A second opinion on the numbers, from someone with no stake in the closing.

The problem

If you're buying Japanese property from overseas, every number you've been shown — the price, the projected rent, the yield on the listing sheet — comes from someone who gets paid when you close. The listing agent's pro-forma is a sales document. English portals list properties; they don't underwrite them. And the public data that could check those numbers is in Japanese, scattered across government databases most buyers never see.

I check the numbers. That's the whole service.

What you get

You send a listing. I return a structured underwriting memo on that specific property:

• Price check — where the asking price sits in the actual recorded transaction history for that district and property class (MLIT reinfolib database), stated factually against the distribution.
• Achievable rent, quoted as a range — from current comparable listings matched on ward, layout, size, and building age. A range on principle: I blind-tested this comp method on 450 listings, and while typical error is small, tail cases can't be identified in advance — a single point estimate would be overselling precision. You'll see the method and its measured error, not just a number.
• Real net yield — the full drag stack the pro-forma skips: management and repair-reserve fees, property management, vacancy, property taxes, insurance, and all-in acquisition costs — computed at both ends of the rent range, with a line-by-line variance against the seller's numbers.
• Red flags — seismic-standard vintage, repair-reserve health vs. planned major works, land rights, hazard maps, tenancy contract terms, short-term-rental restrictions, financing walls for non-resident buyers — each with severity and the exact question that resolves it.
• Questions for the agent — a ranked list you can send as-is.

Sample memo

A complete memo on a real listing (address redacted), exactly as a client would receive it — method appendix included.

The review process cut our own headline number

The first draft of that sample netted the deal at 2.89–3.71%. Before publishing, the memo went through adversarial review: four cold passes, each read in the shoes of a paying buyer — one an experienced yield investor, one a first-time overseas buyer — scored, and pulled apart for what was missing. The reviews found real gaps. The cost stack had no line for tenant turnover (advertising fees and restoration between tenancies — real money in Japan's rental market). An early draft called the building's repair reserve "healthy" when it in fact sits below the MLIT guideline average. The price finding leaned on the friendlier of two public databases until the review forced both to be reported side by side.

Fixing what the reviews found moved the published number down, not up: 2.53–3.31% net. That is the direction honest review cuts when nobody is paid on the closing. The same review gate now runs on every memo before it is delivered — a memo reaches you only after it has survived a reviewer whose job was to find what I got wrong.

Pricing

Flat fee, paid up front, five days from payment to delivery. If your property's segment falls outside where my comp method is calibrated, I'll say so before you pay — or decline the engagement.

What this is not

Get a memo

Email contact@underwritejapan.com with (1) the listing link, (2) what you need the numbers to do — target net yield, needs-to-pencil, or keep-vs-sell — and (3) tenancy status if you know it. I'll confirm fit and calibration coverage before any payment.