spacepolicies.org
Backcasting
When the Future Refuses to Arrive on Its Own
A recurring pattern in space sustainability work is this: the declared target is ambitious, the technical community agrees it is desirable, the funding announcements accumulate — and the underlying trend barely bends. Debris populations climb. Deorbit compliance lags. The international governance conversation circles. Practitioners who have lived inside these files know the frustration: each new strategy document produces a slightly better version of the same trajectory, because each one reasons forward from where we are rather than backward from where we want to be.
This is not a moral failing. It is a methodological one. Predictive tools are built to extend today’s conditions — capabilities, institutions, incentives — into tomorrow. When tomorrow is merely a continuation of today, that is enough. When tomorrow must be meaningfully different, extension of the present is precisely the wrong operation. Backcasting is the corrective. It asks the analyst to stand in the desired future, look behind, and describe the landscape of prerequisites that, had they been in place, would have produced the target. What follows is an account of why the method exists, what it can do for a space strategist, and where it should not be trusted.
A Planning Tradition Born from Frustration
The method’s modern lineage begins with John Robinson, the Canadian planner who coined “backcasting” in 1990 in the context of energy futures. Robinson had watched the dominant forecasting traditions of the 1970s and 1980s — scenario projections, trend extrapolations, techno-economic models — produce energy futures that extended consumption trajectories and then asked how supply could meet them. His objection was philosophical before it was technical: for problems where the trajectory itself is the problem, the analyst who forecasts participates in normalizing what needs to change. A different operation was required.
Robinson’s answer was to invert the arrow. Begin by specifying a socially desirable end state — carbon intensity, consumption profile, energy mix — and then work backward to identify the decisions, technologies, and behaviors that must be in place for it to exist. The method was never meant to predict. It was meant to design the dependency structure between now and a normatively specified then.
The tradition Robinson was arguing against was not a straw man. Predictive foresight had become canonical in corporate and national planning from the 1960s onward, partly through the scenario work associated with Shell and partly through the systems dynamics school. Those tools remain indispensable for exploratory questions: what might happen if current dynamics continue? Backcasting’s claim is narrower and more specific. For normative questions — how do we reach a state that present dynamics will not reach on their own — the forward-looking tools are structurally biased toward continuity.
Over the following decades, the method was absorbed into sustainability planning (Holmberg and Robèrt in Sweden), energy transition modelling, and long-horizon infrastructure planning. It has been applied to water systems, urban form, climate pathways, and increasingly to policy domains where a target year is politically fixed but the route to it is not. The space community’s exposure to backcasting has come through the sustainability and transition literatures rather than through space-native sources, which is one reason its discipline is not yet fully absorbed into space strategy.
| Dimension | Forecasting | Backcasting |
|---|---|---|
| Central question | What is most likely to happen? | What must be true for a desired future to exist? |
| Direction of reasoning | Forward from present conditions | Backward from a specified end state |
| Best suited for | Exploratory problems | Normative problems |
| Primary artifact | Probability-weighted scenarios | Dependency graph of prerequisites |
| Structural bias | Continuity with present trends | Optimism along the constructed pathway |
What the Method Actually Sees
The characteristic move of backcasting is not working backward — many methods do that. It is treating the backward chain as an analysis of prerequisites, not of probabilities. Forecasting produces a distribution of likely futures weighted by today’s dynamics. Backcasting produces a structure of preconditions: if the target state exists in 2045, what must have been true in 2042? In 2037? In 2030? The artifact it generates is a dependency graph, not a probability curve.
This changes what the analyst is looking for. The question “what is most likely?” disappears and is replaced by “what is binding?” — which milestones sit on the critical path, and which are merely parallel enablers? The method quickly reveals that many elements strategists treat as goals are in fact downstream of earlier, less glamorous conditions. A binding international instrument is not itself a first move; it is the output of a political process that requires, upstream, a particular distribution of operator behavior and state interest. An operator behavior change requires, further upstream, a set of incentives that do not yet exist. The backward chain surfaces the invisible scaffolding.
