Why Sentosa Cove Owners Are Losing Millions: The 2004–2011 Bubble, the Cooling-Measure Shock, and the Rental-Yield Evidence

The striking headline in The Business Times on 12 August 2026 — that nearly two-thirds of Sentosa Cove resales have been loss-making — is best understood not as evidence that Sentosa Cove suddenly became a bad location, but as the delayed consequence of three forces that collided over two decades: extraordinarily rapid price inflation during the enclave’s formative years, a business model unusually dependent on foreign and investment buyers, and cooling measures that progressively dismantled the financing and transaction conditions on which those valuations had been built.

The price and rental charts attached for this analysis make the story especially revealing. The earliest project, The Berth by the Cove, began at roughly S$830 psf in 2004. Within a few years, transactions across the Ocean Drive/Ocean Way cluster were reaching S$2,000–S$2,800 psf. Yet rents never remotely doubled or tripled in the same way. That divergence is the key to understanding what happened.

Our conclusion after matching the transaction history against Singapore’s cooling-measure timeline is therefore more nuanced than saying simply that “Sentosa Cove was a bubble”:

Sentosa Cove was probably not fundamentally overvalued at its original 2004 entry prices. The bubble-like phase developed between roughly 2006 and 2011, as later launches and secondary transactions capitalised increasingly heroic assumptions about foreign demand, cheap leverage and perpetual capital appreciation. The cooling measures did not create that overvaluation; they removed the conditions that had allowed it to persist.

The article is describing an old problem crystallising today

The headline numbers are severe. Mogul.sg’s analysis cited by The Business Times found that 64.5% of Sentosa Cove resales from May 2023 through June 2026 were unprofitable, versus 62.8% in the preceding March 2020–April 2023 period. The average loss among loss-making transactions was S$1.28 million. Meanwhile, the average gain among profitable transactions collapsed to about S$656,000, roughly 62% below the previous period. Importantly, these calculations exclude stamp duties, property tax, legal fees and commissions, so the actual investor return after costs would generally be worse.

Newmark’s alternative calculation tells a similar story. Combined landed and non-landed resale volume fell from 208 transactions in March 2020–April 2023 to 126 in May 2023–June 2026, while the median loss widened from about S$271,000 to S$370,000. For non-landed homes specifically, the median loss increased from roughly S$270,000 to S$378,000.

Yet there is an important paradox: Sentosa Cove prices themselves have recently been recovering. Non-landed resale values increased 5.7% quarter-on-quarter and 7.5% year-on-year in Q2 2026, and were 18.1% higher on a psf basis than in Q1 2021. What disappeared was liquidity: there were 68 non-landed deals in H1 2021 but only 35 in H1 2026.

That distinction matters. A property can have risen from its 2020 trough and still produce a million-dollar loss for somebody who paid a 2007 or 2010 bubble price.

The present loss statistics are therefore partly an archaeological record of the original entry prices.

The Ocean Drive price staircase

The first few Sentosa Cove condominiums reveal how rapidly the market’s valuation framework changed.

The Berth established the original price benchmark

The Berth by the Cove was Sentosa Cove’s first private condominium, launched in November 2004 at an average of approximately S$830 psf. By the market peak around late 2007 and early 2008, transactions were exceeding S$1,700 psf, with one 1,905 sq ft unit reaching S$2,200 psf in November 2007. In other words, the record transaction price was approximately 165% above the project’s original average launch price after only three years.

Today, URA-derived data puts The Berth’s recent average at about S$1,477 psf, 32.9% below that S$2,200 historical high. Its current indicated rent is around S$4.80 psf per month, producing an estimated gross rental yield of roughly 3.9%.

That is extraordinarily informative: someone buying at S$830–S$900 psf was purchasing a very different investment from someone buying around S$1,700–S$2,200 psf.

The Azure showed the repricing continuing

At The Azure, an early developer sale in October 2005 was recorded at S$1,055 psf. That owner eventually sold in 2008 at S$2,400 psf and still made a massive S$3 million profit — important evidence that an early Sentosa purchase was not automatically a bad investment.

