The URA price index, all 51 years
- The URA non-landed index rose 20.1× in 51 years (10.5 → 210.8) — about 6.1% a year, nominal.
- But it fell 10%+ on 4 occasions, worst -47% (1981–1986). The 1996 peak took ~9 years to regain.
- There was a genuine lost decade: 1995–2005 compounded at -1.7% a year — prices ended lower than they began.
- Growth is decelerating by decade: +12.5% → +11.9% → -1.7% → +5.9% → +3.8% → +2.6% — underwrite the recent pace, not the 1980s.
- Since 2004 the suburbs beat prime: OCR +227% vs CCR +109%.
- Best for: anyone sizing long-run expectations, stress-testing a hold, or checking whether “property always goes up” survives the record.
Every Singapore property conversation eventually reaches “property always goes up.” The URA price index is the longest official record we have to test that — 205 quarters, 51 years of non-landed private housing, from 1975-Q1 to 2026-Q1. It does go up. It also fell -47% once, and spent an entire decade going backwards.
Fifty-one years in one line
URA Non-Landed Private Residential Property Price Index
Quarterly, base 2009-Q1 = 100. Shaded bands = periods spent below the previous peak. Dashed lines: mid-80s recession, Asian Financial Crisis, SARS, GFC, ABSD introduced, TDSR, ABSD raised sharply.
How far underwater, and for how long
The line above hides the pain. This is the same data drawn as a drawdown — how far below the previous peak the market was at any moment. Flat at zero means new highs; the valleys are the years buyers spent waiting to get back to even.
Drawdown from prior peak
0% = at an all-time high. The market has spent large stretches of its history below water — the mid-1980s and post-1996 valleys each lasted the better part of a decade.
Every major fall, and how long it took to recover
| Cycle | Peak | Trough | Fall | Qtrs down | Qtrs to recover | Back to peak |
|---|---|---|---|---|---|---|
| Early-80s boom → mid-80s recession | 1981-Q2 | 1986-Q2 | -47.0% | 20 | 20 q | 1991-Q2 |
| 1996 anti-speculation → Asian Financial Crisis | 1996-Q2 | 1998-Q4 | -41.7% | 10 | 36 q | 2007-Q4 |
| Global Financial Crisis | 2008-Q2 | 2009-Q2 | -26.1% | 4 | 4 q | 2010-Q2 |
| TDSR & the cooling-measure grind | 2013-Q3 | 2017-Q2 | -10.2% | 15 | 9 q | 2019-Q3 |
Falls of 10%+ from a prior peak. “Qtrs to recover” counts from the trough back to the old high.
The engine is slowing
Split the 51 years into decades and the story is not “steady compounding” — it is a structural deceleration. The 1970s–80s ran at double digits. Then came a genuine lost decade. Each decade since has grown more slowly than the last.
Annualised growth by decade
Compound annual growth of the non-landed index within each decade. The 1995–2005 decade was negative — prices ended lower than they started.
Prime lost to the suburbs
URA splits the non-landed index by region from 2004-Q1. Rebase all three to 100 at the start and the received wisdom — “prime always outperforms” — inverts completely.
Non-landed index by region, rebased to 100 at 2004-Q1
Same base, same period (2004-Q1 → 2026-Q1). The mass-market Outside Central Region compounded far ahead of the prime Core Central Region.
What this means if you are buying
Three things the record actually supports. One: the long-run drift is real (~6.1% a year over half a century) but it is nominal and it is slowing — underwrite nearer the recent decade's pace, not the 1980s'. Two: drawdowns of 25–45% have happened roughly once a generation and recovery has taken up to a decade, so holding power beats timing — the buyers who lost money were the ones forced to sell. Three: the index is an average; the regional split shows how wide the dispersion around it is. Which project you buy, and what you pay for it, matters more than the index.
The index is the market. Your project is not the index.
Half the gap between a good and a bad outcome is the entry price versus what the property is actually worth. That is what our fair-value model measures, project by project.
See which condos trade below fair value › · New launch vs resale: which made more money?