Grantee Research Project Results
2004 Progress Report: Agent-Based Modeling of Industrial Ecosystems
EPA Grant Number: R829688Title: Agent-Based Modeling of Industrial Ecosystems
Investigators: Andrews, Clinton J. , Axtell, Robert
Institution: Rutgers University - New Brunswick , Brookings Institution
EPA Project Officer: Hahn, Intaek
Project Period: July 1, 2002 through June 30, 2005 (Extended to June 30, 2006)
Project Period Covered by this Report: July 1, 2004 through June 30, 2005
Project Amount: $334,146
RFA: Corporate Environmental Behavior: Examining the Effectiveness of Government Interventions and Voluntary Initiatives (2001) RFA Text | Recipients Lists
Research Category: Environmental Justice
Objective:
Multiagent simulation (MAS) is an increasingly popular bottom-up computer modeling technique in which purposive agents (e.g., employees) interact so that aggregate performance (e.g., of a firm) is an emergent property of the system. The objectives of this research are to: (1) investigate behavioral and organizational questions associated with environmental regulation of firms, and (2) test specifically whether an MAS approach that highlights principal-agent problems offers new insights and empirical validity.
Progress Summary:
In this project, we propose to create multiagent computer simulation models of firms and ground them empirically in case studies and aggregate statistical evidence. The first 2 years of the project focused on case study development and the conceptual design of models. During Year 3 of this project, we focused chiefly on implementing the models, of which there now are three.
PolyModel was the first model built. It was an MAS model of the interactions among employees within a polymer processing firm. It was created inductively, drawing on case studies of actual firms studied in the previous year. The case studies documented the flow of work, organizational structures and reporting relationships, technological processes, human resource management decision points, and external conditions affecting four firms. Both the case studies and the model were informed by the socialization theory of Granovetter, which claims that some activity in the marketplace and within economic organizations is determined socially rather than economically. em>PolyModel thus tracked social network relationships as well as formal reporting relationships, and the utility functions of employees led them to pursue both money and friendship.
PolyModel simulations in Year 3 of the project explored human resource management strategies that might reduce pollution. The model proved to be useful for clarifying the strengths and weaknesses of strategies of: (1) quality management that relied on training and supervision to reduce worker errors and hence pollution, and (2) selective hiring and firing to increase the percentage of employees interested in pollution reduction working on production lines. The quality strategy proved to be more profitable, although its environmental benefits were less pervasive. The lesson for environmental policy is that internal management decisions can affect strongly a firm’s environmental profile regardless of external influences such as regulations or factor prices. Although not a new insight, PolyModel provides a flight simulator for both the U.S. Environmental Protection Agency (EPA) regulators and company managers who need to better understand this interlocking set of issues.
TeamPollution was the second model built. It was a multiagent system representing a highly stylized firm and it sought to abstract as much as possible from PolyModel in pursuit of general insights. Employees were purposive agents with imperfect incentives to serve the interests of their principals and to reduce pollution. As in Holmstrom’s team production formulation that features agents contributing to production noncooperatively, as long as pollution abatement requires nonzero levels of employee effort and the firm lacks a contractual basis for punishing nonabaters, then employees will shirk and pollution will continue. A two-employee firm was enough to confirm this outcome.
The most interesting part of TeamPollution was its ability to compare rigid and flexible environmental policies. Rigid, command-and-control policies often are thought of as inefficient, whereas flexible, market-based, or performance-based policies are preferred widely by economists. Once the firm is viewed as a collection of self-interested agents instead of a profit-maximizing machine, however, then shirking, ignorance, and misaligned incentives open up the possibility that command-and-control policies are in fact the superior policy choice.
TeamPollution simulations in Year 3 showed that an extremely parsimonious multiagent representation of a firm was fully adequate to cast doubt on the profit maximization hypothesis underlying much of regulatory economics. This powerful result suggests that the EPA should pay increased attention to the insights emerging from behavioral economics, and not rely on conventional wisdom regarding the relative efficacy of command-and-control, and have more flexible regulatory approaches.
EcoNiche was the third model built. It was a multiagent system representing a highly stylized product market. Instead of assuming a priori that the market reaches equilibrium, this model allowed consumers, firms, and government to interact in search, dealmaking, and regulatory processes, so that the market clearing price and product choices were emergent system properties. The focus of EcoNiche was on the entry of new green products into an existing market for conventional products. Consumers held varying preferences for green product attributes and received information about product attributes only by expending search efforts through their social networks and queries to firms and government. Firms chose whether to perform research on the environmental attributes of their current and future products, and cumulative production experience helped reduce costs and hence prices offered. Government policies, set exogenously, could be laissez faire, provide information to consumers and firms, impose regulations, or provide incentives.
EcoNiche is becoming a useful tool for investigating how niche products such as hybrid cars or green buildings can enter the mainstream. It frames the innovation diffusion process as one that is constrained by inadequate information flows. Thus, consumer preferences, search strategies, environmental research, ecolabeling efforts, and network topology all potentially influence the rate of diffusion, in addition to more traditional income factors and price/attribute tradeoffs. Given adequate information to differentiate green and conventional products, can a small wedge of dedicated and/or wealthy green consumers support adequate production levels to drive prices down to competitive levels and allow the niche product to enter the mainstream? The model clarifies the conditions under which this occurs. Additional work is needed to calibrate this very general model to a specific product market.
EcoNiche simulations in Year 3 showed that an information-based theory of green product diffusion produces plausible results and that the model provides a potentially useful test bed for evaluating policy interventions that emphasize information provision. This is a potentially important methodological advance for innovation-oriented environmental policy analysis. It suggests that EPA should pay systematic attention to the dynamics of market transformation using not only theory but also modeling tools.
Future Activities:
We requested a 1-year no-cost extension to the grant period to finish the work. During the next reporting period, the focus of this project will be on: (1) finalizing and documenting PolyModel; (2) finalizing and documenting the EcoNiche model; (3) applying the EcoNiche model to a specific product market; (4) publishing the modeling results; and (5) finalizing the lessons about the feasibility and value of building operational multiagent simulation models of firms.
Journal Articles:
No journal articles submitted with this report: View all 21 publications for this projectSupplemental Keywords:
pollution prevention (green chemistry, life-cycle analysis, alternatives, sustainable development, clean technologies, innovative technology, renewable, waste reduction, waste minimization, environmentally conscious manufacturing), public policy, decisionmaking, psychological, preferences, sociological, social science, industry (polymer processing, plastics), NAICS codes 3261, 325991, 326199, team production, team pollution, local increasing returns, increasing returns to scale, Nash equilibrium, under-abatement, hypothetical cost savings, explicit cost savings, heterogeneous marginal costs of abatement, non-optimal firm behavior, sub-optimal firm behavior, inefficient market-based regulation,, RFA, Scientific Discipline, Sustainable Industry/Business, cleaner production/pollution prevention, Corporate Performance, Economics and Business, Social Science, environmental policy case studies, corporate decision making, cleaner production, corporate environmental policy, sustainable development, computer generated alternative synthesis, environmentally conscious manufacturing, government intervention, clean technology, computer simulation modeling, government-industry interaction, environmental behavior, agent based modeling, pollution prevention, clean manufacturing designs, corporate environmental behaviorRelevant Websites:
http://www.brookings.edu/dynamics Exit
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.