http://hdl.handle.net/2451/14093 Search the Channel search http://archive.nyu.edu/simple-search http://hdl.handle.net/2451/14632 Title: What Explains Superior Retail Performance?<br/><br/>Gaur, Vishal; Fisher, Marshall; Raman, Ananth<br/><br/>Abstract: We analyze the performance of retail firms for the period 1978-97 usingpublic financial data. Our performance measures are long-term stockreturns and whether the firm filed for bankruptcy in the period ofstudy. We assume that over a long time period of at least five years,stock returns are a reasonable measure of the overall success of a firm.We have found a very wide disparity in performance between firms. On theone hand, retailers like Wal-Mart, the Gap and Circuit City have hadphenomenal success (nineteen year compounded stock returns of 31.2%,29.5%, and 34.5%, respectively), while on the other, 17% of the publicretail firms filed for bankruptcy. We investigate how the followinglevers managed by the CEO of a retail firm affect performance: return onassets, sales growth, inventory turns, gross margin, financial leverage,and selling, general, and administrative expenses. The nature of theanalysis is contemporaneous, providing insights into managerial actionsthat correlate with success as measured by stock returns, but not aprediction of future stock returns. We find that (1) return on assets,sales growth, standard deviation of return on assets and financialleverage explain more than 50% of the variation in stock returns forperiods of ten years or more; (2) retailers in different segments&acirc;apparel, department stores, grocery and convenience stores, drugs andpharmaceuticals, jewelry, consumer electronics, home furnishings, toys,and variety stores&acirc; achieve similar return on assets and returnon equity by following very different strategies with respect to theirgross margins and inventory turns; (3) even within the same segment,high gross margin correlates with low inventory turns, and with highselling, general, and administrative expenses; (4) risk of bankruptcy isrelated to the mismatch between how fast a firm attempts to grow versuswhat growth rate it realizes. We also test for a negative correlationbetween sales growth and return on assets, which is widely believed tobe true but is not borne out by our data. http://hdl.handle.net/2451/26274 Title: Understanding Supplier Quality in Decentralized Supply Chains<br/><br/>Sheopuri, Dr. Anshul; Zemel, Dr. Eitan http://hdl.handle.net/2451/26275 Title: The Greed and Regret Problem<br/><br/>Sheopuri, Anshul; Zemel, Eitan http://hdl.handle.net/2451/26282 Title: Supply Contracts with Financial Hedging<br/><br/>Caldentey, Ren'e; Haugh, Martin B.<br/><br/>Abstract: We study the performance of a stylized supply chain where two firms, aretailer and a producer, compete in a Stackelberg game. The retailerpurchases a single product from the producer and afterwards sells it inthe retail market at a stochastic clearance price. The retailer,however, is budget-constrained and is therefore limited in the number ofunits that he may purchase from the producer. We also assume that theretailer's profit depends in part on the realized path or terminal valueof some observable stochastic process. We interpret this process as afinancial process such as a foreign exchange rate or interest rate. Moregenerally the process may be interpreted as any relevant economic index.We consider a variation (the flexible contract) of the traditionalwholesale price contract that is offered by the producer to theretailer. Under this flexible contract, at t = 0 the producer offers amenu of wholesale prices to the retailer, one for each realization ofthe financial process up to a future time . The retailer then commits topurchasing at time a variable number of units, with the specificquantity depending on the realization of the process up to time. Becauseof the retailer's budget constraint, the supply chain might be moreprofitable if the retailer was able to shift some of the budget fromstates where the constraint is not binding to states where it isbinding. We therefore consider a variation of the flexible contractwhere we assume that the retailer is able to trade dynamically between 0and in the financial market. We refer to this variation as the flexiblecontract with hedging. We compare the decentralized competitive solutionfor the two contracts with the solutions obtained by a central planner.We also compare the supply chain's performance across the two contracts.We find, for example, that the producer always prefers the flexiblecontract with hedging to the flexible contract without hedging.Depending on model parameters, however, the retailer may or may notprefer the flexible contract with hedging. Finally, we study the problemof choosing the optimal timing, of the contract, and formulate this asan optimal stopping problem. http://hdl.handle.net/2451/14638 Title: Staffing and Control of Large-Scale Service Systems with MultipleCustomer Classes and Fully Flexible Servers<br/><br/>Gurvich, Itay; Armony, Mor; Mandelbaum, Avishai<br/><br/>Abstract: We study large-scale service systems with multiple customer classes andmany statistically identical servers. The following question isaddressed: How many servers are required (staffing) and how does onematch them with customers (control) in order to minimize cost ormaximize profit, subject to quality of service (QoS) constraints? Wetackle this question by characterizing scheduling and staffing schemesthat are