Publications

We study regressions with multiple treatments and a set of controls that is flexible enough to purge omitted variable bias. We show that these regressions generally fail to estimate convex averages of heterogeneous treatment effects—instead, estimates of each treatment’s effect are contaminated by non-convex averages of the effects of other treatments. We discuss three estimation approaches that avoid such contamination bias, including the targeting of easiest-to-estimate weighted average effects. A re-analysis of nine empirical applications finds economically and statistically meaningful contamina- tion bias in observational studies; contamination bias in experimental studies is more limited due to idiosyncratic effect heterogeneity

This paper considers the problem of offering a scarce object with a common unobserved quality to strategic agents in a priority queue. Each agent has a private signal over the quality of the object and observes the decisions made by other agents. We first show that, under the widely-used first-come-first-served sequential offering mechanism, herding behavior emerges: initial rejections create an information cascade resulting in inefficient waste. To address this issue, we then introduce a class of batching mechanisms. Agents in each batch report whether they would be willing to accept or reject the object based on their private signals and prior information. If the majority opts to accept, the object is randomly allocated within that batch. We prove that suitable batching mechanisms are incentive-compatible and improve efficiency. A key property of the mechanism is the gradual increase of the batch size after each failed allocation; the size is chosen so that it elicits as much information as possible without distorting the incentives of agents to report truthfully. Additionally, from a healthcare policy perspective, our results can shed light on the large wastage in organ allocation. In particular, wastage that arises due to herding may be reduced by applying adaptive simultaneous offering mechanisms.

This paper examines the current state of social innovation and entrepreneurship programming, courses, and centers within schools of public health through a survey data analysis. This report presents a cross-sectional survey conducted among faculty members of public health schools in the United States. The survey aims to determine the availability and current state of student-centered programs and courses centered around social innovation and entrepreneurship within schools of public health. Insights were drawn from 19 professionals across 15 schools of public health. Uncertainties surround the sustainability of current programs, with insufficient funding, human resources, and the need to teach more pressing topics identified as the most significant obstacles. Key areas identified as opportunities for growth were faculty engagement, expertise, and funding to expand more structured programming.

Using data from 15 countries, this article investigates whether descriptive and prescriptive gender norms concerning housework and child care (domestic work) changed after the onset of the COVID-19 pandemic. Results of a total of 8,343 participants (M = 19.95, SD = 1.68) from two comparable student samples suggest that descriptive norms about unpaid domestic work have been affected by the pandemic, with individuals seeing mothers’ relative to fathers’ share of housework and child care as even larger. Moderation analyses revealed that the effect of the pandemic on descriptive norms about child care decreased with countries’ increasing levels of gender equality; countries with stronger gender inequality showed a larger difference between pre- and post-pandemic. This study documents a shift in descriptive norms and discusses implications for gender equality—emphasizing the importance of addressing the additional challenges that mothers face during health-related crises.

n response to intense pressure from policy makers and the public, technology companies have enacted a range of policies aimed at reducing the spread of misinformation online 1 - 3 . The enforcement of these policies has , however, led to technology companies being regularly accused of political bias 4 - 6 . W e argue that even under politically neutral anti - misinformation policies, such political asymmetries in enforcement should be expected , as there is a political asymmetry in the sharing of misinformation 7 - 12 . We support this argument with an analysis of Twitter data from 9,000 politically active users during the U.S. 2020 presidential election . While users on the political right were indeed substantially more likely to be suspended than those on the left , users on the right also shared far more links to low quality news sites – even when news quality was determined by politically - balanced groups of laypeople , or groups of only Republicans – and were estimated to have a far higher likelihood of being bots. We find similar associations between conservatism and low quality news sharing (based on both expert and politically - balanced layperson ratings) in seven other dat asets of sharing from Twitter, Facebook, and survey experiments, spanning 2016 to 2023 . These results demonstrat e that political im balance in enforcement need not imply bias, and should not dissuade technology companies from taking action against the spread of misinformation

This paper provides a framework in which a multiproduct ecosystem competes with small single-product firms in both price and innovation. The ecosystem is able to use data collected on one product to improve the quality of its other products. We study the impact of regulation which either restricts the use of data across the ecosystem’s business units, or which requires the ecosystem to share data with small firms. This regulation induces small firms to innovate more and set higher prices; it also dampens data spillovers within the ecosystem, reduces the ecosystem’s incentive to innovate, and potentially increases its prices. As a result, the impact of data regulation on consumers is ambiguous. Small firms do not necessarily benefit from sharing data with each other via a data cooperative, because doing so triggers more aggressive pricing by the ecosystem. A data cooperative can also harm consumers by inducing the ecosystem to innovate less.

