Oskar Morgenstern, Klaus Knorr, and Klaus P. Heiss, Long-Term Projections of Power (1973)

Morgenstern, Knorr, and Heiss write a methodological study of political, economic, and military forecasting rather than a simple prediction of future great-power rankings. Their central concern is how governments can reason about long-term power without mistaking projection for knowledge. Power is treated as a composite and unstable phenomenon: economic capacity, technological change, military organization, political will, administrative adaptability, and social cohesion all matter, but they do not become predictable in the same way or to the same degree.

Our work has impressed us with the inherent difficulties of forecasting although we realize that objects of forecasting vary greatly in their conjecturability.

This sentence states the book’s governing discipline. The authors do not reject forecasting; they reject false precision. Their argument depends on disaggregating the future into elements that can be modeled with some confidence and elements that resist extrapolation. Inventories, production capacities, demographic tendencies, and engineering programs may support conditional estimates. Scientific discovery, political resolve, institutional learning, and strategic surprise cannot be handled as though they were merely longer trend lines.

The distinction between engineering development and scientific breakthrough is especially important. Existing science can support projections about the timing and cost of known systems, but major discoveries may alter the entire structure within which planners are reasoning.

They all point in the same direction—engineering prediction on the basis of present science is possible within wide bounds of error regarding timing and cost. But prediction of the far more decisive scientific developments is not.

For defense planning, this means that the most consequential changes may be precisely those least available to routine projection. Long lead times compel states to forecast, yet the assumptions that make a forecast orderly may be overturned by invention, doctrine, alliance shifts, or political crisis. The authors therefore present forecasting as an aid to judgment, not a substitute for it.

Their treatment of military power is similarly anti-reductive. Armed forces, weapons, logistics, and industrial capacity are necessary, but they are not sufficient. Military power exists only when capabilities are organized, directed, and politically usable.

But even if a state has sizable armed forces, there can be no military power without the will to employ them.

This emphasis links military analysis to the broader political theory of power. Capability is relational and purposive; it cannot be read mechanically from budgets or force tables. A state with impressive material resources may fail to act, while another with fewer resources may gain leverage through resolve, organization, or adaptability. Long-term projections of power must therefore include estimates of institutions and decision-making, not only material stocks.

The book’s recurring warning is that surprise is not an accidental defect in forecasting but a structural feature of the subject. Analysts can reduce avoidable surprise by refining concepts and methods, but they cannot eliminate uncertainty from political and technological change.

Unless our understanding becomes more refined, we will most certainly experience similar surprises in the future—which means more surprises than are necessary.

Kan-Hua Young’s appended telecommunications model illustrates the authors’ balanced position. Quantitative modeling has real value, especially where systems have measurable components and discernible constraints. Yet such models require interpretation within a wider theory of uncertainty. The appendices reinforce the work’s character as both a planning instrument and a critique of overconfident planning.

The final lesson is practical skepticism. Forecasts should be explicit, conditional, revisable, and institutionally usable; planners must examine not only projected outcomes but also the assumptions that generated them. In this sense, Long-Term Projections of Power defends disciplined forecasting while warning that the future of power depends on variables that cannot be reduced to trend lines: invention, will, organization, and the capacity to recognize when inherited assumptions no longer hold.