Informs Annual Meeting 2017

TE54

INFORMS Houston – 2017

4 - A Study on Prospect Theory and the Newsvendor Problem Abhishek Sharma, PhD Scholar, Indian Institute of Management Rohtak, Indian Institute of Management, MDU Campus, Rohtak, Haryana, Rohtak, 124001, India, aabhisheksharma01@gmail.com, Sanjay Jharkharia It is identified from the experimental studies on newsvendor problem that decision makers do not choose theoretical optimal order quantities, while using some heuristics to take their decisions. Prospect theory, as a theory for decision framing under uncertainty, has been used in literature to explain such preferences of decision makers. The recent literature shows that how these empirical findings can be explained through original prospect theory (OPT), when the reference point differs from the status quo. In this paper, we further study this argument and apply cumulative prospect theory (CPT) that accounts for the better modeling of psychological phenomenon of diminishing sensitivity. 362A Economics Contributed Session Chair: Jing Xu, University of Pennsylvania, Philadelphia, PA, United States, xjing@sas.upenn.edu 1 - Moral Perceptions of Advised Actions Alexander Gotthard Real, Assistant Professor, Pontificia Universidad Javeriana, Calle 59 # 3A - 24 Apt 504, Bogota, Colombia, agotthard@javeriana.edu.co Can an organization avoid some blame for an unpopular action by hiring an adviser to advise them to do it? We present experimental evidence suggesting this is the case advice to be selfish substantially decreases punishment of being selfish. Further, this result is true despite advisers conflicts of interest, known to all: Through a relational contract, advisers are motivated to tell the principals what they want to hear. In follow-up treatments, we show advice does not decrease punishment solely by affecting beliefs of how necessary the selfish action was. Finally, when advisers are available, selfish principals act more selfishly. 2 - Surfing or Working? Productivity Responses to Changes in Workload Erina Ytsma, Post-doctoral Associate, Massachusetts Institute of Technology, MIT IDE, 77 Massachusetts. Avenue, Cambridge, MA, 02139-4301, United States, eytsma@mit.edu, George F. Westerman We study the effect of workload on productivity, hours worked and internet use while at work using big data on insurance claims examiners. We find that, for every 1% increase in workload, productivity increases by 0.3% to 0.4%, tardiness decreases by 0.06 to 0.13 percentage points, and quality does not falter. Hours worked increase slightly, but extra work time can explain at most 10% of the increase in productivity. Non-productive internet use decreases with workload, while productive uses of internet increase. These findings are consistent with the presence of slack and an increase in efficiency in response to workload surges. 3 - Bargain Network and the Matchmaker Jing Xu, University of Pennsylvania, 209 S.33rd Street, Math Dept, Philadelphia, PA, 19104, United States, xjing@sas.upenn.edu I study a dynamic market with an intermediary that matches participants connected in a network to bargain over prices of heterogeneous goods. Players who reach an agreement pay a commission fee to the intermediary and exist from the network without replacement. The intermediary can choose its favorite matching pair each time before participants make the bargain. I establish the existence of Markov perfect equilibria (MPEs) of the infinite horizon game and illustrate the equilibrium matching probabilistic distributions adopted by the middleman. Moreover, when payoffs are homogeneous, any MPE achieves constrained efficiency. 362B Quantitative Methods in Financial Engineering Sponsored: Financial Services Sponsored Session Chair: Ning Cai, Hong Kong University of Science & Technology, Kowloon, Hong Kong, Hong Kong, ningcai@ust.hk 1 - Recursive Method for Static Options Replication of Complex Structured Products Kyoung-Kuk Kim, Associate Professor, KAIST, Dept of Industrial & Systems Engineering, 291 Daehak-Ro, Daejeon, 34141, Korea, Republic of, catenoid@kaist.edu, Dong-Young Lim This study aims to extend static options replication of financial derivatives to complex products with autocallable features and barrier features. The key idea is TE55 TE54

based on the combination of calendar-spread approach and strike-spread approach in a recursive manner. Such replicating strategies are exact, and we utilize the theory of integral equations. Our numerical examples include complex equity-linked structured products traded in Europe, US, and Korea. 2 - Investment Decisions and Falling Cost of Data Analytics Jussi Keppo, National University of Singapore, Mochtar Riady Building, BIZ 1 8-69, 15 Kent Ridge Drive, Singapore, 119245, Singapore, keppo@nus.edu.sg, Hong Ming Tan, Chao Zhou We study how the cost of data analytics and the characteristics of investors and investment opportunities affect investment decisions and their data analytics. We show that the falling cost of the data analytics raises investors’ leverage, financially constrained or highly risk-averse investors use less data analytics, the quantity of data analytics is highest with mediocre investment opportunities and it is lowest with a high expected return or a low risk opportunities. Due to the increased leverage, the falling cost of data analytics may lead to more frequent crises or higher losses during the crises. 3 - Susceptibility and Predictability in the NASDAQ Order Book James Primbs, California State University Fullerton, Fullerton, CA, United States, jprimbs@exchange.fullerton.edu, B. Ross Barmish, Sean Warnick, Scott Condi In this work, we analyze the structure and predictability of the NASDAQ order book based on ITCH transaction level data. Specifically, we explore the shape of the bid and ask sides of the order book over time, and apply various statistical and machine learning related tools to address questions of predictability and susceptibility to manipulation by adverse agents. 4 - A Structural Model for Liquidity Risk Ning Cai, Hong Kong University of Science & Technology, Room 5559D, Industrial Engineering, and Logistics Management, Kowloon, Hong Kong, Hong Kong, ningcai@ust.hk In this paper, we develop a structural model for liquidity risk and credit risk. Empirical studies illustrate that our model captures the features of the market observed credit spreads quite well. 362C Logistics Contributed Session Chair: Satya Malladi, Georgia Institute of Technology, Atlanta, GA, United States, mss@gatech.edu 1 - Tracking Technologies for the Efficient Operations of Distribution Center Yong-Hong Kuo, Assistant Professor, The University of Hong Kong, Department of Industrial and Manufacturing Systems Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong, yhkuo@hku.hk, Chun Hung Cheng, Matthew Petering Our work was motivated by a real-world problem in a large air mail centre. We have developed an RFID-based location-positioning platform for the application.The RFID-tag attachments to carts and reader mounts at the top of the building allow the facility operators to locate target carts quickly. We have also conducted an initial analysis of data collected using this technological infrastructure to understand issues on cart movements and operational efficiency. 2 - Analyzing the Benefits of Optimal Dispatching in Shipment Consolidation Benhur Satir, Assistant Professor, Dr., Çankaya University, Etimesgut, Ankara, 06790, Turkey, benhur@cankaya.edu.tr, F. Safa Erenay, James H. Bookbinder In practice, most logistics providers dispatch consolidated loads at pre-planned periods, i.e. using a time policy. In this study, we investigate the relative benefit and conditions leading to it when the optimal policy is used instead. The benefit is measured in terms of expected discounted total cost. We model the problem as a Continuous Time Markov Decision Process, and assume compound Poisson arrivals of two customer types, viz. expedited and regular. We prove the existence of quantity-based optimal threshold policies under particular conditions, and develop an efficient solution approach for large problem instances. We provide numerical studies using two industrial cases. TE56

430