All posts
Three-bedroom condos: a series

Published 23 September 2026

Does a 3 bedroom new launch make more money than a resale?

By Realila

Mostly not. Of the 3 bedroom condos bought between 2017 and 2021 and resold from 2022, a resale returned more on the buyer's own cash than a new launch in every purchase year, by 3.2 to 7.6 percentage points a year. The rest of the 3 bedroom picture is on the 3 bedroom condo guide.

That comparison is the one most often quoted, and it is not quite fair. Most resales in it were units well over a decade old. Compared with a resale bought within five years of completion, launch still came out behind in every year, but the gap narrowed, and in 2017 almost closed. Where the answer really depends is the region: in the suburbs a nearly-new resale beat launch every year, and on the city fringe launch came out ahead in more years than not.

Every return here is measured the way a launch is usually argued for: on the buyer's own cash, with the loan drawn in stages as the building goes up.

What is in this note

The comparison everyone quotes

Every 3 bedroom bought in a given year and resold from 2022, launch against resale:

BoughtLaunchResaleGap (pts)
20178.3% (296)13.9% (275)5.6
20188.4% (236)12.9% (313)4.6
201912.3% (368)15.5% (261)3.2
202014.2% (394)17.9% (297)3.7
20219.7% (200)17.3% (237)7.6

Return per year on the buyer's own cash. Pairs in brackets. Gap in percentage points, from unrounded returns.

Resale is ahead in all five years, by between 3.2 points in 2019 and 7.6 points in 2021. No single project drives it: removing the largest project from either side moves no year by more than a point.

On its own, this says a resale made more money. It also compares two different kinds of home. The median price per square foot each side paid:

BoughtLaunch psfResale psf
2017$1,369$1,049
2018$1,499$1,086
2019$1,501$1,145
2020$1,560$1,104
2021$1,715$1,271

Why that is not a fair race

Most resales in that comparison were old. Split by how long after completion the unit was bought:

BoughtNearly-newOlder
201764 (2 yrs)211 (16 yrs)
201892 (3 yrs)221 (15 yrs)
201973 (4 yrs)188 (16 yrs)
202072 (4 yrs)225 (16 yrs)
202137 (4 yrs)200 (16 yrs)

Resales, with median years since completion in brackets. Nearly-new means bought within 5 years of completion.

Roughly three resales in four were units about fifteen years past completion. A launch is a new home. So the comparison everyone quotes mostly sets a new home against a fifteen-year-old one bought at a much lower price per square foot.

The two sides also sit in different places. Prime (CCR) homes were 2.5% to 12.5% of launch pairs in each year, against 9.6% to 17.8% of resale pairs. Launches lean to the city fringe (RCR) and the suburbs (OCR). A pooled comparison mixes age, location and price, not only how the home was bought.

Against a nearly-new resale

The fairer comparison is a launch against a resale bought soon after completion, when the two are close to the same home:

BoughtLaunchNearly-newOlder
20178.3% (296)9.0% (64)15.5% (211)
20188.4% (236)11.6% (92)13.3% (221)
201912.3% (368)14.0% (73)15.6% (188)
202014.2% (394)17.1% (72)17.8% (225)
20219.7% (200)15.3% (37)17.4% (200)

Return per year on the buyer's own cash. Pairs in brackets. Nearly-new means bought within 5 years of completion.

Against a nearly-new resale, launch is still behind in every year. The gap is smaller than against all resales, narrowing from 5.6 points to 0.7 in 2017 and from 3.7 to 2.9 in 2020.

Within three years of completion the resale samples shrink to 11 to 52, and launch is level in 2019 and ahead in 2020 and 2021 on 28 and 11 resales. The five-year group, with more than twice as many, points the other way, and we read it as the answer.

Older resales returned the most of all three, in every year.

Where it depends on the region

The same comparison inside each region, launch against a resale within five years of completion.

