Publications

Nike, as a global leader in athletic footwear and apparel, experiences significant dilemmas when choosing which societal issues to confront. In the contemporary business landscape, stakeholders—including investors, customers, employees, and communities—expect companies to address a range of controversial topics such as racial justice, gender equality, climate change, and voting rights.

The challenges for Nike are multifaceted. Key decisions rest on whether an issue aligns with Nike’s “Purpose Pillars,” which include commitments to people, the planet, and play. These pillars guide Nike’s strategy and operations, helping determine whether to engage with a particular issue. For example, diversity and inclusion directly relate to Nike's core consumers and historical partnerships, making them integral to its brand identity.

However, Nike must navigate the risks associated with taking a public stance. Missteps can impact reputation, consumer loyalty, and sales. The company carefully assesses the potential impact on various stakeholders, including employees and athletes, before deciding to engage. This assessment is critical since different groups may have conflicting expectations or reactions to the company’s position on contentious issues.

A dedicated team, including CEO John Donahoe, evaluates these factors, exploring implications and weighing potential risks and benefits. This thorough vetting process ensures Nike’s responses are consistent with its core identity and capability to make a meaningful impact.

Nike, a global leader in athletic footwear and apparel, is committed to advancing sustainability through innovation. Its purpose is to inspire and innovate for every athlete while moving the world forward by building community, protecting the planet, and increasing access to sport. Responding to the industry's significant environmental impact, Nike has established ambitious goals under its "Move to Zero" initiative, aiming to power all owned facilities with 100% renewable energy by 2025 and divert 99% of its footwear manufacturing waste from landfills.

The Space Hippie project exemplifies Nike's commitment to sustainability by striving to create the lowest carbon footprint shoe in the company's history. In 2019, Nike tasked a small team with designing a shoe that would maintain performance, comfort, and aesthetic appeal while being manufacturable with readily available materials and contributing meaningfully to the company's climate goals. This project challenged Nike to reconsider all aspects of its established practices, from design and materials to manufacturing and marketing, emphasizing the importance of sustainability in innovation.

Networks across many different settings—including social, economic, and natural—are powerful tools for interventions due to the cascading impact of one individual node on others. All networks with degree variation exhibit the friendship paradox phenomenon. We demonstrate its multifaceted nature, and provide its foundations mathematically and empirically. We identify a network property—inversity—and propose network intervention strategies based on the friendship paradox. Inversity uniquely determines the best-performing strategy. These strategies provide a privacy-sensitive approach to obtaining highly connected individuals without knowing the network, and are guaranteed to obtain a greater than average degree for almost any network. Finally, we characterize the value of these strategies theoretically and with real-world networks.

In the optimal cohort partition problem, a planner seeks to distribute a heterogeneous population of individuals into a fixed number of distinct groups to optimize their objective. This problem is prevalent across a wide range of applications, including assigning students to classes, forming teams of experts in organizations, maximizing biodiversity, and designing contests among teams.

In this work, we approach partition problems from a majorization theory perspective and make two main contributions. At a conceptual level, we present a majorization-theoretic optimization framework for objectives that take a part-additive form, i.e., they can be written as a function of the summation vector of a partition. We show that the optimization of several common objectives reduces to a part-additive problem and characterize the structure of the optimal cohort partition policy across various objectives and applications.

At an application level, we focus on a common educational task---how to optimally partition students into classes in the presence of peer effects---and derive theoretical guarantees and new insights for this problem, thus making a theoretical contribution to the chiefly empirical literature on peer effects in education. In particular, we employ the workhorse empirical model, namely the linear-in-means (LIM) model, and study two different behavioral microfoundations proposed in the literature: the LIM spillover model and the LIM conformist model. We theoretically characterize the optimal partition for four distinct objectives---performance, welfare, diversity, and inequality.

We derive several policy-relevant insights. First, in both behavioral models, we show that as the planner's highest priority shifts from the top-performing class to the middle-performing class and then to the least-performing class, the optimal partition becomes progressively less assortative, ranging respectively from fully assortative to upper-uniform and then to fully uniform. Second, we show that, while both behavioral models yield identical optimal partition policies with respect to performance, diversity, and inequality, the welfare-optimal partition policy differs between the two LIM microfoundational models. Within the same behavioral model though, trade-offs are not always inevitable. For example, when student behavior aligns with the LIM conformist model and the planner weakly prioritizes the performance of the least-achieving classes, the uniform partition is optimal in most practical scenarios and across all four objectives. Third, we find that the uniform partition, despite achieving lower inequality than any integral partition, does not achieve perfect equality. We further illustrate these theoretical findings using calibrated simulations with student data from a novel dataset of public high schools.

