1) The problem with ward averages
Averages hide composition effects: a shift in layout mix or vintage can move the mean without any real price change. Neighborhood borders are porous; commuter time and line access dominate more than ward lines.
2) The station as a micro-market
Each station’s reach defines a natural catchment. We normalize comps by walking minutes andbuilding age to compare assets fairly across the network.
Core idea: Price per m² = P₀ × e^(−β × WalkMinutes), adjusted for vintage and layout.
3) What you can do with station-level models
- Spot mispricings vs distance-decay fair values.
- Quantify “age drag” and time-to-refurbishment.
- Segment yield bands by layout (1K vs 1LDK vs 2LDK).
- Underwrite exit scenarios with sensitivity to ±50 bps yield.
4) From data to decisions
Start with station-tagged sold comps, normalize, and fit decay/age coefficients. Use the fits to compare current listings, and stress-test NOI under realistic rent assumptions. The result: cleaner pricing, better risk control, and repeatable decision-making.
Continue reading: Distance-to-Station: the price premium