Algos are reading headlines, but what if they're wrong?
Algos are all over the newswires and twitter. The straightforward way is a race to trade on news and economic data.
As a result of an arms race that's lasted the past 15 years; it's gotten hyper-competitive and now the moves are unbelievably fast. The next frontier was sentiment and that's been mined heavily for the past five years.
The next frontier may be gaming the system.
If there are algos trading on twitter sentiment, then what's to stop someone from making a trade and then bombarding social media with commentary about it. Or gaming news services.
It's something Marko Kolanovic is worried about. He's one of the earliest algo/quant professionals on Wall Street and the global head of derivatives and quantitative strategy at JPMorgan.
"f someone is creating fake tweets to hurt your strategy, are you allowed to defend yourself by throwing off that algorithm? Where's the limit of market manipulation vs. defense?," he said in a Bloomberg profile.
At the same time, he's worried about fake news that simply designed to throw of people, but can also throw off machines. He said recent market moves may have been amplified by fake news.
There are "specialized websites" that present a blend of real and fake news and distorted write-ups of financial research, he said, without citing the specific sites.
"If we add to this an increased number of algorithms that trade based on posts and headlines, the impact on price action and investor psychology can be significant," Kolanovic told CNBC.
The Bloomberg profile is particularly interesting:
There's this fragility in the marketplace that came with the new structure of liquidity, with electronic market-making, computers, and growth in passive. Passive assets and quant assets will grow, and computers and AI will have a bigger role in Âmarket-making. At some point that's going to end up badly-most likely when the next recession hits.