Hasta La Vista, Baby!
Do you remember the thrill of pitting man versus machine in the Terminator? The narrative suggests the humans ultimately win the epic battle. In the real world, however, Sarah Connor would have been toast. The battle has long been won by machines. The resulting possibilities are far reaching and for the most part enriching for all aspects of human life, in particular with regards to investing money.
In this piece, allow us to indulge you how and why a machine will generate better portfolio returns for you than its human counterpart and why, if you haven’t upgraded your investment management to 21st century capabilities, you should seriously consider moving on from the Stone Age.
First, let us address a common –and emotional- argument put forward by humans in defense of their species: “I don’t want a machine to manage my money. I don’t trust machines.” Deep down, we mirror your feelings, after all both evolution and the Bible make us humans, and not the machines, masters of the earth.
However, consider this. We trust machines with our lives already in many ways. Why wouldn’t we trust machines with our money?
Take air traffic for example. As we have a history of offending at least one profession in each of our pieces, here we go with pilots. Pilot error has been the leading cause of commercial airline accidents by a wide margin for many years and is the primary obstacle to 100% flight safety. As a result, machines have become more and more involved in flight, precisely because humans can’t be trusted.
Grove et al analyzed 136 independent scientific studies which examined whether or not human experts were better than systematic models (i.e. machines) at predicting outcomes in many different professions. These situations were highly diverse and involved, among others, criminal recidivism, medical diagnoses, and academic achievement. In 96% of the academic studies, humans experts did a worse job than a systematic decision making model in making accurate predictions.
What about the world of investing? Could it be that humans are worse than machines in predicting successful investment outcomes? In fact they are. Human emotion and subjectivity tend to get in the way of objective decision making and lead to suboptimal outcomes.
Many studies have been written on the subject. In one example, Columbia Professor Greenblatt compared the outcome over a two year period of clients who followed his investment model in a managed account vs. clients who followed his investment model in a discretionary account, where they could pick stocks from a preapproved list based on his model (which theoretically should have given them an edge over the market). The managed accounts outperformed by 25% over the discretionary accounts, and in fact, the discretionary accounts even lagged the market slightly during that time frame. The moral of the story: Models outperform experts.
We can fly to the moon, build massive structures, transplant organs and experience powerful emotions such as love. You might wonder how our otherwise awe-inspiring species can be so incompetent at something as mundane as managing money?
Over the last 20 years a new frontier of knowledge generation in our industry has gained traction. This academic discipline is called Behavioral Finance and deals with how investor psychology impacts investment decisions and as a result, investment success.
What we can conclude is that as soon as humans are involved, decision-making is not just based on facts, but also on emotional biases. We have written on the subject before here. At the core, there are a handful of issues with how our brains are wired that consistently lead us to the wrong conclusions when it comes to making investments. Amongst others, they include the following:
- We are an optimistic species. As we collect more and more information on a pending decision, we get more and more confident with the conclusion we are about to draw. Yet, the accuracy of the conclusion never improves, only the way we feel about it does. Systematic models do not have emotions and don’t let confidence with a conclusion interfere with the accuracy of a conclusion.
- We are a flexible species. The same information can lead us to different conclusions based on our psychological state, when the same inputs should actually result in the same outputs. Systematic models, however, do not let unrelated stimuli influence their conclusions. The same input leads to the same output, regardless of us perhaps having just lost a loved one, how the rest of its day went or how the other people in the office are acting.
- We are a busy species relying on rules of thumb. To save time, we are inherently wired to make decisions based on heuristics instead of empirical evidence that takes into account all available data. Systematic models do not need to take these shortcuts as they truly can process all relevant data.
Of course, a systematic investment model is not a crystal ball. It doesn’t know what will happen in the future either. However, there is overwhelming evidence that systematic investing is far superior to discretionary investing over longer periods of time. It increases your edge statistically so that over a long time horizon you make more money. It’s like having a professional card counter play every hand for you at the blackjack table vs. having your Aunt Betty play every hand for you at the blackjack table.
Surprisingly, trillions of investor money is still being managed as if our industry had not progressed from the Stone Age. At Carden Capital we fail to understand why investors still voluntarily opt for a deck that is stacked against them when that is entirely avoidable.
Unless you are into charitable giving to your investment manager, you might want to make sure that 100% of all decisions in your portfolio are being made by systematic computer-driven models. As a side note, most firms (even very large ones) do not have this capability, as it requires a new breed of investment managers who view themselves as data science nerds rather than hotshot money managers.