Case study 03 · Public policy
A better analysis starts with being corrected.
An urban-policy retrospective combined documentary research with expert review to challenge assumptions and explore alternative paths.
Retrospective + expert review- source documents + expert interview
- 6 + 1
- initial errors corrected through review
- 4
The decision
What could a decision-maker have seen differently at the start?
A long-running urban development initiative provided a documented decision to examine in hindsight. The aim was to reconstruct the original choices, not attribute the initiative’s historic results to Arcana.
The approach
From evidence to next steps.
Read the record
Six source documents formed the initial evidence base. Figures were connected to sources and calculations so the reasoning could be examined.
Invite correction
An expert interview exposed four initial errors, including assumptions about public land ownership and the type of housing issue. The analysis was revised.
Rebuild the decision
The revised evidence was used to explore plausible alternative paths using the information available at the time, with hindsight kept separate.
What the work revealed
Expert context changed the recommendations.
An early recommendation depended on an incorrect assumption about land ownership. Once that assumption was corrected, the recommendation was withdrawn. The review also surfaced context absent from the written record.
The next move
Keep the expert-review loop inside the analysis. Separate documented facts, witness accounts and counterfactual possibilities before using any of them to support a recommendation.
Evidence, in context
What this case does—and doesn’t—show.
- This was a retrospective, not a live policy intervention or a forecast validated in advance.
- Expert accounts added context but were not all independently corroborated.
- Alternative paths are plausible scenarios, not implemented policies or proven outcomes. Historic development gains are not Arcana impact metrics.
Arcana policy research and expert review. Location, project and participant identities withheld. This anonymized summary reflects the supplied engagement report; outcomes have not been independently audited.