Models, Muddles and Megapolicies: The IIASA Energy Study as an Example of Science for Public Policy

This paper is offered as a contribution to general methodological reflection within applied systems analysis. It is about several linked questions: (i) the nature of intrinsic structural bias in the very activity of formal modeling; (ii) the pitfalls involved in attempting to be objective by artificially abstracting physical and technical aspects of an issue from institutional dimensions; (iii) the underlying structural correspondence between particular modes of policy analysis and of the policy process itself; (iv) the problems of proper self-representation of policy analysis, given the inevitable conflation of informal judgement and formal calculation involved; and (v) the ambiguous connections between the pragmatic role and practice of policy analysis, and the processes of quality control. The IIASA energy study happens to be a good example of several general problems and confusions that require further development of methodological reflection already under way. A major point of this paper is that the necessary acceptance of analysis as a craft skill, like conventional science (i.e., not completely specifiable in terms of its rules of inference, logic, etc.), must not be allowed to justify laissez-faire with respect to standards of proper practice in such basic matters as documentation and sensitivity analysis. Although the IIASA ESP suffered problems in these respects and over demarcating the boundaries between formal and informal modes of analysis, it is by no means unique, as this paper shows. [A recent important paper has come to my attention unfortunately too late to assimilate and discuss here. This appears to contain striking similarities--even to the extent of its independently formulated title--but some significant differences of approach to my own presented here: "The Energy Model Muddle," P. Brett Hammond, Policy Sciences, 16(1984) 227-243.] The overall conclusion is that if it is to be meaningful. methodological reflection and change within applied systems analysis requires corresponding systematic attention to the policy process and institutional contexts in which analysis and decision making are conducted. In an important sense, analysis is a symptom of a given policy process, rather than an input to it