Active management has faced fierce competition from passive strategies. Even with the pressures active managers are under, this strategy remains the largest in terms of assets under management. A panel of industry leaders discuss the broadening approach to using data to find alpha.
- Active investment managers are broadening their search for alpha, driven by pressure on expense ratios and performance.
- Technology and data now play significant roles in the investment process, leading to greater use of quant research techniques among asset managers.
- The best service providers will ultimately help to lower the cost of finding alpha in data.
Representatives of the world’s largest data providers, at a global financial services conference in New York, recently discussed the approaches that active managers need to consider for the challenge of finding alpha in data.
As the chart below shows, active managers are up against considerable competition from passive strategies, especially in the United States. But the investment strategy is still the largest in terms of assets under management.
Given the pressure on expense ratios and performance, active managers are understandably broadening their search for alpha.
Tim Gaumer, Director of Research at Refinitiv, told the conference : “Proving skill as an investment manager has never been harder. A lot of what was once considered alpha can now be explained by exposure to factors.”
More traditional methods of finding alpha are proving to be more difficult.
Technology and data are stepping in to take more significant roles in the investment processes, which is reflected in the interest in and increasing use of quant research techniques among traditional asset managers.
Tim recently asked more than 300 delegates at the CFA Institute’s annual conference how many had adopted some element of quant research. Over 80 percent said they had.
As a result of needing to use traditional quant techniques on large data sets, many firms are shifting towards data science to play a role in the investment process.
Tim says: “The search for alpha has to move into more sophisticated analytics.
“The big growth area I’m seeing is that intersection between traditional fundamental asset management and some of the things quants have been doing for a while — some call the intersection between the two quantamental research.”
Best approaches to get value out of data
Finding alpha in data means doing more than just sorting through an immense amount of data.
Catherine Clay, senior vice president of Cboe Global Markets, said: “Alpha is not out there just to be easily harvested — that if you just look at a ton of data, it’s there.
“The data set you have access to or how you use the data are really the only two ways you’re going to find alpha. So, you’d better have a unique, unknown dataset with a long history, or you’d better be doing something with the data that’s somewhat unique.
“It could be speed, algorithms, machine learning or AI that differentiates the way you use data.
“There is alpha in data, but know that about 50 percent of those who hunt for it are probably going to lose and about 50 percent are probably going to find something in the hunt.’
London Stock Exchange Group’s CEO David Schwimmer supports this.
He said: “An area that’s a huge challenge today is just sifting through the volume of data. The data ingestion process is a significant challenge and one that’s going through change and transformation.”
The cost of finding alpha in data
Another challenge in dealing with large data sets is its quality.
This is supported by a recent AI/machine learning survey conducted by Refinitiv. Out of the nearly 500 investment professionals surveyed, 43 percent said poor data quality was the biggest obstacle.
Some of the world’s largest hedge funds may consider their ability to work with messy data to be a competitive advantage.
Most firms, however, are probably better off working with a data partner who has already cleaned up a lot of the data, has the concordance right, and has mapped it back to a particular company identifier.
And, it’s not just the quality of data, it’s dealing with multiple vendors and multiple contracts, and assessing if the vendor is a sustainable business and if it is GDPR-compliant.
Catherine points out that in the mathematical equation of alpha — which is beta plus or minus alpha, minus cost — this is about the cost part of the equation.
She said: “To the extent you can lower the cost of getting the data, accessing the data, and studying the data, what you’ve done is a service to increase alpha. The cost equation of alpha is absolutely huge, and it’s controlled through the quality of the data.”
Alternative data sets
The Refinitiv AI/ML survey found that 70 percent of the financial institutions polled are using some source of alternative data.
Tim notes the days of signing a contract to get data history that you have to download and install may soon be behind us. “We’re hearing that quants want to work differently.
“We have four innovation labs located in San Francisco, New York, London and Singapore. In response, one thing we’re working on, especially out the London office, is a Data Science Accelerator.
“The data is in the cloud. So, you won’t have to work with IT to install anything — you can just go in and play with the data. I think that’s the way of the future.”
Returning to Catherine’s first point about different unique datasets, we have seen a growing, industry-wide interest in alternative data sets from a broad variety of sources — anything from satellite images to ESG data, or social media sentiment to credit card transactions or web-scraped data.
David said: “ESG is an area where there’s information out there, but I’m not sure I’d call it structured. We have people trying to scrub that data and structure it, and use it in a systematic way.”
Solutions to suit all
The best service providers in the data space will ultimately help to lower the cost in finding alpha, through their data offering, one of the main parts to alpha generation is how cost effective its rate of return can be achieved.
But as David has identified, there is not just one type of customer.
“If you look at it, some customers are cutting edge and happy to take sophisticated or unstructured data and do a lot of work with it, and others happy to take an Excel file, and we have to cover that full range.”
