A planning department can be very good at describing the city that exists and still be badly prepared for the city that arrives next.
That problem appears whenever a change is visible before it is measurable in the usual datasets.
A new mobility technology appears in a few cities.
Insurance withdrawals start in one hazard market.
A major employer changes where its workforce lives.
Household size shifts.
Artificial intelligence changes demand for offices, power and data infrastructure.
A new migration pattern emerges.
A building type once considered niche becomes mainstream.
The planner faces a difficult question:
Is this merely interesting, or is it the beginning of a change that should alter land, infrastructure or regulation now?
The American Planning Association’s 2026 Trend Report for Planners gives this question unusually high prominence. APA describes foresight not as prediction but as the discipline of understanding forces outside planners’ control, preparing for possible consequences and recognising when it is time to act. Its 2026 report tracks more than 100 existing, emerging and potential future trends, while the APA Trend Universe separates signals into categories such as matters to act on now, matters to prepare for and matters to keep watching.
That is not the same job as a comprehensive plan.
It is the early-warning layer before the plan becomes stale.
The reader job for this article is therefore:
How does a planning authority build a repeatable foresight system that scans change, distinguishes weak signals from noise, converts uncertainty into scenarios, defines signposts and triggers, and feeds credible new information back into plans, capital budgets and development controls without chasing every fashionable prediction?
This article owns that question.
It does not replace the existing eduKateSG owners for the Plan Stack, Plan Integration Scorecard, Digital Shadow, Data Gap, Time Layer, Housing Observatory or Robust Decision Making. Those owners explain hierarchy, coordination, simulation, missing information, sequencing, monitoring and decision quality. The Foresight Signal Board owns the process before a future condition is established strongly enough to become ordinary planning data.
1. Foresight is not forecasting with more imagination
Forecasting usually asks:
What is most likely to happen?
Foresight asks a broader set of questions:
- What is changing?
- What might plausibly happen?
- What would matter if it did?
- What should we watch?
- Which decisions should stay flexible?
- What evidence would make us change course?
This difference is foundational.
Forecasts are useful when relationships are reasonably stable.
Foresight becomes valuable when the drivers themselves may change.
2. The purpose is preparedness, not prophecy
A foresight team should not be rewarded for correctly predicting one dramatic future.
It should be rewarded if the planning system is less surprised.
That means preserving options, identifying emerging constraints and reducing the cost of changing direction.
A city can make good foresight decisions even when the most alarming scenario never occurs.
Insurance, fire exits and reserve land all work on the same logic.
Preparation has value before prediction is proven.
3. Start with a planning horizon
Different decisions have different lifetimes.
A temporary street pilot may last one year.
A school may last sixty.
A sewer trunk may shape development for a century.
A zoning code may be rewritten every decade but create parcel expectations for much longer.
The foresight horizon should therefore follow the decision.
Ask:
How long will today’s choice constrain tomorrow’s options?
Long-lived, hard-to-reverse investments deserve deeper foresight.
4. Foresight is strongest where reversibility is low
If a policy can be changed cheaply next year, uncertainty is manageable.
If a decision commits billions of dollars and fixes a corridor for fifty years, uncertainty matters much more.
High-foresight decisions commonly include:
- transport corridors;
- major utility systems;
- growth boundaries;
- coastal defences;
- strategic industrial land;
- public land disposal.
The more irreversible the choice, the more important it is to examine multiple plausible futures.
5. The signal board begins with horizon scanning
Horizon scanning is the deliberate search for changes that could matter later.
Sources can include:
- research;
- government policy;
- technology;
- market behaviour;
- demographics;
- climate science;
- court decisions;
- industry practice;
- community observation.
The aim is not to collect everything.
It is to detect developments outside the normal planning dashboard before they become obvious.
6. A signal is smaller than a trend
A signal is an observation that may indicate change.
A trend is a pattern sustained strongly enough to have direction.
Example:
One insurer withdraws from a flood market.
That is a signal.
Repeated withdrawals across several years and jurisdictions may become a trend.
Planning should not confuse one anecdote with a trend.
It should not ignore the anecdote either.
7. Weak signals deserve a place to wait
A weak signal often lacks enough evidence for policy.
If the city discards it, the idea may be forgotten.
If the city immediately regulates it, the city may overreact.
The signal board provides a third state:
watch.
This is institutionally important.
Good planning needs a place between “irrelevant” and “adopt a law.”
8. Every signal should contain a planning mechanism
Do not store vague headlines.
Record:
- what changed;
- why it might matter spatially;
- which planning systems could be affected;
- what evidence would strengthen or weaken the signal.