The suburbs (OCR)

BoughtLaunchNearly-newOlder
20178.3% (132)10.9% (34)17.5% (103)
20187.7% (104)11.8% (65)15.2% (101)
201912.4% (220)14.7% (50)17.7% (88)
202012.5% (213)18.5% (54)19.0% (117)
202111.9% (97)17.3% (21)19.5% (108)

In the suburbs, a nearly-new resale beat launch in every year, by 2.3 to 6.0 points, and the same holds within three years of completion. Removing the largest launch project lowers the launch figure further in four of the five years; in 2017 it raises it to 9.6%, still below the nearly-new resale's 10.9%. This is the firmest result in the note, and the suburbs are about half of every year's sales.

The city fringe (RCR)

BoughtLaunchNearly-newOlder
20177.6% (147)4.3% (13)15.0% (76)
20189.4% (126)8.1% (20)12.8% (79)
201912.2% (136)10.5% (18)13.5% (80)
202017.6% (165)10.0% (12)17.0% (77)
20218.0% (78)14.3% (14)15.8% (64)

On the city fringe, launch came out ahead of a nearly-new resale in four of the five years. The nearly-new side is 12 to 20 resales in each year, so these are thin. Removing the largest project from each side, 2017 and 2020 still favour launch clearly, 2019 becomes level, and 2018 turns to resale. Older resale was ahead of launch in every year except 2020, and ahead in that year too once the largest launch project is set aside.

Prime (CCR)

Prime is too thin to call. Nearly-new 3 bedroom resales there number between 2 and 17 a year, and the launch side is concentrated in a handful of projects, one of which is 59% of the 2017 launch pairs. We make no claim for either side in prime.

What the financing rate does

Everything above assumes a 2.5% mortgage rate for the whole hold. The rate matters more to a resale buyer, who borrows the whole loan on day one, than to a launch buyer, whose loan is drawn in stages over about three years. So the higher the rate, the better a launch does.

The rates these buyers actually paid ran from about 1% to just under 5%: low when they bought, then above 4% through 2023 and 2024. Against that range:

  • Against all resales, launch would have needed a rate above 6%. At the rates buyers actually paid, resale was ahead in every year.
  • In the suburbs (OCR), against a nearly-new resale, the same holds: launch needed 5.0% to 9.4%.
  • Across the whole market, against a nearly-new resale, launch would have come out ahead only above 3.3% for 2017 purchases and above 4.4% for 2019, both at the top of what buyers paid.
  • On the city fringe (RCR), launch came out ahead of a nearly-new resale at almost any rate in four of the five years; in 2021 it needed 9.4%.

The breakeven rate is the mortgage rate at which launch and resale would have returned the same. Read it against your own rate: for a 2018 purchase against a nearly-new resale, the figure is 6.3%, so a buyer borrowing below 6.3% did better with the resale, and one borrowing above it did better with the launch.

BoughtAll regionsSuburbsCity fringe
20173.3%5.6%Any rate
20186.3%7.6%1.0%
20194.4%5.0%0.5%
20205.6%9.4%Any rate
20218.3%8.3%9.4%

Launch comes out ahead if your mortgage rate is above the figure shown, against a resale bought within 5 years of completion. "Any rate" means launch was ahead even at 0%.

Why a launch starts behind

Three things work against a launch from the start. For the 2020 buyers:

  • It costs more per square foot. A median launch was $1,560 psf for 958 sqft. A median resale was $1,104 psf for 1,238 sqft.
  • It earns no rent while it is built. A launch buyer here goes 38 months without rent. A resale buyer is renting it out from the first month.
  • It does not save cash upfront. By the end of the first year a launch buyer has paid the same quarter of the price in cash as a resale buyer, plus the same stamp duty and legal fees.

What a launch does defer is the loan. By month 12 a launch buyer had drawn 30% of the price in total, cash and loan together, rising to 45% by month 24 and 85% by month 36. A resale buyer had drawn all of it on day one. Less borrowed means less interest, which is why a higher rate helps a launch.

What this tells you

A 3 bedroom resale returned more than a new launch for buyers from 2017 to 2021, in every year, on the buyer's own cash. Against all resales the gap was 3.2 to 7.6 points a year.

