Cape Town's 50% Rule: Stay Under, or Go Commercial?
This is about a draft by-law, not law. The City of Cape Town’s draft Short-Term Letting By-Law, 2026 is open for comment until 5 October 2026, and the threshold, the way it is measured and the rates themselves may all change. The figures below are worked examples, not advice. Take professional advice on your own property before you decide.
Read the proposal: the full draft by-law (PDF) · how to comment before 5 October
Most of what has been written about Cape Town’s draft by-law, including our own guide to the threshold, is about how to stay under 50%. That assumes staying under is the right answer. For a lot of properties it is. For some it is the expensive option.
The threshold is not a cap. Cross it and the property is rated as Business & Commercial instead of residential. Nobody stops you letting. So the real question for every owner is a sum:
Does closing enough nights to stay under cost you more or less than paying commercial rates?
The two sides of the sum
What staying under costs you
On the reading we plan for (availability, not bookings; see the threshold guide for why), you stay under by offering no more than half your annual room-nights. For a whole-house let that means your property can be open for roughly 182 nights a year, whatever its size.
So staying under costs you the bookings you would have taken on the nights you close. Not the nights themselves: the bookings. A closed night in July that would never have sold costs you nothing. A closed night in a busy shoulder month costs you a stay.
That makes the cost of staying under:
the nights you have to close × the chance each one would have sold × what a sold night is worth to you after costs
Close the cheapest, quietest nights first and this number can be surprisingly small.
What going commercial costs you
Rates are charged on your property’s municipal valuation, not your income. Two things change when you are reclassified:
- The rate goes up. The 2026/27 residential rate is 0.006428 rand per rand of value. Reported figures for the business and commercial rate differ: 0.015106 in coverage of the City’s draft rates policy, and around 0.0128 in coverage of the draft budget. Check the City’s final tariff sheet or your own rates account. We show both below.
- You lose the residential exclusion. For 2026/27 the first R500,000 of a residential property’s value is reported to be rates-free. Commercial property does not get it.
So the extra you pay is:
commercial rate × valuation − residential rate × (valuation − R500,000)
Because it is driven by valuation, a very valuable property that lets modestly pays the same extra rates as one that lets flat out.
Example 1: the busy family house
A three-bedroom house with a municipal valuation of R3 million, listed all year on two platforms.
| Season | Nights open | Nights sold | Nightly rate |
|---|---|---|---|
| High | 100 | 90 | R4,500 |
| Shoulder | 120 | 80 | R2,800 |
| Low | 145 | 50 | R1,600 |
| Year | 365 | 220 | R709,000 gross |
To stay under, it can be open 182 nights. Keep all 100 high-season nights and 82 shoulder nights open. Close the other 38 shoulder nights and all 145 low-season nights.
- Shoulder sells two nights in three, so the 38 closed shoulder nights would have sold about 25.
- Every one of the 50 low-season bookings goes.
Assume 15% channel commission and R500 a night for cleaning, laundry and consumables. A shoulder night is worth about R1,880 after costs and a low-season night about R860:
| Lost bookings | After costs |
|---|---|
| 25 shoulder nights | R47,000 |
| 50 low-season nights | R43,000 |
| Cost of staying under | about R90,000 a year |
To go commercial: residential rates are (R3,000,000 − R500,000) × 0.006428 = R16,070. Commercial rates are R3,000,000 × 0.015106 = R45,318, or R38,400 at 0.0128. The extra is R22,000 to R29,000 a year.
Verdict: go commercial. Staying under would cost this house roughly R61,000 to R68,000 a year more than paying the higher rates. It is a busy property, and the nights it would have to close are nights that sell.