For example:
Signal: autonomous freight pilots expand.
Possible planning mechanism: changes loading geometry, depot design and labour demand.
This makes the signal actionable later.
9. Signals should be tagged by domain
Useful domains may include:
- housing;
- mobility;
- energy;
- climate;
- economy;
- technology;
- health;
- governance;
- demographics;
- culture.
Cross-tagging is valuable.
A data-centre trend is simultaneously:
- digital;
- energy;
- water;
- industrial land.
Foresight becomes powerful when planners see interactions between domains.
10. Signals should also be tagged by planning scale
Some changes operate at:
- building;
- parcel;
- neighbourhood;
- city;
- region;
- national system.
This prevents the city from using a parcel tool for a regional problem.
Climate migration, for example, may begin with households but reshape metropolitan housing and infrastructure.
Scale should be explicit.
11. Signal age matters
A signal first observed three years ago and repeatedly reinforced deserves different attention from a new one.
Record:
- first seen;
- latest confirmation;
- frequency.
This creates a history of evidence.
Foresight should have memory.
12. Signal strength should not be reduced to popularity
A topic can dominate media and still have weak planning evidence.
Another can receive little attention while producing strong regulatory consequences.
Signal strength can consider:
- quality of source;
- repetition;
- geographic spread;
- magnitude;
- relevance to long-lived decisions.
Do not use social-media attention as a proxy for planning importance.
13. Source diversity improves confidence
Ten articles copying the same report are not ten independent signals.
A stronger pattern appears when:
- research;
- official statistics;
- market behaviour;
- local observation;
point in the same direction.
Track source lineage where possible.
This reduces echo-chamber foresight.
14. APA’s Act–Prepare–Learn logic is useful
APA’s Trend Universe distinguishes issues that planners should act on, prepare for, or continue learning about.
A local government can adapt this structure:
ACT
- evidence strong;
- material current impact.
PREPARE
- credible emerging change;
- long lead time.
WATCH
- plausible but uncertain.
The categories should trigger different institutional responses.
15. “Act” does not necessarily mean regulate
Action can include:
- update data;
- reserve land;
- commission study;
- modify capital design;
- add flexibility.
A signal may justify preparation without justifying prohibition.
This distinction prevents foresight from becoming a factory for premature rules.
16. “Prepare” is the most important category
Cities often understand “do nothing” and “adopt policy.”
Preparation sits between them.
Examples:
- design a station so platforms can expand;
- protect an optional utility corridor;
- require conduit for future charging;
- preserve public land.
These measures create option value.
They are often cheap before demand materialises.
17. “Watch” needs an owner
A watch list without assigned responsibility becomes a graveyard.
For each signal, assign:
- department;
- review date;
- indicator.
A staff member should know what evidence would move it to Prepare or close it.
Foresight is a workflow, not a folder.
18. Signals should be retired
Some predictions do not develop.
A signal may weaken because:
- technology fails;
- policy changes;
- market adoption stalls.
Mark it closed.
Keep the record.
A foresight system should learn which signals were misleading and why.
19. Retired signals teach calibration
If a planning department repeatedly overreacts to technology announcements, it can change weighting.
If demographic signals repeatedly mature faster than expected, it can pay more attention earlier.
The signal board becomes a learning system.
Accuracy improves through post-mortem review.
20. Trend scanning should include local anomalies
National trends are useful.
Local departures may matter more.
A city may see:
- sudden school enrolment decline;
- unusual office vacancy;
- rapid electricity connection applications;
- insurer behaviour.
An anomaly can be the first local expression of a wider transition.
Foresight should connect global scanning with local observatories.
21. The Housing Observatory can supply local signposts
The existing TPW Housing Observatory tracks supply, vacancy, rent, pipeline and need.
Foresight can watch for inflection:
- household formation changes;
- migration;
- tenure shift.
The observatory measures.
Foresight interprets whether a change may represent a new trajectory.
These roles should remain separate but connected.
22. The Data Gap article supplies another discipline
Weak signals often live in incomplete data.
The Data Gap owner asks how planners act when the map is incomplete.
Foresight adds:
Is the incompleteness itself a sign that the system is changing faster than the data architecture?
For new technologies, the official classification may lag the phenomenon.
This is common.
23. Scenarios translate signals into coherent futures
A list of trends is not enough.
Scenarios ask what happens when multiple trends interact.
For example:
- high migration;
- slow construction;
- climate insurance retreat;
- strong job growth.
Together, these may create a housing future different from any single trend.
Scenarios should describe systems, not collections of headlines.
