Knowledge that Transforms

Relative to remote work, working downtown facilitates valuable interactions with other in-office workers but entails commuting costs. The resulting coordination mechanism can lead to multiple stationary equilibria with varying levels of remote work. Temporary reductions in commuters, as in the COVID-19 pandemic, can lead to persistently large fractions of remote workers. Cell phone-based mobility data for the United States shows commuting trips in the largest cities, which are more likely to exhibit multiplicity, have stabilized at only 60 percent of prepandemic levels, while they are fully back in smaller cities. Cities with permanently low commuting experience average welfare losses of 2.3 percent.

The random utility model (RUM), a cornerstone in economics, is typically studied under the assumption that choice frequencies of all alternatives are observable. In practice, however, some alternatives have unobservable choice frequencies and are commonly aggregated into a single category called an outside option. We study RUM in such environments and derive a finite, nonredundant system of inequality constraints on observed choice frequencies that characterizes RU-rationalizability. We show that the conventional practice of aggregating unobserved alternatives can miss key information leading to incorrect conclusions such as that observed choices are rationalizable, even when no RUM is consistent with them.

We study how public sector workers balance their professional motivations with private economic concerns, focusing on police arrests. Arrests made near the end of an officer's shift typically require overtime work, and officers respond by reducing arrest frequency but increasing arrest quality. Days in which an officer works a second job after their police shift have higher opportunity cost, also reducing late-shift arrests. Combining our estimates in a dynamic model identifies officer preferences over workplace activity and overtime work. Our results indicate that officers' private costs of arrests have a first-order impact on the quantity and quality of enforcement.

Mechanisms that aim to reduce environmental degradation at low cost can be undermined when participants' conservation actions are not marginal to the incentive—or “additional”—as the lowest-cost participants may not be the highest social value. We investigate this challenge in the Conservation Reserve Program's auction mechanism for ecosystem services, linking bids to satellite-derived land use. Three-quarters of marginal auction winners are not additional. The heterogeneity in counterfactual land use introduces adverse selection. We develop a model of bidding and additionality to quantify welfare implications. Alternative auctions increase efficiency by using scoring rules that incorporate expected land use impacts.

This paper studies how managers plan under uncertainty. In a new panel survey of German manufacturing firms, we observe both forecasts of sales growth and a quantitative measure of subjective uncertainty. We show that subjective uncertainty reflects change experienced by firms: It is high when growth is either unusually low or unusually high. Subjective uncertainty is more than conditional volatility; while more volatile firms are typically more uncertain, experienced change alone increases uncertainty, consistent with a model of learning. Uncertainty over our sample is mostly idiosyncratic but varies substantially over time and leads managers to plan lower employment and prices.

We study the role of victim reporting in the production of public safety. We examine the Secure Communities program, a crime-reduction policy that involved police in detecting unauthorized immigrants and increased deportation fears in immigrant communities. We find that the policy reduced the likelihood that Hispanic victims report crimes to police and increased offending against Hispanics. The number of reported crimes is unchanged, masking these opposing effects. We show that reduced reporting drives the offending increase and provide the first elasticity of offending to victim reporting in the literature, calculating that a 10 percent decline in reporting increases offending by 7.9 percent.

Most organizations rely on managers to identify talented workers. However, managers who are evaluated on team performance have an incentive to hoard workers. This study provides the first empirical evidence of talent hoarding using personnel records and survey evidence from a large firm. Talent hoarding is self-reported by three-fourths of managers, is detectable in manager ratings of worker talent, and occurs more frequently under stronger hoarding incentives, proxied by performance-related pay, team size, and talent visibility. Using quasi-random exposure to talent hoarding, I show that hoarding deters internal job applications, inhibiting career progression and altering talent allocation in the firm.

We document the effects of a comprehensive set of mergers of consumer packaged goods manufacturers on prices, quantities, and product assortment. Across specifications, we find a small average price effect of mergers (−0.5−1.1 percent) but substantial heterogeneity, with a standard deviation between 3.9 and 7.6 percentage points. Through a model of enforcement, we find that agencies act as if they challenge mergers they expect would increase prices more than 4.8 to 6.3 percent. Increases in stringency would reduce prices and the prevalence of completed price-increasing mergers, with minimal impacts on blocked price-decreasing mergers, at significantly greater agency burden.

This paper studies the productivity of nurse practitioners (NPs) and physicians, two professions performing overlapping tasks but with starkly different backgrounds, training, and pay. Using quasi-experimental variation in patient assignment to NPs versus physicians in Veterans Health Administration emergency departments, we find that, on average, NPs use more resources and exhibit a higher 30-day preventable hospitalization rate than physicians. However, the NP-physician performance difference varies by case complexity and severity. Importantly, even larger productivity variation exists within each profession, leading to substantial overlap between the productivity distributions of the two professions; NPs outperform physicians in 38 percent of random pairs.

We study the problem of designing a dynamic reassignment mechanism when agents' preferences over objects change over time. In the context of Norway's system for (re)assigning patients to general practitioners (GPs), we provide direct evidence of misallocation under the current mechanism—patients waiting for each others' GPs, but who cannot trade—and estimate a structural model of GP-switching behavior to evaluate alternatives. Introducing top trading cycles (TTC) would, on average, reduce waiting times and increase patient welfare. However, patients endowed with less desirable GPs would be harmed. Prioritizing these patients can avoid these harms while preserving most of the gains from TTC.