Break-even occupancy in Dubai

What occupancy does a Dubai holiday home need to beat an annual tenancy?

Each bar is that publisher's observed rate and occupancy run through this page's own cost model. Zero is the annual tenancy, not zero profit — every one of the three clears the break-even the competing pages print, and every one lands below the tenancy it replaced.

The rule

Yield tells you what a property earns when everything goes right. Break-even occupancy tells you how much has to go wrong before it costs you money, and it is the more useful number because vacancy is the risk that actually happens. Prices fall occasionally. Tenants leave routinely.

Every page that defines it gives the same formula, and on a Dubai property that formula answers the wrong question twice. It puts costs on the wrong side of the division, because one of the statutory charges here falls on occupied nights rather than on the year. And it measures against zero, when the thing an owner gave up was never zero.

The split that decides the answer

Break-even occupancy is fixed costs divided by what a fully occupied year leaves after the costs that scale. Everything turns on which costs go where, and the test is a single question: does this line stop when the unit is empty.

The same illustrative unit as net rental yield: AED 1,500,000, 900 sq ft, AED 105,000 a year in rent, service charge at AED 18 a foot. Figures are illustrative and the assumptions are stated, but the structure is not.

The service charge is the clearest case, and it is set by an instrument rather than by a landlord: Law No. (6) of 2019 puts the charge on the owner of the unit, which is why an empty flat in a serviced tower keeps billing. Service charges and reserves takes that apart properly. A district cooling capacity charge passes the same test for the same reason: it is billed on contracted capacity rather than on consumption, so it does not stop when the unit empties either, and the chiller bill prices the void it accrues.

So a fully let year collects AED 105,000, pays AED 5,250 in management, and has AED 99,750 to cover AED 22,950 of fixed cost. Break-even occupancy is 23.01%. The unit can sit empty for nine months of the year and still not cost its owner money.

That is the number for a property with no mortgage. It is also the number almost nobody actually has.

What a mortgage does to it

A mortgage instalment is fixed cost in its entirety. It does not care whether a tenant is present, and it moves the ratio further than any other single line.

At the 75% loan to value this site's worked example assumes, the loan is AED 1,125,000, and the Central Bank sets the outer term: "The maximum tenor of the mortgage loan is 25 years". Over that term, at each rate:

Read the bottom three rows carefully, because they are not a rounding. At 5% the property cannot break even at any occupancy at all. A unit let for all 365 days of the year, with no void, no bad tenant and no rent free period, still does not cover its own costs and its own instalment.

The reason is arithmetic rather than opinion. Fully let, after management, the unit leaves AED 76,800 for a lender. An instalment larger than that cannot be paid out of the property, whatever the occupancy. The highest mortgage rate this unit's own rent can carry at 75% loan to value is 4.72%.

Above that rate the property is not an income asset that happens to be geared. It is a leveraged position on price with a monthly subsidy attached, and it is worth describing that way when deciding whether to keep it. Nothing here says the position is wrong. It says the position should be named accurately, and that the cash to fund it has to come from somewhere the property is not.

This is also the honest answer to a question mortgage capacity raises and does not close: the Central Bank's ratios ask whether the borrower can service the loan, not whether the property can. Those are different tests and they can disagree.

The holiday home, and the benchmark nobody uses

Now the version of this question people actually search for, which is how many nights a Dubai holiday home has to sell before it is worth running.

Short let versus long let sets out the full cost stack and the instruments behind it. The relevant structure here is that the costs fall into three blocks, not two, and only one of them is recovered by occupancy.

At the illustrative AED 750 a night, one more night sold brings in AED 750, of which AED 150 goes to the operator and AED 60 goes on the tourism dirham and the share of a clean. Each night contributes AED 540, and the fixed block of AED 55,300 is recovered at that rate.

Which gives the first break-even. AED 55,300 at AED 540 a night is 103 nights, or 28.2% occupancy. Below that the unit loses money. Above it, the unit makes money.

