Environmental Economics

Agent-Based Modeling of Industrial Ecosystems

  • Abstract
    The objectives of this research are to investigate behavioral and organizational questions associated with environmental regulation of firms, and to test specifically whether a bottom-up approach that highlights principal-agent problems offers new insights and empirical validity. This project will investigate key questions in the new field of industrial ecology using an innovative, agent-based modeling approach. Agent-based computer simulation modeling has matured in recent years to the point that it can provide a safe laboratory for investigating how economic agents interact with one another and their environment, and for exploring alternative organizational structures and contracting arrangements that might reduce adverse impacts on the natural environment. Approach: The project includes the following tasks: (1) create an empirical foundation for models by exploiting the New Jersey hazardous chemical Release and Pollution Prevention Report (RPPR) listing environmental release, waste transfer, throughput, and pollution prevention progress information; coupled with in-depth case studies of several respondent firms to document operating histories and organizational parameters; (2) develop a general multi-agent representation of a single-facility industrial firm; (3) derive initial conditions of specific models from the case study evidence, drive the models using exogenous conditions identified in the case studies, and validate the models against outcomes from the case studies; (4) adapt the generic model and selected specific instantiations to the case of the branch plant (vertical relationships within multi-facility firms); and in a more idealized manner, (5) insert the generic model and selected specific instantiations into the supply chain; and (6) insert the generic model and selected specific instantiations into a competitive sector. Expected Results: Expected results of the project include a suite of empirically validated models that allow research-level "what-if" investigations of the motivations behind corporate behavior. The project will add a new method to the environmental social science toolkit, and develop programming routines that will be useful for agent-based modeling of wide variety of questions in the fields of organizational behavior, public policy, and innovation studies.
  • Metadata
    Principal Investigators:
    Andrews, Clinton J.
    Technical Liaison:
    Research Organization:
    Harvard University
    Funding Agency/Program:
    EPA/ORD/Corporate Performance
    Grant Year:
    Project Period:
    July 1, 2002 to June 30, 2005
    Cost to Funding Agency:
  • Project Reports

  • Project Status Reports