Example 2: the sea-view apartment
A two-bedroom apartment with a municipal valuation of R4.5 million, also listed all year, but its owner uses it in winter and it sells mostly in summer.
| Season | Nights open | Nights sold | Nightly rate |
|---|---|---|---|
| High | 100 | 70 | R3,800 |
| Shoulder | 120 | 40 | R2,200 |
| Low | 145 | 10 | R1,400 |
| Year | 365 | 120 | R368,000 gross |
To stay under, the same 182 nights: all of high season and 82 shoulder nights. Closing 38 shoulder nights loses about 13 bookings, since shoulder sells one night in three. Closing low season loses 10.
| Lost bookings | After costs |
|---|---|
| 13 shoulder nights | R17,800 |
| 10 low-season nights | R6,900 |
| Cost of staying under | about R25,000 a year |
To go commercial: residential rates are (R4,500,000 − R500,000) × 0.006428 = R25,712. Commercial rates are R67,977 at 0.015106, or R57,600 at 0.0128. The extra is R32,000 to R42,000 a year.
Verdict: stay under. It saves this apartment roughly R7,000 to R17,500 a year, and most of the nights it closes were never going to sell.
Example 3: the family home with a garden cottage
A family home with six bedrooms and a two-bedroom garden cottage on the same erf, with a municipal valuation of R25 million. The family lives in the house. The cottage is let short-term and listed all year.
First, what counts. Under the draft, the property is the erf, and it includes every building on it: one title deed, one property. So the threshold is counted on all eight bedrooms, not the cottage’s two. (If the cottage is a separate sectional title unit with its own deed, it is its own property and this changes; check your title.) A Cottage on the Same Erf goes through this in detail.
| Bedrooms | Room-nights a year | |
|---|---|---|
| Main house | 6 | 2,190 |
| Cottage | 2 | 730 |
| The erf | 8 | 2,920, so the 50% line is at 1,460 |
The cottage listed every night of the year uses 2 × 365 = 730 room-nights, which is 25%. It cannot reach the line on its own, however it is let.
Staying under costs nothing. No night needs closing. There is even room to let the main house for about 120 nights a year on top before the erf reaches 50%.
Going commercial would cost a lot, because the whole erf would be reclassified, family home included, and rates follow the valuation:
- residential: (R25,000,000 − R500,000) × 0.006428 = R157,486
- commercial: R320,000 at 0.0128, or R377,650 at 0.015106
- the extra: R162,000 to R220,000 a year
Verdict: stay under, comfortably. There is nothing to decide here, only something to protect: keep the main house’s letting inside its budget.
The reading that would change it. If the City counted only the part that is let, the cottage’s two bedrooms, the cottage listed all year would be at 100% and over from day one. Even then, staying under still wins. Say the cottage sells 90 of 100 high-season nights at R2,200, 60 of 120 shoulder nights at R1,500 and 30 of 145 low-season nights at R1,000, and costs R350 a night to turn round. Closing it down to 182 nights (all of high season, 82 shoulder nights) loses about 19 shoulder and 30 low-season bookings, roughly R33,000 a year after costs. That is far less than R162,000 of extra rates. Either way, this property stays under. Only how many nights the cottage may be open changes.
A high-value family home is exactly the property for which commercial rates are hardest to justify. The value is in the house the family lives in, not in what the cottage earns.
What decides it
The three examples show the pattern:
- Valuation pushes you under. The more the property is worth, the more commercial rates cost, whatever it earns.
- Off-season demand pushes you over. The more your quiet months actually sell, the more closing them costs.
- Size matters less than you think for a whole-house let. On the availability reading, a whole house is capped at about half the year whether it has two bedrooms or five. It matters a lot if you let by the room, or a cottage on the same erf.
A busy, moderately valued house usually does better commercial. An expensive property with a strong summer and a quiet winter usually does better under.
Four things that can flip the answer
The final by-law might count bookings instead. If the 50% ends up measured on nights sold rather than nights offered, the maths changes completely. The family house above sells 660 of its 1,095 room-nights, which is 60%. It would be over however it managed its calendar, unless it turned guests away. The apartment sells 240 of 730, which is 33%, and it would be comfortably under without closing a thing.
VAT. VAT registration is compulsory once taxable supplies pass R1 million in any twelve months, and short-stay accommodation is then charged at 15%. It has nothing to do with the rates classification, but going commercial usually means selling more nights. If you are close to R1 million, the extra nights can cost you more in VAT than they earn. See running it as a business.