Part of that is because most resales were old. Three in four were about fifteen years past completion, bought at a much lower price per square foot, and those older resales returned the most of anything here. Against a resale bought within five years of completion, launch was still behind in every year, but by less: 0.7 to 5.6 points.

Where it really depends is the region. In the suburbs (OCR), about half the market, a nearly-new resale beat launch in every year. On the city fringe (RCR), launch came out ahead of a nearly-new resale in four of five years, though on 12 to 20 resales a year, and only two of those years hold clearly once the largest project on each side is removed. Prime (CCR) is too thin to call.

A launch does not save cash upfront. What it defers is the loan, so the higher the mortgage rate, the more a launch gains on a resale. At the rates these buyers actually paid, resale came out ahead everywhere except the city fringe. The same comparison for a 2 bedroom gives the same answer: resale ahead overall and in the suburbs, launch competitive on the city fringe.

What this does not tell you

  • What happens with a second property. Every figure assumes a first property with no additional buyer's stamp duty.
  • What a different buying window would show. These are homes bought from 2017 to 2021 and sold from 2022, a period of strong resale prices. A different window could give a different answer.
  • What each buyer actually paid. Every figure uses one 2.5% rate for the whole hold. Real mortgage rates were near 1% for many of these buyers when they bought, and above 4% through 2023 and 2024. The breakeven rates show how much the answer depends on that.
  • What a major renovation would do. The preparation cost we tested covers readying a unit for a tenant, not a full renovation. An older resale that needs a new kitchen or bathroom costs more than we allowed for.
  • What any particular project did. These are medians across many projects. A single launch or resale can do much better or much worse.
  • Why older resales returned the most. A lower starting price per square foot is the likely reason, but this data does not show it.
  • How these compare with other return figures in this series. The returns in who-buys-a-three-bedroom are price gains alone, without financing, rent or costs, over a different set of sales. The two are not comparable.
Method and data notes. Data to 23 August 2026.

Each figure is a pair: the same 3 bedroom unit bought and later resold, from URA caveat data for private condominiums and apartments, excluding collective sales and sub-sales. A pair counts when the unit was bought between 2017 and 2021, held four to six years, and resold from 2022. A launch is a unit bought as a new sale; a resale is a unit bought as a resale. A nearly-new resale is one bought within three, or within five, years of the project's completion.

Return is the annualised internal rate of return on the buyer's own cash. The buyer pays 25% of the price in cash and borrows 75% at 2.5% a year. A resale buyer pays the cash and draws the whole loan at purchase. A launch buyer pays the cash first, a 5% booking fee and 15% at the option to purchase followed by the first progressive payment, with the rest drawn from the loan in stages to completion. Buyer's stamp duty is charged at the rates in force, legal fees of $3,000 on each purchase and sale, and no additional buyer's or seller's stamp duty. Rent is the median rent psf for a 3 bedroom in the region, net of vacancy, maintenance, property tax and letting fees; a resale earns it from the first month, a launch from two months after completion.

Net rent already deducts ongoing maintenance, vacancy and letting fees across the hold. We also tested a one-off cost to prepare each unit for its first tenant: two months' rent for a nearly-new resale, four for an older one, and one month for a launch at completion. It lowers every return by up to a point and changes none of the comparisons in this note.

The rates buyers actually paid are from Global Property Guide's series of quarterly average mortgage rates in Singapore.

Figures are medians. A figure is printed only when it rests on at least 11 pairs. For each figure we also recomputed the result with the largest project removed, and the text says where that changes the answer.

Data to 23 August 2026.

More from Realila

Every number in this post comes from Realila Research, dated to when it was true. For now we publish research notes from the platform to answer the community's questions; the platform itself opens to the public later.

Join the waitlist

Get each new note by email, and be first in when the research platform opens.

Ask for a research note

A question about the market, or data you would like to see.

We read every request. We cannot write a note on every one, but the ones we do write will appear on the Feed, and we will email you if yours does.