Assessing the Value of Information Dissemination Programs
Using Agent-Based Models

    Information dissemination programs are an increasingly popular instrument of environmental policy.  The presumption is that such information dissemination programs are valuable because the choices of firms and other economic actors do not generate the optimally efficient level of environmental performance because these actors lack critical information, such as the performance of new efficiency-improving technologies.  Information dissemination programs provide incentives for actions which help generate and disseminate such information.  Current Integrated Assessment models used in the evaluation of environmental policies generally do not consider the effects of imperfect information.  Thus, these models are unable to assess the value of information dissemination programs nor evaluate their role as part of  a portfolio of regulatory and policies.  In this work, we will use agent-based models of technology diffusion to assess the role of information dissemination programs in a robust strategy for mitigation of greenhouse gas emissions.  In particular, the model focuses on the effects of imperfect information on the patterns of adoption of new technologies by agents with heterogeneous preferences.  In general, there will be considerable uncertainty about many of the parameters and structural relationships in such models, which we address via robust adaptive planning (RAP), a new method of decision-making under conditions of deep uncertainty, particularly well-suited for exploiting agent-based models.  In a RAP analysis we compare alternative policies over a wide range of scenarios defined by the range of uncertainties in the input parameters for the model.  We then seek robust policy strategies, that is, ones which perform reasonably well compared to the alternatives across a wide range of scenarios.  This RAP method allows us to draw conclusions from the reliable information contained in the agent-based models, even if there is significant uncertainty in some portions of it.


Robert Lempert
RAND
Lempert@rand.org