Airbnb in Bali vs Japan: Canggu vs Kyoto Compared (2022-2026)

Two tourism booms, two different scales: visitors 2022-2025

Both Bali and Japan broke their all-time visitor records in 2025: 6,948,754 international arrivals for the Indonesian island (BPS Bali) and 42.7 million for Japan (JNTO). But neither the scale nor the trajectory compares.

Year	Bali (intl. arrivals)	Japan (intl. arrivals)	Kyoto (foreign visitors)
2019 (pre-Covid benchmark)	6.28M	31.9M
2022	2.16M	3.83M (borders closed until October)	0.58M foreign guests only
2023	5.27M (+144.6%)	25.1M	5.36M foreign guests (+830%)
2024	6.33M (+19.4%)	36.9M (+47.1%)	10.88M foreign visitors (+53.5%)
2025	6.95M (record)	42.7M (record)	hotel occupancy 80.6% (record)

Three takeaways:

• In 2025 Japan welcomed six times more international visitors than Bali, running +34% above its pre-Covid level (Bali: +11%).
• Kyoto alone received more foreign visitors (10.88M in 2024) than the entire island of Bali (6.33M the same year).
• A historic first: in 2024, foreign guests staying overnight in Kyoto (8.21M) outnumbered Japanese guests (8.09M) for the first time since records began in 1958. In 2025, 66.4% of Kyoto hotel guests were foreign.

And this clientele spends: a foreign visitor spends roughly three times more in Kyoto than a Japanese visitor (¥71,661 vs ¥23,809 in 2023, i.e. ~€478 vs ~€159). That demand is the bedrock of any rental property purchase in Japan.

ADR, occupancy, revenue: the Canggu vs Kyoto scorecard

To compare like with like, we use the same market-data source (AirROI, trailing 12 months 2025-2026) for both cities.

Metric (same source, same period)	Canggu	Kyoto
ADR (average nightly rate)	$215 (~€198)	$194 (~¥29,000, ~€179)
Average occupancy rate	37.5%	52.1%
Median occupancy	38%	56%
Top 10% of listings	82%+	89%+
RevPAR (revenue per available night)	$84 (~€77)	$106 (~€98)
Average annual revenue per listing	$22,855 (~€21,000)	$34,042 (~€31,300), i.e. +49%
Revenue trend, year on year	−19.2%	−9.6%
Seasonal peak	July: $221, 50.6%	October: $243, 74.8%
Seasonal trough	February: $188, 32.6%	February: $159, 37.0%

The nightly rate is a deceptive draw: a Canggu villa and a Kyoto machiya (machiya, traditional townhouse) charge almost the same per night. The whole gap lies in occupancy: Kyoto's winter trough (37%) equals Canggu's annual average, and Kyoto's October peak (74.8%) beats Canggu's best month by 24 points.

In fairness: data providers disagree on absolute levels (Airbtics, which filters out semi-dormant listings, shows ~63% for Bali and ~81% for Kyoto). Whichever source you pick, Kyoto's lead holds at +14 to +18 points. Canggu's trajectory is a correction: operators reported around 65% occupancy in 2022, ~50% in 2023, 37-41% today. To see how Kyoto ranks within Japan, read our city-by-city Airbnb profitability comparison for Japan.

Why the gap keeps widening: free-for-all supply vs licensed supply

The tourism boom is real on both sides. What differs is how fast housing supply responded.

Bali: the tap wide open

• 38,640 active listings in Bali as of January 2026, up +27.7% in one year and +107% in three years: supply more than doubled since early 2023.
• Over the same period, international arrivals grew "only" +32% (5.27M in 2023 → 6.95M in 2025).
• The mechanical result island-wide: occupancy down ~16% and median revenue per listing down ~20% over three years; average host discounts rose from 15% to 19%.
• The construction moratorium announced in late 2025 specifically spares Badung (Canggu, Seminyak, Uluwatu) and Gianyar (Ubud): precisely where building is most intense. The only self-correction so far: Canggu's supply shrank 6.9% in a year… by eliminating the losers.

