Does The Price-To-Index Ratio Still Work When We Measure Percentage Return?
In the previous article, I tested a simple idea:
Can we compare a condo's entry PSF against the URA regional price index to see whether the buyer was paying a stretched price?
The first version was useful, but it had one obvious weakness. It looked heavily at dollar profit and dollar loss.
That matters because dollar profit can be distorted by size. A large apartment can lose $1 million and look dramatic. A small apartment can make $150,000 and look boring. But percentage return tells a cleaner story of capital efficiency.
So this is the follow-up test.
If the price-to-index ratio is useful, then higher ratios should still produce weaker results even when we stop looking at absolute profit and measure percentage return instead.
The formula stays the same:
Price-to-index ratio = entry PSF / URA regional non-landed price index at purchase quarter
In plain English: how much PSF did the buyer pay, relative to where the regional market already was at that point in time?
A lower number means the buyer entered at a less stretched price. A higher number means the buyer paid a higher PSF while the regional index was already elevated.
This is not an industry rule. It is a working theory. The question here is whether it survives a tougher test.
The cleaner test: percentage return
I rebuilt the test using private condo resale transactions, joined against the URA regional non-landed price index for CCR, RCR, and OCR.
For each resale transaction, I worked backwards:
- when was the property bought?
- what was the entry PSF?
- what was the URA regional index in that purchase quarter?
- what was the buyer's price-to-index ratio?
- what percentage return did the seller eventually get?
Then I filtered to transactions where the owner held for at least 5 years.
That gave 16,830 usable transactions:
| Region | Transactions |
|---|---|
| CCR | 2,385 |
| RCR | 4,280 |
| OCR | 10,165 |
The first broad result was already clear: higher ratio bands generally produced lower percentage returns in all three regions.
But broad comparisons can still be unfair. Some buyers held much longer than others. Some bought much larger units. So the more useful test is the matched-hold version: only sellers who held between 5 and 10 years.
That is where the ratio becomes more interesting.
Test 1: same holding period band
Here is the 5-to-10-year hold slice.
| Region | Ratio Band | Txns | Typical Ratio | Typical Entry PSF | Avg Return | Typical Return | Annualised Return | Win Rate | Typical Hold | Typical Size |
|---|---|---|---|---|---|---|---|---|---|---|
| CCR | Lowest | 221 | 11.23 | $1,458 | 34.8% | 34.6% | 4.1% | 99.1% | 7.5 yrs | 1,378 sqft |
| CCR | Low-mid | 222 | 14.01 | $1,822 | 20.9% | 20.2% | 2.5% | 94.1% | 7.5 yrs | 985 sqft |
| CCR | High-mid | 221 | 16.57 | $2,150 | 9.8% | 8.4% | 1.2% | 78.7% | 7.4 yrs | 850 sqft |
| CCR | Highest | 222 | 19.69 | $2,579 | -1.0% | -0.1% | 0.0% | 49.1% | 7.6 yrs | 850 sqft |
| RCR | Lowest | 594 | 8.68 | $1,244 | 40.0% | 39.3% | 4.7% | 100.0% | 7.4 yrs | 1,152 sqft |
| RCR | Low-mid | 595 | 10.63 | $1,564 | 31.1% | 30.2% | 3.8% | 97.8% | 6.8 yrs | 829 sqft |
| RCR | High-mid | 594 | 11.67 | $1,719 | 27.8% | 28.3% | 3.7% | 99.2% | 6.6 yrs | 732 sqft |
| RCR | Highest | 595 | 13.17 | $1,936 | 16.9% | 16.3% | 2.2% | 92.6% | 7.0 yrs | 678 sqft |
| OCR | Lowest | 1,274 | 4.86 | $766 | 67.7% | 67.6% | 6.8% | 99.9% | 8.0 yrs | 1,109 sqft |
| OCR | Low-mid | 1,275 | 5.54 | $908 | 62.2% | 60.2% | 6.6% | 99.9% | 7.6 yrs | 1,055 sqft |
| OCR | High-mid | 1,275 | 7.27 | $1,193 | 36.1% | 34.9% | 4.3% | 99.8% | 7.3 yrs | 915 sqft |
| OCR | Highest | 1,275 | 8.75 | $1,448 | 26.5% | 24.9% | 3.2% | 99.1% | 7.3 yrs | 678 sqft |
The answer is yes, directionally.
The ratio still works when we measure percentage return.
