AI Takes on Research03.21.2017 By Rob Daly Editor-at-Large
Although MiFID II’s research unbundling mandate will end the longstanding research relationship between the buy and sell-side, it has sown the seeds for the next wave of research generation.
The buy side’s scrutiny of sell-side research is only going to increase under the regulations that go into effect January 2018, noted Brennan Carley, global head, enterprise proposition & product at Thomson Reuters.
“When firms actually are paying hard dollars for it instead of soft dollars for it, they will want to make sure that they are getting value for that dollar,” he said in a Market Talks segment, which is a Nasdaq-Markets Media partnership.
Payment for research will only add to the pressure on the sell-side research teams to differentiate their offering, he added.
Research generated by artificial intelligence and machine learning look to be the next evolutionary step in producing research.
“Nasdaq started years ago automating the actual matching process at the center of trading,” explained Carley. “As it rippled out from there, you saw the automation of the trading function performed on the exchanges in terms of algorithms and so forth. Now that has rippled out to the research and the portfolio management that actually that ultimately drives that trading activity.”
The technology and best practices necessary to enable this change are already in place, he noted.
“MiFID’s regulations has driven a requirement for everybody to up their game in their ability to aggregate and get a hold of data in a consistent form,” said Jon Robson, executive director at First Derivatives/KxSystems. “The next wave is how to draw intelligence from it and how to generate alpha from it, or create differentiation from it.”
Broker-dealers already have begun investing in the necessary technology, and the buy side will look to adopt the same model internally, according to Carley.
How much of a research analyst role AIs and machine learning will consume is open for debate.
“If you look at a lot of the processes- analytics, data analysis, and so forth- machines are doing a lot of the work, but people are guiding them as an active part of the feedback loop,” said Carley.
However, some see analysts playing a less and less role in the research generation process eventually.
Mike O’Rourke, global head of machine intelligence and data services at Nasdaq envisions an environment where end users only need to express their intent to an AI. In turn, the AI would find the appropriate data set and use the proper analytics to answer the users’ questions rather than having users mold their workflows and gather the data themselves.
By Greta Zhou and Andy Cheung, APAC AES, Credit Suisse
With Julien Messias, Founder, Head of Research & Development, Quantology Capital Management
Firm is testing a capability to retrieve more unstructured data from alternative investment documentation.
Asset managers are increasing budgets for alpha-generating technologies.
Machine learning promotes significant efficiencies in portfolio management.