Kyoto: the tap under licence

• Only ~4,110 active listings for 10.9M foreign visitors.
• Registered minpaku (minpaku, home-sharing short-term rental) is capped at 180 nights per year nationally, and Kyoto restricts it to about 60 days per year (mid-January to mid-March) in residential zones. Full details in our guide to the 180-day minpaku licence.
• Serious operations therefore run on the kan'i shukusho licence (kan'i shukusho, simple lodging under the hotel law), valid 365 days a year: Kyoto counts 3,198 of them (May 2026), still below the pre-Covid peak of 3,337. Licensed supply has not even recovered its 2019 level while demand explodes. Our article on Airbnb regulation in Kyoto covers zones and conditions.

The ratio that says it all: Bali has about 5.6 listings per 1,000 foreign visitors; Kyoto about 0.4: fourteen times less supply per visitor. In Bali the barrier to entry is a chequebook; in Kyoto it is a licence. It is demanding to obtain, and that is exactly what protects the returns of those who hold one.

Real returns: what the brochures leave out

This is where the two markets differ most from their own sales pitch.

Canggu: promises at 80% occupancy, reality at 37.5%

Bali developers routinely market projected ROIs of 15-22% per year, built on 80-90% occupancy assumptions. Measured data says 37.5% on average, and the industry itself admits commercial projections overstate reality "by 30 to 50%". Credible ranges: 5-8% net for an average property, 9-13% for the best professionally-managed ones, before accounting for lease-value erosion (next section).

Kyoto: 5-7% net, on an appreciating asset

In Kyoto, a licensed machiya guesthouse typically nets 5-7%, with the best operations exceeding 10% by combining a smart purchase with optimised management (the top 10% of listings run at 89%+ occupancy). A well-located whole machiya rents for ¥35,000-55,000 a night (~€233-367) in Gion or Higashiyama. Meanwhile the land appreciates: +3.6% a year for residential Kyoto in 2025 (+8.7% commercial), while the machiya stock has shrunk from ~47,000 to ~34,000 units in fifteen years, scarcity works for the owner.

Entry tickets

• Canggu: 2-bedroom leasehold villa $135,000-370,000 (~€124,000-340,000), 3-bedroom $264,000-639,000, i.e. ~$2,200-2,700/m² built.
• Kyoto: machiya to renovate ¥10-40M (~€67,000-267,000), full renovation ¥400,000-500,000/m²; licensed, ready-to-run guesthouses from ~¥31M (~€207,000) in the city centre: exactly the kind of properties we track in Les pépites, our curated and analysed listings. To understand this niche, read our guides to buying a machiya in Kyoto and the 365-day lodging licence.

Ownership rights: a melting lease vs perpetual freehold

This is the structural chasm between the two destinations: one that no occupancy rate can offset.

Criterion	Bali (Indonesia)	Japan
Freehold for a foreigner	Never. Hak Milik (freehold) is reserved for Indonesian citizens under the 1960 Agrarian Law No. 5	Yes, perpetual, land + building, no visa or residency required
Common structure	Hak Sewa (leasehold) of 25-30 years: the asset's value trends to zero at lease end	Ownership recorded in the land registry, inheritable and mortgageable
Resale	Very difficult under 15 years remaining, near-impossible under 10	Normal resale market
Lease extension	$30,000-120,000 depending on the area, never guaranteed	Not applicable
Corporate structure	PT PMA required to operate legally (investment plan ≥ IDR 10bn, ~€590,000)	None required: purchase in your own name
Short-term rental licence	Pondok wisata reserved for Indonesians → foreigners need a PT PMA + villa licence (6-12 months of paperwork)	Kan'i shukusho licence open to foreigners; note it does not transfer on resale

The one-liner: in Canggu, your $500,000 villa will be worth zero when the lease expires; in Kyoto, your machiya sits on land you own forever. Japan is one of the few Asian countries where a foreigner, even without a visa, buys under exactly the same conditions as a national: our guide to buying in Japan without a visa walks through the process, and total purchase costs stay at or below 6% of the price (see the full buying guide). One key reminder for both countries: buying property grants no visa, neither in Japan nor in Indonesia.

Non-resident taxation: taxed on gross in Bali, on net in Japan

Taxation widens the gap further: in a way few investors anticipate.

