Pay Attention To The Man Behind The Curtain

robot-img

This week S3, a part of Amazon Web Services (AWS), had an outage that affected many large internet companies that we all use in our regular routine. Netflix, Spotify, and Pinterest to name a few, had problems related to their use of the Amazon cloud services. Things we’ve started taking for granted over just the last few years were suddenly removed from us, or at least moved slower. Amazon soon announced that this wasn’t a computer glitch, but actually human error. It wasn’t that long ago that everything on the internet moved slowly. Dial up modems have given way to high speed data movement that allows us to not only post pictures quickly and easily, but to even watch movies without a glitch in front of our computers, or with an iPad in our hands, and this has become the norm.

The use of computers in our daily routine, or for use in very critical parts of our lives, is something we not only take for granted but depend on with increasing importance and breadth. We all know that our cars will soon be self-driven, and the new wave of investment management is known as robo-investing. Not a niche, but something offered by major investment houses. And all those robo-investing actions are the result of human programming.

This trend interests me greatly, as I’ve been both an active trader as well as having back tested thousands of automated trading ideas to build robust automated strategies. I recently read a piece on LinkedIn about the lack of gray hair on trading desks across the securities industry. The trend is toward hiring young ‘quants’ and programmers to create computer models that make money by analyzing huge amounts of data and deciding what to buy or sell and when. Don’t get me wrong, I started using computers myself to do this over 20 years ago. Not as a quant, but as someone who believed a human only has the capacity to monitor a very small amount of markets or securities on their own.

So is there a problem with this? Well, let’s go back to those cars for a moment. In the 2004 movie “I, Robot” a police detective is saved in a car crash at the expense of a 12 year old girl. The robot behind this action decided whose life was more valuable. And our self-driving cars will at some point need to make just those types of decisions. But maybe that 12 year old girl that is lost in our real world was going to cure cancer, figure out space travel through worm holes. At the time of the accident she’s just a kid getting ready for the awkwardness of adolescence. Who knows? The next Bill Gates. The next Steve Jobs.

And this has what to do with your robo-investments? One more movie reference and I’ll tell you. I watched the movie “Sully” yesterday. In the movie, Tom Hanks has a line that prompted me to write this blog entry: “Everything is unprecedented until it happens for the first time.” This is not only true in our lives, it’s true in the investment world. We call them Black Swan events. In quant investments, people are using statistical analysis to find the best odds of mean reversion. The place where things are so out of whack that they have to return to normal.

This has led to major profits for those that can program this math in ways that analyzes all that data. But, it’s also led to spectacular losses. In 1998, Long Term Capital Management (LTCM) used analysis techniques developed by Myron Scholes and Robert Merton to put on large positions across security classes around the globe. These were Nobel winning smart people. And the hedge fund had outstanding returns. Until it all blew up and we the taxpayers bailed them out. Heard that before? What happened? That unprecedented event.

I’ve mentioned an early student of mine, a physics major, who was amazed at the frequency of six sigma moves in commodities. Six standard deviations away from normal market action. The standard bell curve doesn’t even go out that far. They’re generally drawn to 3 standard deviations because in statistics, that includes 99.7% of the probable outcomes.

Standard Deviation

Those points all the way to the right and left are the tails. They just don’t happen. Until they do. Like Sully noted, they happen for the first time. And this goes on with more regularity in the markets than we realize. It can be something financial like the implosion of LTCM due to a foreign currency crises. Or a political event, like Brexit or the unpredicted election of Donald Trump as President.

The point is, the programmers often don’t take this into account because it hasn’t happened…yet. Or maybe just a human mistake slips in, like this week with S3. All the computer models, even those involved in robo-investing for the average investor designed to level the playing field and supposedly take out the human factor of your chosen advisor are programmed by people.  And all of those people have human thoughts and biases that go into those models. It can’t be avoided. There are forks in the decision process that must be taken. Not like Yogi Berra, “When you come to a fork in the road, take it,” but one direction or the other. And as time goes on, the models may learn and get ‘smarter,’ but they still start somewhere, programmed by someone.

And the lack of gray hairs is what gives me pause. How many Black Swans have the programmers experienced? We can’t plan for everything, and that’s when the unplanned event usually occurs. Experience has always been of great value in the investment world. The more unexpected events an advisor or trader has lived through, the more they will take this into account in their decisions. Remember, risk/reward, in that order. And if all you’ve experienced is a bull market over the last 8 years, do you really take into account what happened at LTCM? Do you even think about the internet bubble that burst in 2000? The mortgage crisis is the most recent large Black Swan, and many of these investment and trading strategy programmers were kids. Are they thinking about auto debt risk? Student loans? Whatever might be the next event trigger.

I’m not trying to imply that automation of trading decisions doesn’t work. I had success doing it both for myself and clients. It broadens the ways you can allocate money in just recently unforeseen ways. What I am saying is be careful. Because tail events happen. There are Black Swans swimming all the time. We just need to be sure to look out for them.

One thought on “Pay Attention To The Man Behind The Curtain

  1. Indeed, Bill, I am principally interested in automating their estimation. The asymmetry of returns on tail events is quite attractive. It’s also why I own so much Monero.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s