Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TB23

3 - Correcting for Price Endogeneity in Demand Models: Linear Regression v. Discrete Choice

4 - A Tale of Timescales: Surge Pricing and Matching for Hotspot Demand Shock in Ride-Hailing Networks Zhe Liu, Columbia Business School, 3022 Broadway, New York, NY, 10027, United States, Philipp Afeche, Costis Maglaras We study a ride-hailing network that matches price- and delay-sensitive riders with strategic drivers. We consider surge pricing and spatial matching in response to an uncertain demand shock at a hotspot. We characterize optimal centralized and decentralized policies, and show how their performance depends on three key timescales, rider patience, shock duration, and drivers’ travel delay to the hotspot.

Stacey Mumbower, University of South Carolina, 1014 Greene Street, Columbia, SC, 29208, United States, Pelin Pekgun, Mark Ferguson The combination of capacity controls and price changes in the airline industry results in airfares that are heavily influenced by demand signals, leading to price endogeneity in demand models. Using an airline dataset to model flight-level demand as a function of daily prices, we demonstrate and compare two modeling approaches where price endogeneity can be corrected with instrumental variables. More specifically, we compare a linear regression modeling methodology to a discrete choice methodology and compare price elasticity estimates from both models. We highlight several key differences between the two approaches and discuss challenges that researchers encounter in model estimation. 4 - A Product-focused Approach to Dynamic Programming Decomposition for the Network Revenue Management and Pricing Problems Ravi Kumar, PROS Inc, Houston, TX, United States, Wei Wang, Darius Walczak We consider a dynamic programming formulation of the network revenue management problem. Due to the curse of dimensionality, various approximate dynamic programming methods based on the classical resource-based decomposition approach are used in practice. We revisit our earlier alternative idea of the product-based network decomposition and develop a novel dynamic model for the problem that relies on simultaneous solving of sub-problems. We present numerical analysis to compare the revenue and computation performance of the new method with the more classical ones. On-Demand Transportation Services Sponsored: Revenue Management & Pricing Sponsored Session Chair: Philipp Afeche, University of Toronto, Toronto, ON, M5S 3E6, Canada Co-Chair: Costis Maglaras, Columbia University, New York, NY, 10027, United States 1 - Ride Solo or Pool: The Impact of Sharing on Optimal Pricing of Ride-sharing Services Jagan Jacob, University of Rochester, Simon Business School, 4-349 Carol Simon Hall, Rochester, NY, 14627, United States, Ricky Roet-Green Using on-demand ride-sharing service providers (RSPs) such as Uber and Lyft, passengers can either ride solo and pay full fare, or share the ride with a fellow passenger (pooling) and pay a reduced fare. Though pooling is less expensive, cost of sharing is incurred from sharing the space with a possible stranger. We use stochastic modeling to study the passengers’ ride-choices at equilibrium, the RSP’s revenue-maximizing pricing strategies, the impact of the number of cars available, and social welfare. 2 - Economic Challenges in Designing a Sustainable Urban Mobility Marketplace Ragavendran Gopalakrishnan, Postdoctoral Associate, Cornell University, Ithaca, NY, United States, Chamsi Hssaine, Siddhartha Banerjee, Samitha Samaranayake Mobility-as-a-Service is becoming increasingly more efficient, convenient, and ubiquitous in urban areas, with the proliferation of multiple service providers across several modes. Instead of private service providers operating independently, a centralized marketplace in which they participate along with a public transit agency to collectively clear demand using multi-modal trips would enjoy a better operational efficiency, by increasing the quality of the mobility solutions offered to commuters. In this talk, we discuss the challenging task of finding an efficient mechanism/contract structure that would stabilize such a n TB22 North Bldg 130

n TB23 North Bldg 131A Financial Risk Management

Sponsored: Finance Sponsored Session Chair: Christoph Frei, University of Alberta, Edmonton, AB, T6G 2G1, Canada Co-Chair: Abel Cadenillas, University of Alberta, Edmonton, AB, T6G 2G1, Canada 1 - Bail-ins And Bailouts: Incentives, Connectivity, and Systemic Stability Benjamin Bernard, University of California, Los Angeles, 1440 Veteran Ave, Unit 344, Los Angeles, CA, 90024, United States, Agostino Capponi, Joseph E. Stiglitz We develop a framework for analyzing how banks can be incentivized to make contributions to a voluntary bail-in consortium. At the heart of the issue lies the credibility of the regulator’s threat to not bail out insolvent banks when no solvent bank agrees to contribute. We show that credible bail-in strategies exist if and only if the network hazard does not exceed a certain threshold. The threat is more credible and incentive-compatible contributions by banks are larger in more concentrated networks, making more them more desirable in the presence of bail-ins than more diversified networks. 2 - The Optimal Government Debt Ceiling When Interventions Are Bounded Abel Cadenillas, University of Alberta, Department of Mathematical Sciences, Edmonton, AB, T6G 2G1, Canada We develop a government debt management model to study the optimal government debt ceiling when the ability of the government to generate primary surpluses to reduce its debt ratio is bounded. We obtain an analytical solution for the optimal debt ceiling. [This is a joint work with Ricardo Huaman-Aguilar]. 3 - Contracting For Financial Execution Christoph Frei, University of Alberta, Mathematical and Statistical Sciences, CAB 621, Edmonton, AB, T6G 2G1, Canada, Markus Baldauf, Joshua Mollner Financial contracts often specify reference prices whose values are undetermined at the time of contracting, which introduces the possibility that one side manipulates them to the detriment of the other. Motivated by this, we study a model of financial contracting with the potential for manipulation of this nature, finding that a simple contract based on the volume-weighted average price (VWAP) emerges as uniquely optimal solution to a principal-agent problem between a client and her broker. This result explains the popularity of guaranteed VWAP contracts in equity trading and suggests a direction for improvement upon the status quo in other asset classes. 4 - Real Options With Performance-sensitive Debt Benoit Chevalier-Roignant, King’s College London, 30 Aldwych, London, London, WC2B 4BG, United Kingdom, Alain Bensoussan, Alejandro Rivera We consider a firm that decides on when and by how much to expand its production capacity. This firm is not financed solely by equity, but also by a performance-sensitive debt (PSD) instrument (Manso, Strulovici and Tchistyi, 2010). Debt financing induces here the possibility for shareholders—-to which benefits the management decides—-to default on their debt obligations (Leland, 1994). We investigate the interactions between the performance-sensitive coupon and the firm’s capacity expansion strategy as well as the extent to which PSD can mitigate the underinvestment problem on both the timing and sizing of the investment.

centralized mobility marketplace from an economic perspective. 3 - Fairness of Pricing and Service Priority in Last Mile Transportation Systems Yiwei Chen, University of Cincinnati, Cincinnati, OH, United States, Hai Wang

We consider a last-mile transportation system (LMTS) that consists of regular-type passengers and special-type passengers, such as senior. A special-type passenger who has a higher valuation of service usually has a lower waiting disutility. We show that with the objective of maximizing profit or social welfare, the LMTS operator always has incentive to give the special-type passengers the least service priority and charge them more than the regular-type passengers. This entails the necessity to impose fairness constraints on the price discount and service priority for the special-type passengers. We use real data in Singapore transportation system to quantify the analysis.

283