Bali’s branded residences are selling the wrong thing.

A developer pricing a branded residence in Bali today is selling two things at once, and only one of them is holding its value. The first is a unit — a leasehold apartment, badged with a hotel name, sold off-plan at a 25–35% premium to its unbranded neighbour. The second is an operating platform: the management, the compliance, and the rental machine that turns the building into income long after the buyers have moved on. Across every resort market that has run the cycle Bali is now entering, the price eventually stops following the badge and starts following the platform — and the developers who understood that survived the correction while the ones who sold the badge did not.

The surface of the market could not look healthier. Bali drew 6.95 million international arrivals in 2025, a record and a 10% gain on the year,¹ and the capital chasing those visitors has built more than 70 hospitality-managed developments, with branded residences now around 10% of active for-sale supply.¹ The wider Asian branded-residence pipeline has reached roughly $40 billion across some 50,000 units, up 30% in value year on year, and Indonesia’s slice of that — about $1.4 billion of launched product, a quarter of it in Bali — is still growing.² The names are escalating to match: Aman set the local template at Nusa Dua in the 1990s, Banyan Tree, Bvlgari and Karma followed through the 2000s, and Raffles Residences Bali is due to open in 2026.¹ Underneath that surface, though, the economics have already turned, and they turned while tourism was still breaking records, which is precisely the signal worth paying attention to.

Record arrivals, falling returns

When a destination sets a visitor record and its hotels still earn less per room, the explanation is rarely weak demand. In 2025 Bali’s hotels averaged 73.2% occupancy, easing about 2.5 points from 2024’s elevated level even as arrivals hit their record; the average daily rate rose 2.4% in rupiah but slipped about 2% in dollars on currency movement, and revenue per available room held broadly flat in rupiah, which — with the currency drag — means it fell in dollar terms.¹ The softening was sharpest in the budget and mid-economy tiers, where the report ties weaker performance directly to expanding supply. The property market is registering the same pressure: overall prices fell 5% in the second quarter of 2025 and apartments specifically fell 8%, even as freehold villas edged up 2.2%.³ Supply is growing faster than demand can absorb it, and it is doing so before the bulk of the pipeline has even completed.

The badge is being charged on the weakest product

The branded segment concentrates its premium on exactly the product the market is repricing downward. Villas are just 18% of Bali’s supply — up from 13% a year earlier — which leaves apartments and condominiums as the great majority of what is being built, better than four in five units,¹ sold at built-up prices of roughly IDR 50–65 million per square metre (about $3,100–4,100) and carrying the 25–35% brand premium that developers justify by pointing to global benchmarks.¹ Savills’ 2025 report puts the worldwide branded premium at about 33%, with established cities such as Dubai closer to 30%.⁴ The difficulty is that those benchmarks were largely earned in markets where the brand sits on freehold and resells through a deep secondary market, and Bali offers neither — foreigners cannot hold freehold at all. The one time anyone published local resale data, at the end of 2018, branded condominiums changed hands on the secondary market at 32% below their primary price while villas held a 7% premium,⁵ and the dominant product type today is the condominium, not the villa. Completed stock now trades at a 22% premium to off-plan,³ which is the clearest possible sign that buyers are nervous about what the pipeline will deliver, since they will pay materially more to see the finished building than to trust the rendering.

The yield that survives the spreadsheet

The pitch holding the whole structure together is the rental return, and it is where honest underwriting and marketing diverge most sharply. Developers advertise gross yields of 8–15%, and some market projected returns of 15–20%, but those are gross figures in a market where operating costs — management fees of 15–25%, platform commissions of another 15–18%, plus tax, maintenance, utilities and vacancy — consume 40–50% of revenue before the owner sees anything.⁶ Strip those out and the realistic net yield on a well-managed prime villa in Canggu or Uluwatu lands at 7–10%, while the island-wide average across ordinary, less professionally run stock sits closer to 3–6%,⁶ and even those figures are usually quoted against the purchase price alone rather than the true all-in capital of purchase plus furnishing plus transaction costs, which lowers them further still.

This matters because a return has two halves, income and capital, and both are now under pressure at the same time. The income side compresses as new supply floods the rental pool, and the capital side, which the total-return pitch quietly assumed would appreciate, is already falling for the apartment product that dominates branded supply. A leasehold makes that worse rather than better, because the asset is not land held in perpetuity but a wasting term, typically 25–35 years in the managed branded segment,¹ whose value bleeds mathematically toward zero as the clock runs down.

The return, as marketed vs as realistic	Marketed	Realistic
Gross rental yield	8–15%	—
Operating cost load (% of gross)	—	40–50%
Net yield — prime, well-managed villa	15–20% “ROI”	7–10%
Net yield — island-wide average	—	3–6%
Capital, apartments (YoY, Q2 2025)	appreciation assumed	−8%

In an oversupplied resort market, the market eventually stops valuing the unit and starts valuing the operating platform.

Five markets have already run this film

None of this is unprecedented, and that is the most useful thing a developer can know, because the same sequence has played out across continents and decades with enough consistency to treat as a base case rather than a coincidence.