Levy	Bali (Indonesia)	Japan
Tax on a non-resident's rent	PPh 26: 20% of GROSS revenue, no deductible expenses (reducible to ~10% under some tax treaties)	Possible 20.42% withholding, then final taxation on NET income at progressive rates: effective tax often falls below 10% of gross rent
Depreciation	Not deductible (gross-basis taxation)	Wooden buildings over 22 years old: depreciable over 4 years, enough to wipe out much of the rental income for tax purposes
Local tax on stays	10% (PB1 hotel tax, Badung regency)	Lodging tax paid by the guest (in Kyoto: steeply raised on 1 March 2026, from ¥200 up to ¥10,000 per person per night depending on room rate)
Annual holding taxes	PBB (local property tax)	Fixed-asset tax 1.4% + city-planning tax 0.3% of the assessed value (far below market price)

Gross-basis taxation is a silent yield killer: 20% of turnover, not of profit. On a villa running 37.5% occupancy whose costs already absorb half the revenue, the Indonesian tax can swallow 40% of actual profit. Japan's mechanism (net-basis taxation, deductible expenses, accelerated wood depreciation) is detailed in our article on Japan rental income tax for non-residents.

Case study: 300,000 euros in Canggu or in Kyoto

An illustration using the average market assumptions above (rounded orders of magnitude; every project deserves its own numbers).

Option A: 2-bedroom Canggu villa, ~27-year leasehold, ~$330,000 (~€300,000)

• Revenue at average ADR ($215) and average occupancy (37.5%): ~$29,400/year (~€27,000).
• Costs: 10% hotel tax (~$2,900), management 20-25% (~$6,600-7,400), cleaning/platforms/utilities ~15-20% (~$4,400-5,900) → pre-tax net ~$13,000-15,500.
• Non-resident tax PPh 26: 20% of GROSS = ~$5,900 (or ~$2,900 where a tax treaty applies).
• After-tax net: ~$7,000-12,500 (~€6,500-11,500), i.e. ~2-4% net: on an asset whose residual value declines every year toward zero at lease expiry.

Option B: Licensed machiya guesthouse in Kyoto, freehold, ~¥45M all-in (~€300,000)

• Property at ~¥42.5M + purchase costs ≤ 6% ≈ ¥45M. Cash purchase (Japanese mortgages are reserved for residents employed in Japan).
• Revenue at average ADR (~¥29,000) and average occupancy (52%): ~¥5.5M/year (~€36,700).
• Short-term-rental operating costs 40-50% (management 20-30%, cleaning, platforms, utilities, upkeep) → pre-tax net ~¥2.8-3.3M (~€18,700-22,000), i.e. ~6-7%.
• Net-basis taxation with wooden-building depreciation: tax often modest in the early years → ~5.5-6.5% net after tax, plus land appreciation (+3.6%/year in Kyoto in 2025) on an asset held in perpetuity.

Case verdict: for the same stake, Kyoto nets roughly double, on an asset that appreciates instead of melting. Run your own assumptions (price, ADR, occupancy, costs) through our yield simulator, calibrated on real Japanese market data.

Common mistakes to avoid

On both sides of the China Sea, the same traps recur:

• Believing brochure occupancy rates. 80-90% promised in Canggu, 37.5% measured. Always demand independent market data, never the seller's business plan.
• Forgetting the leasehold melts. A Bali villa loses value every year by legal construction; resale becomes very hard under 15 years remaining and the extension ($30,000-120,000) is never guaranteed.
• Assuming the licence follows the property in Kyoto. The kan'i shukusho licence is not transferable: the new owner must reapply, under whatever rules apply at the time of resale. Verifying re-licensability BEFORE signing is checkpoint number one.
• Ignoring zoning. Minpaku capped at ~60 days in Kyoto's residential zones; villa licences impossible outside tourism zoning in Bali.
• Neglecting local operations. Kyoto enforces the kaketsuke yōken (駆け付け要件, the duty to reach the property within 10 minutes): an operator must be stationed within ~800m during every stay, a local manager is unavoidable.
• Comparing Bali gross yields with Japanese net yields, while forgetting Indonesia's 20% gross-basis tax.
• Thinking a purchase grants a visa. Neither Japan nor Indonesia gives buyers any right of residence.