CCR is the clearest warning. In the lowest ratio band, sellers averaged a 34.8% return with a 99.1% win rate. In the highest ratio band, the average return was slightly negative and the win rate dropped to 49.1%.
That is not a small difference. That is the difference between almost everyone making money and a coin flip.
RCR was less brutal, but the compression was still obvious. The lowest ratio band averaged a 40.0% return. The highest averaged 16.9%. Buyers still generally made money, but they gave up a lot of upside by entering at the stretched end.
OCR was the most forgiving region. Even the highest ratio band still averaged a 26.5% return. But the pattern still held: the lowest ratio band averaged 67.7%.
So the ratio is not just a dollar-profit trick. It still shows up when we measure capital efficiency.
Test 2: remove the very small and very large units
There is another possible objection.
Maybe the ratio is just picking up unit size. The highest ratio bands often include smaller apartments, while the lowest bands may include older, larger units bought at lower PSF.
So I ran a tighter slice: only 700 to 1,500 sqft units, also with 5-to-10-year holding periods.
| Region | Ratio Band | Txns | Typical Ratio | Typical Entry PSF | Avg Return | Typical Return | Annualised Return | Win Rate | Typical Hold | Typical Size |
|---|---|---|---|---|---|---|---|---|---|---|
| CCR | Lowest | 116 | 11.47 | $1,488 | 34.4% | 33.3% | 4.0% | 100.0% | 7.2 yrs | 1,104 sqft |
| CCR | Low-mid | 117 | 13.59 | $1,796 | 23.6% | 22.1% | 2.9% | 97.4% | 7.5 yrs | 1,001 sqft |
| CCR | High-mid | 116 | 16.26 | $2,117 | 13.0% | 12.9% | 1.8% | 86.2% | 7.4 yrs | 974 sqft |
| CCR | Highest | 117 | 19.31 | $2,566 | 1.8% | 2.3% | 0.3% | 59.8% | 7.4 yrs | 969 sqft |
| RCR | Lowest | 342 | 8.33 | $1,209 | 42.8% | 41.4% | 5.1% | 100.0% | 7.3 yrs | 1,141 sqft |
| RCR | Low-mid | 342 | 10.34 | $1,499 | 35.1% | 34.7% | 4.2% | 98.5% | 6.9 yrs | 958 sqft |
| RCR | High-mid | 342 | 11.26 | $1,673 | 34.8% | 33.9% | 4.7% | 100.0% | 6.4 yrs | 904 sqft |
| RCR | Highest | 343 | 12.65 | $1,864 | 22.0% | 22.0% | 2.9% | 94.5% | 7.1 yrs | 861 sqft |
| OCR | Lowest | 929 | 4.84 | $762 | 68.9% | 68.7% | 6.9% | 100.0% | 8.0 yrs | 1,098 sqft |
| OCR | Low-mid | 929 | 5.33 | $843 | 70.8% | 69.7% | 7.0% | 100.0% | 7.8 yrs | 1,055 sqft |
| OCR | High-mid | 929 | 6.58 | $1,070 | 42.9% | 42.3% | 5.0% | 99.9% | 7.4 yrs | 1,033 sqft |
| OCR | Highest | 930 | 8.26 | $1,353 | 34.4% | 34.2% | 4.2% | 99.6% | 7.3 yrs | 915 sqft |
This tighter slice does not destroy the theory.
It softens it, but it does not destroy it.
In CCR, the highest-ratio band no longer goes negative once the extreme sizes are removed, but the return still collapses from 34.4% to 1.8%.
In RCR, the highest-ratio band still does worse than the lower-ratio bands.
OCR is the interesting one. The lowest and low-mid bands are almost tied, which tells us the ratio is not a perfect ordering tool. But once the ratio moves into the higher bands, the return drop is still clear.
This is probably the most useful interpretation:
The price-to-index ratio is not precise enough to rank every property neatly from best to worst. But it is useful enough to tell you when the entry price is starting to look stretched.
What this means by region
CCR is where the ratio is most dangerous.
The reason is not just that CCR properties are expensive. It is that the CCR index itself has not grown as strongly as RCR or OCR over the same long period. If you buy into a high ratio in a region where the index does not move much, the future resale market has very little room to rescue your entry price.
That is why CCR's highest band looks so ugly.
RCR is different. A stretched entry still hurts, but the region had more index tailwind. The highest-ratio band did not collapse into losses the way CCR did. It simply produced much weaker returns.
OCR is the most forgiving. The region's index growth did a lot of heavy lifting. Even buyers who entered at higher ratios generally still made money, but their returns were meaningfully lower than the lower-ratio cohorts.
