Financial markets face the final stretch of the year with a new reminder of how quickly certainties can change. After a first day of increases after the Democratic victory in the midterm elections in the United States, the stock markets suffered sharp falls on Thursday, especially sharp in values linked to artificial intelligence. The same electoral outcome that was initially received with relief gave way to fear over its possible implications for the technology sector. It is tempting to find a convincing explanation for each movement once it has occurred. The difficult part was anticipating both before they happened.
Artificial intelligence continues to be the main narrative and economic driver of the market, although the enormous investments necessary to maintain this technological race are beginning to raise doubts. This uncertain present contrasts with how easy it is to organize the past. Seen from today, it seems evident that artificial intelligence was going to be one of the major investment themes in recent years and that the companies that manufacture chips, build data centers or develop large models would be among the main beneficiaries to date. Once you know the route, everything falls into place.
The problem is that we now know how history has developed. I like a phrase that summarizes this illusion well: “When the future becomes the present, chance disguises itself as destiny.” I think that few sentences better describe one of the most frequent mistakes that investors make.
In psychology, retrospective bias is called our tendency to consider foreseeable, once known, events that before they occurred were subject to enormous uncertainty. It's the classic "I already knew it." Once we know the outcome, we reconstruct the story, mentally eliminating a good part of the alternative paths that could have been taken. The future is a tree full of branches. The past, on the other hand, only preserves one.
Let's think about what happened with artificial intelligence. Looking back, it's tempting to draw a straight line from the launch of ChatGPT to today's spectacular investments in data centers, semiconductors, and computing power. But along the way there were questions about monetization, capacity restrictions, technological competition, regulatory changes and episodes that called into question who the winners would really be. Even today we still don't fully know. However, each kilometer traveled makes the destination seem more inevitable.
In the markets this illusion is especially dangerous because we confuse explanation with prediction. We are extraordinarily good at explaining what happened yesterday. Much worse at predicting what will happen tomorrow. And that difference matters.
Hindsight bias fuels overconfidence; It also makes it difficult to learn from our decisions. If we believe that what happened was obvious, we end up convinced that next time we will also be able to see it coming. We will believe we are better predictors than we really are. And, if we remember our expectations as if they had always coincided with the outcome, we lose the necessary reference to review what we were right and where we were wrong. Misremembering the uncertainty of the past prevents us from learning correctly from it.
In March 2020, it was unknown how long the pandemic would last, whether there would be effective vaccines, which companies would survive, how much the economy would fall, or what capacity governments and central banks would have to contain the damage. That was the starting point, not the certainty of a recovery. Seen now, however, it seems almost evident that the confinements were going to benefit large technology companies, that central banks would respond with extraordinary stimuli, that effective vaccines would soon arrive and that the stock markets would end up recovering strongly. Knowing the outcome, we have unintentionally erased all those futures that also seemed possible. We remember what happened, but we forget how uncertain it was that it would happen.
Maybe that's why investing while looking in the rearview mirror is so comfortable. Everything seems orderly. The winners were recognizable, the bubbles were evident, the crises were announced and the opportunities were waiting for someone to take advantage of them. There is only one small drawback: you cannot invest in the past.
When we look ahead, that comfortable feeling of predetermination disappears. Probabilities, alternative scenarios, decisions with incomplete information and the possibility of making mistakes are back. We can be convinced that artificial intelligence will have an enormous impact without knowing who will ultimately capture that value. Conviction about a technology does not solve the question.
Perhaps a good discipline for an investor is to write down what he expects to happen, what other things could happen, and what probability he assigns to each scenario. When the future comes, you will be able to contrast the outcome with what you really thought, not with what you think you remember. This will give you some chance to honestly confront your past self. Because the future was never written: it was always a leafy tree that the present reduced to a single branch.
