Filtering: An Approach to Generating the Information Base for Collective Choice

Nonmarket resource allocation decisions require some means for generating information on relevant preferences. This is part of the role of a collective choice process. Implicit in the optimization techniques central to MS/OR is the assumption that makes collective choice equivalent to the optimization of a social preference function. This paper proposes an alternative approach to collective choice, called "filtering," and compares it with the general optimization approach. Though not foreclosing the use of optimization techniques, filtering represents a fundamental departure from the perspective on collective choice implicit in these techniques. The fundamental advantage of filtering is that it increases the likelihood of collective agreement on specific plans, while minimizing the strategic use of misinformation. The general logic of filtering may be implemented via a variety of interactive methods. The information generated can contribute to more stable and implementable plans. Filtering responds to a societal context characterized more by diversity of interests than uniformity, and as much concerned with the possible adverse consequences of plans proposed by "experts" as with their promised benefits. The more modest objective of collective choice supported by filtering is to identify collectively acceptable plans for interdependent social groups whose goals and preferences may not be fully known. This supports the need for participation by relevant interests in the planning process. In general, filtering provides a framework for "joint charting" of the future collaboratively, by aiming at the evolution of a "reasonable consensus" from a base of what may appear initially to be "unreasonable vetoes."