24. Scenarios are not predictions
A scenario should not say:
This will happen.
It should say:
If these conditions interact, this is a plausible future we should be able to handle.
This lowers political defensiveness.
The city can stress-test plans without endorsing a forecast.
25. Use a small number of scenarios
Too many scenarios create paralysis.
Three or four are often enough to expose major uncertainty.
A useful set might include:
- baseline;
- high-growth;
- low-growth;
- disruptive transition.
The specific axes should come from the city’s most consequential uncertainties.
26. Scenario axes should be truly uncertain
Do not use one axis where the outcome is already established.
Choose variables that are both:
- uncertain;
- consequential.
Examples:
- migration intensity;
- energy cost;
- remote-work persistence;
- climate damage rate.
Scenario work should focus attention where certainty is genuinely low.
27. A scenario should have a spatial expression
Where does the future happen?
If migration increases:
- which neighbourhoods receive people?
If logistics changes:
- which freight corridors intensify?
If sea level rises faster:
- which assets cross thresholds?
Map scenarios.
Town planning needs geography, not only narrative.
28. The Digital Shadow can test spatial scenarios
The existing Digital Shadow owner explains digital twins and simulation.
Foresight supplies:
- plausible future conditions.
Simulation tests:
- physical consequences.
These are complementary tools.
A digital model without foresight can optimise one assumed future too precisely.
29. Foresight should stress-test adopted plans
For each scenario, ask:
- Does the housing plan still work?
- Does the transport plan still work?
- Does the capital plan still work?
- Which policies fail first?
This is where foresight becomes useful to TPW-0201’s Plan Integration Scorecard.
The scorecard tests consistency.
Foresight tests durability under changing conditions.
30. Identify brittle policies
A brittle policy works only under a narrow forecast.
Example:
A growth area depends on:
- one population forecast;
- one bridge;
- one funding source.
If any assumption shifts, delivery fails.
Foresight should highlight brittleness.
The city can then add flexibility or contingency.
31. Identify robust policies
A robust policy performs reasonably across several futures.
Examples can include:
- mixed-use land;
- connected street networks;
- adaptable buildings;
- protected utility corridors.
Robust does not mean perfect.
It means useful under a range of plausible conditions.
32. Preserve optionality where uncertainty is highest
Optionality is the capacity to change later without enormous cost.
Planning tools include:
- reserve land;
- oversize conduit;
- flexible zoning;
- modular infrastructure;
- staged capital.
The Reserve Map owner explains option value in land.
Foresight tells planners where preserving options is worth the cost.
33. Foresight should identify no-regret actions
A no-regret or low-regret action provides value across many futures.
Examples:
- better data;
- asset maintenance;
- safer walking;
- adaptable public space.
These are good early responses when uncertainty is high.
They buy time without wasting resources.
34. Not every uncertainty should delay action
Waiting for certainty can create lock-in.
If a coastal defence takes twenty years to plan and build, the city cannot wait until failure is imminent.
Foresight should compare:
- decision lead time;
- uncertainty resolution time.
When lead time is longer, action may need to begin before uncertainty disappears.
35. Signposts are observable indicators
A signpost tells planners whether the world is moving toward a scenario.
Examples:
- insurance premiums;
- migration applications;
- office vacancy;
- electricity connection queues;
- school enrolment.
Signposts should be measurable.
A scenario without signposts cannot be monitored.
36. Trigger points connect monitoring to decisions
A trigger is a pre-agreed condition that prompts review or action.
Example:
If annual housing completions remain below 60% of target for three years, initiate zoning/infrastructure diagnostic.
Triggers reduce political delay.
The action is considered before the crisis.
37. Triggers should not be automatic where judgment is necessary
A threshold can trigger:
- review.
It does not always need to trigger a predetermined outcome.
This is important when indicators are noisy.
The planning system can preserve professional and democratic judgment while still avoiding inaction.
38. A trigger needs a data owner
If nobody reliably measures the indicator, the trigger is fictional.
For each trigger record:
- dataset;
- frequency;
- custodian;
- quality.
Foresight should connect directly to the city’s data governance.
39. Trigger lag matters
Official statistics may arrive a year late.
A trigger based on stale data can activate too late.
Pair slower authoritative data with faster proxies where appropriate.
For housing:
- permits;
- listings;
- utility connections;
may provide earlier signals than census data.
40. Leading indicators deserve special value
A leading indicator changes before the final outcome.
Building applications may change before completions.
Insurance pricing may change before property abandonment.
Grid connection requests may change before industrial construction.
Foresight is strongest when it monitors the leading edge.
