.webp)
Every projection you have been shown for a Zanzibar apartment describes a good year, and that is precisely why this article describes a bad one. A rental property's worst-case occupancy, the year when the bookings soften and the calendar shows more white than a brochure ever would, is the number an honest investor actually underwrites, because anyone can own an asset through a record season. This piece takes a real Paje studio, runs it below the published worst case, and shows every line of what happens. In this guide, we'll explore why we are publishing the case against ourselves, what genuinely causes a bad year here, the stress test line by line, the worked numbers, what history says about how this market behaves under pressure, what separates resilient units from fragile ones, and how to build your own margin of safety before you buy.
Why we are publishing the case against ourselves
The suspicion is fair and common: projections only show sunshine, and the developer's spreadsheet always ends in a happy number. The suspicion is also useful, because it points at the correct question, which is not "how good can this get" but "how bad can this get while still working."
There is a self-interested reason to answer it openly, and stating that reason is part of the honesty. Buyers who purchase on the stretch case and meet a soft season become unhappy owners, and unhappy owners are expensive in every way that matters to a developer with a long-term operation on the island. Buyers who purchase on the downside case and then experience a normal year become the owners who refer their friends. The published Vela Breeze ROI simulation already runs a worst case at 50 percent occupancy rather than quoting only its best case, and this article simply extends that logic one uncomfortable step further, below the floor the simulation itself uses. The same stress test can be applied to your own numbers, using the assumptions and cost structure outlined below.
Defining a bad year for a rental property
A stress test is only as honest as its scenario, so the first job is to name what actually produces a bad year in Paje, rather than gesturing at vague misfortune. Three mechanisms cover nearly all of it.
Soft demand
The ordinary version is a demand dip: a weaker European travel season, an economic wobble in the source markets that supply most of the island's guests, or simply a year when the weather, the news cycle, or the exchange rate nudges bookings elsewhere. The island's baseline is strong, with annual hotel occupancy averaging around 62 percent and peak season exceeding 90 percent, but averages contain lean months, and a soft year is the version where the lean months multiply. Soft demand trims occupancy and pressures nightly rates at the same time, which is why the stress test below moves both.
It helps to be precise about where a bad year actually happens in the calendar. The quiet months are quiet every year, and a rainy April earns little whether the market is strong or weak, so the low season is not where a bad year is made. A bad year is a weak high season: a December and January that book at 70 percent instead of above 90, at rates that needed defending. The annual model concentrates most of the income into a handful of months, which means the stress scenario is really a story about those months underperforming, and an owner reading their statements in a soft year should judge the year by its peaks, not its troughs.
Regional shocks
The rarer version is the shock: the event that stops travel broadly, of which the pandemic was the extreme case. Shocks are not forecastable, which is exactly why they belong in an underwriting scenario rather than in a probability argument. What can be examined is how this market behaved through the last one, and that record is covered in the history section below, because it is one of the more reassuring facts the library holds.
New supply pressure
The third mechanism is competition. The oversupply worry, that a wave of new apartments arrives and drags everyone's occupancy down, deserves respect rather than dismissal. The current picture is a modest pipeline, on the order of a few hundred new units a year island-wide against a fast-growing visitor base, and the full supply analysis sits in the Zanzibar property market outlook for the year ahead. The honest framing for a stress test is that supply pressure, if it comes, arrives gradually and shows up as exactly the softer occupancy and rate discipline this article models. In other words, the 40 percent scenario below is also the oversupply scenario.
The stress test at worst-case occupancy, line by line
The published simulation's scenarios run at 50, 65, and 85 percent occupancy. The stress test drops below the floor.
The 40 percent occupancy scenario
At 40 percent occupancy, a unit sells 146 nights a year, against 183 nights at the simulation's own worst case of 50 percent. That is a fifth less revenue than the published floor, in a year where the market has already told you demand is soft. It is worth pausing on how pessimistic this is as an assumption for a professionally managed beachfront unit: the island's poorly run tail, the generic listing with amateur photos and no management, averaged around 41 percent occupancy in a 2024 island-wide survey. The scenario, in plain terms, assumes a professionally managed unit performs like an unmanaged one for a full year. That is the right kind of assumption for a stress test: not impossible, and comfortably worse than anything the operating model is built to produce.
Rate cuts and their limits
A soft year attacks price as well as volume, so the honest test trims the nightly rate too. Management responds to soft demand by yielding on rate to defend occupancy, and there is real room to do that in a market where the high season carries premium pricing. But the test should also respect the limit: cutting rates below the cost of servicing a stay buys occupancy that loses money, so a disciplined operator lets the calendar go quiet rather than chase the bottom. The stress scenario therefore assumes both a thinner calendar and a softer achieved rate, which compounds the revenue decline beyond the occupancy drop alone.
The costs that do not fall
Here is the mechanism that makes bad years dangerous, and it deserves plain language. The proportional deductions scale down with revenue: the 5 percent agency fee, the operator's 40 percent share of the pool, and the 15 percent withholding on the owner's share all shrink as the gross shrinks. The fixed lines do not. The annual service fee is the same in a record year and a soft one, and the token ownership charges, the ground rent at $0.35 per square metre and the $22 annual property tax, arrive regardless. The full cost anatomy is set out in the true cost of owning a Zanzibar apartment, every fee explained. The structural consequence: as revenue falls, the fixed service fee consumes a growing share of it, which is why net yield falls faster than occupancy does. That is the honest shape of the downside, and any model that shows net falling exactly in line with occupancy has forgotten its fixed costs.
