This dissertation presents a prototype model and methodology for validating a simple agent-based model against the basic Hotelling monopoly model and a few basic extensions. Hotelling's Rule identifes the expected behavior of a market for a nonrenewable resource. The statement is simple - marginal profit will increase at the prevailing rate of interest - but the implications are far-reaching and not broadly understood. Agent-based modeling is a computer modeling and simulation methodology. It has its origins in biology and physics, but has become a powerful tool in the social sciences for examining systems in which the well-understood behaviors of individuals result in unanticipated outcomes. Validation of the basic Hotelling monopoly model is a necessary step in wider acceptance of agent-based modeling as a predictive and analytical tool in natural resource economics. An agent-based model is a valid predictive tool if, given rules to express preferences, it is possible to predict the large-scale outcomes of the choices made by individuals over time. An agent-based model is a valid analytical tool if it provides a means to explore the behaviors that lead to known results, much like nonlinear regression. This simple agent-based model is found to be valid for the basic Hotelling monopoly model. The agent-based model is validated with caveats for the Hotelling monopoly model with extensions to include basic production technologies. The caveats are based on small deviations, the magnitudes of which depend on the specific form of costs associated with a production technology. It is argued that those deviations are not unlike the deviations that a human would make. An extension of the Hotelling monopoly model to a small oligopoly exhibits emergent cooperation-like properties, despite the absence of explicit interagent communication. Depending on the number of producers and the initial distribution of resource stocks, the behavior is either collusion-like or Cournot-like. The Cournot-like outcome occurs when only some of the producers lower production, resulting in a rise in the market price, which causes the other producers to experience a Hotelling's Rule increase in marginal profit without reducing their own production. This continues at each time step, so that the latter producers maintain a constant, higher production level while the others continue to decrease production. The outcomes of the agent-based models are reassessed with the costs previously arising from production technologies replaced by taxes associated with fiscal policies. Each fiscal regime is evaluated in terms of its efficacy and the unintended consequences of the policy. The policy goals examined are preservation for future generations, internalization of externalities (Pigouvian taxation), and revenue generation. This comparison uses data from the agent-based models, examining agent error as one of the unintended consequences of fiscal policy. The basics of agent-based modeling are presented. This methodology is suited to problems in which the immediate preferences of the agents can be stated as equations or rules but complexity in their interactions with the environment or each other make predicting the outcomes difficult.
Level of Degree
Department of Economics
Chermak, Janie M.
First Committee Member (Chair)
Second Committee Member
Cullen, Bradley T.
Nonrenewable Resources, Hotelling's Rule, Agent-Based Modeling
Dixon, David S.. "Computational Methods in Natural Resource Economics: Agent-Based Modeling and Hotelling's Rule." (2011). http://digitalrepository.unm.edu/econ_etds/4