asymptotically optimal in the limit, as system load grows toinfinity. The main asymptotic regime considered is the many-serverheavy-traffic Quality and Efficiency Driven (QED) regime. TheEfficiency Driven (ED) regime is also studied. In the QED regime, whichwas formally introduced by Halfin and Whitt, a delicate balance isobtained between server efficiencies and quality of service. Thisbalance is enabled by the economies of scale associated with the systemsize. Our main findings are: a) Decoupling of staffing and control,namely (i) Staffing disregards the multi-class nature of the system andis analogous to the staffing of a single class system with the sameaggregate demand and the lowest priority class cost and QoS parameters,and (ii) Class level service differentiation is obtained by using asimple threshold-priority (TP) control (with state-independentthresholds), b) Robustness of the staffing and control rules: In theQED regime, our proposed Square-Root Safety (SRS) staffing rule and TPcontrol are asymptotically optimal with respect to various problemformulations and model assumptions. c) The QED and ED regimes areobtained as solutions of the joint staffing and control problem ratherthan as assumptions. http://hdl.handle.net/2451/14630 Title: Offshore &amp; Onshore contracts<br/><br/>Sheopuri, Anshul; Zemel, Eitan<br/><br/>Abstract: We study the buyer-seller relationship in a supply chain from theperspective of quality. In particular, we examine situations wherequality levels of the product can be observed, yet, conformance cannotbe enforced unless the buyer engages in costly, legally-bindingactivity, which we call audit. We assume that the cost of audit is borneby the buyer. We examine three types of remedies that can be taken incase an audit confirms the existence of quality problems, namely &acirc;Full Rebate&quot;, &acirc; Compensate&quot; and &acirc; Repair&quot;contracts. Under the Full Rebate regime, the seller rebates the buyerthe purchase price of defective units. Under the Compensate regime(Expectation Damages), the buyer is compensated for his loss of profitdue to defects. Finally, under the Repair regime (Specific Performance),the seller is made to repair all the defective units at his own expense.We examine the effects of these contracts on the performance of thesupply-chain, and point out the implications for offshore and onshoresupply networks, respectively. Our results indicate conditions where thebuyer is not motivated to &acirc; squeeze&quot; all the profits out ofthe supply chain, even in a setting of complete information, and even ifthe relationship is short-term. We extend the analysis to examinesupply-chains which include more than one seller or buyer. Also, as partof the analysis, we study a problem, which we call the &acirc; Crimeand Punishment&quot; problem that is interesting in its own right. http://hdl.handle.net/2451/26284 Title: Market Incompleteness and Super Value Additivity: Implications for Securitization<br/><br/>Gaur, Vishal; Seshadri, Sridhar; Subrahmanyam, Marti<br/><br/>Abstract: In an incomplete market economy, all claims cannot be priced uniquelybased on arbitrage. The prices of attainable claims (those that arespanned by traded claims) can be determined uniquely, whereas the pricesof those that are unattainable can only be bounded. We first show thattighter price bounds can be determined by considering all possibleportfolios of unattainable claims for which there are bid/offer prices.We provide an algorithm to establish these bounds. We then examine how aprice-taking agent can &ldquo;package&rdquo; new assets in order to takeadvantage of the incompleteness since the market places a premium onclaims that improve its spanning. In particular, we prove that a firmwith a new investment opportunity can maximize its value by&ldquo;stripping away&rdquo; the maximal attainable portion of the cashflow, for which prices are determined uniquely, and selling the balanceto investors at prices that preclude arbitrage. Our framework hasseveral applications in financial economics to problems ranging fromsecuritization to the valuation of real options. http://hdl.handle.net/2451/14633 Title: Linking Finance and Operations in Retailing<br/><br/>Fisher, Marshall; Gaur, Vishal; Raman, Ananth http://hdl.handle.net/2451/14636 Title: Estimating Demand Uncertainty Using Judgmental Forecasts<br/><br/>Gaur, Vishal; Kesavan, Saravanan; Raman, Ananth; Fisher, Marshall L.<br/><br/>Abstract: Measuring demand uncertainty is a key activity in supply chain planning.Of various methods of estimating the standard deviation of demand, onethat has been employed successfully in the recent literature usesdispersion among experts&acirc; forecasts. However, there has beenlimited empirical validation of this methodology. In this paper weprovide a general methodology for estimating the standard deviation of arandom variable using dispersion among experts&acirc; forecasts. Wetest this methodology using three datasets, demand data at item level,sales data at firm level for retailers, and sales data at firm level formanufacturers. We show that the standard deviation of a random variable(demand and sales for our datasets) is positively correlated withdispersion among experts&acirc; forecasts. Further we use longitudinaldatasets with sales forecasts made 3-9 months before earnings reportdate for retailers and manufacturers to show that the effects ofdispersion and scale on standard deviation of forecast error areconsistent