Working Papers

We propose a new estimator for the average causal effects of a binary treatment with panel data in settings with general treatment patterns. Our approach augments the twoway-fixed-effects specification with the unit-specific weights that arise…

We study a prominent energy regulation affecting large Chinese manufacturers that are part of broader conglomerates. Using detailed firm-level data and difference-in-differences research designs, we show that regulated firms cut output and…

We study the problem of model selection for contextual bandits, in which the algorithm must balance the bias-variance trade-off for model estimation while also balancing the exploration-exploitation trade-off. In this paper, we propose the first…

We propose a new diffusion-asymptotic analysis for sequentially randomized experiments, including those that arise in solving multi-armed bandit problems. In an experiment with n time steps, we let the mean reward gaps between actions scale to…

It has become increasingly common for data to be collected adaptively, for example using contextual bandits. Historical data of this type can be used to evaluate other treatment assignment policies to guide future innovation or experiments.…

COVID-19 drove a mass social experiment in working from home. We survey more than 30,000 Americans over multiple waves to investigate whether WFH will stick, and why. Our data say that 20 percent of full workdays will be supplied from home after…

Individuals face non-linear budget constraints in myriad situations. We test a fundamental assumption of economic analysis in such settings: that individuals display stable preferences when facing linear and non-linear incentives. We use a…

The article offers information about the platform annexation, and the logic using basic principles from platform economics. It analyzes the platform annexation to the traditional antitrust categories in the market. It mentions that a platform…

Computationally efficient contextual bandits are often based on estimating a predictive model of rewards given contexts and arms using past data. However, when the reward model is not well-specified, the bandit algorithm may incur unexpected…

In medicine, randomized clinical trials are the gold standard for informing treatment decisions. Observational comparative effectiveness research is often plagued by selection bias, and expert-selected covariates may not be sufficient to adjust…

Vaccinating the world’s population quickly in a pandemic has enormous health and economic benefits. We analyze the problem faced by governments in determining the scale and structure of procurement for vaccines. We analyze alternative approaches…

We study how fathers’ access to workplace flexibility affects maternal postpartum health. We use variation from a Swedish reform that granted new fathers more flexibility to take intermittent parental leave during the postpartum period and…

When the government commits to a debt policy, the future value of government primary surpluses at all horizons is dictated by the debt dynamics under the risk-neutral measure. We compare the present discounted value of future surpluses implied by…

The health care system commonly relies on information about family medical history in the allocation of screenings and in diagnostic processes. At the same time, an emerging literature documents that treatment for “marginally diagnosed”…

We posit a standard model of an asymmetric double auction with interdependent values in which each trader observes a private signal about a hidden state before submitting a bid or ask price for a unit demand or supply. The state and signals are…

We study the dynamics of participation and health care consumption in the Affordable Care Act’s health insurance marketplaces. Unlike other health insurance contexts, we find individuals commonly drop coverage midyear — roughly 30% of…

Choice screen auctions have been recently deployed in 31 European countries, allowing consumers to choose their preferred search engine on Google’s Android platform instead of automatically defaulting them to Google’s own search engine. I show…

Tractable contextual bandit algorithms often rely on the realizability assumption — i.e., that the true expected reward model belongs to a known class, such as linear functions. We investigate issues that arise in the absence of realizability and…

Alongside the outbreak of the novel coronavirus, an “infodemic” of myths and hoax cures is spreading over online media outlets and social media platforms. Building on the literature on combating fake news, we evaluate experimental interventions…

In the urgent setting of the COVID-19 pandemic, treatment hypotheses abound, each of which requires careful evaluation. A randomized controlled trial generally provides the strongest possible evaluation of a treatment, but the efficiency and…