Is Singapore Property in a Bubble? Four Tests and Every Bust Since 1975

A bubble is not just high prices. It is prices pulling away from what the homes earn and what buyers earn, kept going by credit that cannot last. So the useful questions are four, and each has an official number behind it:

Sources: URA's price and rental indices; HDB's resale index; SingStat; MAS; URA. Figures and methods in the sections below.

Put together, private property looks expensive rather than clearly in a bubble: two tests show strain, but prices are rising at an ordinary pace for Singapore and the credit rules are tight. HDB resale has risen about as fast as private homes since 2015 (+50.2%), and faster since 2020 (+51.5%).

This is a reading of the past, not a forecast. Bubbles are easier to name after they burst.

URA has tracked private home prices every quarter since 1975. That history gives 166 ten-year windows to compare:

• The index was at least twice as high ten years later in 53 of the 166 windows. Every one of them began in 1991Q1 or earlier: the last ran from 1991Q1 to 2001Q1.
• The median ten-year rise since 1975 is +60%. The latest ten years (+56.7%) sit close to it.
• The current level: 219.4 in 2026Q2, +41.9% above the 2013 peak and +60.6% above the 2017 low.

Records alone say little. The index has set 77 new record highs since it began in 1975, yet went 14 years, from 1996 to 2010, without one. The pace is the better signal, and today's pace is ordinary by Singapore's own history. The quarterly detail is on PropKaki's private residential market overview.

Every fall of 10% or more in the private index, measured from a local peak:

Source: URA's private residential price index (all private homes including ECs), 1975Q1–2026Q2. The 2000 and 2008 falls started while the index was still below its 1996 high.

HDB resale prices have had 2 such falls since 1990:

Source: HDB's resale price index.

Two patterns stand out. Falls last: the private index fell for 4 to 15 quarters each time. And recovery can take a decade or more: buyers at the 1996 peak waited until 2010 to see the index back at that level. The deepest fall was already under way before the Asian financial crisis began in July 1997: by 1997Q2 the index was 118.2, 8.9% below its peak.

In a bubble, buyers pay more and more for each dollar of rent a home can earn. The simplest check divides URA's private price index by its rental index. The ratio only means something against its own history:

Source: URA's private price index and rental index (whole island). Both are indices; the ratio compares their paths, not the rent a particular home earns.

The ratio peaked at 1.53 in 2021Q1, then fell to 1.19 in 2023Q2 as rents surged. Since then it has climbed back to 1.35: prices are up +12.9%, while the rental index has slipped from 163.2 to 162.5. Over the latest year, prices rose +2.9% and rents +1.7%.

How dear that looks depends on the yardstick. The ratio is higher than in 121 of 146 quarters since 1990, and +14% above its 1990–2026 average (1.18). But that long average is pulled down by the early 1990s, when the ratio averaged 0.49; against the 2000–2026 average (1.30) it is only +4% higher. PropKaki's rental market overview has current rents by region.

SingStat changed its household income measure in February 2026, to "market income", which includes income from investments and rent as well as work. The new series starts in 2015: "Data on household market income are only available from 2015 onwards" (SingStat). So the comparison can run over ten years:

Income: SingStat's median monthly household market income among resident households (2025 preliminary). Prices: yearly averages of URA's and HDB's quarterly indices; the 4-room median is from every 4-room resale registered in the year.

The start year changes the answer. Measured to 2025 from each year of the series:

From 2016 or 2017, when private prices were near their last low, private homes ran well ahead of incomes (for 2017, +55.7% against +37.9%). From 2020 or 2021 the two moved together, and from 2022 onwards incomes rose faster. HDB resale prices ran ahead of incomes from every start year from 2015 to 2023; only from 2024 did incomes rise faster (+7.7% against +6.4%).

For a flat buyer, the clearest number is the last line of the first table. A median 4-room resale flat cost 3.8 years of the median household's income in 2015 and 3.8 in 2020, and 4.2 in 2025, at a median price of $630,000. That is a real squeeze, but a gradual one.

Bubbles need cheap, plentiful credit and buyers who expect to flip. Singapore's rules are built against both:

• Debt limits. A borrower's total debt servicing ratio (TDSR) "should be less than or equal to 55%" of gross monthly income (MAS). For an HDB flat, or an EC still within its minimum occupation period, the mortgage servicing ratio is "capped at 30%".
• A stress-tested rate. Banks must calculate that ratio on "the higher of a 4% floor or the thereafter interest rate" for residential loans (MAS), so a borrower can keep paying "even when interest rates increase".
• Loan limits. A first housing loan can cover at most "75% or 55%" of the property's value; with one loan outstanding, "45% or 25%"; with two or more, "35% or 15%" (MAS).
• Stamp duties. Citizens pay 20% additional buyer's stamp duty on a second home and 30% on a third or later; permanent residents 5%, 30% and 35%; foreigners 60% (IRAS). Under free trade agreements, nationals of the United States, and nationals and permanent residents of Iceland, Liechtenstein, Norway and Switzerland, get the same treatment as citizens (IRAS). Selling within four years of buying (for purchases from 4 July 2025) carries seller's stamp duty (IRAS).
• Supply. The Government Land Sales Confirmed Lists for 2026 total 9,320 homes, which URA says is "more than 50% higher than the annual average Confirmed List supply over the past 10 years" (URA). URA expects "about 60,600 private residential units (including ECs)" to be completed over the next few years, and counted 14,929 uncompleted private homes unsold in 2026Q2, excluding ECs (URA). See GLS sites and tender results.

These rules do not stop prices falling in a recession: they mainly stop households borrowing beyond their means on the way up.

