Building a Defensible Comparable-Sales Set in Japan

Start with completed evidence

An asking price records a seller's position. A completed transaction records an agreement, though public records may be aggregated, rounded or delayed. Begin with MLIT's real-estate information library and professionally available completed evidence. Keep listings as current competition in a separate table.

Record source, observation date, transaction period and any reporting limitation. Never combine an asking-price average and a completed-price average under one label.

Define the subject before searching

Write the subject's tenure, land rights, building, construction date, structure, unit area, floor, aspect, layout, condition, station route, management charges, reserve contribution, lease status and special rights. For a house, add land area, road, frontage, boundaries, topography and legal rebuildability.

A comparable can only be judged against a defined subject. Missing subject facts are due-diligence tasks, not permission to widen the set.

Use a hierarchy

For a condominium, search the same building first, then genuinely similar nearby buildings. Expand time before discarding physical comparability, but show the market movement used to interpret an older sale. For land and houses, prioritise road, zoning, shape, size, topography and effective access—not straight-line distance alone.

Set objective inclusion bands, then explain exceptions. A sale is not comparable merely because it is in the same ward or shares a station name.

Control the denominator

Confirm whether area means exclusive internal area, wall-centre area, registered area, land area or total building floor area. Japanese records and marketing material can use different definitions. State the denominator beside every price per square metre.

Do not adjust mechanically for every difference. Separate observed facts from professional judgment. Floor, view, renovation and tenant status may matter, but the adjustment needs paired evidence or a range—not intuition disguised as arithmetic.

Reconcile outliers

Map the set and inspect high and low observations. Possible explanations include renovation, distress, family transaction, leasehold, occupied sale, unusual land rights, severe condition, superior view or a different reporting basis. Remove an observation only with a recorded reason.

Show median, range and count. With a small sample, false precision is dangerous; a wider supported value range is more honest than a single number to the nearest yen.

Connect value to the decision

Translate the range into offer price, required conditions and walk-away point. Then add acquisition costs, immediate work and capital risks. A property at the low end of peers may still be expensive after correcting a road or association problem.

For an investment, reconcile comparable value with sustainable net income and an exit case. If the methods diverge, investigate rather than averaging them automatically.

Challenge the preliminary conclusion

After forming a range, search deliberately for evidence that would disprove it. Widen the time window, inspect the nearest superior and inferior buildings, and test whether one large transaction or one unusually small unit is driving the median. Compare the result with sustainable rent and current competing supply without mixing those measures into the completed set.

Ask a second reviewer to reproduce the set from the written criteria before seeing the proposed offer. Disagreement should be resolved by attributes and evidence, not by choosing the number closer to the asking price. If the range moves materially after one observation is removed, describe the result as fragile and reduce confidence.

Keep an audit sheet

For every observation retain location, date, price, area definition, price per area, building or land attributes, source and inclusion decision. Include a map, photographs where lawful, and a short adjustment narrative. Date the final range because a comparable set expires.

The set is defensible when another reviewer can reproduce the search, understand omissions and see which facts would change the conclusion.

Primary sources

• MLIT — Real-estate information library and transaction-price data: https://www.reinfolib.mlit.go.jp/
• MLIT — Land General Information System methodology and guidance: https://www.mlit.go.jp/tochi_fudousan_kensetsugyo/tochi_fudousan_kensetsugyo_tk17_000001_00003.html
• East Japan REINS — Market trend reports: https://www.reins.or.jp/library/trend/

Reviewed 23 August 2026. Public data have coverage and disclosure limits; subject documents and matched completed evidence control.