Agency models in supply chain management

Jun-Yeon Lee, Purdue University


In this thesis we examine several business scenarios in which the owner (principal) of an inventory system, who is risk-neutral, delegates its design and/or implementation to an agent (internal manager, external supplier, or consultant) whose hidden effort influences the duration of the item's replenishment leadtimes. In the (Q, r) agency model, we examine two scenarios: the one-time-effort scenario and the combined-effort scenario. For both scenarios we demonstrate that if the agent is risk-neutral, then a simple contract achieves first-best. In the one-time-effort scenario, if the agent has an exponential utility function and the leadtime is drawn from a normal distribution, then we provide the parameters of the optimal linear contract. In the (S-1, S) agency model with Poisson demand and stochastic replenishment leadtimes, we provide the first-best optimal payment scheme, in which the agent is completely insured against the random timing and the random size of his payoffs. For the second-best case, if the agent has an exponential utility function and the leadtimes are drawn from a normal distribution, then we provide a salary-and-linear-bonus heuristic payment scheme, in which the agent receives a fixed payment in each of his consumption periods and a history-independent and linear incentive payment for each replenishment order. In the periodic-review agency model, we consider a standard, T-period inventory system, where a replenishment order is delivered immediately or one-period later, depending probabilistically on the effort expended by the agent. We provide the form of optimal payment scheme, and suggest that a state-dependent base-stock policy is optimal. We also find that, unlike in the corresponding traditional multi-period newsvendor model, given the same initial inventory, the optimal order-up-to levels may not be monotonic across time periods. For each of the three agency models, we conduct a computational study, and observe that ignoring the possible influence of the agent on replenishment leadtime can be significant; and that the magnitude of agency loss for the principal due to the agent's hidden effort appears to be small except in the periodic-review agency model with an agent who has a high risk-aversion. ^




Major Professor: Leroy B. Schwarz, Purdue University.

Subject Area

Business Administration, Management

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server