Did When You Bought Matter More Than What You Bought?

There's a real answer to "does market timing matter," and in this dataset it's an emphatic yes — the year you bought was the single most decisive thing about whether you profited.

Across 267,641 matched private resale pairs — the same unit bought, then later sold — we sorted every home by the era its owner bought in. The odds of profit didn't wobble a little between eras; they collapsed and soared:

• Bought pre-2000: just 60.2% profitable — the only era where a large chunk of owners barely cleared their purchase price.
• Bought 2005-2008: 93.5% profitable, a median gross gain of +41.5%.
• Bought 2019 or later: 97.4% profitable — the highest of any band (but read with care, below).

That is a 37.2-percentage-point spread between the best and worst era to have bought — and it is the widest odds swing of any single factor in this data. It beats the gap opened by entry price, by region, by freehold-versus-leasehold tenure, by property type, and even by holding period. The home you bought mattered; when you bought it mattered more.

The catch — and it's a big one — is that you can't change the vintage of a home you already own, and the newest, best-looking band is partly a mirage. So the rest of this guide does two things: it walks the full era-by-era table and explains why the odds moved, and it turns a number you can't act on retroactively into a lesson you can use at the point of purchase. To see how this sits alongside the other five factors, start with the pillar, how to tell if your property will be profitable; to model your own unit, use the Profitability model.

Before the eras, here's exactly what sits behind the numbers, because a claim this strong lives or dies on the method.

• Matched pairs, not an index. We paired private homes with a recorded purchase and a later sale, so we compare the same unit's buy price to its own sell price. An index averages different homes selling in different periods; matched pairs remove that noise and let us group cleanly by purchase year.
• Grouped by when you bought. Each pair is filed under the era of its purchase — the entry vintage — not the year it sold. That's the whole point: we're isolating the effect of the market you walked into.
• Private only. URA private caveats — condos, apartments, executive condominiums and landed — not HDB, which behaves differently.
• Gross, not net. "Profit" means the sale price beat the purchase price. It does not subtract agent commission, buyer's stamp duty, any Seller's Stamp Duty, legal fees or loan interest. That gap is real, and it matters more for the thin-margin eras below.
• Genuine holds. Pairs that turned over in under 180 days are excluded as sub-sale noise, so this is real ownership, not flipping.
• 267,641 pairs, current to June 2026 — large enough that an era's pattern isn't the fluke of one project or one district.

One structural quirk is worth flagging up front, because it shapes how you read the newest eras: a pair only exists if the unit has already been sold again. Homes bought recently that haven't resold yet simply aren't in the data. That under-counts recent buyers who are still holding — and, as you'll see, it flatters the recent bands. Read every figure as "what usually happened to homes bought in this era," not "what will happen if you buy now."

Here is the full picture, straight from the matched pairs, grouped by the era the owner bought in:

Trace it top to bottom and two stories jump out. The first is the odds of profit: dismal for pre-2000 buyers at 60.2%, then a step-change to 85.9% for 2000-2004, and comfortably above 88% for every era after. Once you're past the millennium, the chance of clearing your purchase price was high and fairly stable; it's the pre-2000 vintage that stands apart as the era where a large minority of owners lost.

The second story is the size of the gain, and it doesn't track the odds. The fattest median gains belong to the two 2000s bands — +46.4% for 2000-2004 and +41.5% for 2005-2008 — not to the newest, safest-looking eras. The 2000-2004 buyer had merely average odds (85.9%) but, if they profited, typically walked away with the biggest gain of any era. The 2005-2008 band pulled off the rare double: high odds (93.5%) and a big median gain (+41.5%), at a remarkable +7.6% median annualized — the strongest yearly return in the table.

So "when you bought" split into two advantages depending on the era: older buyers who profited won big, while newer buyers won often. We unpack the two extremes next.

The odds of winning and the size of the win peaked in different decades.

Every other era in the table sits above 85% profitable. The pre-2000 band sits at 60.2% — roughly four in ten of those owners eventually sold at or below what they paid. It's not just the worst odds in the table; it's the only era where "most owners did fine" barely holds. And the gain, when it came, was thin: a +10.4% median gross gain over a typical 10-year hold works out to just 1.1% a year — the lowest annualized figure of any era by a wide margin.