And that is the number every competing page prints, and it is not the number that matters.

Because the alternative to a quiet holiday home is not an empty flat. It is the annual tenancy the owner gave up in order to run one, which on this unit nets AED 78,550 with no nights to sell, no guests to manage and no forecast to believe. To beat a contract that pays AED 78,550, the short let has to recover the fixed block and that amount: AED 133,850 at AED 540 a night.

That is 248 nights, or 67.9% occupancy.

The gap is 145 nights. Almost five months of the year, in which the holiday home is profitable on every page written about it and is nevertheless doing worse than a signed tenancy that took an afternoon to arrange. An owner running at 50% occupancy is well clear of break-even in the sense the word is normally used, and is losing roughly a month and a half of income against the thing they stopped doing.

Where the statute actually sits, which is not where people look

The Dubai specific costs are the ones owners expect to be the obstacle. They are not, and the arithmetic says so precisely.

The holiday home permit under Executive Council Resolution No. (49) of 2014 is AED 300 per bedroom, capped at AED 1,200 a year. On this one bedroom that is AED 300, which is 0.5% of the AED 55,300 fixed block. Furnishing, amortised, is fifty times larger. The permit is not what makes this hard.

The tourism dirham matters structurally rather than in size. Executive Council Resolution No. (2) of 2014 sets it at AED 10 per occupied room per night, and the words "per occupied room per night" are the whole point: it is charged on nights sold, so it belongs with the operator's share and the cleaning, reducing what each night contributes. A page that files it beside the permit as an annual cost of compliance has put it on the wrong side of the division and will understate break-even occupancy for every unit that is busy.

At 248 nights the tourism dirham comes to AED 2,480. It is a real cost and it is not the reason the answer is 67.9%.

The frontier, which is the part to keep

One nightly rate gives one answer, and an operator's forecast will not use the rate assumed here. So the useful form of this page is the pair: at each nightly rate, the occupancy that covers costs and the occupancy that beats the tenancy.

The top row is the one worth sitting with. At AED 500 a night there is no occupancy in the year that beats the annual tenancy, including 365 nights out of 365. Each night contributes AED 340 after the operator's share and the per night charges, and a full year of that does not reach AED 133,850. Below AED 534 a night the question of occupancy does not arise at all, and a forecast quoting a rate under that is answering a question the owner did not ask.

What Dubai holiday homes actually achieve

Everything above runs on an illustrative AED 750 a night. Three firms sell observed Dubai data, and each publishes an emirate-wide average nightly rate and an emirate-wide average occupancy. Those are the numbers an owner is actually quoted, so the honest test of this page is to put them through the frontier in place of an assumed rate.

Dollar rates are converted at AED 3.6725, the midpoint of the Central Bank's published automatic intervention rates of 3.672 buying and 3.673 selling. The dirham is pegged, so a market quote here would be spurious precision rather than accuracy.

Two of those occupancies are about seventy percent and one is forty, on the same city over overlapping years. The spread is not the finding on its own, because the periods differ and one of the three is counting a different denominator. The finding is what each pair does to the arithmetic.

Each pair through this page's own model

The cost stack is the illustrative one bedroom's, unchanged: AED 55,300 fixed, twenty percent to the operator, AED 60 a night. Only the nightly rate moves.

All three land short. Every one of them clears the break-even that every competing page prints, the one measured against zero, with room to spare. And every one of them, on this unit's costs, leaves between AED 14,094 and AED 20,700 less than the twelve month tenancy that took an afternoon to sign.

Note what AirROI's row does, because it is the one that looks like the exception. Its nightly rate is the highest of the three by sixty percent, which drags the occupancy needed to beat the tenancy all the way down to 47.1%, the only figure in that column a normal year could plausibly reach. And its own published occupancy is 39.9%, which is below it. A high rate does not rescue a low occupancy; it lowers the bar and then the occupancy fails to clear the lowered bar.

The denominator, which is where the seventy percent goes

AirDNA states its definition, and it is not the one most readers will assume: occupancy is the share of available nights that get booked. That is not the share of the year. A unit blocked by its owner for two months, or listed for only nine, can report a high occupancy on a modest count of nights sold, and a break-even measured against a twelve month tenancy is measured against twelve months.

Where a publisher gives both an annual revenue and a nightly rate, the calendar occupancy its own figures imply can be recovered by division: revenue over a full year at that rate. That is an identity, not an estimate.

Airbtics' figures agree with each other. AirDNA's do not, and its own definition says why. Sixty nine percent of available nights is 57.3% of the year on AirDNA's own revenue figure, a difference of 43 nights. An owner handed "69% occupancy in Dubai" and reading it as 251 nights sold is reading it as something AirDNA did not say.

AirROI is absent from that table deliberately. Its rate, occupancy and RevPAR agree with one another, since USD 286 at 39.9% is about USD 114 a night against a published RevPAR of USD 112. But its published annual revenue of USD 20,441 is roughly half of what that RevPAR produces over a year, and the page does not say which of the two sits on a different basis. One of those figures is not comparable with the other, there is no way to tell which from the page itself, and so neither is used here as a calendar occupancy.

The assumption that cuts the other way

The model above assumes a three night average stay, and cleaning is charged per stay rather than per night. AirROI publishes the observed figure: guests in Dubai stay 7.7 nights on average. Longer stays mean fewer cleans for the same nights sold, and that is money back to the short let. Running the same three pairs at the observed stay length is therefore the strongest case the published data can make for short letting, which is the reason to run it.

The gap narrows and does not close. The best of the three still finishes AED 5,994 behind the tenancy on its own published figures, with the one assumption that favours short letting set as generously as the observed data allows. Which is the honest statement of the finding: not that a Dubai holiday home cannot beat an annual tenancy, but that none of the three published market averages does, on this unit, once the comparison is made against the tenancy rather than against zero.

What this is not

The weak joint is the pairing, and it is worth naming exactly rather than burying. These are emirate-wide averages across every listing size, run against one illustrative one bedroom's cost stack. Two things make that defensible rather than arbitrary. AirROI reports that one bedroom units are 51.1% of Dubai's active listings, and that one and two bedrooms together are 77.1%, so the modal listing is close to the unit modelled here. And this unit's AED 105,000 tenancy and AED 18 a foot service charge are stated rather than hidden, so a reader with a real service charge and a real achievable rent can redo every row.

What the finding does not survive is substitution without arithmetic. A unit with a materially lower service charge, an achievable nightly rate above the city average, or no annual tenancy available to it at all is a different calculation. The frontier table above exists to be re-run, not quoted.

What to do with somebody else's forecast

A city average is not a forecast. An operator's forecast will use a rate and an occupancy for a specific building in a specific season, and it should, because that is the observed market data worth paying for and nothing above substitutes for it.

What the frontier does is tell you what to do with it once you have it. Take the two numbers any operator or data provider gives you, the nightly rate and the occupancy, and find them in the table. If the pair sits below the fourth column, the forecast is describing a unit that would earn more on a twelve month contract, however profitable the forecast says it is. That is a question the forecast will not raise on its own, because the firm making it is not paid to.

Three follow-ups worth asking in the same conversation. Whether the quoted occupancy is nights sold or nights available after owner use, because the two differ by however long the owner intends to stay, and because at least one of the three publishers above means the second and says so. Whether the quoted rate is before or after platform commission, since a 20% difference moves every row in that table. And whether the building permits short letting at all, which is the developer's or the owners' association's decision and not the operator's.

Reading the answer

Under 50%. The property carries itself through a bad year. Typically unleveraged, or leveraged lightly against a modest service charge.