We provide large-sample evidence on whether U.S. publicly traded corporations opportunistically use voluntary disclosures about their commitments to employee diversity. We document significant discrepancies between companies' disclosed commitments and their hiring practices and classify firms that discuss diversity more than their actual employee gender and racial diversity warrants as “diversity washers." We find diversity-washing firms obtain superior scores from environmental, social, and governance (ESG) rating organizations and attract investment from institutional investors with an ESG focus. These outcomes occur even though diversity-washing firms are more likely to incur discrimination violations and pay larger fines for these actions. Our study highlights the consequences of selective ESG disclosures on an important social dimension of employee diversity, equity, and inclusion.

Several firms claim to be socially responsible. We confront these claims with data using the most notable recent proclamation, the Business Roundtable’s (BRT) 2019 Statement on the Purpose of a Corporation. The BRT is a large, influential business group containing many of America’s largest firms; the Statement proclaimed a corporation’s purpose as delivering value to all stakeholders, rather than only shareholders. However, we find no evidence that signatories – who voluntarily signed – engaged in such stakeholder-centric practices before or after signing. Relative to peers, signatories violate environmental and labor laws more frequently, have higher carbon emissions, rely more on government subsidies, and are more likely to disagree with proxy recommendations on shareholder proposals. We also do not observe post-signing improvements along these dimensions, suggesting that the Statement was not a credible commitment to improve. Our results suggest that firms’ proclamations of stakeholder-centric behavior are not backed up by hard data.

This paper examines the effect of state-level monitoring on municipal governance, focusing on outcomes in financial reporting quality, local corruption, political entrenchment, and municipal financial soundness. I exploit the staggered adoption of fiscal monitoring policies that entail a regular review of municipal financial reports for signs of fiscal distress. I find that introducing these monitoring policies is associated with an increase in the proxies for reporting quality, a decrease in the number of corruption convictions, and a reduction in re-election likelihood for incumbent politicians. Consistent with the purpose of the policies, my evidence shows that fiscal health ratios of municipalities improve after initiating state monitoring. Collectively, my results are consistent with state fiscal monitoring improving several important aspects of municipal governance.

This paper proposes a tractable model of the Global Financial Cycle and study its welfare implications for emerging market economies (EMEs). When local firms issue debt denominated in dollars, central banks must increase their policy rate as the U.S. tightens in order to offset balance sheet effects stemming from the depreciation of their currency. When global financial mar- kets are imperfect, this synchronized policy response has negative spillovers: all individual countries seek to attract capital inflows at the expense of one another, exacerbating the Global Financial Cycle. This congestion externality requires further tightening and results in inefficiently low levels of output and employment in EMEs, and generates gains from coordination. On the con- trary, discouraging debt issuance in dollars through macroprudential policy has positive spillovers, and does not necessarily require coordination between EMEs. Its optimal use dampens the Global Financial Cycle and its inefficien- cies.

We introduce a model of dynamic pricing in perishable goods markets with competition and provide conditions for equilibrium uniqueness. Pricing dynamics are rich because both own and competitor scarcity affect future profits. We identify new competitive forces that can lead to misallocation due to selling units too quickly: the Bertrand scarcity trap. We empirically estimate our model using daily prices and bookings for competing U.S. airlines. We compare competitive equilibrium outcomes to those where firms use pricing heuristics based on observed internal pricing rules at a large airline. We find that pricing heuristics increase revenues (4-5%) and consumer surplus (3%).

We investigate an unexplored mechanism of earnings management: income shifting from not-wholly-owned subsidiaries to help the parent company avoid losses at the expense of subsidiaries. Consolidated net income attributable to the parent company (i.e., net income) increases through this mechanism, as the parent company enjoys the full amount of the shifted earnings rather than sharing them with minority investors. We design an empirical model to directly estimate the amount of income shifted from subsidiaries to parent firms. Employing this measure, we find that firms opportunistically decrease earnings of their not-wholly-owned subsidiaries to manage net income upward to avoid losses. The results are stronger for firms with high noncontrolling ownership, firms with large subsidiaries, firms with strong influence over not-wholly-owned subsidiaries, and firms with a high level of related-party transactions. Our results are robust to alternative research designs, including controls for within-firm variations, alternative earnings thresholds, propensity score matching, and entropy balancing techniques. Our mechanism of earnings management is generalizable to other earnings management scenarios, such as share pledging.

