing much about the other horses in the race. That seems to be what’s happening when PE professionals get immersed in the details of their own deals.
Bizarrely, Lovallo initially expected to show the opposite. He thought PE professionals would be better at estimating returns for their own exhaustively researched deals, especially if they were given the chance to evaluate them in light of similar “reference-class” deals. After all, part of the job when underwriting your own transaction is evaluating analogous deals in the marketplace,
But this is precisely where the test subjects failed. Having developed tunnel vision around their own projects (what Lovallo and Kahneman call the “inside view” in this paper they published in the Harvard Business Review Delusions of Success: How Optimism Undermines Executives’ Decisions (hbr.org)), they became overly optimistic and didn’t properly incorporate an “outside view” that took into account reference-class deals. Until, that is, they were informed of this fact, in which case they promptly lowered their forecasts.
That suggests one approach to address this problem of overoptimism and inside bias that seems common to those with specialized knowledge, which is simply to remind ourselves of the critical value of outside “comps” while guarding against tunnel vision as we immerse ourselves in our own projects. Optimism is important in private equity, but so is realism.
Another approach we may have in the future is using AI to remedy the problem. In this context, AI might serve as a disinterested third party. These algorithms learn by training on data which is outside of, but analogous to, the AI’s ultimate tasks, making them a logical fit for incorporating the outside view (though AI is not immune to its own biases and even hallucinations at times).
Either approach can potentially improve the accuracy of our forecasts and the ultimate profits from executed deals. And perhaps combining both approaches will give us the best results of all, man and machine working together for better returns. As I continue in AI research mode, I will keep sharing compelling insights and findings relatable to the world of private equity and investing. In the meantime, if you want to learn more about private equity, we have plenty of material here at ASM. Drop me a note if this topic interests you and happy reading!