41. Coincident and lagging indicators still matter
A coincident indicator confirms the condition is happening.
A lagging indicator confirms consequences.
Use all three types:
- leading;
- coincident;
- lagging.
This helps separate weak signal from established trend.
42. Trigger thresholds should be revisited
A threshold chosen in 2026 may become obsolete.
Review:
- data quality;
- policy relevance;
- false alarms.
Adaptive monitoring applies to the foresight system itself.
43. False alarms are not automatically failure
A trigger may initiate review and reveal no action is needed.
That can still be useful.
The cost of review may be low compared with missing a high-consequence transition.
Threshold design should consider the cost of:
- false positive;
- false negative.
44. Foresight should feed the capital programme
Long-lived infrastructure is where foresight has greatest financial value.
A signal may suggest:
- oversize a pipe;
- reserve land;
- delay one investment;
- design for modular expansion.
The capital plan should have a mechanism to receive foresight recommendations.
Otherwise scanning never reaches concrete.
45. Capital options can be staged
Instead of building a maximum system immediately:
- protect land;
- build Phase 1;
- define expansion trigger.
This can reduce stranded investment.
The Time Layer owner explains phasing.
Foresight supplies the uncertainty logic behind the phases.
46. Procurement can preserve adaptability
A public contract can require infrastructure that can:
- expand;
- retrofit.
This may cost slightly more today.
The value depends on future uncertainty.
Foresight can justify adaptability where it has real option value.
47. Land disposal should pass a foresight check
Selling public land can be irreversible.
Before disposal ask:
- Could this site matter for future schools?
- utilities?
- climate adaptation?
- housing?
The Reserve Map owner explains safeguarding.
Foresight supplies emerging needs not yet in formal plans.
48. Regulatory foresight can prevent code obsolescence
New uses may not fit old definitions.
Examples:
- micro-fulfilment;
- battery storage;
- co-living.
A signal board can identify code areas likely to need future adaptation.
This is better than waiting for a permit dispute.
49. The zoning code should not chase every emerging use
Foresight can also prevent overreaction.
A niche activity may disappear.
Use flexible definitions and administrative interpretation where suitable.
Reserve full code amendments for changes that become durable.
50. Legal change is a signal too
Higher courts, national law or state legislation can transform local planning.
The Planning Preemption Map owner addresses higher-level override.
Foresight should monitor emerging legislation before enactment when the implications could require rapid local response.
51. Finance is a planning signal
Interest rates, insurance, credit and construction finance shape what gets built.
A city cannot control them.
It can observe them.
A pipeline that looks strong in permits can collapse when financing conditions change.
Foresight should integrate market signals without pretending to forecast markets precisely.
52. Labour markets can change land demand
A major employer may shift:
- remote work;
- automation;
- hiring.
This can affect:
- offices;
- commuting;
- housing.
Economic-development staff should be part of trend scanning.
Foresight needs cross-department intelligence.
53. Demography is slow until it is not
Birth rates change gradually.
Migration can change quickly.
Ageing is highly predictable in aggregate but uncertain in local movement.
Use different foresight methods for different demographic drivers.
Some future conditions are uncertain in timing.
Others are almost certain but under-planned.
54. Climate trends combine slow change and sudden shocks
Sea-level rise is gradual.
Storms are sudden.
Heat trends are long-term.
Wildfires can transform perception overnight.
Foresight should connect chronic shifts with shock events.
The Shock Map owner provides the emergency/resilience layer.
55. A shock can accelerate an existing trend
Remote work expanded rapidly during a global shock.
A flood can accelerate relocation.
An energy crisis can accelerate efficiency.
After a shock, ask:
Did this event create a temporary deviation or advance a transition already under way?
This is a classic foresight question.
56. Social values are planning signals
Attitudes toward:
- density;
- cars;
- housing types;
- climate risk;
can change.
Planning is partly a legitimacy system.
A technically good policy can fail if public values move dramatically.
Foresight should scan culture as well as technology.
57. Values should not be confused with polling snapshots
One survey is not a long-term social shift.
Look for:
- repeated behaviour;
- policy change;
- market choice;
- civic participation.
Values are difficult to measure.
Use humility.
58. Community observation can contribute signals
Residents often notice changes early:
- new flood patterns;
- street behaviour;
- rental pressure.
The Listening Town owner provides the participation framework.
A foresight system can include structured channels for community-observed signals.
Do not reduce foresight to expert scanning.
59. Front-line staff are excellent signal scouts
Permit reviewers, inspectors, transit staff and school planners see patterns before central analysts do.
Create a simple way to submit:
- repeated anomalies;
- emerging requests.