Violence against civilians has killed nearly one million people worldwide and displaced
millions more over the past three decades. This paper examines the economic forces driving
civilian persecution in Myanmar. Using a difference-in-differences approach, we show that
violence against civilians increases in rice-suitable townships when rice prices rise, the op-
posite pattern of that documented in the literature on two-sided conflicts. We argue that
large power asymmetries inherent in civilian persecution explain this difference. Higher re-
turns from expropriating rice harvests and rice-growing inputs during these periods drive
the pattern we observe in Myanmar, which we corroborate with an original survey of Ro-
hingya refugees forcibly displaced to Bangladesh. Our work demonstrates how to generate
systematic and representative evidence on civilian persecution in politically sensitive and
data-poor contexts.

Elemental Excelerator is a not-for-profit fund established in 2017 by Dawn Lippert. With $36 million of funding, the organization focuses on investing in for-profit businesses that aim to mitigate climate change while embedding equity and social justice into their solutions. 

The fund faces a dilemma in selecting the last companies for its 2023-2024 cohort of climate tech businesses. With only one or two slots remaining, Elemental must decide how to balance its portfolio. Options include prioritizing technologies in the electric power and transportation sectors favored by the Inflation Reduction Act, or continuing to diversify in buildings, agriculture, and industry, sectors with few sustainable climate solutions. Another consideration is whether to focus on companies that already attract venture capital or support those overlooked by investors.  Yet another concern is balancing sustainability with social justice. Six very different companies are presented as finalists.

The DMA raises tantalising opportunities for app stores innovation, making it the most exciting area of digital regulation.

Facing the 2018-19 school year, Michael Garcia, the budget director of the Spring Valley School District in California (a disguised name), confronted challenges in changing the district’s accounting system to comply with the federal Every Student Succeeds Act (ESSA).

A successor to the Elementary and Secondary Education Act (also known as the No Child Left Behind Act), ESSA was signed by President Obama in 2015. ESSA governed multiple programs providing federal funding to states, as well as requirements for school district reporting to the U.S. Department of Education. Among other revisions, ESSA required for the first time that districts report the per-pupil spending for each school in the district, including allocated central costs. It also required that the data be available to the public. This requirement was designed to give parents and community members more knowledge about relative spending for the schools within the district.

Before ESSA, Spring Valley, like many districts, had done its budgeting on a functional basis, dividing district-wide spending into various operational areas, like instruction, maintenance, and central office costs. However, it only had to report a consolidated number for district-wide costs: the total district-wide costs divided by the number of students in the district. Faced with the new school-level reporting requirement, Garcia was inclined to track the school-level costs in as much detail as possible. Spring Valley schools were a heterogeneous collection of old and new facilities in a city with widely disparate income levels and educational needs, and Garcia believed that additional information could only help the district and its students.

However, Jessica Adams, the newly installed superintendent, clashed with Garcia on how to implement several ESSA accounting requirements. In her short tenure, Adams had already faced criticism from parents at several schools. She did not want the school-level spending report required by ESSA to become another divisive flashpoint in the already contentious school board meetings. In private conversations, she told Garcia that the new allocation scheme should show as little spending difference as possible between the schools.

Garcia and Adams had to resolve several allocation issues. One concern was allocating staff salaries and related benefits to individual schools. These costs were a large portion of the district budget – Spring Valley spent around 81% of its total operational budget on employee salaries and related benefits for instructional, support, and administrative costs.

Another area of concern was allocating the district-wide costs related to the maintenance of plant facilities. Should they allocate maintenance of plant facilities by school square footage, or look for more granular indicators of the actual costs in each school? Finally, what about busing costs? Should they allocate costs to each school on a per-pupil basis across the system or assign busing costs to each location?

Garcia understood Adams’s reluctance to implement the more detailed approach, but he still thought a more granular approach could provide valuable information. The Spring Valley School District faced several important issues in the next few years. A new contract with the teachers’ union was looming. Declining enrollments meant that the district might need to close one or more schools. Would a more precise calculation of costs for each school lead to better decisions?

We demonstrate how to obtain the distribution of consumer willingness to pay (WTP) for digital subscription products, where consumers pay a fixed price each period for potentially unlimited usage, for example, music streaming like Spotify. Typically, in such applications, usage data are observed and is critically valuable for the method here. We demonstrate how variation in usage and subscription choice together can identify the WTP distribution in the absence of price variation. Our framework accommodates and builds upon a range of utility specifications for usage, which is related to subscription decisions. We provide the conditions required on exogenous variation impacting usage, and prove how these lead to identification of the WTP distribution. We also investigate the conditions under which usage variation is not equivalent to price variation. We apply our method to an empirical application using the data from a music streaming service. Using the estimated WTP distribution, we obtain the revenue maximizing prices for different consumer segments.