Reference Projection

Definition

A reference projection is a deliberate extrapolation of a system and its environment into the future assuming no intervention and no change in the trends already at work. It is explicitly not a forecast of what will happen — since interventions are likely — but "more a forecast of what is not likely to happen than of what will." Its purpose is diagnostic: to expose when and how the current trajectory produces a breakdown, so that interventions can be planned in advance rather than improvised once the system is already in crisis, when "interventions under crisis conditions seldom provide effective solutions to problems."

In the Book

Chapter 7 is a full worked reference projection of U.S. urban transportation to the year 2000, extrapolating population growth, automobiles-per-adult, average miles driven per vehicle, income growth, and suburbanization from historical data (Landsberg et al., Lansing and Hendricks, Bureau of Public Roads figures) under the assumption that no one changes course. Ackoff uses the resulting trajectory — automobiles already carrying 85 percent of urban passenger travel and set to keep rising — to show the current system heading toward gridlock and unsustainable resource demand well before 2000, and argues this kind of projection is what should trigger planning, not the crisis itself. He is explicit that a reference projection's second purpose is generative: seeing precisely how and where the unconstrained system fails suggests types of intervention (redesigning the automobile, changing work-week structure, decentralizing government offices — the same set he separately proposes to Mexico City planners in Chapter 2) that would not occur to planners only reacting to today's complaints.

Why It Matters

A reference projection turns "we'll deal with it when it becomes a problem" into a falsifiable, dated claim about the future, which is what makes proactive planning possible instead of reactive firefighting: you cannot plan an intervention against a crisis you haven't yet let yourself see coming. It also reframes forecasting's job — not to predict what will actually happen, since intervention is the point, but to make the do-nothing baseline visible enough that its failure mode can be designed around now.