over time. http://hdl.handle.net/2451/27725 Title: Dynamic Pricing for an M/M1 Make-To-Stock System with Controllable Backlog<br/><br/>Vulcano, Gustavo http://hdl.handle.net/2451/26279 Title: Duality Approaches to Economic Lot-Sizing Games<br/><br/>Chen, Xin; Zhang, Jiawei<br/><br/>Abstract: We consider the economic lot-sizing (ELS) game with general concaveordering cost. In this cooperative game, multiple retailers form acoalition by placing joint orders to a single supplier in order toreduce ordering cost. When both the inventory holding cost andbacklogging cost are linear functions, it can be shown that the core ofthis game is non-empty. The main contribution of this paper is to showthat a core allocation can be computed in polynomial time. Our approachis based on linear programming (LP) duality and is motivated by the workof Owen [19]. We suggest an integer programming formulation for the ELSproblem and show that its LP relaxation admits zero integrality gap,which makes it possible to analyze the ELS game by using LP duality. Weshow that, there exists an optimal dual solution that defines anallocation in the core. An interesting feature of our approach is thatit is not necessarily true that every optimal dual solution defines acore allocation. This is in contrast to the duality approach for otherknown cooperative games in the literature. http://hdl.handle.net/2451/26273 Title: Costly Enforcement of Quality Standards in Decentralized Supply Chains<br/><br/>Sheopuri, Anshul; Zemel, Eitan<br/><br/>Abstract: We consider a supply chain where the quality level can be observed bythe buyer(s) only after the purchase is completed (experience good). Ifthe delivered quality is below the levels agreed in the contract, thebuyer(s) may take action to seek remedy, but this effort is costly.Obviously, this fact can be exploited by the seller. In the case of asingle buyer, we show that the buyer may not be motivated to appropriateall the channel profits but, rather, is able to pay a higher purchaseprice in order to induce the seller to provide higher quality. The setof purchase prices that support trade split into different regions. Weidentify regions where counterintuitive behavior is exhibited, forexample, the buyer prefers higher purchase price, while the sellerprefers lower purchase price. For the case of more than one buyer, weexamine and contrast the behavior of two remedy regimes: In the case ofindividual enforcement, each buyer must work individually to enforce herown quality. In the case of joint enforcement, all buyers benefit fromenforcement by a single buyer. We examine the externalities that arisein this supply chain and identify how the market share, enforcementcosts and purchase price affect the quality level. http://hdl.handle.net/2451/26277 Title: Cost Allocation for Joint Replenishment Models<br/><br/>Zhang, Jiawei<br/><br/>Abstract: We consider the one-warehouse multiple retailer inventory model with asub modular joint setup cost function. The objective of this model is todetermine an inventory replenishment policy that minimizes the long-runaverage system cost over an infinite time horizon. Although the optimalpolicy for this problem is still unknown, a class of easy-to-implementpower-of-two policies are 98% effective. This paper focuses on how thecost, under an optimal power-of-two policy, should be allocated to theretailers. This question generates an interesting cooperative game. Weprove that this cooperative game has a non-empty core. The key to ourresult is a strong duality theorem for the one-warehouse multipleretailer problem under power-of-two policies. http://hdl.handle.net/2451/14637 Title: Convexity Properties and Comparative Statics for M/M/S Queues withBalking and Reneging<br/><br/>Armony, Mor; Plambeck, Erica; Seshadri, Sridhar<br/><br/>Abstract: We use sample path arguments to derive convexity properties of an M/M/Squeue with impatient customers that balk and renege. First, assumingthat the balking probability and reneging rate are increasing andconcave in the total number of customers in the system (head-count), weprove that the expected head-count is convex decreasing in the capacity(service rate). Second, with linear reneging and balking, we show thatthe expected lost sales rate is convex decreasing in the capacity.Finally, we employ a sample-path sub-modularity approach to comparativestatics. That is, we employ sample path arguments to show how theoptimal capacity changes as we vary the parameters of customer demandand impatience. We find that the optimal capacity increases in thedemand rate and decreases with the balking probability, but is notmonotone in the reneging rate. This means, surprisingly, that failure toaccount for customers&acirc; reneging may result in over-investment incapacity. Finally, we show that a seemingly minor change in systemstructure, customer commitment during service, produces qualitativelydifferent convexity properties and comparative statics. http://hdl.handle.net/2451/14634 Title: Asymmetric Consumer Learning and Inventory Competition<br/><br/>Gaur, Vishal; Park, Young-Hoon<br/><br/>Abstract: We develop a model of consumer learning and choice behavior in responseto uncertain service at the marketplace. Learning could be asymmetric,i.e., consumers may associate different weights with positive andnegative experiences. Under this consumer model, we characterize thesteady-state distribution of demand for retailers given that eachretailer holds constant in-stock service