Adapted from Valuing Financial Data by Maryam Farboodi of MIT Sloan, Dhruv Singal and Laura Veldkamp of Columbia Business School, and Venky Venkateswaran of NYU Stern School of Business.
Key Takeaways
- Data is increasingly being viewed as a new asset class, with great value for all kinds of stakeholders, including investors and policymakers. But unlike other asset classes, for which the market settles on a “correct price,” data has different values to different users.
- Just a few inputs are required to value a stream of data: the wealth of a user, the price impact of that user’s trades, and the user’s style, or strategy, of investing.
- As stakeholders become accustomed to viewing data as an asset class, its sensitivity to the price impact of trades might emerge as a new source of fragility in the financial system.
The paper from Columbia Business School, “Valuing Financial Data,” explores the inputs that might go into a price model of data. The study was co-authored by Maryam Farboodi of MIT Sloan, Dhruv Singal and Laura Veldkamp of Columbia Business School, and Venky Venkateswaran of NYU Stern School of Business.
Why the research was done: It’s easy to see why data is as important to our knowledge economy now as oil, gold, and other commodities have been throughout history. What’s less clear is how to price it. As data begins to be viewed and traded as an asset class, that will become more important.
It may seem challenging to construct a valuation model for data, which has different uses to different users across a wide range of characteristics, and it may even have different values to the same user in different circumstances. But new research suggests it might be fairly straightforward. Just a few inputs — wealth of data users, the price impact of their trades, and their trading style — may be all that is needed to construct a price model.
Researchers found that the wealth of a user is important because it determines how much risk that person is able to tolerate, which in turn informs how much they value data. The impact that user’s trades have on the price of data also matters: if they move prices significantly, it “absolutely quashes” how they value that data, as Columbia’s Laura Veldkamp puts it. Finally, trading styles or strategies clearly influences the value of data for a user: someone with a mandate to invest in small-cap companies will value such information much more than data on large-caps, for example.
These inputs also matter the most for pricing other, very different assets, that fall within a portfolio, the researchers found. One example might be an advertising budget: how do you allocate dollars to different types of promotion for various users with diverse interests?
What the researchers found: As data becomes increasingly important to the knowledge economy, having a way to value it also takes on increasing importance. Data’s varying use cases across different types of users and all kinds of circumstances make it trickier to value than some assets, but new research suggests just a few key pieces of information—such as a user's wealth, market impact, and trading style—might be sufficient to model the value of data.