Phuket is the closest analogue and runs 2 to 3 years ahead of Bali on the same curve. The island carries 40,600 units across 343 active projects, 83% of them condominiums,⁷ and it now sits on 10,159 completed but unsold units worth about THB 77 billion even as developers launched a further 10,613 units in a single year, a jump of almost 80%.⁸ Villa absorption collapsed from 5.5% a month to 1.8% in the first half of 2025, which implies roughly 4 years to clear the standing stock at the current pace,⁹ and the 30+30+30 leasehold renewal structure that underpinned a decade of sales is drawing growing legal scrutiny over how enforceable those renewals actually are.⁷ What held value through all of this was not the badge but the differentiated, well-located, genuinely operated product, and the established developers responded by moving upmarket and competing on build quality and management rather than on yield promises.¹⁰

Tulum is the same story unfolding in real time on another continent. The town carries somewhere between 6,600 and 11,500 active short-term rental listings depending on the data provider, with market-wide occupancy stuck below 50%, and the local real-estate association estimates a 40% fall in buyer interest since the post-pandemic peak while audited figures show new-home sales down 32–35% over two years against more than 10,000 units sitting in resale.¹¹ The split there is already stark, with professionally operated villas clearing 90% occupancy in high season while generic one-bedroom condos collapse, two outcomes drawn from the same tourism market.¹² And the tourism tailwind that justified the building — the new airport, the train — is itself now weakening, with scheduled flight capacity down more than a fifth year on year,¹² which is the point worth holding onto: demand growth did not rescue the generic product.

Miami ran the condo-hotel version of this in 2008, and its lesson is the most pointed of all, because when the boom broke the unsold projects sat for years and the way they were eventually salvaged was by conversion, into traditional hotels or into rental buildings, since the value that survived was in the operation rather than the unit sale.¹³ The roughly 22,000 units of overhang took about 5 years to clear, and the generic, interchangeable product cleared last. The Costa del Sol tells the European version across a longer arc, with coastal prices falling 30–40% from their 2007 peak — and nearer 50–60% in the worst-built fringe towns — transaction volumes not returning to pre-crisis levels until around 2017 and prices themselves not regaining their old highs until the mid-2020s, a recovery that split cleanly between the prime, well-located stock that fell least and recovered first and a residue of generic coastal product that the valuers at Tinsa effectively wrote off as obsolete.¹⁴

Two further markets complete the picture rather than repeat it. Dubai, the most mature branded-residence market in the world with more than 43,000 units, has resolved into an explicit verdict on what the premium is actually for: residences run by a genuine hotel operator delivering real service command and hold their premium, while those carrying a non-hospitality brand licensed onto the door increasingly do not — a distinction the developer Nick Candy put bluntly when he said that without a true operator the brand means nothing, and one Savills echoes in its own finding that brand alone is not enough.⁴ And the leasehold mathematics underneath all of it are not a matter of opinion, since Singapore’s widely used Bala’s Curve values a lease with 60 years left at about 80% of freehold and one with 30 years left at about 60%, with the decay accelerating sharply as the term shortens.¹⁵ That particular cliff is steepened by Singapore’s own financing rules and will bite more gently in cash-driven Bali, but the underlying logic — that a finite income stream is worth less every year it shortens — applies to a 25-year Bali leasehold with full force.

Market	The oversupply	What survived
Phuket (≈2–3 yrs ahead)	10,159 unsold; 83% condos; villa absorption 5.5%→1.8%/mo	Differentiated, operated, prime; renewal promise exposed
Tulum (now)	Buyer interest −40%; sales −32–35%; occupancy below 50%	Operated villas at 90% occ; generic condos collapsing
Miami (2008)	~22,000 unsold; projects broken 3+ yrs	The operation — projects converted to hotels and rentals
Costa del Sol (2008)	Prices −30–40%; recovered only by mid-2020s	Prime, well-located; generic stock written off as obsolete
Dubai (mature)	Two-speed market	Real-operator brands; “name on the door” did not
Singapore (leasehold math)	Bala’s Curve	60 yrs = 80%, 30 yrs = 60% of freehold; finite stream decays

The guaranteed-return promise that often accompanies these sales deserves its own caution, because it too has a documented history across Malaysia, the United Kingdom, Australia and India: the guarantee is typically funded out of the buyer’s own marked-up purchase price, underwritten by a thin special-purpose company rather than the developer itself, and structured to expire just as a wave of identical units reaches the rental market at once.¹⁶ A guaranteed yield is a marketing cost dressed as an investment return, and it is the first thing to disappear when the cycle turns.

What the pattern actually says

Read together, the precedents say one thing clearly enough to build a strategy on, which is that in an oversupplied resort market the price eventually stops following the unit and starts following the operating platform. Every one of these markets bifurcated, and the split was never simply branded against unbranded but operated, differentiated and well-located against generic, interchangeable and fringe. The badge on its own protected no one, the operating platform protected everyone who had one, corrections ran anywhere from 5 to 15 years, the worst generic stock sometimes never recovered at all, and tourism growth rescued none of it.