Working Papers

Multi-armed bandit (MAB) algorithms are efficient approaches to reduce the opportunity cost of online experimentation and are used by companies to find the best product from periodically refreshed product catalogs. However, these algorithms face…

Decisions take time, and the time taken to reach a decision is likely to be informative about the cost of more precise judgments. We formalize this insight using a dynamic model of optimal evidence accumulation. We provide conditions under which…

Labor economists regularly analyze employment data by fitting predictive models to small, carefully constructed longitudinal survey datasets. Although modern machine learning methods offer promise for such problems, these survey datasets are too…

Under what conditions can members of poor disenfranchised communities survive and even foster entrepreneurship in environments where violence is cheap? How do such conditions alter ethnic identities and political institutions? In this paper, we…

Modern digital advertising platforms allow ads to be targeted in a variety of ways, and generally aim to match the ad being shown with either the user or the content being shown. In this study, we examine the effect of matching in emotional…

How does a failure in a retailer’s mobile app impact shoppers’ purchases in its online (website and app) and offline (brick-and-mortar store) channels? Our main hypothesis is that an app failure has a negative effect on offline purchases and no…

For individuals who take on the responsibility of caring for another person due to illness, disability, or declining abilities, it can often be challenging, lonely, costly, and exhausting. As the U.S. continues to address the impact of the…

We revisit the classic result on the (non-)existence of pure-strategy Nash equilibria in the Generalized First-Price Auction for sponsored search advertising and show that the conclusion may be reversed when ads are ranked based on the product of…

We present a data-driven algorithm that advertisers can use to automate their digital ad-campaigns at online publishers. The algorithm enables the advertiser to search across available target audiences and ad-media to find the best possible…

“Buy Now Pay Later” (BNPL) is a largely unregulated FinTech innovation that provides consumers with easy access to credit for specific retail purchases. The BNPL market is projected to reach $1 trillion by 2025, but we know little about the…

This study finds that voluntary non-earnings disclosures substitute for redacted proprietary contract information. When firms redact contract information, they provide more voluntary disclosures and have higher information uncertainty and…

In this study, we causally examine complementarity in usage across a set of related software products from a multi-product firm. Digital contexts are characterized by little price variation, bundled pricing plans, and infrequent purchase or…

Predicting the responses of a cell under perturbations may bring important benefits to drug discovery and personalized therapeutics. In this work, we propose a novel graph variational Bayesian causal inference framework to predict a cell’s gene…

Privately issued money often bears devaluation risk that create monetary transaction frictions. We evaluate the real effects of supplying a new type of safe money in the historical context of the U.S. in 1863. We instrument for the change in…

Legislative and regulatory reforms often contain various forms of complexity — multiple contingencies, exemptions and alike. Complexity may be desirable if it better satisfies the needs of political constituencies, and if these benefits are…

How can we assess whether macro-prudential regulations are having their intended effects? If these regulations are optimal, their marginal benefit of addressing externalities should equal their marginal cost of distorting risk- sharing. These…

We develop a new approach to identify different categories of depositors during periods of uncertainty and quantify their compensation to remain in the bank. We isolate withdrawals due to liquidity needs, deterioration of fundamentals, and…

We study an economy with traders whose payoffs are quasilinear and their private signals are informative about an unobserved state parameter. The limit economy has infinitely many traders partitioned into a finite set of symmetry classes called…

We consider the problem of a rational, Bayesian agent receiving signals over time for the purpose of taking an action. The agent chooses when to stop and take an action based on her current beliefs, and prefers (all else equal) to act sooner…

We propose a new adaptive empirical Bayes framework, the Bag-Of-Null-Statistics (BONuS) procedure, for multiple testing where each hypothesis testing problem is itself multivariate or nonparametric. BONuS is an adaptive and interactive knockoff-…