Institutional foresight should use the whole organisation.
60. External expert networks improve horizon scanning
APA’s Trend Scouting Foresight Community illustrates the value of distributed scouting.
A city can build smaller networks with:
- universities;
- utilities;
- employers;
- community groups.
No one office sees the whole future.
Foresight benefits from distributed sensing.
61. The board should distinguish evidence from interpretation
Record:
Evidence
- three insurers exited.
Interpretation
- property risk pricing may change.
This prevents assumptions from becoming facts.
Good foresight is explicit about inference.
62. Confidence should be stated qualitatively
A practical scale:
- low;
- medium;
- high.
Confidence can reflect:
- evidence quality;
- repetition;
- causal understanding.
Do not turn uncertain futures into fake decimal probabilities unless the method supports them.
63. Impact and confidence should be separate
A low-confidence signal can have enormous consequence.
Example:
A potential regulatory ban on a major infrastructure technology.
The city may choose:
- monitor closely;
- preserve options.
Risk management depends on both likelihood/confidence and impact.
64. A 2×2 signal matrix can prioritise attention
Axes:
- evidence strength;
- planning consequence.
High evidence + high consequence:
- Act.
Low evidence + high consequence:
- Watch closely / prepare options.
High evidence + low consequence:
- integrate routinely.
Low evidence + low consequence:
- archive.
Simple matrices can improve institutional focus.
65. Foresight should have a monthly and annual rhythm
Monthly:
- add/review signals.
Quarterly:
- check priority indicators.
Annual:
- update scenarios and major assumptions.
The rhythm should be light enough to survive staff workload.
A system too elaborate to maintain is not resilient.
66. The annual foresight report should be short
The signal database can be large.
Decision-makers need:
- top changes;
- changed assumptions;
- action triggers.
The APA Trend Report works because it translates a large trend universe into usable planning relevance.
Local government should do the same.
67. Every foresight recommendation should name a decision
Avoid:
Monitor AI.
Use:
Monitor office vacancy and data-centre load because both affect the 2030 employment-land and grid plan.
Tie the signal to an actual planning choice.
This prevents trend watching from becoming intellectual entertainment.
68. Foresight should identify decisions that can wait
One of its most valuable outputs is restraint.
If a decision is reversible and uncertainty will resolve soon:
- defer irreversible commitment.
Waiting can be strategic.
The key is knowing what information is expected and when.
69. Waiting needs a review date
“Wait and see” without a date becomes avoidance.
Specify:
- next review;
- signpost.
Adaptive delay is deliberate.
Passive delay is not.
70. The board should identify decisions that cannot wait
If infrastructure lead time is twenty years, delay can remove options.
Mark these decisions clearly.
Foresight helps leaders understand why action is justified before the public can see the final need.
71. A worked example: office demand
Signal:
- persistent hybrid work.
Scenario:
- office demand falls 25%.
Potential impact:
- central office vacancy;
- conversion opportunity;
- tax-base change.
Signposts:
- lease renewals;
- vacancy;
- transit ridership.
Actions:
- update conversion rules;
- preserve mixed-use flexibility.
The city does not need to predict remote work exactly.
72. A worked example: climate insurance
Signal:
- insurers raise premiums in hazard zones.
Potential impact:
- mortgage difficulty;
- property-value change;
- disinvestment.
Signposts:
- premium increases;
- policy nonrenewal.
Preparation:
- map insurability;
- strengthen adaptation;
- review growth allocation.
The existing Insurability Map owner takes over once the signal becomes a durable planning system.
73. A worked example: school enrolment decline
Signal:
- three years of falling enrolment.
Possible drivers:
- ageing;
- housing cost;
- smaller families.
Scenario:
- school consolidation.
Planning implications:
- facility reuse;
- housing geography.
Foresight connects demographic signal to land before a closure decision arrives.
74. A worked example: rapid in-migration
Signal:
- rental applications and school enrolment rise before census data.
Potential issue:
- housing and services.
Preparation:
- temporary capacity;
- housing pipeline review.
The next TPW article on Climate Mobility develops one major source of population shifts.
75. The Foresight Signal Board workflow
Step 1 — Define planning horizons.
Step 2 — Identify long-lived decisions.
Step 3 — Build scanning network.
Step 4 — Capture signals with sources and mechanisms.
Step 5 — Classify Act / Prepare / Watch.
Step 6 — Group signals into drivers.
Step 7 — Build 3–4 coherent scenarios.
Step 8 — Map spatial consequences.
Step 9 — Stress-test plans and capital projects.
Step 10 — Identify robust and brittle policies.
