Marko Kolanovic is one of the best-known and well respected macro forecasters in business today. He has built a fearsome reputation for predictive power as the chief of macro quantitative and derivatives research at JPMorgan Chase & Co. (JPM, Financial).
In this research note, we discuss some of Kolanovic’s latest insights on a hot topic: algorithmic trading.
Rise of the algos
Kolanovich has been involved in quant trading for many years. He cut his teeth in derivatives and at JPMorgan he has become a foremost expert on the industry - and its future. But he has always been keenly aware of the limitations inherent in the pure quant approach, as he made clear in a recent interview with Bloomberg:
“If you’re running a strategy just based on the formulas, you’ll miss a lot of bigger-picture issues, whether it’s macro developments around central banks or politicians, or geopolitical risks, or a lot of behavioral biases. There are limitations as to how far a quantitative approach can get you.”
This is a problem ingrained in any program or model. A model is only as good and as accurate as the data being fed into it. Without the correct data points, a model will never work. But models are also susceptible to leaving out variables that might prove extremely important.
Still, the quantitative trading strategies have seen explosive growth in the past 10 years, and their programs have grown in sophistication, speed and accuracy. While good for the traders, that might not be good for markets.
Increasing dangers
During his Bloomberg interview, Kolanovic spoke extensively about the issues emerging thanks to the rise of algorithmic trading. One major problem is the growth of abuses as programs get faster and human traders get shut out:
“There’s more algorithmic trading, where algos are going through headlines or sorting through earnings statements or going through social media in real time and trading. What are the consequences for investors?
We’re seeing reaction time get shorter and shorter for releases, which can also incur costs or take advantage of slower human investors. There are signs of potential abuses with social media posts and headlines. That’s going to get worse and worse and be more of an impediment for human investors to make money. It’s going to cause more confusion in the marketplace.”
Kolanovic sees algorithmic traders as a growing threat to the marketplace. He should know, having observed the growth of quantitative trading from the inside over the past decade. The efforts to build increasingly fast and accurate trading programs has led to a virtual arms race in the high-frequency trading industry.
This worries Kolanovich for a number of reasons. One is simply abuse, with traders exploiting market sentiment - or even weaknesses in dominant algorithmic trading programs - via headlines and social media. We might think of it as a 21st century pump-and-dump scheme.
Another more dangerous problem is that, as computers supplant humans in trading and market-making, there will be increased room for further abuses. Ultimately, that could make markets less efficient and potentially expose the system to greater shocks. We are already seeing the first signs of this. In lieu of governmental action, or a severe turn of the market, this pattern looks set to continue.
Verdict
Kolanovich is clearly worried about the implications of computer-driven trading. Though he is not willing to put a timeline on when the music will stop, he is only willing to forecast that it will stop eventually. But when it does, it could be really ugly:
“At some point that’s going to end up badly—most likely when the next recession hits.”
We have no better crystal ball than Kolanovich when it comes to market timing. But we are confident the risks seeded by the rise of algorithmic and systematic trading operations are very real indeed.
Read more here:
