Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TA23

general nonconvex costs, and allows using general parametric price functions. The flexibility of our scheme allows finding prices that are typically more economically-efficient and less discriminatory. We supplement the proposed method with a polytime approximation algorithm for finding the optimal quantities and parameters. 2 - Efficiency and Information Aggregation in Heterogeneous Markets Yaarit Even, Columbia Business School, New York, NY, United States, Alireza Tahbaz-Salehi, Xavier Vives This paper studies how heterogeneous preferences shape the informational and allocative efficiency of centralized markets with asymmetric information. We show that introducing agent-level heterogeneity to the standard rational expectations equilibrium models reduces price informativeness. This reduction in price informativeness in turn manifests itself as an informational externality: in the presence of heterogeneity, agents do not internalize the impact of their trading decisions on the information revealed to others via prices. We conclude by investigating the welfare implications of market segmentation in the presence of this informational externality. 3 - Social Learning with Product Choice Stefano Vaccari, LUISS University, Via Pietro Ceccato, 15, Roma, 00156, Italy, Costis Maglaras, Marco Scarsini We study a model where consumers infer the quality of a set of products through online reviews, and make their subsequent product choice according to an MNL model. Consumers that purchase, write reviews based on their experienced quality. We study the impact of choice on the learning outcome and the speed of learning. We explore the platform display order optimization problem in a setting where consumers have search costs. Sponsored: Finance Sponsored Session Chair: Zachary Feinstein, Washington University in St. Louis, St. Louis, MO 1 - Systemic Risk in a Multilayered Network Zachary Feinstein, Washington University in St. Louis, St. Louis, MO, 63130, United States In this talk we provide a general framework for modelling financial contagion in a system with obligations in multiple illiquid assets, e.g., currencies. In so doing, we develop a multi-layered financial network that extends the single network of Eisenberg & Noe (2001). In particular, we develop a financial contagion model with fire sales that allows institutions to both buy and sell assets to cover their liabilities and act as utility maximizers. Under standard assumpsions, equilibrium portfolio holdings and market prices exist which clear the multi-layered financial system. 2 - A Time-dynamic Eisenberg-Noe Network Model Tathagata Banerjee, Washington University in St. Louis, St. Louis, MO, United States, Alex Bernstein, Zachary Feinstein We will consider a generalized extension of the Eisenberg-Noe model of financial contagion to allow for time dynamics in both discrete and continuous time. We will provide the derivation and the interpretation of the financial implications. Emphasis will be placed on the continuous-time framework and its formulation as a differential equation driven by the operating cash flows. Mathematical results on existence and uniqueness of firm wealths under the discrete and continuous- time models will be provided. We will discuss the financial implication of the time dynamics with the focus on how the dynamic clearing solutions differ from those of the static Eisenberg-Noe model. 3 - Systemic Risk and Trade Compression Stephan Sturm, WPI, Department of Mathematical Sciences, 100 Institute Road, Worcester, MA, 01609, United States Since the financial crisis 2008 the practice of trade compression has taken off in over-the counter derivatives markets. We analyze the impact of different trade compression schemes on the systemic risk in derivatives markets. 4 - Systemic Risk Measures and Portfolio Choice Alexey Rubtsov, Global Risk Institute, Toronto, ON, Canada, Agostino Capponi, Alex LaPlante We develop a model for the optimal portfolio choice in presence of systemic risk. In our modelling approach we use two risk measures: VaR and CoVaR (Adrian and Brunnermeir (2011)). Our investor maximizes portfolios expected returns conditioned on a systemic risk index being at (at most at) its VaR level and portfolios returns below their CoVaR level. Under some assumptions the optimal investment strategy is derived in closed form. The proposed methodology is flexible in balancing the significance of portfolios variance and portfolios correlation with systemic risk index. n TA23 North Bldg 131A Systemic Risk and Network Models

n TA21 North Bldg 129B

Using Estimation Techniques and Field Experiments to Optimize Operations and Marketing Decisions Sponsored: Revenue Management & Pricing Sponsored Session Chair: Spyros Zoumpoulis, Paris, 75005, France 1 - On the Learning Benefits of Resource Flexibility Mihalis Markakis, Universitat Pompeu Fabra, Barcelona, Spain, Jiri Chod, Nikolaos Trichakis We consider a firm that acquires resources to sell two products/services. We compare the demand learning under two designs: one with non-flexible resources and one with a single flexible resource. Our analysis reveals that resource flexibility improves learning, unless the products/services have significantly different demand characteristics or economic parameters, leading to an important managerial implication: flexibility is not just the “ability to react;” it also provides better demand information, which can be used pro-actively for better resource planning, pricing, and assortment decisions. 2 - Setting Retail Staffing Levels: A Methodology Validated with Implementation Santiago Gallino, Dartmouth College, 100 Tuck Hall, Hanover, NH, 03755, United States, Marshall L. Fisher, Serguei Netessine We describe a three-step process that a retailer can use in setting retail store sales staff level. First, use historical data on revenue and planned and actual staffing levels by store to estimate how revenue varies with staffing level at each store. We disentangle the endogeneity between revenue and staffing levels by focusing on randomly occurring deviations between planned and actual labor. Second, using historical analysis as a guide, validate these results by changing the staffing levels in a few test stores. Finally, implement the results chain-wide and measure the impact. We describe the successful deployment of this process with a large specialty retailer. 3 - Targeting Prospective Customers Using Field Experiments: Efficient Causal Inference and Robustness of Machine Learning Methods to Data Challenges Spyros Zoumpoulis, INSEAD, Fontainebleau, France, Duncan I. Simester, Artem Timoshenko We investigate how firms can use the results of field experiments to optimize the targeting of promotions when prospecting for new customers. We propose an approach to designing and analyzing field experiments that has substantive efficiency advantages over standard approaches for experimenting and evaluating targeting policies. We also evaluate seven widely used machine learning methods using a series of two large-scale field experiments, and discuss how well the methods address common data challenges. 4 - Spatial Pricing in Ride-sharing Networks Kostas Bimpikis, Stanford Univeristy, Ozan Candogan, Daniela Saban We explore spatial price discrimination in a ride-sharing platform that serves a network of locations.Riders are heterogeneous in their destinations and their willingness to pay. Drivers decide whether, when, and where to work to maximize their earnings, given the platform’s prices. Our findings highlight the impact of the demand pattern on prices, profits, and the induced consumer surplus. We establish that profits and consumer surplus are maximized when the demand pattern is balanced across the network. Also, we show that they both increase with the balancedness of the demand If demand is not balanced, the platform can benefit substantially from pricing rides differently based on their origin.

n TA22 North Bldg 130 Design of Online Marketplaces Sponsored: Revenue Management & Pricing Sponsored Session Chair: Kostas Bimpikis, Stanford Univeristy 1 - Optimal Pricing in Markets with Non-convex Costs

Navid Azizan, California Institute of Technology, Pasadena, CA, United States, Su Yu, Krishnamurthy Dvijotham, Adam Wierman We consider a market run by an operator, who seeks to satisfy a given consumer demand by purchasing it from a group of suppliers with non-convex cost functions. The operator knows the cost functions and announces a price function for each supplier. Each supplier then makes an individual decision about how much to produce. We propose a new pricing scheme, which is applicable to

251