January 18, 2026
By the HalfKey team
Why reikin doesn't apply on a 90-day Tokyo furnished stay
Plaza Homes lists reikin at one to two months' rent on the renting-in-Tokyo page. Their own monthly-apartments page says reikin: none. Both are correct. The reason is a different cash-flow problem, not a courtesy to foreigners.
Plaza Homes' "renting in Tokyo for foreigners" page lists 礼金 (reikin — "gratitude money," non-refundable payment to the landlord at signing) at "通常家賃の1～2ケ月分," meaning one to two months' rent. The same Plaza Homes site, on its monthly-apartments page, runs a comparison table with reikin marked as "なし," meaning none.
Both pages are accurate. They describe two different products that share the word rental and almost nothing else.
The midterm furnished segment did not waive reikin as a courtesy to foreigners. It removed reikin because the math broke. A 6-to-12-turnover-per-year unit and a 24-month tenancy face different cash-flow problems. The instrument that solves one does not solve the other.
Start with what reikin is solving for on a traditional lease.
A landlord rents an unfurnished 1K to a Japanese salaryman on a 2-year contract. The tenant pays ¥110,000 a month. Over 24 months, the landlord collects ¥2,640,000 in rent. They will turn the unit over once during that window. The cost of getting the next tenant in is real: agent's marketing, unit refresh, credit-check apparatus, broken-equipment risk that accrued over two years of one occupant. Real, but spread across a long denominator.
The landlord's risk is not turnover. It is everything else. Vacancy between leases. Unrecovered damage at exit. Late-payment friction during the term. Reikin is the instrument that prepays those risks at signing. One or two months of rent, ¥110,000 to ¥220,000, non-refundable, collected before the tenant moves a box. The landlord's exposure on the 24-month bet just dropped by that much. So did their motivation to second-guess the tenant.
That is what reikin does economically. It is a risk premium charged once, against a long-tail bet, payable at the moment of maximum tenant motivation (they want the unit). The 1-to-2-months figure is calibrated around the landlord's loss distribution across two years.
Now run the same accounting on a furnished midterm operator.
The same 1K, same building, but subleased by the operator at ¥125,000/month under a 5-year master lease. The operator lists the unit at ¥165,000 plus a ¥50,000 departure cleaning fee. Their average stay length runs 60 nights. So in any 12-month window, this unit turns over roughly six times.
Six turnovers a year, not one every two years. That is a 12x increase in turnover events versus the long-lease product, on the same physical box, at the same upstream sublease cost.
Reikin's whole architecture assumes a long denominator. Charging a non-refundable risk premium once at the start of a 60-day stay would mean choosing between two bad options. Option one: size it for a single turnover's risk. The number can't exceed maybe ¥20,000. Too small to bother. Option two: size it for the operator's full annual risk on the unit. Now the first guest of the year is paying for the wear-and-tear of the next five guests they will never meet.
Neither version survives the competitive booking funnel. Guests compare three operators on Suumo's price-per-night display. The first listing with a ¥150,000 line item called reikin loses the click.
So the operator did the only thing the math allowed. They moved the risk premium from a once-at-signing line item to a per-turnover line item.
That line item is the departure cleaning fee.
The fee structure looks different from reikin but does the same job.
Reikin on a 2-year ¥110,000 lease: ¥110,000 to ¥220,000, paid once, covers two years of risk. Per-month allocation: ¥4,600 to ¥9,200.
Departure cleaning on a 60-night ¥165,000 stay: ¥50,000, paid once, covers one turnover. Per-month allocation: ¥25,000 across the two-month stay. Of that ¥50,000, the cleaner gets roughly ¥18,000 and the operator keeps ¥32,000 as turnover margin.
Run the math at the operator's annual scale. Six turnovers × ¥32,000 of operator margin per turnover = ¥192,000 per year per unit, in turnover-margin specifically. On the long-lease side, the landlord collects roughly ¥110,000 of reikin every 24 months, or ¥55,000 per year amortized.
Two ways the same building generates risk-premium revenue. The midterm operator pulls about 3.5x what the landlord on a 2-year lease pulls from the same physical unit. They are turning it over 12x as often. The size of each individual cut is smaller. The frequency does the work.
