An Analytics Approach to Guide Randomized Controlled Trials in Hypertension Management

Blood pressure (BP) is a significant controllable risk factor for cardiovascular disease (CVD), the leading cause of death worldwide. BP comprises two interrelated measurements: systolic and diastolic. CVD risk is minimized at intermediate BP values, a notion known as the J-curve effect. The J-curve effect imposes fundamental trade-offs in simultaneous management of systolic and diastolic BP; however, assessing a comprehensive set of joint systolic/diastolic BP treatment thresholds while explicitly considering the J-curve effect via randomized controlled trials (RCTs) is not feasible because of the time and cost-prohibitive nature of RCTs. In this study, we propose an analytics approach to identify promising joint systolic/diastolic BP threshold levels for antihypertensive treatment. More specifically, using one of the largest longitudinal BP progression data sets, we first build and fit Brownian motion processes to capture simultaneous progression of systolic/diastolic BP at the population level and externally validate our BP progression model on unseen data. We then analytically characterize the hazard ratio, which enables us to compute the optimal treatment decisions. Finally, building upon the optimal joint BP treatment thresholds, we devise a practical and easily implementable approximate policy. We estimate the potential impact of our findings through a simulation study, which indicates that the impact of explicitly considering the J-curve effect and joint systolic/diastolic BP in treatment decisions could be substantial. Specifically, we estimate that between approximately 3,000 and 9,000 premature deaths from cardiovascular disease in the United States could be prevented annually, a finding that could be tested empirically in randomized trials. This paper was accepted by Stefan Scholtes, healthcare management.

Experimentation and Start-up Performance: Evidence from A/B Testing

Recent scholarship argues that experimentation should be the organizing principle for entrepreneurial strategy. Experimentation leads to organizational learning, which drives improvements in firm performance. We investigate this proposition by exploiting the time-varying adoption of A/B testing technology, which has drastically reduced the cost of testing business ideas. Our results provide the first evidence on how digital experimentation affects a large sample of high-technology start-ups using data that tracks their growth, technology use, and products. We find that, although relatively few firms adopt A/B testing, among those that do, performance improves by 30%–100% after a year of use. We then argue that this substantial effect and relatively low adoption rate arises because start-ups do not only test one-off incremental changes, but also use A/B testing as part of a broader strategy of experimentation. Qualitative insights and additional quantitative analyses show that experimentation improves organizational learning, which helps start-ups develop more new products, identify and scale promising ideas, and fail faster when they receive negative signals. These findings inform the literatures on entrepreneurial strategy, organizational learning, and data-driven decision making. This paper was accepted by Toby Stuart, entrepreneurship and innovation.

Demand Estimation Using Managerial Responses to Automated Price Recommendations

We provide a new framework to identify demand elasticities in markets where managers rely on algorithmic recommendations for price setting and apply it to a data set containing bookings for a sample of midsized hotels in Europe. Using nonbinding algorithmic price recommendations and observed delay in price adjustments by decision makers, we demonstrate that a control-function approach, combined with state-of-the-art model-selection techniques, can be used to isolate exogenous price variation and identify demand elasticities across hotel room types and over time. We confirm these elasticity estimates with a difference-in-differences approach that leverages the same delays in price adjustments by decision makers. However, the difference-in-differences estimates are more noisy and only yield consistent estimates if data are pooled across hotels. We then apply our control-function approach to two classic questions in the dynamic pricing literature: the evolution of price elasticity of demand over and the effects of a transitory price change on future demand due to the presence of strategic buyers. Finally, we discuss how our empirical framework can be applied directly to other decision-making situations in which recommendation systems are used. This paper was accepted by Omar Besbes, revenue management and market analytics.

Dynamic Pricing for New Products Using a Utility-Based Generalization of the Bass Diffusion Model

The Bass Model (BM) has an excellent track record in the realm of new product sales forecasting. However, its use for optimal dynamic pricing or advertising is relatively limited because the Generalized Bass Model (GBM), which extends the BM to handle marketing variables, uses only percentage changes in marketing variables, rather than their actual values. This restricts the GBM’s prescriptive use, for example, to derive the optimal price path for a new product, conditional on an assumed launch price, but not the launch price itself. In this paper, we employ a utility-based extension of the BM, which can yield normative prescriptions regarding both the introductory price and the price path after launch, for the new product. We offer two versions of this utility-based diffusion model, namely, the Bass-Gumbel Diffusion Model (BGDM) and the Bass-Logit Diffusion Model (BLDM), the latter of which has been previously used. We show that both the BGDM and BLDM handily outperform the GBM in forecasting new product sales using empirical data from four product categories. We discuss how to estimate the BGDM and BLDM in the absence of past sales data. We compare the optimal pricing policy of the BLDM with the GBM and derive optimal pricing policies that are implied by the BLDM under various ranges of model parameters. We illustrate a dynamic pricing approach that allows managers to derive optimal marketing policies in a computationally convenient manner and extend this approach to a competitive, multiproduct case. This paper was accepted by Gui Liberali for the Management Science Special Issue on Data-Driven Prescriptive Analytics.

Public Concern About Immigration and Customer Complaints Against Minority Financial Advisors

We examine the relation between public concern about immigration and customer complaints against minority financial advisors in the United States. We find that minority advisors are more likely to receive complaints in periods of high public concern about immigration than in other periods, relative to their white colleagues from the same firm, at the same office location, and at the same point in time. This result holds for both complaints with merit and dismissed complaints and is more pronounced in counties where residents likely hold stronger anti-immigration views. We also find that minority advisors are more likely to face regulatory actions or leave their firms after customer allegations in periods of high public concern about immigration than in other periods. Overall, our study provides descriptive evidence of a positive relation between public concern about immigration and customer dissatisfaction with minority advisors. This paper was accepted by Suraj Srinivasan, accounting.