Against its historical high of S$2,400 psf in 2008, the recent URA-derived average of around S$1,710 psf remains approximately 29% below its peak. With rents averaging about S$5.00 psf per month, the property currently offers an indicative rental yield of around 3.5%.

Again, the loss problem is overwhelmingly an entry-basis problem.

The Oceanfront captures the speculative acceleration particularly well

The Oceanfront @ Sentosa Cove provides one of the clearest historical comparisons for this bubble. In July 2006, a 5,242 sq ft unit was acquired at S$1,099 psf and resold exactly one year later at S$1,780 psf, generating a profit of approximately S$3.5 million for the investor in just 12 months.

More expensive units went much further. A 2,077 sq ft apartment purchased in July 2007 at S$2,799 psf was acquired near the project’s peak, with the highest recorded price reaching S$2,843 psf in November 2010. When the owner eventually sold in October 2012 at S$2,190 psf, the transaction resulted in a loss of approximately S$1.3 million.

With a historical high of S$2,843 psf versus today’s average of approximately S$1,781 psf, The Oceanfront is currently trading about 38% below its peak. Same development, but a vastly different investment outcome depending on the entry price. With a rental price of $5.01 psf, the yield works out to be around 3.37%.

The Coast shows how far the late-2006 market had already moved

An apartment at The Coast was purchased for S$1,639 psf in November 2006. The unit was then subsequently sold at the historical record price of S$2,600 psf in December 2007. Its recent average is approximately S$1,572 psf, about 39.5% below that record, with current indicated rent near S$4.74 psf per month and an estimated gross yield around 3.6%.

Notice the progression:

The first four figures are supported by URA-derived transaction records extracted by PropNex Protrend. For W Residences, The Business Times reports that only 20 units were sold when it was first launched in 2010 at an average S$2,793 psf; when CDL relaunched the remaining stock in 2024, the average was S$1,780 psf — 36% lower — and 65 units were sold.

W Residences is perhaps the closest thing we have to a controlled experiment. The building and location remained essentially the same, yet the market-clearing capital value fourteen years later was roughly one-third lower.

That is difficult to reconcile with the proposition that 2010 pricing was primarily supported by enduring housing utility.

The cooling-measure event study

The timing is even more important than the measures themselves.

A simplistic version of Sentosa Cove’s history would say: “Government introduced cooling measures, therefore prices fell.”

The evidence points to a subtler mechanism. The first measures slowed speculation, but they did not immediately reverse prices. The sustained break occurred as Singapore progressively attacked every leg of the investment trade: deferred financing, short holding periods, high leverage on additional properties and, finally, foreign investment demand.

The September 2009 change removed the cheap option on future appreciation

On 14 September 2009, MAS disallowed both the Interest Absorption Scheme and interest-only housing loans for private residential property. The government explicitly noted that these arrangements allowed buyers to put down roughly 10–20% upfront while making little or no significant payment until completion, which could encourage speculation in a rapidly rising market.

The context was extraordinary. Developers sold 10,017 units in the first seven months of 2009, already more than twice the 4,260 units sold during all of 2008.

For an investment market like Sentosa Cove, IAS was especially powerful. A buyer could effectively acquire exposure to a several-million-dollar uncompleted luxury apartment while deferring much of its carrying cost. In financial terms, it made the purchase behave more like a leveraged option on future property appreciation.

But the PropNex price chart shows why September 2009 cannot be labelled the moment that “caused” Sentosa’s collapse: The Berth, Azure, Coast and Oceanfront all rebounded strongly into roughly 2010–2011.

Official national data tell the same story. In February 2010, the government observed that prices had risen particularly quickly in the second half of 2009 and that mortgage lending had been growing around 12% year-on-year. It introduced SSD and reduced the maximum bank LTV to 80%, explicitly describing the intention as acting before a property bubble formed.

August 2010 started attacking the repeat investor

The next round was materially more relevant to wealthy Sentosa purchasers. On 30 August 2010, SSD was extended from one year to three years. For borrowers who already had a housing loan, the minimum cash payment increased from 5% to 10%, while LTV fell from 80% to 70%.

Even then, the government reported that private residential prices had increased 11% in the first half of 2010 and had exceeded the previous 1996 historical peak.