Finally, we showcase how our theoretical framework extends beyond educational contexts by studying the optimal assignment of experts to teams, the design of Tullock contests, and non-linear peer effect models.

Reclassification risk is a major concern in health insurance where contracts are typ- ically one year in length but health shocks often persist for much longer. While most health systems with private insurers pair short-run contracts with substantial pricing regulations to reduce reclassification risk, long-term contracts with one-sided insurer commitment have significant potential to reduce reclassification risk without the negative side effects of price regulation, such as adverse selection. We theoreti- cally characterize optimal long-term insurance contracts with one-sided commitment, extending the literature in directions necessary for studying health insurance markets. We leverage this characterization to provide a simple algorithm for computing optimal contracts from primitives. We estimate key market fundamentals using data on all under-65 privately insured consumers in Utah. We find that dynamic contracts are very effective at reducing reclassification risk for consumers who arrive to the market in good health, but they are ineffective for consumers who come to the market in bad health, demonstrating that there is a role for the government insurance of pre-market health risks. Individuals with steeply rising income profiles find front-loading costly, and thus relatively prefer ACA-type exchanges. Switching costs enhance, while myopia moderately compromises, the performance of dynamic contracts.

Should policymakers adapt their macroprudential and monetary policies when the financial sector is vulnerable to belief-driven boom-bust cycles? I develop a model in which financial intermediaries are subject to collateral constraints, and that features a general class of deviations from rational expectations. I show that distinguishing between the drivers of behavioral biases matters for the precise calibration of policy: when biases are a function of equilibrium asset prices, as in return extrapolation, new externalities arise, even in models that do not have any room for policy in their ratio- nal benchmark. These effects are robust to the degree of sophistication of agents re- garding their future biases. I show how time-varying leverage, investment and price regulations can achieve constrained efficiency. Importantly, greater uncertainty about the extent of behavioral biases in financial markets reinforces incentives for preven- tive action.

In late 2023, Orange Grove Bio co-founder and CEO Marc Appel considered the future of his private drug development investment firm. Orange Grove Bio, (OGB) had finished its fourth- year nurturing promising early-stage drugs for new treatments of cancer, autoimmune, and inflammatory disease. The possibility of an influx of new capital was good. The company was thinking about launching additional rounds of private equity financing. Furthermore, Appel expected three of its portfolio companies to be cleared by the Federal Drug Administration for initial tests on people by 2025.

Teleconferencing with his executive team, CEO Appel gamed out possible future strategic directions for OGB. One possibility was building the company’s infrastructure. OGB could buy a contract research organization (CRO) to do bench work, sift through enormous data sets from lab tests, and conduct animal and human clinical trials on novel therapeutics. Having this technology in-house instead of paying outside companies for it would expand OGB’s capabilities and cut costs. Or the company could simply invest in building more wet labs for its portfolio companies.

Another possibility was building more relationships to increase OGB’s pipeline. The company had been built on the premise that valuable research to in-license was being ignored in universities located away from the biotech hot spots of Boston and San Francisco. OGB had invested in establishing connections with scientists in universities away from these locations, and even moved its headquarters to Cincinnati to take advantage of this opportunity. The team also had started looking internationally, considering relationships with universities and biotech firms outside the United States. However, creating productive alliances required resources and time. OGB had to consider the importance not only of the breadth but also the depth of its scientific alliances.

Appel also discussed the idea of starting an associated venture capital fund to leverage the firm’s relationships with investors and drug companies. OGB had been established to nurture promising scientific work ready for clinical trials on people. Its business model was to in-license biological discoveries from academia to build businesses around, with the plan of selling these for large profits to pharmaceutical companies when the underlying molecules showed promise of improving human health in tests on people. The company could also look further ahead in the development cycle and in-license medicines approaching the end of human clinical testing and possibly reap profits on the next blockbuster drug.

Or the company, for now, could just stand on its original business model and develop its existing portfolio of eight early-stage, preclinical subsidiaries. What would attract the most investor interest? OGB was building toward an eventual IPO, but in 2023, market conditions for biotech firms were deteriorating. The U.S. S&P 500 Biotechnology Industry index fell 50 percent in the third quarter of 2023 from its high in 2021.5 The sector was in its worst shape in 20 years, some analysts said.  And venture capital funding had sunk to a six-year low amid worries about high interest rates and inflation. Given the market conditions, what would be the smartest move?

