Cap Rate, BRRRR and Which Property Metrics Travel

There are three numbers every property-investing course teaches, and one strategy that chains them together. The numbers are arithmetic. They work in Johor, in Jakarta, in Cleveland, in any currency, on any building. The strategy is not arithmetic. It is a bet on a particular tax code, and it was written in a country that taxes property in a way Singapore very deliberately does not.

That distinction matters more than it sounds, because the courses do not draw it. They teach the metrics and the strategy in the same breath, in the same confident tone, and a reader in Singapore or Kuala Lumpur reasonably concludes that both travel. Only one does.

Cap rate is a price tag with the mortgage taken out

Capitalisation rate is net operating income divided by what the property costs. Net operating income is the rent you actually collect after vacancy, minus what it costs to run the building — management, maintenance, insurance, property tax, the sinking fund. A property throwing off $48,000 a year in net operating income at a price of $800,000 has a cap rate of 6%.

The important thing about cap rate is what it deliberately leaves out: your mortgage. No interest, no principal, no loan at all. That looks like an omission and it is the entire point. Cap rate is built to compare buildings, not financings. Two investors can buy the same block on completely different terms and, if the rent and the running costs are the same, the cap rate is identical. It is a property-level number, and it is the closest thing property has to a yield you can put beside a bond.

What it quietly assumes is that you counted the operating expenses honestly. This is where most amateur cap rates go wrong, and it is not usually deliberate. People forget vacancy, or budget it at a flattering 2%. They forget that the agent takes a month's rent. They forget that a twenty-year-old unit needs a water heater eventually. Every one of those omissions inflates net operating income, and because cap rate divides by a fixed price, every dollar of forgotten cost shows up as pure yield.

Cash-on-cash is the one that moves when you borrow

Cash-on-cash return asks how much of the money that actually left your bank account comes back each year. Annual pre-tax cash flow — that is net operating income minus the mortgage payment — divided by the cash you put in. Down payment, stamp duty, legal fees, renovation, all of it.

Because it includes the mortgage, it moves with leverage in a way cap rate cannot. Borrow more and the cash you put in falls, so the same cash flow divides by a smaller number and the percentage rises. This is the mechanism behind every claim that property "returns 25% a year". Leverage does exactly this.

It also cuts the other way, and that half rarely makes it into the pitch. A high cash-on-cash return on a thin cash position is fragile by construction: the same leverage that multiplies the return multiplies a bad month. Two tenants leave, the interest rate resets, and a number that read 25% goes negative without the building having changed at all. Cap rate would barely have moved.

The two are not rivals; they answer different questions. Cap rate tells you whether the asset is any good. Cash-on-cash tells you whether your deal on it is. A strong cap rate can be wrecked by bad financing, and clever financing cannot rescue a building that does not earn.

Neither of these is "rental yield"

Listings across the region quote rental yield, and it is almost always gross yield: annual rent divided by price, with no costs deducted at all. Gross yield and cap rate look like the same shape of number and are not comparable, because one has had the running costs taken out and the other has not.

The gap is not a rounding error. On a unit where costs eat a third of the rent — not unusual once management, maintenance, insurance and property tax are counted — a 4% gross yield is a 2.7% cap rate. If you compare a listing's gross yield against a cap rate you calculated properly, you will conclude the listing is a better deal than the one in front of you, and you will be wrong by the entire cost of running a building.

BRRRR is not a metric. It is a loop.

Buy, Rehab, Rent, Refinance, Repeat. The strategy is to buy something under-priced because it is in poor condition, fix it, let it, then refinance against the new higher valuation and pull your original cash back out. If the refinance releases everything you put in, you own an income-producing asset having ended with the same money you started with, and you go and do it again.

That last clause is the whole product. BRRRR is less a way to buy one property than a way to fund the next. Every part of the loop exists to serve the fifth step. Which is why the honest measure of a BRRRR deal is not its yield but a much blunter question: how much of my cash came back out?

When the answer is "all of it", cash-on-cash return stops being a percentage. Divide a positive cash flow by zero cash remaining and the return is not a large number, it is undefined upward — the "infinite return" the courses advertise. That is not marketing. It is what the arithmetic does when the denominator empties, and our BRRRR calculator reports it as INFINITE for exactly that reason.

What the loop needs, then, is for the cash you sank to be recoverable by refinancing. And a refinance lends against the appraised value of the building. Anything you spent that did not become building value cannot come back out. Renovation becomes building value. Legal fees mostly do not. And tax — tax never does.

The fifth step meets Singapore's tax code

Singapore taxes the purchase of residential property twice. Buyer's Stamp Duty is banded and everyone pays it: 1% on the first $180,000, rising through 3% and 4% to 6% on the part of a residential price above $3 million. On a $1.2 million purchase it comes to $32,600, and at 2.7% of the price it is an ordinary transaction cost of the kind every market has.

Additional Buyer's Stamp Duty is the one that matters here, and it is charged flat on the whole price according to who you are and how many residential properties you already own. Same property, same price, entirely different bill:

Read that column against the strategy. Repeat is the step BRRRR is named for, and in Singapore the second property costs a citizen twenty per cent of its price before anything has been rented, painted or valued. The third costs thirty.