This matters for causal interpretation: the first several cooling rounds were pushing against a very powerful post-GFC liquidity cycle. They did not immediately destroy the market.

January 2011 effectively killed the short-term flipping model

On 14 January 2011, Singapore imposed SSD rates of 16%, 12%, 8% and 4% for sales during the first through fourth years respectively, while cutting the LTV for an individual with an existing housing loan from 70% to 60%.

At that point, an investor buying a second luxury home could no longer rely on extremely high leverage and a quick exit. Selling within a year could incur a 16% SSD regardless of whether the property had appreciated or depreciated.

That fundamentally changed the risk/reward ratio of buying a S$4 million or S$6 million Sentosa apartment for capital appreciation.

December 2011 was the straw that broke Sentosa’s back

The most consequential phase of the cooling measures began on 8 December 2011 with the introduction of ABSD, followed by the Total Debt Servicing Ratio (TDSR) framework in June 2013.

Under the initial ABSD regime, foreigners purchasing Singapore residential property faced an additional 10% stamp duty, while PRs buying their second or subsequent property and Singapore citizens buying their third or subsequent property were subject to an additional 3%. This was particularly significant because foreign buyers’ share of private residential transactions had risen sharply—from about 7% in H1 2009 to 19% in H2 2011.

For a typical suburban owner-occupier market, restricting foreign investment demand would already have had a meaningful impact. For Sentosa Cove, however, the effect was far more structural. The precinct had been conceived, promoted and marketed internationally to high-net-worth buyers, making it disproportionately dependent on foreign capital.

Then came TDSR in June 2013, which introduced a system-wide limit on how much of a borrower’s gross monthly income could be committed to servicing total debt obligations. This was particularly relevant to Sentosa Cove, where large absolute property values meant correspondingly large mortgage requirements. Even buyers with substantial assets could now find their borrowing capacity constrained by income-based debt-servicing limits. Combined with tighter LTV requirements, this substantially increased the amount of equity required to purchase high-value properties.

The PropNex chart is consistent with this cumulative-policy interpretation: prices across the major early Sentosa Cove projects broadly peaked around 2010–2011 before entering a prolonged correction through much of the following decade.

It would therefore be overly simplistic to attribute Sentosa Cove’s correction to any single cooling measure. It was the cumulative sequence that mattered:

IAS removal weakened speculative financing → SSD discouraged the quick flip → lower LTV increased upfront equity requirements → ABSD reduced the marginal foreign buyer pool → TDSR further constrained borrowing capacity, particularly for high-value properties.

Sentosa Cove was therefore hit from both sides: the pool of eligible buyers became smaller, while the amount those buyers could leverage became more restricted.

The rental-yield test points to where the bubble actually was

Price alone does not establish a bubble. A waterfront home has consumption value, scarcity value and prestige value that rental income cannot fully capture.

But rental yield is extremely useful for asking a simpler question:

How much of the purchase price was being justified by the property’s actual income-producing capacity, and how much depended on continued capital appreciation?

The two PropNex charts supplied for this research provide a useful stress test. The sales chart gives each project’s maximum annual average selling price, while the rental chart gives both its through-cycle average and maximum annual average rental rate from 2009–2026.

The gross-yield formula is:

Gross rental yield ≈ monthly rent per sq ft × 12 ÷ purchase price per sq ft

Using the supplied PropNex figures:

Calculations use the PropNex price and rental tables in the chart. They are valuation stress tests, not unit-matched historical yields.

The second yield column is particularly revealing.

We deliberately used the highest annual average rent recorded over the entire 2009–2026 series against the historic peak annual average capital value. This is generous to the old valuation, because several of those maximum rents occurred much later, particularly during the extraordinary 2022–2023 rental upswing visible in the supplied chart.

Yet even with that advantage, the implied gross yield at the old peak capital values is only 2.36% to 3.45%.

And that is gross, before maintenance charges, property tax, vacancy, furnishing, repairs, agency fees and financing costs.

Another way to show the disconnect is to ask how much rent would have been required for a 4% gross yield at those peak annual average capital values: