Publications

Student loans with income-contingent repayment insure borrowers against income risk but can reduce their incentives to earn more. Using a change in Australia’s income-contingent repayment schedule, I show that borrowers reduce their labor supply…

We use administrative corporate tax data from the IRS to study a particular form of tax avoidance for US multinational corporations (MNCs). This strategy relies on cost sharing agreements (CSAs), which govern joint R&D efforts conducted with…

Computational theories of reinforcement learning suggest that two families of algorithm—model-based and model-free—tightly map onto the classic distinction between automatic and deliberate systems of control: Deliberate evaluative responses are…

Load balancing across parallel servers is an important class of congestion control problems that arise in service systems. An effective load balancer relies heavily on accurate, real-time congestion information to make routing decisions. However…

Sovereign debt yields have declined dramatically over the last half-century. Standard explanations, including aging populations and increases in asset demand from abroad, encounter difficulties when confronted with the full range of evidence. We…

Roe vs. Wade (1973) recognized a constitutional right to abortion under various circumstances and, in doing so, facilitated a political environment in which politicians could endorse more restrictive abortion policies that …

Since the 2000s, U.S. public pension funds have actively shifted their risky investments away from public equities and toward alternative assets like private equity and hedge funds—some much more than others. We explore a range of possible…

This research examines how the movements an interface requires of a consumer—that is, its “kinesthetic properties”—can alter what a consumer attends to when responding and, in turn, change the response itself. We compare the kinesthetic…

Bank capital requirements are based on book values, which are slow to reflect losses. In this article, we develop a dynamic model of banks to study the interaction of regulation and delayed accounting. Our model explains four stylized facts: book…

The Segment Anything Model (SAM) was fine-tuned on the EchoNet-Dynamic dataset and evaluated on external transthoracic echocardiography (TTE) and Point-of-Care Ultrasound (POCUS) datasets from CAMUS (University Hospital of St Etienne) and Mayo…

Modern language-model deployments must often balance competing objectives—for example, helpfulness versus harmlessness, cost versus accuracy, and reward versus safety. We introduce Conformal Arbitrage, a post-hoc framework that…

Continuous-time random matching with a large (continuum) population is widely exploited in the literature, but has not had a rigorous formulation, nor a demonstration of its key assumed properties. This paper provides the first probabilistic…

Drawing on the Significance Quest Theory (Kruglanski et al., 2022), we used the Honor Dictionary (Gelfand et al., 2015) in a word frequency textual analysis (Pennebaker et al., 2007) to investigate extreme rhetoric. We thus conducted two studies…

It remains a continuing challenge for companies worldwide to reliably assess the greenhouse gas emissions incurred in connection with their operations. Here we argue that financial accounting offers an architectural template for corporate carbon…

While companies today are less exposed to defined employee benefits than they were in the past, the lessons from historical restructuring transactions in these areas can and have been applied to other types of nondebt liability claims. These…

An agent acquires information dynamically until her belief about a binary state reaches an upper or lower threshold. She can choose any signal process subject to a constraint on the rate of entropy reduction. Strategies are ordered by “time risk…

Banks’ lending technology hinges on their handling of soft and hard information in dealing with different types of credit demand. Through assembling a novel dataset on banks’ investment in information technologies (IT), this paper provides…

We introduce a framework for calibrating machine learning models to satisfy finite-sample statistical guarantees. Our calibration algorithms work with any model and (unknown) data-generating distribution and do not require retraining. The…

Governments worldwide have recently launched policy support programs for hydrogen, where the level of support is to be tied to the carbon intensity of the hydrogen produced. Here we analyze the impact of alternative accounting rules for assessing…

We study how international environmental agreements can take advantage of domestic time-inconsistency problems. Policymakers often prefer future policies to be sustainable, but are tempted to invest less when being in office. We find the…