Step 11 — Define signposts and triggers.
Step 12 — Assign data owners and review dates.
Step 13 — Feed decisions into plans/budgets.
Step 14 — retire or strengthen signals as evidence changes.
76. A foresight audit
Ask:
- Does the department have a formal scanning process?
- Are global and local signals included?
- Is source quality recorded?
- Is first-seen date recorded?
- Is planning mechanism stated?
- Are signals tagged by domain?
- Are they tagged by spatial scale?
- Are Act/Prepare/Watch categories used?
- Does every Watch signal have an owner?
- Are signals retired when evidence weakens?
- Are scenarios coherent rather than trend lists?
- Are scenario axes consequential and uncertain?
- Are scenarios mapped spatially?
- Are adopted plans stress-tested?
- Are capital projects stress-tested?
- Are brittle policies identified?
- Are low-regret actions identified?
- Is optionality preserved where valuable?
- Are signposts measurable?
- Are triggers pre-defined?
- Does every trigger have a data owner?
- Are leading indicators used?
- Are lagged datasets acknowledged?
- Is the board reviewed on a schedule?
- Does foresight change actual decisions?
- Are post-mortems performed?
- Can future staff reconstruct why a signal mattered?
77. Foresight should not become a licence for speculation
Planning has public authority.
That authority should remain grounded.
A foresight paper can explore speculative futures.
A zoning restriction should still have:
- evidence;
- lawful rationale.
Keep the boundary between exploration and regulation explicit.
This protects both creativity and due process.
78. Foresight should make the plan more stable, not less
This seems paradoxical.
If a plan includes:
- triggers;
- optional pathways;
- contingency sites;
it can survive change without full rewrite.
A brittle plan requires emergency amendments.
A foresight-informed plan has flexibility designed in.
79. The goal is planned adaptability
Adaptability is not constant change.
It is the ability to change when evidence crosses a meaningful threshold.
This is why the signal board naturally connects to adaptation pathways.
One watches emerging futures.
The other sequences action when thresholds arrive.
80. The deepest test is whether the city can name what would make it change its mind
A plan often states:
This is our policy.
A foresight-ready plan also states:
If these conditions change, we will revisit it.
That sentence is a sign of maturity, not weakness.
Planning operates in a world that continues after adoption.
The Foresight Signal Board succeeds when uncertainty stops being an excuse for either paralysis or prediction theatre and becomes a managed stream of evidence, options and decision triggers.
Sources and further reading
- American Planning Association, 2026 Trend Report for Planners, January 28, 2026: https://www.planning.org/publications/document/9323378/
- American Planning Association, APA Foresight: https://www.planning.org/foresight/
- American Planning Association, Trend Universe: https://www.planning.org/foresight/trends/
- American Planning Association, Adaptation Pathways for Climate Resilience, PAS QuickNotes 115: https://www.planning.org/publications/document/9317990/
- OECD, Unlocking the Potential of Intermediary Cities for Regional Development, June 18, 2026: https://www.oecd.org/en/publications/unlocking-the-potential-of-intermediary-cities-for-regional-development_55b14922-en.html
- UN-Habitat, Innovate4Cities 2026 concludes in Nairobi with renewed momentum for urban climate action, June 25, 2026: https://unhabitat.org/news/25-jun-2026/innovate4cities-2026-concludes-in-nairobi-with-renewed-momentum-for-urban-climate
- New Zealand Ministry for the Environment, National Adaptation Framework, updated July 15, 2026: https://environment.govt.nz/what-government-is-doing/areas-of-work/climate-change/adapting-to-climate-change/national-adaptation-framework/
Continue reading: Strategic planning · Plan integration · Adaptation pathways · Full Town Planning Series Index.
81. Foresight needs a formal relationship with statutory planning
A strategic foresight team can identify change long before the comprehensive plan is due for review.
What happens then?
There should be a defined route.
Possible responses include:
- technical note;
- plan monitoring report;
- interim policy clarification;
- accelerated plan amendment;
- evidence reserved for next review.
The route should depend on consequence and urgency.
Without a formal handoff, foresight remains intellectually interesting but institutionally weak.
82. The comprehensive plan should name its critical assumptions
A future population.
A freight demand.
A climate horizon.
A technology assumption.
Plans often contain these assumptions implicitly.
Foresight works better when they are stated.
Then the monitoring system can ask:
Is this assumption still credible?
A plan can survive many changing details.
It becomes vulnerable when a foundational assumption breaks.
83. Assumption failure should have a response class
Possible classes:
Minor
- no policy change.
Material
- update implementation.
