Is alternative data becoming, well, less alternative?
A recent Greenwich Associates report points out that close to half of all investment managers surveyed are dipping their toe into alternative data, with some diving headfirst. Add to this another 25% planning on getting their feet wet, and the one-time elite investing approach is now firmly approaching the mainstream. In fact, among the notable trends is a move among discretionary, fundamental investors entering the alternative domain that was mostly dominated by early adopter quantitative fund managers. But with this new usage wave comes an underlying truth that points to those who have figured out how to benefit from the data science discipline. Finding alpha can remain illusive and is done through experience.
The big question is: What are the best processes and methodologies to integrate alternative data into an investment approach?
What's fueling dramatic growth?
The rapid growth of alternative data is being largely fueled by a new crop of investors seeking to harvest unique alpha type, the Greenwich report showed. Alternative data spending was up a whopping 76% in 2017 and 52% in 2018 with little sign of slowing down, the report surmised. More than half of all quantitative and fundamental investors surveyed recently added alternative data to their strategy toolbox while existing users are going deeper, spending four times more than new market entrants.
A wide majority of investment managers, 72%, said that alternative data was improving enhanced their signal quality in an arena where filtering out signal noise. Of those who are implementing an alternative data strategy, more than one-fifth claim to have received 20% or more of their alpha from the practice. But looking deeper into the data reveals insight into who is benefiting.
Finding alpha with alternative data is not always obvious. Richard Johnson, a principal in Greenwich’s Market Structure and Technology practice who authored the report, told ValueWalk there is likely a learning curve before repeatable success is unlocked.
Nearly one-third of investment managers who have been leveraging alternative data been doing so for more than three years. Such “power users,” many from a quantitative background, often utilize multiple data sources and are deploying it across diverse strategies and often want the data in raw format. New entrants, most often fundamental in their analytic approach, tend to prefer the data wrapped in a ready to use package after being cleansed.
To Tammer Kamel, CEO of Toronto-based alternative data provider Quandl, the usage progression from quantitative to fundamental asset managers is a natural evolution. Quantitative managers have built in processes for acquiring, cleansing, structuring and manipulating alternative data. “For quants, the jump to alternative data is trivial,” he told ValueWalk. “Fundamental managers love the concept of alt data, but they are ill-equipped to handle it.”
While alternative data providers are not strategy consultants, they often provide access to in-house data scientists to explain the data source, a service desired by 54% of those surveyed, while 20% want technical support for data integration. Some data providers are taking the data and formatting it all the way to insight. After many of the major auto makers went from monthly sales reporting to quarterly, Quandl developed a pre-packaged product that culls information from insurance policy adoption and state government motor vehicle statistics to infer new car sales on a monthly basis.
The most significant usage of alternative data is found in web-scraped and crowd-sourced data, followed by credit card and point of sale systems. Social media sentiment, search trends and web traffic are meaningful, while drowns, data from computer-linked wareables and Internet of Things sensors are on the list if currently trailing in adoption rates, for now at least.
The real alpha to be discovered in fields that are not being obviously plowed. The question is how to set up an alternative data program?
Kamel recommends fundamental managers start the process of getting their feet wet but first identifying a valuable investment question that cannot be answered with public domain data. Next assemble a data science team that can speak and operate in the world of fundamental investments. “They need to talk in a fashion a fundamental portfolio manager can understand,” he said. "They should be bridging the gap between the two vastly different paradigms of fundamental and quantitative." This can mean building a data science team that is capable of interrogating data to answer exactly the questions the fundamental PMs have. bridging the gap between the two vastly different paradigms of fundamental and quantitative. Build a data science team that is capable of interrogating data to answer exactly the questions the fundamental PMs have. "The leader of that data science team has to truly speak both languages."
Just like searching for a needle in a haystack, however, there is no guarantee that mining alternative data will answer the specific question asked. Often corollary data points are not strong enough to infer causation. But at times, when seemingly unintuitive correlations can be validated, unique alpha can be found. It doesn’t happen overnight and can require a long-term commitment, but when little known insight is discovered the gold mine can run deep. While alternative data might be obtaining mainstream attention, finding that rare alpha remains an elite achievement.
This article first appeared on ValueWalk Premium