Ankit Bhutani

The Missing Value of Data (with Guillermo Ordoñez and Laura Veldkamp)

Data assets are increasingly vital in modern economies, yet macroeconomic measurement is not well-adapted to capturing its value. Part of the problem is that data is an intangible asset: investments in data are missed in national accounts and losses due to depreciation are missed in firms’ balance sheets. Another problem, unique to data, is that it constitutes means of payment in the modern economy: consumption bartered for data is also missed in national accounts. We propose an output-based approach to measure the missing value of data. We treat data as an asset, measure it with the quality of firms’ revenue forecasts, and endogenously determine its depreciation. We then capitalize data value and explore what measured GDP would be if data were treated and transacted similarly to a physical asset. Our findings suggest that the aggregate value of data is about 3% of GDP, increasing in the last decade to around 4.5%.

Market Power and Risk Sharing in Production Networks

I study the role of trade credit in production networks when trade credit terms are determined endogenously as part of an optimal risk-sharing contract between suppliers and buyers. A supplier with market power extends trade credit by borrowing against her own profits, exposing herself to the buyer's demand risk in exchange for higher procurement volume. Market power shapes trade credit provision through two opposing forces: greater market power reduces the need to extend credit to attract buyers, but higher profits lower the supplier's cost of capital, making credit cheaper to provide. I explore how the resolution of this trade-off determines trade credit volume across the network and shapes the economy's response to demand and financial shocks, in contrast to frameworks that take financial linkages between firms as exogenously given.

The Externalities of Going Public (with Sonakshi Agrawal)

We document spillovers to peer firms' cost of capital and price informativeness arising from IPO completions in the United States. Exploiting a stacked difference-in-difference design, we show that when a firm successfully completes an IPO, its existing public peers experience a significant decrease in the cost of equity. We are in the process of developing a theoretical model which embeds various mechanisms like signaling through IPO completion, enhanced information production, and changes in investor attention. The model will be disciplined using empirically identified moments, including the cost of capital and a rich set of information measures. We further use the model to explore the macroeconomic implications of the well-documented listings gap. Taken together, our findings identify a previously underexplored externality of IPO activity and underscore its importance beyond the issuing firm.

Standardized Test Scores in U.S. College Admissions (with Navin Kartik)

In this work, we study how US Colleges use test scores from multiple attempts on standardized tests. Allowing students to improve their score by retesting provides more data about their ability. However, many talented students do not have the financial resources to test multiple times. Given this tradeoff, we use a mechanism design approach to characterize the optimal incentive-compatible mechanism to combine scores for a wide class of payoff functions for the College. We show that in general, the optimal mechanism may not resemble any of the commonly used mechanisms like the most recent score or highest score. We derive two sets of sufficient conditions – one for the optimal mechanism to take the form of average score and second for the optimal mechanism to take the form of most recent score.