Taking every 40-quarter span in URA's index since 1975:

• At least twice as high ten years later: 53 of 166 windows, the last from 1991Q1 to 2001Q1.
• Reached double at some point inside the ten years, even if it fell back: 61 windows, the last starting in 1993Q1.
• The typical ten-year rise has been +60%; the latest ten years produced +56.7%.
• The fastest four-year rise was +264%, from 1977Q2 to 1981Q2. The run into the 1996 peak was +150% (1992Q2 to 1996Q2), just before the deepest fall. Since 1998 the fastest four-year rise has been +62% (2009Q2 to 2013Q2).

So a ten-year doubling was common from the 1970s to the early 1990s, and has not happened in any ten-year window that began after 1993Q1. Whether prices will double from here is a forecast, and the section below explains why this page makes none.

Counting completed deals by calendar month, per day so that a short February isn't penalised:

HDB: 291,413 resale registrations, 2013–2025 (HDB dates resales by registration since March 2012). Private: 161,693 URA resale caveats, 2013–2025, by contract date.

The quiet stretch runs from December to February, around year-end holidays and Chinese New Year, which falls in January or February. The pace picks up from March to May for private homes, and in mid-year for HDB flats.

The averages include 2020, when the circuit breaker cut HDB resales to 424 in April and 363 in May. Leaving 2020 out changes neither the quietest nor the busiest months, though HDB's April moves from +1% to +8% against the average.

Two cautions. Fewer completed deals is not the same as lower prices or a harder sale; it can simply mean fewer buyers and sellers are active. And the gaps are modest: 19% at the private low. As PropKaki's guide to timing a condo sale puts it, treat seasonality as a pattern, not a timing strategy.

"Crisis" means different things in different cities. By the measures this page uses:

• Resale HDB flats got harder to afford. The median 4-room resale price rose from $420,000 in 2020 to $630,000 in 2025, while the median household income rose +36.8%.
• Private homes kept pace with incomes from 2020 (+39.0% against +36.8%), though not from 2015 (+49.2% against +40.8%).
• Empty homes: URA put the vacancy rate of completed private homes (excluding ECs) at 6.4% at the end of 2026Q2, up from 6.2% the quarter before (URA).
• The latest quarter edged down for HDB: its resale index moved -0.3% in 2026Q2.

For new flats, the Government's main lever is supply: see PropKaki's guides to the BTO launch schedule and to how public housing works.

The four tests on this page are also the warning signs to watch, quarter by quarter:

• Prices against rents. A ratio climbing fast and far above its history. It stands at 1.35, +4% above its 2000–2026 average and up from 1.19 in 2023Q2.
• Prices against incomes. Prices outrunning household income year after year. Over 2015–2025: private prices +49.2%, income +40.8%.
• The pace. Rises far above the long-run norm. The latest ten years: +56.7%, against a median of +60%.
• Credit. Looser borrowing limits. Today's limits are in the section on what guards against a bubble.

The 5 busts since 1975 are listed above. None of the signs is a timer: a ratio can stay high for years, and a fall can come without one. PropKaki's market overview updates URA's price and rental indices each quarter.

Nobody knows. A forecast can only be judged afterwards; PropKaki is tracking one in this commentary.

What the index's history can say:

• Doubling by the end of 2030 from 2026Q2 would take about 17% a year. The index last rose that fast over four years from 1993Q1 to 1997Q1 (+104%); it peaked inside that span, in 1996Q2, and then had its deepest fall. Since 1998, its fastest four-year rise has been +62%.
• Falls happen: 5 of 10% or more since 1975, the most recent 2013Q3 to 2017Q2, -11.6%.
• The latest numbers: URA's private index was 219.4 in 2026Q2, +0.5% on the quarter and +2.9% on the year; HDB's resale index was 202.8, -0.3% on the quarter.

URA's own words in its latest quarterly release: "The macroeconomic outlook remains highly uncertain. Households are advised to exercise prudence when purchasing property and taking out mortgage loans" (URA).

Whether now suits you depends on things only you know: how long you will stay, what you can comfortably repay at a test rate of at least 4%, and whether a better-fitting home is available at a price you can support. For the long-run odds of selling above what you paid, see the market overview; for your own numbers, the property financial planner.

What we used.

• URA's private residential price index (1975Q1 onwards) and rental index (1990Q1 onwards), all private homes including ECs, as held in PropKaki's copy of URA's quarterly series.
• HDB's resale price index (1990 onwards), and every 4-room resale registered in 2015, 2020 and 2025.
• SingStat's median monthly household market income among resident households, 2015–2025 (Key Household Income Trends, 2025, and SingStat's Table Builder; 2025 preliminary). The series starts in 2015.
• HDB resale registrations and URA private resale caveats, 2013–2025, for the seasonality table.

How we calculated.

• Busts: falls of 10% or more from a local peak (a quarter higher than the next) to the lowest point before the index regained that peak.
• Ten-year windows: every 40-quarter span in the index; "twice as high ten years later" compares the end of each span with its start.
• Records: quarters above every earlier quarter.
• Price ÷ rent: URA's price index divided by its rental index, quarter by quarter.
• Prices against incomes: yearly averages of the quarterly indices against SingStat's yearly medians.
• Seasonality: each month's deals divided by its days, averaged over the 13 years.

What we did not claim. We do not forecast prices, name a bubble or a crash in advance, or say when to buy or sell. An index is not the price of any one home, and a record is not a verdict.

Rules are as of 19 September 2026: verify with MAS and IRAS before relying on them. Read more about how PropKaki works with data on our methodology page. This page is for information only; it is not legal or financial advice.