level. We then consider anon-cooperative game at the steady-state between two retailers competingon the basis of their service levels. Our model yields a unique purestrategy Nash equilibrium. We show that asymmetry in consumer learninghas a significant impact on the optimal service levels, market sharesand profits of the retailers. When retailers have different costs, italso determines the extent of competitive advantage enjoyed by the lowercost retailer. http://hdl.handle.net/2451/14770 Title: Assortment Planning and Inventory Decisions Under a Locational Choice Model<br/><br/>Gaur, Vishal; Honhon, Dorothee<br/><br/>Abstract: We consider a single-period assortment planning and inventory managementproblem for a retailer, using a locational choice model to representconsumer demand. We first determine the optimal variety, productlocation, and inventory decisions under static substitution, and showthat the optimal assortment consists of products equally spaced out suchthat there is no substitution among them regardless of the distributionof consumer preferences. The optimal solution can be such that somecustomers prefer not to buy any product in the assortment, and such thatthe most popular product is not offered. We then obtain bounds on profitwhen customers dynamically substitute, using the static substitution forthe lower bound, and a retailer-controlled substitution for the upperbound. We thus define two heuristics to solve the problem under dynamicsubstitution, and numerically evaluate their performance. This analysisshows the value of modeling dynamic substitution and identifiesconditions in which the static substitution solution serves as a good approximation. http://hdl.handle.net/2451/26276 Title: A Unified Theorem on SDP Rank Reduction<br/><br/>Man-Cho So, Anthony; Ye, Yinyu; Zhang, Jiawei<br/><br/>Abstract: We consider the problem of finding a low{rank approximate solution to asystem of linear equations in symmetric, positive semidefinite matrices.Specifically, let A1; : : : ;Am 2 Rn&pound;n symmetric, positivesemidefinite matrices, and let b1; : : : ; bm &cedil; 0. We show that ifthere exists a symmetric, positive semidefinite matrix X to thefollowing system of equations: http://hdl.handle.net/2451/26281 Title: A Stochastic Programming Duality Approach to Inventory Centralization Games<br/><br/>Chen, Xin; Zhang, Jiawei<br/><br/>Abstract: In this paper, we present a unified approach to study a class ofcooperative games arising from inventory centralization. Theoptimization problems corresponding to the inventory games areformulated as stochastic programs. We observe that the strong duality ofstochastic linear programming not only directly leads to a series ofrecent results concerning the non-emptiness of the cores of such games,but also suggests a way to find an element in the core. The proposedapproach is also applied to inventory games with concave ordering cost.In particular, we show that the newsvendor game with concave orderingcost has a non-empty core. Finally, we prove that it is NP-hard todetermine whether a given allocation is in the core for the inventorygames even in a very simple setting. http://hdl.handle.net/2451/26278 Title: A Heavy Traffic Approximation for Queues with Restricted Customer-Server Matchings<br/><br/>Caldentey, Ren&eacute; A.; Kaplan, Edward H.<br/><br/>Abstract: We consider a queuing system with n customer classes and m servers. Foreach class i there is only a subset S(i) of servers that are able toprocess customer' i requests and they do that using afirst-come-first-serve discipline. For this system, we are primarilyinterested in computing Pij , the steady-state fraction of class-icustomers that are served by server j. We also look at stabilityconditions and standard performance measures like waiting times andqueue lengths. Under the assumption that the system is heavy loaded, weapproximate Pij as well as the other performance measures. Computationalexperiments are used to show the quality of our approximations. http://hdl.handle.net/2451/27726 Title: A Column Generation Algorithm for Choice-Based Network Revenue Management<br/><br/>Bront, Juan Jose Miranda<br/><br/>Abstract: In the last few years, there has been a trend to enrich traditionalrevenue management models built upon the independent demand paradigm byaccounting for customer choice behavior. This extension involves bothmodeling and computational challenges. One way to describe choicebehavior is to assume that each customer belongs to a segment, which ischaracterized by a consideration set, i.e., a subset of the productsprovided by the firm that a customer views as options. Customers choosea particular product according to a multinomial-logit criterion, a modelwidely used in the marketing literature. In this paper, we consider thechoice-based, deterministic, linear programming model (CDLP) of Gallegoet al. [6], and the follow-up dynamic programming (DP) decompositionheuristic of van Ryzin and Liu [16], and focus on the more generalversion of these models, where customers belong to overlapping segments.To solve the CDLP for real-size networks, we need to develop a columngeneration algorithm. We prove that the associated column generationsubproblem is indeed NP-Complete, and propose a simple, greedy heuristicto overcome the complexity of an exact algorithm. Our computationalresults show that the heuristic is quite effective, and that the overallapproach has good practical potential and leads to high quality solutions.