Patent Thickets, Stock Returns, and Conditional CAPM

Patent thickets, a phenomenon of fragmented ownership of overlapping patent rights, hamper firms’ commercialization of patents and thus deliver asset pricing implications. We show that firms with deeper patent thickets are involved in more patent litigations, launch fewer new products, and become less profitable in the future. These firms are also associated with lower subsequent stock returns, which can be explained by a conditional Capital Asset Pricing Model (CAPM) based on a general equilibrium model that features heterogeneous market betas conditional on time-varying aggregate productivity. This explanation is supported by further evidence from factor regressions and stochastic discount factor tests. This paper was accepted by Karl Diether, finance.

Antiracist Curriculum and Digital Platforms: Evidence from Black Lives Matter

In this paper, we examine the impact of racially charged events on the demand for antiracist classroom resources in U.S. public schools. We use book requests made by teachers on DonorsChoose, the largest crowdfunding platform for public school teachers, as a measure of intent to address race-related topics in the classroom. We use the precise timing of high-profile police brutality and other racially charged events in the United States (2010–2020) to identify their effect on antiracism requests relative to a control group. We find a significant increase in antiracism requests following the killing of George Floyd in 2020 and a null effect for all other events in the decade. We also find an increase in requests for books featuring Latinx, Asian, Muslim, and Jewish cultures, suggesting that a focus on equality for one group can spill over and yield culturally aware dialogues for other groups as well. Event studies suggest that local protests played a role in motivating some of the teachers to post these requests. In just four months following George Floyd’s death, $3.4 million worth of books featuring authors and characters from marginalized communities were successfully funded, reaching more than half a million students. Text analysis of impact notes posted by teachers suggests that hundreds of thousands of young students are being engaged in discussions about positive affirmation and cross-cultural acceptance. This paper was accepted by D.J. Wu, information systems.

Tax-Induced Inequalities in the Sharing Economy

The growth of sharing economy marketplaces like Airbnb has generated discussions on their socioeconomic impact and lack of regulation. As a result, most major cities in the United States have started to collect an “occupancy tax” for Airbnb bookings. In this study, we investigate the heterogeneous treatment effects of the occupancy tax policy on Airbnb listings, using a combination of a generalized causal forest methodology and a difference-in-differences framework. While we find that the introduction of the tax significantly reduces both listing revenues and sales, more importantly, these effects are disproportionately more pronounced for residential hosts with single shared-space (nontarget) listings versus commercial hosts with multiple properties or entire-space (target) listings. We further show that this unintended consequence is caused by customers’ discriminatory tax aversion against nontarget listings. We then leverage these empirical results by prescribing how hosts should optimally set prices in response to the occupancy tax and identify the discriminatory tax rates that would equalize the tax’s effect across nontarget and target listings. This paper was accepted by Victor Martínez-de-Albéniz, operations management.

Online Learning and Optimization for Revenue Management Problems with Add-on Discounts

We study in this paper a revenue-management problem with add-on discounts. The problem is motivated by the practice in the video game industry by which a retailer offers discounts on selected supportive products (e.g., video games) to customers who have also purchased the core products (e.g., video game consoles). We formulate this problem as an optimization problem to determine the prices of different products and the selection of products for add-on discounts. In the base model, we focus on an independent demand structure. To overcome the computational challenge of this optimization problem, we propose an efficient fully polynomial-time approximation scheme (FPTAS) algorithm that solves the problem approximately to any desired accuracy. Moreover, we consider the problem in the setting in which the retailer has no prior knowledge of the demand functions of different products. To solve this joint learning and optimization problem, we propose an upper confidence bound–based learning algorithm that uses the FPTAS optimization algorithm as a subroutine. We show that our learning algorithm can converge to the optimal algorithm that has access to the true demand functions, and the convergence rate is tight up to a certain logarithmic term. We further show that these results for the independent demand model can be extended to multinomial logit choice models. In addition, we conduct numerical experiments with the real-world transaction data we collect from a popular video gaming brand’s online store on Tmall.com. The experiment results illustrate our learning algorithm’s robust performance and fast convergence in various scenarios. We also compare our algorithm with the optimal policy that does not use any add-on discount. The comparison results show the advantages of using the add-on discount strategy in practice. This paper was accepted by J. George Shanthikumar, big data analytics.

Digitization, Prediction, and Market Efficiency: Evidence from Book Publishing Deals

Digitization has given creators direct access to consumers as well as a plethora of new data for suppliers of new products to draw on. We study how this affects market efficiency in the context of book publishing. Using data on about 50,000 license deals over more than 10 years, we identify the effects of digitization from quasi-experimental variation across book types. Consistent with digitization generating additional information for predicting product appeal, we show that the size of license payments more accurately reflects a product’s ex post success, and more so for publishers that invest more in data analytics. These effects cannot be fully explained by changes in bargaining power or in demand. We estimate that efficiency gains are worth between 10% and 18% of publishers’ total investments in book deals. Thus, digitization can have large impacts on the allocation of resources across products of varying qualities in markets in which product appeal has traditionally been difficult to predict ex ante. This paper was accepted by Joshua Gans, business strategy.