The operation also forces an honesty that forward-looking methods often evade. Because every milestone must be expressed in terms specific enough to check, wishful steps cannot survive the procedure. “By 2040, states have aligned on a common framework” is not a milestone; it is an outcome masquerading as one. The method requires decomposition: which states, aligned through which mechanism, producing which observable output. Vagueness is flushed out by the demand that each step enable the next.
What backcasting does not do is question the target itself. The target is given, and the method analyzes the pathway. This is a deliberate design choice — separating the normative from the analytical — but it requires a pairing with methods that interrogate desirability, or the analysis becomes the engineering of a destination no one has earned.
Finally, the method insists on two artifacts rather than one. The backward chain is the analytical product: a reverse-chronological set of prerequisites with their dependencies. The forward roadmap is the executable product: actor-by-actor, year-by-year, expressed in decisions and resources. The temptation is to collapse them. The discipline is to keep them separate, because they answer different questions. The backward chain answers “what must be true?” The forward roadmap answers “what must we do?” Elision causes the roadmap to inherit the speculative character of the backward chain, or the backward chain to inherit the defensive self-censorship of the roadmap.
Walking Back from a Lunar Governance Regime
Suppose the target is a binding international regime governing lunar resource extraction, with a functional dispute-resolution mechanism, operational by 2045. This is a useful example because it is neither fanciful nor assured.
The first operation is honest gap analysis. Today, lunar resource rights are addressed by permissive interpretations of existing treaty law, national statutes in a handful of jurisdictions, and non-binding multilateral instruments. A dispute resolution mechanism specific to lunar activity does not exist. The gap, at the level of dimensions, is primarily institutional (no body with jurisdiction), political (no coalition mobilized for binding rules), and behavioral (no accumulated operator practice establishing claims worth arbitrating). It is not primarily technological.
Working backward: for a regime to be operational in 2045, treaty or treaty-equivalent negotiations must have concluded roughly five years earlier, around 2040, because the ratification-to-operation lag for this class of instrument is rarely shorter. For conclusion in 2040, a critical mass of states accepting binding governance must have formed by approximately 2035, because negotiation of binding text cannot precede the political willingness to bind. For that willingness to crystallize, disputes of real economic consequence must have occurred during the 2028–2033 window — early enough that the costs of ungoverned competition became politically legible, late enough that the lunar operator population was large enough to generate friction.
That last step is the pivot the method produces. A strategist reasoning forward would likely concentrate effort on legal drafting and on building diplomatic coalitions. The backward chain reveals a different center of gravity: the pathway depends not primarily on lawyers, but on whether early lunar operations generate visible, costly, unresolved disputes that create political demand for binding rules. Effort directed at drafting without that demand-generation layer produces instruments without ratifying constituencies. This is the characteristic non-obvious output of the method: a reordering of which upstream conditions are load-bearing.
A disciplined analyst would not stop there. Each milestone should be stress-tested. What happens if early operations are conducted by a small number of well-resourced actors who internalize disputes privately through commercial arrangements? That scenario breaks the demand-generation step, and the pathway to 2045 stalls. What happens if a non-kinetic incident — an operational interference, not a legal one — creates political demand earlier than expected? That scenario compresses the timeline but may produce a regime designed for the wrong kind of dispute. Either branch reveals a vulnerability in the pathway that the reverse-chronological chain did not fully expose without interrogation.
The deliverable, in the end, is not a prediction about lunar governance. It is a map: here are the conditions that must hold for the target to exist, here is the sequence that binds them, here are the points where the pathway is most exposed, and here is what effort should be directed toward each to move the probability of arrival upward.
Where the Method Earns Its Keep, and Where It Limps
Backcasting shines at a particular class of problem: normatively specified, high-horizon, and requiring non-incremental change. It breaks trend fatalism. It forces target specification at a level of granularity that ambitions rarely receive. It surfaces dependencies that forward planning buries in assumption. And it produces, at its best, the reordering insight — the realization that the apparent goal is downstream of a less visible one.
It limps elsewhere, and these weaknesses matter.
Within the library, backcasting pairs most usefully with scenario planning for pathway robustness, horizon scanning for unexpected branches, and red team analysis to break the optimism. Practitioners who treat it as a standalone tool will get a cleaner deliverable and a less trustworthy one.