What makes an entire era of buyers underperform this badly? The honest answer is interpretation, not a data claim — the pack tells us what happened, and the market cycle offers the most plausible why. Singapore private prices ran up hard into a mid-1990s peak; buyers who entered around that high paid rich prices and then held through a long, punishing stretch — the Asian Financial Crisis, then the dot-com slump and SARS — that took years to work off. Someone who bought at the top of that cycle could hold for a decade and still be waiting for the market to overtake what they'd paid. The 60.2% figure is what a peak entry looks like in the rear-view mirror.

The lesson isn't "the 1990s were cursed" — it's what a peak entry costs you in any era. Buy when prices have run far ahead of fundamentals and you don't just risk a smaller gain; you risk being one of the four-in-ten who, years later, still hasn't cleared the purchase price. That's the enduring warning inside the oldest, dustiest row of the table.

Pre-2000 is the receipt for buying at a peak: a decade held, and barely ahead.

The bottom row is the one to be most careful with. On its face, buying in 2019 or later delivered a 97.4% hit rate — nearly flawless, the best of any era. It is also the single most misleading number in this article if you take it at face value.

Here's the trap. Every pair in this dataset needs both a purchase and a completed resale. For a home bought in 2019 or later, the only way it appears here is if the owner has already sold it again — within a handful of years. Who sells that fast? Disproportionately, people sitting on a quick gain: they bought well, watched the price jump, and cashed out. The recent buyers who are underwater mostly haven't sold — you don't crystallise a loss if you can avoid it — so they're absent from the data. The 2019+ band, in other words, is survivor-skewed: it over-represents recent quick-flip winners and under-counts everyone still holding. The n of 16,190 is real, but it's a self-selected slice, not a representative one.

This is why the pack labels the recent bands directional, and why you should too. It does not mean buying in 2019+ was a 97% safe bet; it means that among the recent buyers who have already resold, almost all did well — close to a tautology. The same caution applies, in softer form, to 2014-2018. The older bands (pre-2000, the 2000s) have had decades for the full range of outcomes to resolve, so their numbers are far more trustworthy as base rates.

97.4% isn't the odds of buying today — it's the odds among those who already got out. Read it as a hint, not a guarantee.

This is the question the maxim "time in the market beats timing the market" tries to settle in one line — and the data says the truth is a partnership, not a winner.

Timing the entry mattered enormously. That's the whole finding above: the era you bought in swung the odds of profit 37.2 points, the widest gap of any factor. Walk in at a cycle peak like the mid-1990s and 40% of your cohort was still underwater years later; walk in after a correction and almost everyone cleared their price. The market you entered wrote the first draft of your outcome.

But time in the market was the rescue. Look again at the pre-2000 row: those buyers had the worst odds, yet over a median 10-year hold, 60.2% still profited despite entering at the worst possible moment. A long enough hold gives even a peak entry the runway for the next up-cycle to overtake the purchase price — exactly the pattern the holding-period study found from the other direction, where the 10-year-plus hold was in a different league because a decade usually contains a full cycle. Entry vintage and holding period are the same cycle logic viewed from opposite ends: one is the hand you were dealt, the other is how long you stay at the table.

So which won? Neither, cleanly. Timing set the odds; holding could bend them back. You can't retroactively change when you bought — that ship sails the moment you sign — so the entry lesson is purely forward-looking: don't buy at a peak. But once you own, the hold is a lever you still control, and it's the one that can rescue a mistimed entry. The worst outcome is the pincer: a peak entry and a forced early sale, paying a rich price and then not giving the cycle time to bail you out.

You can't re-pick your vintage — but a long enough hold can forgive a bad one.

It's a fair challenge, and the honest answer is: partly, yes — and you should know where the two stories bleed together.

