Timing: what you can know about the cycle, and what you cannot

Every year somebody publishes a guide to investing in Bali next year. It is the same guide. The reason it can be rewritten annually without new information is that it never contained any.

Timing is not one question. It is three, and they have different answers. Some things about a purchase are observable today, a few are measured by somebody other than the person selling you the villa, and one is not knowable by anybody. This sorts them.

What you can observe on a single scheme

Almost everything genuinely actionable about timing is scheme-level, not market-level.

Construction stage is visible. A scheme at foundations and a scheme at roof level are different risks, and they are priced differently. Understand what that gap is before you treat it as a discount: an early-stage price is compensation for carrying the developer's completion risk and funding their build, not a bargain the market has failed to notice. Stage seven covers what to monitor once you are in.

Phase pricing is published and dated. A developer releasing units in tranches leaves a record of what they asked in phase one, phase two and phase three, and how long each took to clear. That record tells you more about the local demand curve than any island-wide statistic will.

Unsold stock in the last completed phase is observable if you ask, and it is the single most useful question about a developer's pricing that nobody asks. A scheme that is launching phase four with eleven units of phase two unsold is telling you something.

Pre-release timing is ours, and we are explicit about it: our pre-release list at /pre-release shows what we have accredited before public launch, fourteen days ahead. What that gives you is first choice at a stated price. What it does not give you, and what we will not claim, is knowledge of what the next phase will price at.

What is measured, and by whom

Indonesia's statistics agency, BPS, publishes Bali tourism monthly. It is the only regular, methodologically stated, publicly available series that bears on this asset class at all, and it measures hotels rather than villas.

As reported from the BPS Bali releases:

• Foreign arrivals to Bali in the first half of 2026 were 3,203,156, down 2.42% on the 3.28 million in the first half of 2025.
• July 2026 arrivals were 697,809, up 0.10% on July 2025. The head of BPS Bali said February and July were the only months of 2026 to rise year on year.
• Star-hotel room occupancy in June 2026 was 64.87%, up 3.71 points on May and 0.21 points on June 2025. Five-star hotels ran at 72.34% and one-star at 45.96%.
• Australia supplied 802,837 arrivals in the first half of 2026, just over a quarter of the total. China, India, the United Kingdom and the United States follow, none of them above 300,000.

Two honest readings of that. Arrivals have stopped growing and occupancy has not fallen with them, which is consistent with a market where the top end is holding and the bottom end is not. And Australia is a quarter of the demand, so an Australian recession or a fuel price shock on that route is a Bali occupancy event.

Then the caveats, which are as important as the numbers. These are hotels, not villas. They are the whole island, not your street. They are lagging by four to eight weeks. And a figure about June 2026 is not a figure about what your villa earns in 2029. Use them as a floor under your own assumptions, not as a forecast.

What is not measured at all

This is the part competitors skip, and it is the part that should change how you read everything else.

There is no register of private villas in Bali. BPS counts classified accommodation; the pool of privately owned villas let through online platforms is not in any official series. The only supply counts in circulation come from commercial scrapes of short-let listings. Some of them are decent work. We have not licensed one and we will not quote a supply figure we cannot show you the methodology for.

There is no sold-price series. No MLS, no land registry price feed, no portal history. Every statement of the form "Bali property has appreciated X% a year" is either asking prices, one agency's own book, or invention. Ask whoever shows you one where it came from, and watch what happens.

There is no achieved-rent series either. Yields quoted in brochures are modelled from an assumed nightly rate and an assumed occupancy. Our yield calculator at /guides/yield-calculator is the same kind of model. The difference is that it shows you every assumption and lets you change them, and it is still a model.

So when someone tells you the market is about to move, ask what series they are reading. There isn't one.

What is announced, and can be diarised

Policy is the one part of the cycle that arrives with dates on it, which makes it the one part you can actually track.

Two dated events frame where Bali's supply politics currently sits. On 9 January 2025 the governor publicly rejected a blanket moratorium on hotel and villa construction, saying he would legislate strict control instead. On 14 September 2025, after floods that killed 18 people, the provincial government announced a moratorium on converting productive agricultural land and water-catchment areas to hotels and villas.

As at 1 September 2026 we have not found a numbered provincial instrument enacting either. We are stating our search, not the law, and the distinction matters commercially: an announced moratorium and a gazetted one behave very differently at a permit counter, and the difference decides whether the scheme you are buying into gets its approvals renewed. If your developer's answer to "what is the status of the moratorium" is a press cutting, they do not know.

The register at /guides/regulation carries the rules we have sourced to a primary text, each with an effective date and a review date. That is the format to hold anyone to, including us.

The part nobody knows

Nobody can tell you where the top is. Anyone who says they can is selling something, and in this market they are usually selling the thing they are forecasting.

Two specific unknowns deserve naming, because they are large and they are yours.

The exchange rate over the build. A twenty-four month construction period is twenty-four months of currency exposure between what you earn in and what you pay in, on a payment schedule you have already committed to. You can hedge the instalments. You cannot hedge the rental income that arrives four years later.

What gets approved next door. Zoning varies plot to plot rather than area to area, and the plot that decides your view and your rate is not the one you bought. Ask what is approved within five hundred metres. Nobody can tell you what will be approved.

What actually sets your timing

Since the market will not tell you when to buy, the decision has to come from your own position. Three things do most of the work.

Your holding period. The round trip is expensive and slow. Going in costs 8 to 12% on top of the price once tax, notary and, if you need one, company formation are counted. Coming out costs about 7.5%, being the 2.5% final tax on the transfer and around 5% to an agent, and it takes six to twelve months. Run your own numbers in /guides/cost-of-ownership. If your honest holding period is under five years, the costs and the lease decay will eat most of the case before the market gets a vote.

Your structure. A PT PMA takes six to ten weeks to incorporate. Reserving a unit and then starting the company is how people miss payment deadlines and lose reservation fees. If the structure is the constraint, the structure sets the date.

The lease clock. Establish whether your term starts at signature or at handover, because the deed decides and the brochure does not. Losing the first two years of a thirty-year lease to a build costs about 2% of its value, which is small. That number is small precisely because the front years are cheap, and it is the same arithmetic that makes the back years brutal: two years off the last five costs more than a third of what is left. Model it at /guides/lease-decay.

Instead of a forecast

• Ask what series any market claim comes from, and stop reading when there isn't one.
• Track the scheme, not the island: phase prices, unsold stock in the last phase, permit status, and the developer's completed projects.
• Diarise the policy, because it is the only part of the cycle with dates attached, and check whether an announcement has become an instrument.
• Buy when your holding period, your structure and your currency exposure are all ready. That is a date you can know.