And that money cannot come back through a refinance, because stamp duty does not become building value. A valuer does not appraise your tax receipt. It is a permanent withdrawal from the loop, not just an expensive input to it.

A second constraint arrives at the same moment and compounds it. Borrowing limits step down by how many housing loans you already carry: 75% loan-to-value with none outstanding, 45% with one, 35% with two or more. So the second purchase demands a far larger deposit at exactly the point the tax bill arrives, and the eventual refinance is capped at the lower ratio too, which shrinks the release at both ends.

What that does to one deal

Take the same $1.2 million property twice: once as a citizen's first purchase, once as the second. Assume a $60,000 renovation, and that it appraises afterwards at $1.32 million.

The second deal strands six hundred and thirty-six thousand dollars more than the first, and three things account for it precisely: $240,000 in ABSD, an extra $360,000 in deposit forced by the lower loan-to-value, and $36,000 of refinance release lost to that same ratio. Nothing about the building changed. Nothing about the rent changed. The cap rate is identical in both columns, which is precisely what cap rate is for.

Recovering 5.4% of your capital is not a BRRRR. Our own calculator would call it a weak one and suggest you consider whether the deal stands up as a plain buy-and-hold instead, which is the right answer. The loop has not been made harder; it has been broken. The capital it is meant to recycle is gone.

And the exit is taxed too

If the response is to sell rather than hold, Seller's Stamp Duty is waiting. For residential property bought on or after 4 July 2025 it runs 16% in the first year, then 12%, 8% and 4%, reaching zero only after four years. On our $1.2 million example that is $192,000 to sell inside twelve months.

The schedule changed on that date; property bought earlier sits on the older three-year ladder of 12%, 8% and 4%. Two nearly identical flats can therefore carry different exit taxes purely by purchase date, which is worth checking before you assume anything about a quick sale.

So the strategy is squeezed from both ends. Accumulating is taxed at entry. Trading out is taxed at exit. What is left in the middle is holding — which is a perfectly good thing to do, and is not the strategy anyone sold you.

What still works

The metrics do. All of them. Cap rate, cash-on-cash and net operating income are arithmetic and carry no assumptions about jurisdiction; use them exactly as taught. It is only the loop that fails, and it fails for a specific, findable reason rather than because the region is somehow unsuited to property.

Three things follow. The first property is structurally different from every one after it, so the cheapest leverage you will ever have is the one you already qualify for — a citizen's first purchase carries no ABSD and the highest loan-to-value, and that gap never comes back. Second, non-residential property is a different tax animal: ABSD does not apply to it at all, which changes the arithmetic of the loop completely and is worth understanding before dismissing commercial or industrial as exotic. Third, if the plan is to accumulate residential units, the honest model is not BRRRR. It is slower, it is funded by new savings rather than recycled capital, and it should be planned as such rather than discovered halfway through.

What this audit found in our own tools

Every guide in this series checks the tools it links. This one found a calculation error, not just a cosmetic one.

Our Singapore stamp duty calculator was charging ABSD on non-residential property. Select "Non-residential" and "Foreigner" at $1.2 million and it reported $720,000 of Additional Buyer's Stamp Duty — a duty that does not exist for commercial or industrial purchases. The correct figure is zero. It was wrong at every buyer profile, and it overstated the cost of a commercial purchase by the whole ABSD amount.

The cause was narrow and the kind of thing that survives review. The function that computes ABSD accepted the property type in its arguments and never read it; only the Buyer's Stamp Duty function used it, to pick between the residential and non-residential bands. The page's own explanatory text already said ABSD counts "residential properties owned" — so the prose on the page was right while the code underneath it was not, which is why nobody reading the page would have caught it. It is fixed, with five regression tests, three of which fail if the fix is removed.

The second finding did not reach a single reader and is worth reporting anyway. A comment in the same file described the loan-to-value ladder as 75/55/45 while the code did 75/45/35. The code was right: 55% is the reduced ratio for a first loan with a long tenure, not a step on the loan-count ladder, and the two had been conflated. It is a comment, so nothing computed wrongly — but the figure in this guide came from reading that file, and 55% was two minutes from being published here as fact. A comment that contradicts the code beneath it is more than a documentation problem. It is a defect with a delay on it.

Also checked and clean, reported because a null result is a result: the BRRRR calculator's handling of the infinite-return case, where recovering all of your capital divides a cash flow by zero. It guards that twice, once in the branch that prints the verdict and once in the formatter itself. Nothing to fix.

Run your own numbers

Our cap rate calculator and cash-on-cash return calculator take the two figures apart so you can see which one your deal is failing. The BRRRR calculator is the one that reports how much capital actually came back out, and it is the number this guide argues you should read first. For the duty that never comes back, the Singapore stamp duty calculator prices BSD, ABSD and SSD by buyer profile, and the rental property ROI calculator puts a whole holding period together.

This is general education about how the arithmetic and the duties work, not personalised investment or tax advice. Stamp duty outcomes depend on facts specific to you — residency, what you already own, how a purchase is structured — and are worth confirming with IRAS or a conveyancing lawyer before you commit to anything.