It was a nearly unanimous conclusion: in an investment environment where information is increasingly ubiquitous, asset managers who mine quantitative data will have a better shot at finding higher returns.
“Introducing data to your process can only give you an advantage,” said Tammer Kamel, founder and CEO of Quandl, a financial and alternative data provider, at the DataDisrupt Financial Services conference in New York on Thursday. Kamel was one of four panelists for a talk entitled “Rise of the Machines,” a discussion of how complex financial data, often shaped by machine learning, will become increasingly important for portfolio managers.
Kent Collier, founder and CEO of Reorg Research, explained during the panel that human judgement still plays a role in the application of data. Reorg Research delivers email alerts when an item is filed in a bankruptcy case, which allows subscribers to move quickly if they’re invested in the company. But the firm also employs journalists and analysts to provide intelligence to clients based on the filings.
Hedge funds that have in-house data scientists are outperforming in the market, said Michael Marrale, CEO of research and analytics firm M Science, during the panel, giving further evidence that alpha is being found in data.
“Data utilization is necessary for any sort of performance, even benchmark performance at this stage, and it’s only going to increase,” Marrale said.
But data sets don’t have endless alpha-generating abilities; to address this, Kamel said his firm limits how many times a data set can be sold to clients. There is also the question of what to do with all that data. One theme that emerged among the panelists is that many investment firms are struggling to incorporate data into their strategies, making the task of hiring and retaining data scientists important.
Many firms have the ability to gather data, but the challenge is how to define what it can be used for, said panelist Seong Lee, a quant strategist at Quantopian, a firm that crowdsources investment algorithms and has attracted backing from Point72 Asset Management’s Steve Cohen.
“People aren’t taking as much time to figure out where technology can be pointed,” he said. For example, a firm may try to use data for alpha generation, but the data may not be suited for that purpose and could be used in another way, he explained.
Marrale acknowledged that the cost of incorporating quantitative data into an investment strategy would make it too expensive for smaller firms to use. Still, he said, portfolio managers need to be aware of the financial data that is readily available about a company. Marrale pointed to athleisure brand Lululemon, which reported fourth-quarter earnings on Wednesday that were below expectations; its stock price subsequently dropped. Investors who were paying attention would have avoided being on the wrong side of the trade, he said.
“Not everyone can afford to be in the game like hedge fund clients,” Marrale said. “But there has to be a level of understanding from the portfolio manager about what the data is showing.”