Firms facing complex objectives often decompose the problems they face, delegating different parts of the decision to distinct sub-units. Using comprehensive data and internal models from a large U.S. airline, we establish that airline pricing is not well approximated by a model of the firm as a unitary decision-maker. We show that observed prices, however, can be rationalized by accounting for organizational structure and for the decisions by departments that are tasked with supplying inputs to the observed pricing heuristic. Simulating the prices the firm would charge if it were a rational, unitary decision-maker results in lower welfare than we estimate under observed practices. Finally, we discuss why counterfactual estimates of welfare and market power may be biased if prices are set through decomposition, but we instead assume that they are set by unitary decision-makers.

We develop a dynamic theory of “Partial Equilibrium Thinking” (PET), which micro-founds time-varying price extrapolation: extrapolative beliefs are present at all times, but only sometimes manifest themselves in explosive ways. To study this systematically, we formalize the distinction between normal times shocks and “dis- placement shocks” (Kindleberger 1978). In normal times, PET generates constant extrapolation, contrarian trading, and price momentum. Instead, following a dis- placement shock that increases uncertainty, PET leads to stronger and time-varying extrapolation with momentum trading, triggering bubbles and endogenous crashes. Our theory sheds light on both normal times dynamics and Kindleberger’s narrative of bubbles within a unified framework.

Using unique hand-collected data covering the political affiliation of 1,045 cities over the last two decades, this paper studies the implications of city-state partisan conflicts on the financing and provision of public goods. Cities with the same political affiliation as the state governor face 8 basis points lower borrowing costs, as compared to misaligned cities. These effects are stronger in states where governors are granted more powers, where cities are more fiscally dependent on the state, and for bonds issued by riskier borrowers. Consistent with state investments substituting for city-specific initiatives, we show that aligned cities reduce their investment in costly hazard preparedness projects when a same-party governor is elected.

Should Pepsi reconsider its long-term focus on the youth market for the company's signature brand?

For decades, PepsiCo, the largest food company in the United States, marketed its Pepsi-Cola brand with a primary focus on the youth demographic, leveraging advertisements featuring teens and young adults. However,  sales of the brand struggled; the company had seen its market share erode over the 2010s, dropping even faster than the overall cola segment. Amidst this backdrop, Todd Kaplan, PepsiCo's new Chief Marketing Officer for U.S. Pepsi, faced the challenge of revitalizing the brand:

Youth Targeting: Persist with Pepsi's traditional approach of appealing to youth, despite the drop in youth consumption of cola products and the inherent risks of marketing missteps. Look for new users to drink their brand.

Increasing Share of Wallet: Look to increasing consumption among existing customers, regardless of their age. 

Kaplan's team considered the options. What marketing strategy should Pepsi take to improve its brand position? Would Pepsi reconsider its long-term focus on the youth market for the company's signature brand? If so, what options would succeed in the market as well as with a skeptical internal audience?

We study personalized pricing in a general oligopoly model. The impact of personalized pricing relative to uniform pricing hinges on the degree of market coverage. If market conditions are such that coverage is high (e.g., the production cost is low, or the number of firms is high), personalized pricing harms firms and benefits consumers, whereas the opposite is true if coverage is low. When only some firms have data to personalize prices, consumers can be worse off compared to when either all or no firms personalize prices.

We consider a persuasion problem between a sender and a receiver where utility may be nonlinear in the latter’s belief; we call such receivers risk conscious. Such utility models arise when the receiver exhibits systematic biases away from expected utility maximization, such as uncertainty aversion (e.g., from sensitivity to the variance of the waiting time for a service). Because of this nonlinearity, the standard approach to finding the optimal persuasion mechanism using revelation principle fails. To overcome this difficulty, we use the underlying geometry of the problem to develop a convex optimization framework to find the optimal persuasion mechanism. We define the notion of full persuasion and use our framework to characterize conditions under which full persuasion can be achieved. We use our approach to study binary persuasion, where the receiver has two actions and the sender strictly prefers one of them at every state. Under a convexity assumption, we show that the binary persuasion problem reduces to a linear program and establish a canonical set of signals where each signal either reveals the state or induces in the receiver uncertainty between two states. Finally, we discuss the broader applicability of our methods to more general contexts, and we illustrate our methodology by studying information sharing of waiting times in service systems.