The overlap is real. Homes bought in the older eras were, on average, bought at far lower prices per square foot — the cheapest-PSF entries in our entry-price study are disproportionately older-vintage buys. So when the pre-2000 and 2000s bands show up here, part of what you're seeing is the cheap entry effect wearing a different hat. The two dimensions are correlated by construction, and it would be dishonest to present "when you bought" as a cleanly separate cause. This is the same caveat the entry-price piece carries in reverse, and the two should be read together.

But vintage is not just price in disguise, and the table proves it. The 2005-2008 buyers didn't enter cheap — they bought into a hot, rising market at elevated prices — yet they posted the best combination in the whole dataset: 93.5% profitable and a +41.5% median gain. Price alone can't explain that; the cycle position of their entry can. They bought before a major run-up and a long recovery carried them, which is a timing story, not a bargain-hunting one. Likewise, the pre-2000 buyers weren't punished for paying a high absolute price so much as for buying at a cyclical peak — a distinction price-per-square-foot bands can't fully capture.

The clean way to hold both ideas: what you pay and when you pay it are two overlapping levers, not one. A cheap entry helps, an early-cycle entry helps, and the older eras happened to offer both — which is exactly why their rows look the way they do. Neither piece is the full story on its own; that's why they're separate spokes that point at each other.

Does the year you buy really change your odds of profit? Yes — more than any other single factor in this data. Across 267,641 matched private resales, owners who bought pre-2000 profited just 60.2% of the time, while those who bought in 2005-2008 profited 93.5% — a swing wider than the gap opened by entry price, region, tenure, property type or holding period. The era you enter sets your starting odds before you've held for a day. Every figure is gross, before commission, duties and interest.

Is buying now safe, given the 2019+ band was 97.4% profitable? Be careful with that number — it's survivor-skewed. A matched pair needs a completed resale, so the only recent buyers in the data are the ones who already sold again, which self-selects for quick winners; recent buyers sitting on a loss mostly haven't sold and aren't counted. So 97.4% is close to "among recent buyers who already got out, almost all did well" — directional, not a promise that buying today is 97% safe. Treat the recent bands as a hint and lean on the older, fully-resolved eras for a real base rate.

If I bought at a bad time, is it too late? No — the entry is fixed, but the hold isn't. Pre-2000 buyers had the worst odds yet still profited 60.2% of the time over a typical 10-year hold, because a long enough hold usually spans a full market cycle that can lift even a peak entry clear of the purchase price. You can't re-pick your vintage, but holding is the lever that can rescue a mistimed one — see the holding-period study. The worst case is a peak entry followed by a forced early sale.

What we counted. 267,641 matched private resale pairs from URA private caveats (condos, apartments, executive condominiums and landed), current to June 2026, grouped by the year the owner bought. A "pair" is one unit with a recorded purchase and a later sale, so we compare the same home's buy and sell price. Pairs that turned over in under 180 days are excluded as sub-sale noise. The buy-year bands, the 85.3% overall profitable rate, the +25.8% overall median gross gain, the 37.2-percentage-point odds spread, and every per-era figure (percentage profitable, median gross gain, median annualized return, median hold and pair counts) are derived from PropKaki's profit_by_dimension_mv, a GROUP BY over resale_pairs_mv sliced to the buy_year_band dimension.

What we did not claim. That any of these numbers is a forecast — they are historical base rates that move with the market. That the recent bands are representative — the 2019+ (and, more mildly, 2014-2018) figures are survivor-skewed: a pair needs a completed resale, so recent quick-flip winners are over-represented and owners still holding are absent, which is why we read those bands as directional, not as odds for buying today. That gross equals net — every gain is before agent commission, buyer's and seller's stamp duties, any Seller's Stamp Duty, legal fees, loan interest and maintenance, and the thin-margin eras (pre-2000's +1.1% a year) barely survive those costs. That vintage is a cause independent of price — cheap-PSF entries skew to older eras, so this dimension overlaps the entry-price story and should be read alongside it. That the why behind any era is a data claim — the cycle explanations (the mid-1990s peak, the crises that followed) are interpretation of the pattern, not something the pairs themselves assert. A pair also needs a prior recorded purchase, so first-sale-only and very old units are under-represented. This is general information, not financial advice. To assess a specific unit, use the Profitability model.