Singapore vs Tokyo
Compare the same six-category planning model in both cities, with the evidence state shown for every component so model anchors are not mistaken for direct observations.
At a glance
The full six-category planning model remains useful for scenario planning. 2 of 6 categories have direct component-level numeric evidence on both sides, while 2 of 6 have some captured numeric calibration on both sides. Broad calibration is weaker than a direct category observation, and model-only categories remain normalized planning inputs.
| Core category | Singapore | Tokyo |
|---|---|---|
| Rent · 1BR non-prime | S$2,900Source-linked model | ¥160,000Source-linked model |
| Utilities + connectivity | S$250Direct component evidence | ¥23,000Direct component evidence |
| Groceries | S$480Source-linked model | ¥45,000Source-linked model |
| Eating out | S$260Source-linked model | ¥18,000Source-linked model |
| Public transit | S$122Direct component evidence | ¥12,000Direct component evidence |
| Household + personal | S$180Source-linked model | ¥12,000Source-linked model |
| Full six-category model midpoint | S$4,220 | ¥272,000 |
What drives the difference?
Rent is the largest modeled component in both cities: about 69% of Singapore's six-category component sum and 59% of Tokyo's. The comparison is strongest where both sides carry direct component evidence. Broad calibration is useful context but is not a direct category measurement; a source-linked model anchor is weaker still. Matching rounded values should not be read as proof that real-world costs are identical.
Evidence behind this comparison
| Core category | Singapore | Tokyo | How to interpret |
|---|---|---|---|
| Rent · 1BR non-prime | Source-linked model | Source-linked model | At least one side is a source-linked model anchor without a captured numeric observation; do not read matching values as measured equality. |
| Utilities + connectivity | Direct component evidenceStarHub · reviewed 2026-09-24 | Direct component evidenceTokyo Metropolitan Government Bureau of Waterworks · reviewed 2026-09-24 | Both sides have direct component-level numeric evidence; the modeled values are still normalized planning anchors. |
| Groceries | Source-linked model | Source-linked model | At least one side is a source-linked model anchor without a captured numeric observation; do not read matching values as measured equality. |
| Eating out | Source-linked model | Source-linked model | At least one side is a source-linked model anchor without a captured numeric observation; do not read matching values as measured equality. |
| Public transit | Direct component evidencePublic Transport Council Singapore · 2026 | Direct component evidenceTokyo Metro · 2026 | Both sides have direct component-level numeric evidence; the modeled values are still normalized planning anchors. |
| Household + personal | Source-linked model | Source-linked model | At least one side is a source-linked model anchor without a captured numeric observation; do not read matching values as measured equality. |
Planning ranges, not single answers
For Singapore, the published one-person model band runs from 3690 to 4750. For Tokyo, it runs from 238000 to 306000. These bands are generated from the same persona and uncertainty framework, so they are more useful for scenario planning than a point estimate alone.
Why some numbers can match across cities
The following categories are model anchors on both sides right now: Rent · 1BR non-prime, Groceries, Eating out, Household + personal. The model can legitimately use the same rounded anchor in two cities when the evidence is incomplete or the normalized persona lands on the same planning value. That does not mean independent market observations produced identical prices. These rows are deliberately labeled so the limitation is visible.
Source depth on each city page
Singapore: 2/6 core components have captured numeric observations and 19 deeper evidence records are published. Tokyo: 2/6 core components have captured numeric observations and 19 deeper evidence records are published. The city pages show the source publishers, observation periods and evidence status component by component.