Your home. If the property is also where you live, trade use reduces the primary residence exclusion when you sell. More letting, more apportionment. It is a real cost, just a deferred one.
It is a decision for a year, not a month. The threshold is judged on a year of room-nights. Opening up for a busy December and closing again in January doesn’t undo nights already counted. Pick a side and plan the whole year’s calendar to it.
Doing it on your own numbers
You need four things per property:
- Your municipal valuation, from your rates account.
- What each season actually sold: nights open and nights booked, across every channel together, not one platform’s view.
- What a sold night is worth after costs: rate, less commission, less cleaning and consumables.
- Which nights you would close to get down to half your room-nights, starting with the cheapest.
The first is on a piece of paper. The other three are the hard part, because most hosts’ booking history is scattered across two or three platforms and a spreadsheet.
What Airflow does today
- Measures where you are. The letting limit gauge counts offered room-nights across every channel you connect, forecasts the date you would cross, and records every day from the moment you switch it on.
- Shows which nights you would close. When you are over, it proposes closing your cheapest season first, and shows every night sorted by rate so you can see that the nights it closes are the cheap ones.
- Shows what each night earned. The per-night value chart shows what every night in the year was offered at and what it earned, which is the “chance it would have sold” and “what it is worth” in one picture.
- Takes your history. Import past bookings from a CSV, a Google Sheet or your channels’ calendars, so all of the above works on the years you already have, not only from today.
Ask the Account Manager for the verdict
AI Flow is the assistant built into your Airflow account. You open it from any page of the portal and choose the agent for the job. The Account Manager is the one that knows your account: your properties, your bookings and your settings. It answers questions about them, and it makes changes for you once you have said yes to what it proposes.
Things you can ask the Account Manager to do:
- Set up a property from your website. It reads your site and fills in the listing: description, location, capacity, rates and photos.
- Build and update your booking website: logo, cover photo, colours, pages such as About and Contact, and publishing it when you are ready.
- Change your prices, including seasonal rates for your high and low seasons.
- Update your property and organisation details: names, descriptions, contact details and your host profile.
- Help you connect your calendars, and find your booking forwarding address.
- Answer questions about your bookings and revenue, property by property.
- Work out the stay-under-or-commercial verdict in this article.
It shows you what it is about to change before it changes it, and it never creates, changes or cancels a booking. A conversation with the Account Manager costs 2 actions, and it is free while a paid account is being set up. AI Flow also has a free Help Agent, a Web Designer and a Market Analyst; more in Meet AI Flow.
How it gets you the verdict
With two years of your history imported, the Account Manager works this out for you, property by property:
- It asks for what it needs, one question at a time: your municipal valuation, roughly what your channels charge in commission, what a night stayed costs you in cleaning and consumables, and any months the property isn’t let at all. It saves each answer, so you are never asked twice.
- If you have less than two years of history, it sends you to import it first. A CSV, a Google Sheet or a channel export. Past stays are recorded as history: no invoices are raised and no guests are emailed.
- Then it runs the sum in this article on your own numbers: which months you would close, the bookings you would lose, what that costs after commission and costs, against the extra rates at your valuation. It tells you which side of the line to plan for, and by how much.
Two years, because one year is a single sample of each season. A quiet winter might have been the weather. Two tell you whether it is the property.
The verdict itself costs 2 actions on top of the conversation, charged only when a verdict comes back. It is arithmetic on your own history, not legal, tax or rates advice.
Where to start
- Open an account and add the property.
- Connect every channel’s calendar, and import your past bookings.
- Switch on the letting limit gauge and choose Cape Town (draft 2026).
- Ask the Account Manager: should I stay under or go commercial?
- Before 5 October 2026, if you think the measure is wrong, comment on the draft.
Related: Staying Under Cape Town’s 50% Letting Threshold, A Cottage on the Same Erf, Short-term Letting in the City of Cape Town, South African Short-Term Rental Tax and Rates and Occupancy Was 78%. So Why Was It a Bad Year?.