This is also why operators do not waive cleaning fees the way landlords occasionally waive reikin in slow rental markets. Reikin is the landlord's once-every-two-years upside; skipping a single instance only costs them one tranche on a long denominator. Cleaning is the operator's cash-flow staple, repeated six times a year. Skipping one is a much bigger fraction of the business. Plaza Homes can write "reikin: none" on their monthly-apartments page because they know what is sitting in the cleaning-fee column doing the same job.
There is a second instrument worth naming. The setup fee.
Some operators charge a 事務手数料 (jimu tesūryō — "administrative handling fee," typically ¥10,000 to ¥30,000 charged once at booking). It covers contract drafting, the identity check, platform integration, and a small contribution to the operator's English-support headcount. Plaza Homes' monthly-apartments table calls this out alongside the cleaning fee. Their phrasing: "退去時清掃費用や事務手数料," exit cleaning costs and administrative fees.
The administrative fee is structurally closer to a Japanese tenant's 仲介手数料 (chūkai tesūryō — "brokerage fee," typically one month's rent paid to the agent on a long lease) than to reikin. It is paying for documented work done at booking, not a risk premium against the future. Some operators bake it into rent. Some itemize it. The midterm-pricing dial moves it around without changing what it covers.
Reikin and the brokerage fee on a 2-year lease together total roughly two to three months' rent. On a ¥110,000 unit, that's ¥220,000 to ¥330,000 of upfront non-refundable cost before the tenant has slept anywhere.
The midterm equivalent looks smaller. Cleaning fee plus administrative fee on a 60-night stay runs ¥50,000 to ¥80,000. Not because the operator is more generous. Because the operator collects the equivalent risk premium six times that year, distributed across six guests, none of whom individually pays what one long-lease tenant pays at signing.
The total annual margin extracted from the unit is comparable. The per-guest invoice is not.
There is one residual case where the boundary blurs.
A small minority of "monthly mansion" operators still charge a non-refundable line item they call 礼金 on stays of 30 nights or longer. These tend to be older Japanese-domestic brands aimed at corporate-relocation customers. The amount is small. Typically ¥10,000 to ¥30,000 flat, not a multiple of rent. The label is legacy.
Functionally it is doing the same work as a setup fee at a different operator. The accounting line on the operator's books reads "reikin." The economic content is administrative-fee shape: a one-time fixed charge against this specific stay. It does not scale with rent. It does not represent two years of landlord risk. It is a domain-name leftover from when the same operator offered both 2-year and short-term contracts and accountants used the same chart of accounts for both.
If you see "reikin" on a midterm furnished quote in 2026, ask the operator what it covers. The answer will be a flat figure under ¥30,000 and a description that sounds like setup work. The Plaza Homes page warning you about 1-to-2 months' rent is describing a different ledger entry under the same Japanese characters.
The reason this structural shift is worth understanding rather than just memorizing.
Read Plaza Homes' main renting-in-Tokyo guide, GaijinPot's foreigner-rental walkthrough, or MailMate's key-money explainer. They are correctly describing the 24-month residential-lease product. The 30-day to 12-month furnished product runs on different math. The same building can house both, under different operators, with different fee architectures. Both are legal under Japanese tenancy law.
The mistake is treating the 24-month-lease numbers as a baseline you should expect to encounter and feel relieved when they're absent. They are not absent. They have been redistributed across departure cleaning, the optional setup fee, and the rent number itself. The operator has done the redistribution math more carefully than you will. The total they extract from the unit's annual cash flow is similar; the shape of what hits your individual invoice is what changed.
A midterm operator's quote might show zero reikin, zero shikikin, a ¥50,000 departure cleaning fee, and a ¥15,000 administrative fee. That is not a deal relative to the long-lease product. It is the same pricing engine running on a different turnover ratio.
The first six search results are warning you about an instrument the midterm market replaced a decade ago. The replacement is sitting on your invoice under a different name, doing the same job for the same kind of math. Ask the operator which line is theirs and which line is the contractor's. They will tell you.
— halfkey lists the cleaning fee, the administrative fee, and the deposit on every unit page before you inquire. Compare those three lines across operators when you read other listings.