The worked numbers on a real studio
The entry-level Bahari studio is the published simulation's own reference unit, so it carries the worked example. Per $100,000 of purchase price, the simulation's published outcomes are these: at the 85 percent best case, a net return of $22,900 a year. At the 65 percent mid case, $17,400. At the 50 percent worst case, $13,300, and it is worth underlining that the simulation's floor scenario still produces a double-digit net yield after every deduction, including the fixed fee and the withholding tax.
Extending below the floor: at 40 percent occupancy with a softened achieved rate, revenue runs at least a fifth below the worst case, and after the fixed service fee takes its unchanging bite, the net lands materially below a proportional reduction, in the region of 9 to 10 percent net per $100,000 rather than the floor's 13.3 percent. That figure is a derivation from the published cost structure, not a published simulation output, and it is flagged accordingly. The reading matters more than the decimal: a year pessimistic enough to assume professional management performs like the island's unmanaged average still leaves the studio profitable, covering its own costs with a single-digit-to-low-double-digit net return, rather than draining its owner. The bad year, fully costed, is a disappointing year, not a dangerous one. For what the good years look like and why, the full yield architecture is in what rental yield can you really expect from a Paje apartment.
What history says, how the market behaved in past downturns
A stress test gains credibility when history has already run a harsher version of it, and here history has. The pandemic was a global stop on travel, the worst realistic shock this asset class faces, and the Zanzibar record through it is documented in the price index: property values held firm through the downturn and then accelerated as tourism rebounded. The rebound itself came fast and kept compounding, from 736,755 international arrivals in 2024, a 15 percent jump on the prior year, to 917,167 in 2025, with the one million annual visitor milestone now formally reached.
Two lessons carry from that episode into any future bad year. Capital values and rental income behaved differently: income dipped with the travel stop, while the asset itself did not devalue, because the underlying scarcity of well-located beachfront did not change. And recovery was demand-led and quick, because the source of demand, the appetite of cold-country travellers for a warm island, is among the most persistent forces in tourism.
The strength of the demand baseline since then is its own datapoint. December 2024 recorded island-wide hotel bed occupancy of 92.4 percent, August 2025 was the busiest month on record, and the growth has run across nearly every month of 2025 and into 2026. A market whose record months keep arriving is a market where a bad year is a deviation with a gravitational pull back toward trend, rather than the start of a new normal. An investor should not confuse that with immunity, but it is the correct backdrop against which to size the stress scenario's probability. None of this guarantees the next downturn repeats the pattern. It does mean the worst shock on record produced lost seasons rather than lost capital, and an investor underwriting a bad year is mostly underwriting a pause in income, not a collapse in value.
What separates resilient units from fragile ones
Bad years are not distributed evenly, and the gap between the units that sail through and the units that suffer is visible in advance. The island's own data draws the line: professionally managed units in the beach hotspots deliver the 12 to 15 percent gross band, while the unmanaged tail averages around 41 percent occupancy at around $50 a night. In a soft year, that gap widens, because scarce demand flows to the listings with the reviews, the photography, the instant response times, and the distribution reach.
Resilience, concretely, looks like this. A location inside a proven demand pool, where the kitesurfing lagoon and the beach keep drawing visitors even in a thin season. Professional management with rate discipline and marketing reach, so the unit competes for whatever demand exists. A cost structure known in advance, with the fixed lines modest relative to a normal year's revenue. And an owner whose finances do not require the high season to arrive on schedule.
The revenue split itself is part of the resilience, and it is worth seeing why. Under the 60/40 structure, the operator's income falls in exactly the months the owner's does, which means a soft season is the operator's emergency too, answered with the tools an individual owner does not have: repricing across a whole portfolio, redirected marketing, and relationships with the agents and platforms that still hold demand. A private landlord weathering a bad year alone has a problem. An owner inside an aligned operation has a partner whose rent depends on solving it. Fragility is the mirror image: an undifferentiated unit, improvised management, unknown costs, and an owner who needs every month to perform. The stress test is kinder to the first owner than the second in every scenario it runs.
How to build your own margin of safety
The margin of safety is built at purchase, not discovered in the downturn, and it has four components. Underwrite at the downside: make your buying decision on the 40 to 50 percent occupancy numbers, and treat everything above them as upside rather than baseline. Keep reserves: a year of fixed costs held in cash converts a bad year from a crisis into an accounting entry. Avoid obligations that assume the best case: if the purchase uses a developer installment plan, size the payments so a soft season does not strain them. And choose resilience on the checklist above, location, management, and known costs, because the cheapest insurance in this market is bought in the selection of the unit itself.
None of this is pessimism. It is the discipline that lets an owner enjoy the good years without fearing the calendar, pole pole, with the downside already priced. The honest summary of the whole exercise: a professionally managed Paje studio, stress tested below its published worst case, still pays its way, and the market it sits in has already survived a harder test than any model needs to imagine. Run the numbers yourself. If the downside still works for you, the upside takes care of itself.
