What key factors impact investors' willingness to pay for data?

New MIT Sloan research offers investors and financial firms a method for data valuation

CAMBRIDGE, Mass., April 26, 2023 – In a new study co-led by MIT Sloan School of Management, researchers have developed a statistical method that investors and financial firms can use to value their existing data and/or a potential data stream that they are considering acquiring.

Authors of the paper, “Valuing Financial Data,” which is currently under peer review, areMaryam Farboodi, assistant professor of finance at MIT Sloan; Laura Veldkamp, professor of finance at Columbia Business School; Venky Venkateswaran, associate professor of economics at New York University Stern School of Business; and Columbia Business School PhD candidate Dhruv Singal. 

“Valuing data is not an easy task, because data is not observable,” says Farboodi. “How much investors value data depends on the profits they can make from it, which in turn depends on who else knows the data, and how aggressively they trade on it.”

Valuations are critically important because they offer information around consumer surplus and welfare, and show how demand responds to changes in price, markups, and market competition. Valuations also show whether prices are efficient. The researchers’ novel statistical approach bypasses the need to understand what other investors know in order to value data.

In their paper, Farboodi and her co-authors use real-world data to measure several components of an existing asset pricing and portfolio choice model, the so-called “noisy rational expectations model,” with the goal of illuminating what the data valuations are and how they vary across investors.

Their statistical method can be applied to any finance-relevant data series or bundle of data series as long as the purchasing investor’s characteristics are known and there is access to a history of market prices and data realizations. The method involves specifying the regressions that should be run to value data. It then outlines the statistics that need to be computed based on the regression results. Lastly, it shows how to aggregate these statistics into a single number that corresponds to the value of the data.  

This methodology enabled the researchers to uncover key determinants of investors’ willingness to pay for data. The research also uncovered key factors that play a part in how investors’ willingness to pay for data can change. These determinants include their style of investing, how big their portfolios are, the data they already have, and market liquidity.

The researchers found that differences in investor characteristics have a considerable impact on how investors value the same data. In other words, data, as an asset, is not equally valuable to all investors. In general, wealthy investors are willing to pay more for data because they may be able to use it to earn a healthy return on their sizable investments. For example, when markets are competitive and a trade has no impact on the market price, the researchers find that an investor with $250 million in initial wealth would be willing to pay almost 300 times that of an investor with only $500,000.

When the researchers applied their approach to the model using real-world data, they found that an investor with $250 million who invests in five different portfolios of S&P 500, growth, value, large and small stocks is willing to pay $1.2 million for I/B/E/S (Institutional Brokers’ Estimate System) forecasts. But an investor with $500,000, they find, will only pay $35,000 for the same data.

 “Large investors value data of each asset more than small investors because data can help them make more money. Growth investors value data more because growth stocks are more uncertain,” Farboodi says. “In both cases, data can improve their expected payoff.”

The researchers also found that all investors are willing to pay less for data when asset markets are illiquid since they cannot incorporate the data into their trading strategy as effectively. Most strikingly, they note that in illiquid asset markets, the amounts by which different investors value data become more similar as wealthy investors exhibit a sharper decline in data value. This could be because, when a large investor places a big order to buy, asset prices move against this purchase, and other investors then demand a higher price to sell to the large investor. With expenses adding up, the large investor, in turn, is less willing to pay for data. This means that liquid asset markets and illiquid data markets go hand in hand.

The findings suggest that, as data becomes an even more important asset for financial firms selling data, these firms’ prices and market value may become more sensitive to market liquidity. “In a world in which data is becoming increasingly abundant,” the authors state, “this new effect could grow much stronger.”

“Data markets are still far from being competitive, as they are very opaque and mostly on a bilateral basis, making it an area that requires additional research,” Farboodi says.

About the MIT Sloan School of Management

The MIT Sloan School of Management is where smart, independent leaders come together to solve problems, create new organizations, and improve the world. Learn more at mitsloan.mit.edu.