Residential Property Assessed Clean Energy (PACE) loans are a new class of financial contract, whereby homeowners borrow to fund green residential projects and repay the loan via their local property tax payments. We assess equity-efficiency trade-offs of PACE using loan-level data from Florida merged to property transaction, tax, and permitting records. Consistent with the program's objectives, borrowers are more likely to obtain permits related to disaster-proofing homes, and loan takeup is concentrated in areas with higher ex ante and ex post natural hazard risk. Such investments are capitalized into home values, but expansions of the property tax base are partially offset by an uptick in tax delinquency rates among borrowers. Although PACE loans are super senior to other debt, lenders expand their provision of mortgage credit in PACE-enabled counties. Enabling PACE loans increases the fiscal income of participating local governments while closing the investment gap in projects which improve the climate resiliency of the housing stock.

We study the classic divide-and-choose method for equitably allocating divisible goods be- tween two players who are rational, self-interested Bayesian agents. The players have additive values for the goods. The prior distributions on those values are common knowledge. We con- sider both the cases of independent values and values that are correlated across players (as occurs when there is a common-value component). We describe the structure of optimal divisions in the divide-and-choose game and identify several cases where it is possible to efficiently compute equilibria. An approximation algorithm is presented for the case when the distribution over the chooser’s value for each good follows a normal distribution, along with a randomized approximation algorithm for the case of uniform distributions over intervals. A mixture of analytic results and computational simulations illuminates several striking differences between optimal strategies in the cases of known versus unknown preferences. Most notably, given unknown preferences, the divider has a compelling “diversification” incentive in creating the chooser’s two options. This incentive leads to multiple goods being divided at equilibrium, quite contrary to the divider’s optimal strategy when preferences are known. In many contexts, such as buy-and-sell provisions between partners, or in judging fairness, it is important to assess the relative expected utilities of the divider and chooser. Those utilities, we show, depend on the players’ levels of knowledge about each other’s values, the correlations between the players’ values, and the number of goods being divided. Under fairly mild assump- tions, we show that the chooser is strictly better off for a small number of goods, while the divider is strictly better off for a large number of goods.

This paper studies an emerging subscription model called ship-then-shop. Leverag- ing its predictive analytics and artificial intelligence (AI) capability, the ship-then-shop firm curates and ships a product to the consumer, after which the consumer shops (i.e., evaluates product fit and makes a purchase decision). The consumer first pays the up-front ship-then- shop subscription fee prior to observing product fit and then pays the product price afterward if the consumer decides to purchase. We investigate how the firm balances the subscription fee and product price to maximize its profit when consumers can showroom. A key finding is the ship-then-shop firm’s nonmonotonic surplus extraction strategy with respect to its prediction capability. As prediction capability increases, the firm first switches from ex ante to ex post sur- plus extraction (by lowering fees and raising prices). However, if the prediction capability increases further, the firm reverts to ex ante surplus extraction (by raising fees and capping prices). We also find that the ship-then-shop model is most profitable when (i) the prediction capability is advanced, (ii) the search friction in the market is large, or (iii) the product match potential is large. Finally, we show that the marginal return of AI capability on the firm’s profit decreases in search friction but increases in product match potential. Taken together, we pro- vide managerially relevant insights to help guide the implementation of the innovative sub- scription model.

A signal is privacy-preserving with respect to a collection of privacy sets, if the posterior probability assigned to every privacy set remains unchanged conditional on any signal realization. We characterize the privacy-preserving signals for arbitrary state space and arbitrary privacy sets. A signal is privacy-preserving if and only if it is a garbling of a reordered quantile signal. These signals are equivalent to couplings, which in turn lead to a characterization of optimal privacy-preserving signals for a decision- maker. We demonstrate the applications of this characterization in the contexts of algorithmic fairness, price discrimination, and information design.

We examine property tax reduction as a tool for increasing housing affordability. Analyzing various tax reduction policies through the lens of property tax incidence reveals a complex relationship between affordability and property taxes, with differential effects across demographic groups. Many policies often fail to improve affordability for young first-time homebuyers and renters, sometimes worsening affordability. We present a new nationwide atlas documenting the prevalence of local measures altering property tax burdens. Quasi-experimental evidence from Georgia's homestead exemption valuation freezes suggests strong capitalization of assessment limits into home values, reinforcing that property tax relief may worsen affordability for first-time buyers.