So the framework should not be used as one universal number.
A ratio that looks dangerous in OCR may not mean the same thing in CCR. Each region has its own normal range.
Danger Zones
| Region | Ratio Range | Txns | Loss Rate | Win Rate | Avg Return | Avg Profit | Avg Size | Read |
|---|---|---|---|---|---|---|---|---|
| CCR | Below 12 | 60 | 0.0% | 100.0% | 39.4% | $703k | 1,260 sqft | Strong |
| CCR | 12-15 | 74 | 2.7% | 97.3% | 27.4% | $586k | 1,226 sqft | Shortlist |
| CCR | 15-18 | 64 | 10.9% | 89.1% | 17.1% | $447k | 1,208 sqft | Risky |
| CCR | 18-21 | 36 | 41.7% | 58.3% | 1.8% | $31k | 1,146 sqft | Walk-away |
| CCR | Above 21 | 15 | 53.3% | 46.7% | -4.9% | -$186k | 1,279 sqft | Walk-away |
| RCR | Below 9 | 187 | 0.0% | 100.0% | 47.0% | $642k | 1,223 sqft | Strong |
| RCR | 9-11 | 228 | 1.3% | 98.7% | 39.4% | $673k | 1,169 sqft | Strong |
| RCR | 11-13 | 159 | 1.9% | 98.1% | 35.7% | $695k | 1,145 sqft | Shortlist |
| RCR | 13-15 | 35 | 11.4% | 88.6% | 18.3% | $425k | 1,174 sqft | Risky |
| RCR | Above 15 | 10 | 10.0% | 90.0% | 13.1% | $351k | 1,179 sqft | Risky |
| OCR | Below 5 | 609 | 0.0% | 100.0% | 67.4% | $590k | 1,175 sqft | Strong |
| OCR | 5-6.5 | 895 | 0.0% | 100.0% | 63.1% | $643k | 1,166 sqft | Strong |
| OCR | 6.5-8 | 454 | 0.0% | 100.0% | 43.5% | $607k | 1,179 sqft | Shortlist |
| OCR | 8-9.5 | 135 | 0.0% | 100.0% | 39.4% | $635k | 1,168 sqft | Caution |
| OCR | Above 9.5 | 20 | 0.0% | 100.0% | 41.0% | $771k | 1,146 sqft | Caution, not walk-away |
Highlighted Transactions
| Case | Project | Region | Size | Purchase Q | Entry Ratio | Hold | Return | Profit / Loss |
|---|---|---|---|---|---|---|---|---|
| Winner | Jadescape | RCR | 1,259 sqft | 2018Q3 | 11.0 | 7.3 yrs | +56.6% | +$1.15M |
| Winner | One Amber | RCR | 1,453 sqft | 2016Q3 | 8.9 | 8.1 yrs | +80.8% | +$1.46M |
| Loser | Beaufort on Nassim | CCR | 1,367 sqft | 2015Q1 | 23.7 | 9.6 yrs | -23.5% | -$1.00M |
| Loser | Marina One Residences | CCR | 1,163 sqft | 2020Q3 | 20.5 | 5.5 yrs | -26.3% | -$810k |
| Exception | Reflections at Keppel Bay | RCR | 1,227 sqft | 2019Q3 | 10.4 | 5.5 yrs | -6.0% | -$120k |
So did the theory hold up?
Yes, but with conditions.
The price-to-index ratio survived the percentage-return test. Higher ratios generally produced lower percentage returns across CCR, RCR, and OCR.
That makes the framework more useful than the first article showed, because it means the signal is not only about large homes creating large dollar losses. The pattern still appears when the outcome is measured as return on capital.
But this is still not a magic formula.
A low ratio does not automatically mean a good buy. Some cheap properties stay cheap because there are structural problems: weak buyer demand, awkward layouts, poor project positioning, thin liquidity, or simply a location that the market does not love.
A high ratio also does not automatically mean disaster. Trophy assets, rare formats, and exceptional locations can sometimes sustain premiums that ordinary projects cannot.
The cleaner way to use the ratio is as a warning light.
If the ratio is low, the entry price has more room to breathe.
If the ratio is high, the question becomes sharper:
What must be true for a future buyer to pay enough to rescue this entry price?
That is the point of the framework.
It does not tell you whether a condo is good or bad. It tells you how much pressure your entry price is putting on the future resale market.
And when we measure percentage return instead of dollar profit, that pressure still shows up.