Scenario Analysis in Population Projection

This paper focuses on the approach to population projections that has been labeled scenario analysis. It looks at the original meaning of "scenario" and then discusses its appropriate usage in population projections with respect to several criteria, especially that of consistency of assumptions. Next, the paper describes the practical considerations and actual experiences in an IIASA effort to define alternative scenarios for 12 world regions to the year 2030 through discussions of a group of experts. Because of differential expertise and an uneasiness of experts to numerically define alternative scenarios, an interactive group process was chosen rather than a larger Delphi. On practical and theoretical grounds it became apparent that individual responsibility and judgement of the authors cannot be replaced by an anonymous "objective" entity making the assumptions. Finally, the paper discusses what kind of alternative variants (or scenarios) the users can handle and do expect. It is exemplified through the specific question whether the UN should change its current practice of making population projections. Pro and contra arguments are listed concerning the proposal to include alternative mortality assumptions into the three main variants that are widely publicized.