Print
Last Friday, we talked about how artificial intelligence (AI) is, at heart, something simple: a set of rules and if-then statements. As such, AI can be very helpful in a simple environment, where relationships and rules remain constant, but it tends to stumble when those rules change. Waze, which is a great example of this, lives in a world of maps. It is very useful, but it fails when the reality doesn’t match up with the map. Any self-driving car will have to take closed roads, weather, and crazy drivers into account, which works well much of the time, but not always.
In practice, because of these limitations, AI is being used as a cognitive amplifier for people, rather than replacing them. That may change over time, as the models get closer and closer to reality. For at least some time, however, we will need humans at the wheel in times of stress.
Rule-based systems
AI is a special case of a much larger tool set, which I refer to as rule-based systems. Like AI, these systems are based on a set of relationships that have held in the past—and are assumed to continue to do so. These systems, however, face the same problems that more formal AI systems do when the rules change and reality simply doesn’t match the map. Investing is an area where rule-based systems have been repeatedly developed and deployed, only to fail when the rules change. Let’s look at some examples.
Nifty Fifty. Back in the 1970s, the Nifty Fifty stocks were a good example of a simple rule-based system. Buy these stocks and you’ll make money! It worked for a while, until the world changed. And then it didn’t.
Portfolio insurance. Similarly, in 1987, we had portfolio insurance. It allowed investors to sleep better and take more risks because of the new computer programs (wow, computers!) that let them manage—and lay off—risk. The programs were written based on a clear set of rules on how the market functioned, and they worked well. Until they didn’t.
Long-Term Capital Management. In 1998, Long-Term Capital Management knew how markets worked and was willing to make bets at 100-1 leverage based on that knowledge. It had Nobel Prize winners! Indeed, the firm did know the rules and made a lot of money. Until the rules changed.
Dot-com boom and bust. Last but certainly not least, the dot-com boom and bust and the housing market in the mid-2000s tell the same story. Nothing could go wrong because we had things figured out, because we knew the rules. Until the rules changed.
At any one of those times, rule-based systems did very well for quite a while. Until the rules changed. In fact, a long period of sustained performance is exactly when rule-based systems come to dominate because they do outperform. They do so because, in a certain environment, they capture what matters most.
What we’re seeing now
Looking at recent economic and market performance, with unprecedentedly smooth and long-lasting economic and market growth, now is exactly the time you would expect to see rule-based systems outperform and start to catch on. In fact, this is exactly what we’re seeing.
Passive investing, in particular, seems to be the rule-based system of the moment. Simple, easy to explain, and demonstrably effective since the crisis, once again large parts of the market are being run on the expectation that we understand the rules. And, as at every point in the past, there is good reason to believe that is right.
Until, of course, the rules change. We will look at how that could happen tomorrow.
This article is distributed for informational purposes only and should not be considered as investment advice or a recommendation of any particular security, strategy, or investment product. No strategy can guarantee that any objective or goal will be achieved.
