We’re wildly, exuberantly irrational organisms. That’s true in our lives generally. In investing specifically, we’re even worse. And it’s hurting our ability to exploit this moment. So we need a latticework of mental models to make decisions.
Without it, we’re hopeless.
Without it, we think it’s better not to lose $20, than to gain $20. So we hang onto investments that don’t make sense, because the pain of losing is twice as powerful as the pleasure of gaining.
Without it, we seek information that confirms what we already think, instead of challenging our assumptions with contrary data.
Without it, we think when something bad happens, it’s more likely to happen again. If we had a car crash last week, we think car crashes are now more likely overall. If we had a portfolio crash last quarter, we think portfolio crashes are more likely overall. We become more risk-averse, instead of treating it as a moment in time with an equal distribution of outcomes from that point forward.
What other cognitive biases are operating right now?
VC and PE are unfairly beaten down. And cognitive biases are getting in the way of exploiting this moment.
VC has had 3 consecutive nasty quarters. So investors assume the 4th is more likely to be nasty. But that’s a base rate fallacy. We underestimate the base rate likelihood that VC will have a bad quarter. So when there’s a bad quarter, we think that’s terrifying evidence of a fundamental shift that will negatively impact the future, instead of just VC having a bad quarter, as it is prone to do.
And we fail to consider that one of the most powerful forces in investing (and life) is reversion to the mean – the likelihood that over time, we’ll return to historical norms – where VC and PE have consistently outperformed the public markets.
Then we further compound the effect of that cognitive bias. Because enough of us suffer from base rate cognitive bias, it creates another: social proof, the bias that what lots of people believe, is more likely to be correct – making pessimism even more widespread.