50% to 70%. Normal for a sensibly geared rental. A month or two of vacancy is absorbed by the margin rather than funded from elsewhere.

70% to 85%. Thin. A single difficult re-letting, or a service charge increase of the kind rent increase caps cannot offset because caps bind rent and not costs, moves the owner to funding the asset.

Above 85%. The property is an income asset in name only.

No answer at all. The rows above that read "Never" are not a failure of the model. They are the model working, and they are common on leveraged Dubai stock at current mortgage rates.

The portfolio version

Run it across everything owned at once rather than unit by unit: fixed costs from every property against total rent at full occupancy, less the fees that scale. That number describes what a market wide soft patch does, which is the case where several units go quiet together rather than one going quiet alone.

It is also the number to have before selling anything, because an owner deciding which unit to release usually reaches for the lowest yield. The unit worth releasing is the one carrying the highest break-even occupancy, and those are frequently not the same property. Selling well covers what that exit costs once the decision is made.

The arithmetic

Where it breaks

• The arithmetic assumes the rent and the nightly rate hold. A re-letting at a lower rent raises break-even occupancy even if the unit is never empty for a day, which is the same risk arriving quietly.
• A repayment instalment is treated here as cost in full, when part of it is capital repaid to the owner. Using interest only makes every ratio on this page look better and is defensible, but the two cannot be mixed inside one comparison.
• Service charges rise, and a special levy for a chiller replacement can move the ratio several points in one letter. A break-even calculated on this year's charge is a snapshot rather than a property of the asset.
• The observed nightly rates and occupancies below are emirate-wide averages across every listing size, run against one illustrative one bedroom's cost stack. That pairing is the weakest joint in the comparison and the page says so: a unit with a materially lower service charge, an achievable rate above the city average, or no annual tenancy available to it is a different calculation. Nothing here is a forecast of what any individual unit will achieve.
• Break-even occupancy says nothing about capital value. A property with an excellent ratio in a falling market is still losing money, just not in cash, and the two losses are not offset against each other anywhere except in the owner's account.
• Portfolio level netting hides a single bad asset inside a healthy average, so the ratio is worth running both ways. A market wide soft patch is the case where several units go quiet together rather than one.
• The developer or community may refuse short letting outright, in which case the holiday home half of this page is arithmetic about something the title deed does not permit. That restriction is established before a furnishing budget, not after.

When to use it

Before borrowing, before furnishing a unit for short letting, and once a year on everything already owned, because break-even occupancy is the number that converts a yield into a month count and a month count is the thing an owner can actually plan against.

Sources

1. Central Bank of the UAE, Regulations Regarding Mortgage Loans (Circular No. 31/2013), Article 3, Important Ratios, consolidated version of 8 April 2020
2. Executive Council Resolution No. (49) of 2014, Schedules 1 and 2
3. Executive Council Resolution No. (2) of 2014, Schedule 1
4. Law No. (6) of 2019 Concerning Ownership of Jointly Owned Real Property in the Emirate of Dubai, Articles 25, 27, 30, 32 and 34
5. Executive Council Resolution No. (30) of 2013 Approving Fees of the Land Department
6. Central Bank of the UAE, Domestic Market Operations, automatic intervention rates USD/AED 3.672 buying and 3.673 selling, retrieved 28 September 2026
7. AirDNA, Dubai short-term rental overview, trailing twelve months to August 2026, page updated 22 September 2026, retrieved 28 September 2026
8. AirROI, Dubai Airbnb market report, 2026 dataset covering August 2025 to July 2026, page updated 12 September 2026, retrieved 28 September 2026
9. Airbtics, Dubai Airbnb data, revenue for February 2025 to January 2026 and rate and occupancy as at January 2026, data dated 12 March 2026, retrieved 28 September 2026

Last reviewed 28 September 2026. Commercial relationship disclosure: the author works in Dubai real estate brokerage. See the disclosure standards. Disclosure standards.