Structural
- reopen strategy.
This helps staff avoid two extremes:
- ignoring major change;
- reopening the entire plan for small deviations.
A mature plan knows which assumptions are load-bearing.
84. Foresight should identify policy lock-in
Some policies make future change harder.
Examples:
- disposing of strategic public land;
- approving housing beside a freight corridor;
- burying a utility in a location that blocks future transit.
Lock-in can be physical, legal or financial.
The signal board should flag decisions that close future pathways.
A cheap action today can create a very expensive constraint later.
85. Lock-in can also be social
A neighbourhood built entirely around car dependence creates:
- habits;
- expectations;
- political resistance.
A city can become locked into behaviour as well as infrastructure.
Foresight should therefore consider path dependence.
Town planning shapes what future choices feel normal.
86. Stranded assets are the financial mirror of lock-in
A city may invest in infrastructure whose demand falls.
Examples:
- oversized parking structures;
- obsolete fuel infrastructure;
- unused office-oriented transit assumptions.
Foresight cannot eliminate stranded assets.
It can reduce exposure by:
- modular design;
- staged investment;
- adaptable use.
Long-lived public assets deserve particular caution.
87. Foresight should test the downside of being wrong in both directions
Suppose the city believes autonomous vehicles will dominate.
If it builds for them and they do not arrive:
- wasted investment.
If it ignores them and they arrive:
- retrofit cost.
Compare both errors.
This two-sided test helps determine how much preparation is rational.
The best policy may be a small reversible step.
88. Option value can be estimated qualitatively
Not every option needs a financial model.
Ask:
- How expensive is preservation?
- How costly would future reacquisition be?
- How plausible is the future need?
- How many other uses can the option support?
A reserved corridor that can also function as green space may have high option value.
A single-purpose reservation with high opportunity cost may not.
89. Reserve maps should be updated from foresight
The existing TPW Reserve Map protects future choices.
The Foresight Signal Board can nominate new potential reserve needs.
Examples:
- future substation sites;
- flood-storage corridors;
- transit depots;
- emergency water.
This is a practical institutional connection.
A weak signal should not automatically freeze land.
It can trigger a reserve study.
90. Public land portfolios should be reviewed against emerging needs
Government often owns land through many agencies.
Before a major disposal round, compare the portfolio against:
- housing need;
- climate adaptation;
- utility expansion;
- community facilities.
The future may need land that no current department claims.
Foresight can make invisible future demand visible enough to study.
91. Regulatory sandboxes can test emerging uses
When a new land use arrives, planners may not know the right permanent rule.
A time-limited pilot can allow learning.
Possible tools:
- temporary use;
- special district;
- controlled demonstration.
The goal is to collect evidence before permanent codification.
The Temporary Use Permit owner provides the legal mechanism where appropriate.
92. Pilots need evaluation criteria before they begin
A pilot should state:
- what is being tested;
- what success means;
- what harm is unacceptable;
- what data will be collected.
Otherwise the pilot produces anecdotes.
Foresight should turn experimentation into structured learning.
93. Pilot success should not guarantee permanent approval
A small demonstration may not scale.
Ten autonomous shuttles are different from 5,000.
A microhub pilot may work at one station and fail citywide.
Scale should be tested explicitly.
Innovation planning needs both experimentation and caution.
94. Regulatory review dates can create adaptive codes
Some emerging-technology ordinances can include a scheduled review.
Example:
Review within three years using permit, complaint and market data.
This prevents the first rule from becoming permanent through inertia.
The city learns as the phenomenon matures.
95. Foresight should distinguish substitution from addition
A new technology may replace an old one.
Or it may add another layer.
E-commerce did not eliminate physical retail completely.
Remote work did not eliminate offices.
Electric vehicles do not eliminate roads.
Scenario design should avoid simplistic replacement stories.
Most urban transitions are mixed for long periods.
96. Transition periods often create the greatest planning complexity
During transition the city may need:
- old infrastructure;
- new infrastructure;
simultaneously.
Example:
petrol stations remain while charging demand rises.
This can temporarily increase land and capital requirements.
Foresight should plan for the transition, not only the assumed end state.
97. The end state may never stabilise
Technology, climate and demography continue changing.
This is why flexible systems matter.
The purpose of foresight is not to identify the final city.
There is no final city.
The purpose is to improve navigation through successive change.
98. Signals can interact nonlinearly
Two modest changes can combine into a large effect.
Example:
- ageing population;
- declining car ownership among younger adults;
- rising heat.
Together they may create much stronger demand for shaded local services than any trend suggests alone.
Scenario work should look for interactions.
