In financial markets, asset-management companies increasingly rely on data to generate profitable trading strategies. However, developing skills in data analysis and processing is a costly and uncertain process, especially as many forms of data emerging in today’s Information Age are entirely new to investors. To use such “alternative” data to make the right trading decisions, investors need to build the appropriate technology infrastructure from scratch and hire data scientists to decode the data. Focusing on investors’ skill acquisition, HKUST’s Yan Xiong and co-researchers offer timely economic insights into the implications of alternative data for financial markets.

Decades ago, stock prices and other accounting figures were considered relatively new and unfamiliar sources of data. Now, however, such data are completely traditional, and a host of new types of alternative data has arisen. “By today’s standards,” the researchers explain, “alternative data ranges from customer/investor sentiment to crop yields calculated from satellite images and from app usage to survey data on construction permits.” In modern financial markets, sophisticated investors (such as hedge funds) are increasingly relying on such alternative data, purchased from data vendors, to generate lucrative trading strategies. What factors determine how accurately they can analyze and how effectively they can trade on the purchased data?

The researchers conduct a series of rigorous statistical analyses to explicitly model how skilled investors extract information from alternative raw data and how data vendors influence the precision of the information they acquire. Their model considers the composition and performance of the investor as well as the vendor’s data-selling behavior. Skilled investors can purge more “noise” from the data, resulting in more precise processed information. Meanwhile, data vendors influence the precision of the information extracted by investors by controlling the amount of data provided; as the researchers note, “a larger data sample allows skilled investors to extract more precise information about asset fundamentals.” In addition, the researchers examine three market variables: price informativeness, the cost of capital, and return volatility.

“Our framework delivers novel economic insights,” they report. For example, differences in investor composition and data vendors can result in different levels of information precision in the alternative data setting. “Our framework is the first to concretize the theoretical notion that the data vendor adds ‘personalized noise’ to the sold information,” the authors explain. Their findings also shed light on how the rise of alternative data is changing the landscape of active funds. Contrary to fears for the future of the funds industry, the researchers find that the data industry and funds industry tend to foster each other, suggesting that “the active funds industry may actually grow with the development of alternative data.”

In today’s financial markets, skills acquisition is an expensive and risky process for investors—especially with the rise of alternative data. This study could provide much-needed guidance for asset-management companies, especially funds, on translating new and unfamiliar data into profit-making investment strategies.