Expert tips

• Aim for the market's top quartile, not the average: in Kyoto the top 10% run at 89%+ occupancy. Location, design and management make all the difference.
• Prefer an already-licensed kan'i shukusho property (after checking its re-licensability), or one located in a zone where licences remain obtainable.
• Budget renovation realistically: ¥400,000-500,000/m² for a full machiya, plus a structural inspection (older machiya predate current seismic standards).

Conclusion: which market for which investor profile

The verdict, criterion by criterion:

Criterion	Advantage
Tourism volume and growth	Japan (6× Bali's volume, +34% vs pre-Covid)
Nightly rate (ADR)	Draw (~€200)
Occupancy rate	Kyoto (+14 to +18 points, all sources)
Revenue per listing	Kyoto (+49%)
Supply control	Kyoto (licensing vs unrestricted building)
Ownership rights	Japan (perpetual freehold vs melting lease)
Effective taxation	Japan (net-basis + 4-year depreciation)
Day-to-day operating simplicity	Bali (ubiquitous delegated management, no kaketsuke equivalent)
Licensed entry ticket	Kyoto (ready-to-run from ~€207,000)

Bali keeps genuine strengths: a lifestyle, a mature property-management ecosystem, and a handful of elite operators who beat the averages. But for a wealth-building investor, the numbers are unambiguous: Bali offers rental yield on an asset you will never own; Kyoto offers the same headline yield (roughly double after tax) on an asset you own forever, in a city that receives more foreign visitors than all of Bali with fourteen times less supply per visitor.

If this market appeals to you, explore what is achievable at your budget, and if you would like guidance from property search to key handover (licence included), discover our personalised support.

Frequently asked questions

Is Airbnb more profitable in Bali or in Japan?

At equivalent nightly rates (~$200), Kyoto runs at 52.1% occupancy versus 37.5% in Canggu and earns 49% more per listing per year ($34,042 vs $22,855, AirROI 2025-2026 data). After tax the gap widens further: Indonesia taxes a non-resident's rent at 20% of gross, while Japan taxes net income.

Can a foreigner own a Bali villa freehold?

No, never: freehold (Hak Milik) is reserved for Indonesian citizens under the 1960 Agrarian Law. Foreigners only access 25-30 year leaseholds, whose value trends to zero, or corporate structures via a PT PMA. In Japan, by contrast, a foreigner even without a visa buys land and building in perpetual freehold.

What is the real occupancy rate of an Airbnb in Canggu and in Kyoto?

Over the trailing 12 months 2025-2026 and with identical methodology (AirROI), Canggu averages 37.5% occupancy and Kyoto 52.1%. Providers that filter out semi-dormant listings (Airbtics) show ~63% for Bali and ~81% for Kyoto. Whichever provider, Kyoto's lead is 14 to 18 points.

What budget for a licensed guesthouse in Kyoto?

Already-licensed machiya guesthouses start around ¥31M (~€207,000) in the city centre, with purchase costs at or below 6% on top. A machiya to renovate runs ¥10-40M, with a full renovation to budget at ¥400,000-500,000/m². Non-residents buy cash, as Japanese mortgages are reserved for residents employed in Japan.

Can you rent short-term all year round in Kyoto?

Yes, but only with the kan'i shukusho licence (simple lodging under the hotel law), valid 365 days a year. Registered minpaku is capped at 180 nights nationally, and Kyoto restricts it to about 60 days per year in residential zones: which is why the lodging licence dominates in this city.

How is a non-resident's rental income taxed in Bali and in Japan?

In Indonesia, PPh 26 withholds 20% of GROSS rent with no deductible expenses (sometimes ~10% under a tax treaty). In Japan, after a possible 20.42% withholding, final tax is computed on NET income: with expenses deducted and wooden buildings depreciable over 4 years, effective tax often falls below 10% of gross.

Does buying property in Bali or Japan grant a visa?

No, in neither country. Buying real estate opens no right of residence in Indonesia or Japan. Visa strategies (business manager, digital nomad, etc.) are entirely separate processes from the purchase itself.