Linear extrapolation misses them.
99. Foresight workshops should include “what would have to be true?”
When someone proposes a future:
Most deliveries will be autonomous.
Ask:
- What must happen technically?
- Legally?
- Economically?
- Socially?
This produces a chain of prerequisites.
Those prerequisites become signposts.
The future becomes testable rather than rhetorical.
100. Reverse scenario analysis can start from an undesirable outcome
Suppose the city wants to avoid:
- widespread uninsurable housing.
Work backward.
What drivers could produce it?
What early warning would appear?
Which decisions could reduce exposure?
This is useful for high-consequence risks.
It turns fear into planning logic.
101. Backcasting starts from a desired future
The city may want:
- net-zero mobility;
- abundant housing;
- low heat exposure.
Backcasting asks what must be true before that future can exist.
Then it identifies:
- infrastructure;
- policy;
- timing.
Foresight can therefore support both defensive resilience and ambitious transformation.
102. Backcasting and forecasting should be used together
Forecasting says:
Where are current trends heading?
Backcasting says:
What path reaches the desired outcome?
The gap between them is the intervention challenge.
This is especially useful for climate and housing policy.
Planning is not merely observing the future.
It is partly shaping it.
103. Foresight should recognise agency
Some forces are outside local control.
Others are influenced by planning.
Classify drivers:
External
- national interest rates.
Influenceable
- regional transit investment.
Controllable
- local zoning.
This prevents cities from promising control over macro trends while ignoring tools they do possess.
104. Uncontrollable drivers still deserve contingency planning
A local government cannot control:
- migration;
- global energy price.
It can control how prepared its housing and infrastructure systems are.
Foresight is especially valuable for external drivers because conventional targets may assume stability.
105. The signal board should include “planning leverage”
For each signal ask:
- What can the city control?
- What can it influence?
- What can it only monitor?
This creates realistic responses.
It also reduces the temptation to write zoning rules for problems that zoning cannot solve.
106. Political attention should not determine signal priority alone
A mayor may care deeply about one trend.
A quieter infrastructure risk may be more consequential.
The foresight method should provide a structured evidence base.
Political leaders still set priorities.
The board ensures they see beyond the loudest issue.
107. Foresight should be nonpartisan in description
APA’s trend work emphasises neutral description of trends and signals.
A local signal record should separate:
- observed change;
- normative judgment.
This is especially important for contested issues like migration, AI or remote work.
Planning staff should provide evidence before recommendation.
108. Recommendations can still be value-laden
Choosing resilience, inclusion or affordability involves public values.
The mistake is not values.
The mistake is hiding them inside claims about what the future “will” be.
State:
- evidence;
- scenario;
- planning objective;
- recommendation.
This makes the decision transparent.
109. Foresight should include distributional questions
A future can benefit the city overall and harm one group.
Automation may increase productivity and reduce some jobs.
Climate adaptation may protect one district while raising costs elsewhere.
Scenario analysis should ask:
- who gains?
- who loses?
- who is exposed?
This connects foresight to environmental justice.
110. Future residents should count too
Planning decisions affect people who do not yet live in the city.
They cannot attend today’s hearing.
Foresight is one way institutions represent future users.
Reserve land, climate adaptation and long-lived infrastructure are all partly decisions for absent future populations.
111. Foresight is also an intergenerational governance tool
Short political cycles reward visible results.
Long planning horizons contain costs and benefits beyond those cycles.
A formal signal and trigger system can preserve long-term evidence through political turnover.
The future should not depend entirely on one administration remembering why a concern mattered.
112. Staff turnover makes documentation essential
The planner who noticed the first signal may leave.
The reasoning should remain.
Record:
- source;
- interpretation;
- status;
- trigger.
Institutional memory is a resilience asset.
A foresight practice should survive personnel changes.
113. The city should publish enough foresight to invite external challenge
Public or expert review can identify:
- missing trends;
- weak assumptions.
Not every working note needs publication.
An annual summary can expose the major logic.
Foresight improves when outsiders can challenge it before capital is committed.
114. External challenge should not turn the signal board into a prediction market
The purpose is not to vote on which future is most popular.
The department still needs professional synthesis.
Public input adds information and values.
The foresight team maintains methodological discipline.
115. The final institutional achievement is earlier adaptation with less drama
A mature city should rarely need to say:
Nobody could have seen this coming.
Some shocks truly are surprising.
Many transitions leave signals for years.
The Foresight Signal Board is the planning system’s memory of what it noticed, its discipline for deciding what matters, and its contract with itself about when new evidence is strong enough to change direction.