01The ten SUVs that hold on hardest
Malaysians buy SUVs on space, height and presence. Almost nobody checks the number that ends up costing the most — how much of the purchase price is still there when it is time to sell. In this segment that number ranges from three-quarters to under half.
72.8%
Best five-year retention
Mazda CX-30, the midpoint of two independent methods
12
SUVs with a defensible figure
Out of 117 SUV cohorts that produced a valid curve
16×
Gap between best and worst
RM2,539 a year against RM41,757, in value lost
| # | SUV | 5-year retention | Curve method | Price-ladder method | Listings |
|---|---|---|---|---|---|
| 1 | Mazda CX-30 | 72.8% | 70.4% | 75.3% | 668 |
| 2 | Perodua Aruz | 71.4% | 65.2% | 77.7% | 1,295 |
| 3 | Honda CR-V | 69.3% | 73.3% | 65.3% | 2,439 |
| 4 | Honda HR-V | 65.8% | 69.2% | 62.5% | 2,347 |
| 5 | Proton X70 | 65.4% | 69.9% | 61.0% | 1,776 |
| 6 | Mazda CX-5 | 64.4% | 63.1% | 65.6% | 2,765 |
| 7 | MINI Countryman | 59.9% | 57.7% | 62.1% | 740 |
| 8 | Lexus RX350 | 57.8% | 51.9% | 63.7% | 481 |
| 9 | BMW X4 | 51.5% | 49.7% | 53.3% | 562 |
| 10 | Volvo XC90 | 49.2% | 48.1% | 50.2% | 416 |
02What each row actually costs to own
Retention is only half the ownership question. The other half is what the vehicle asks of you every year while you keep it — and in this segment the two halves point the same way far more tidily than they do in the market as a whole.
| SUV | Near-new ask | 5-year-old ask | Value lost per year | Annual road tax |
|---|---|---|---|---|
| Perodua Aruz | RM56,877 | RM44,184 | RM2,539 | RM120 |
| Mazda CX-30 | RM104,310 | RM78,509 | RM5,160 | RM440 |
| Proton X70 | RM77,823 | RM47,487 | RM6,067 | RM360 |
| Honda HR-V | RM95,387 | RM59,570 | RM7,163 | RM360 |
| Mazda CX-5 | RM115,373 | RM75,661 | RM7,942 | RM440 |
| Honda CR-V | RM125,152 | RM81,692 | RM8,692 | RM439 |
| MINI Countryman | RM189,644 | RM117,703 | RM14,388 | RM440 |
| Lexus RX350 | RM305,595 | RM194,564 | RM22,206 | RM754 |
| BMW X3 | RM250,780 | RM130,070 | RM24,142 | RM440 |
| BMW X4 | RM291,049 | RM155,194 | RM27,171 | RM440 |
| Volvo XC90 | RM297,627 | RM149,542 | RM29,617 | RM428 |
| BMW X5 | RM359,925 | RM151,140 | RM41,757 | RM1,637 |
The ordering barely changes between the two tables, which is the useful part. In the market as a whole, percentage retention and ringgit lost rank very differently — a cheap car can shed a large share of a small number. Among SUVs the entry price and the retention rate reinforce each other, so the expensive end of the segment is punished twice.
SUVs are taxed on a different scale, and it cuts both ways
Malaysian road tax uses one rate table for saloons and another for SUVs, MPVs, vans and pickups. Below about 2,300cc the non-saloon scale is the more expensive of the two, and above it the cheaper — the two schedules cross at roughly that point. A 2.0-litre SUV pays RM440 a year where a 2.0-litre saloon pays RM380; a 3.0-litre SUV pays RM1,637 where a 3.0-litre saloon pays RM2,130. The rates come from JPJ's licence-rate calculation guideline.
03Why both Mazdas sit high, and what expires
Mazda has a structural advantage here that shows up directly in resale, and it has an expiry date. Every Mazda registered from January 2019 carries a five-year warranty and five years of free scheduled service, capped at 100,000 km, covering parts, labour and lubricants at 10,000 km or six-month intervals.
For a used buyer that is not a marketing line, it is a transferable asset with a countdown on it. A three-year-old CX-30 under 100,000 km still has roughly two years of scheduled maintenance already paid for. A four-and-a-half-year-old one has almost none. Two cars with the same odometer reading and the same asking price can carry very different amounts of remaining cover, and that difference is a legitimate thing to negotiate over.
When the free period ends, Mazda's own extended service plan covers a further four services at the same intervals for RM2,700 to RM3,400 on petrol models. That is a useful anchor for what the next stretch of ownership costs, and it is the number a buyer should have in mind when a car's free-service window is nearly closed.
04The national and Japanese middle is where the volume is
The three deepest-sampled SUVs in the table are also the three most people actually shortlist, and they land within four points of each other. That is the practical centre of this market.
Perodua Aruz — the cheapest SUV to own, by a distance
Second on retention at 71.4%, and first by a long way on cost: RM2,539 of value lost a year and RM120 of road tax on its 1,500cc engine. It also carries a five-star ASEAN NCAP result, scoring 43.25 points for adult occupant protection under the 2017–2020 protocol on the 2018 G variant. Nothing else in this table combines that little depreciation with that much crash-test evidence.
Honda CR-V and HR-V — the deepest evidence, and a recall to check
The CR-V is third at 69.3% and rests on the broadest base in the ranking: 2,439 listings across 127 model-year cohorts, a span of nearly three decades. The HR-V is fourth at 65.8% on 2,347 listings. Honda's published maintenance schedule sets the servicing rhythm, and an independent five-year comparison put the HR-V's scheduled maintenance at RM3,874.45 over 100,000 km — the cheapest of the SUVs in that study, against RM4,326.95 for the Proton X50 and slightly more for the X70.
One thing to verify on any used Honda: the company recalled 87,490 vehicles in Malaysia for fuel-pump replacement. A recall that has been carried out is not a defect — it is a repaired car with a paper trail. An outstanding one is an unfinished job you inherit. Ask for the rectification record and check it against the chassis number before you commit.
Proton X70 — strong sales, mid-table resale
Fifth at 65.4% on 1,776 listings. Its two methods disagree in a direction worth understanding: 69.9% on the fitted curve against 61.0% on the observed price ladder. A steady flow of newer units entering the market at competitive prices compresses what earlier cars can ask. Strong new-car sales and strong resale are not the same thing, and the X70 is the clearest example of that in the segment.
05How we measured it, and the SUVs we refused to rank
We used the same editorial method as our market-wide value-retention table: a decay curve fitted across model-year cohorts inside a single market snapshot, cross-checked against an independently computed price ladder. This is editorial analysis of asking-price patterns, not how a Carvaly valuation is calculated. Understanding your Carvaly valuation explains how to read a result.
Deciding what counts as an SUV is its own problem. Body type is free text in the source data and splits across `SUV`, `SUVs`, `4 Wheels` and worse, and only 195,877 of the 339,761 listings we can attribute to a make and model carry one at all. We normalise every value to a fixed set, take a majority vote per model, and require at least 20 body-typed listings and 70% agreement before calling a cohort an SUV. That resolved a body type for 470 model cohorts.
| Stage | SUV cohorts remaining |
|---|---|
| Model cohorts in the 14 August 2026 snapshot | 1,587 (all body types) |
| Cohorts with a confidently resolved body type | 470 (all body types) |
| SUV cohorts that produced a valid curve | 117 |
| Cleared every depth and fit-quality gate | 15 |
| Corroborated by a second independent method | 12 |
The gates a model had to clear:
- No fallback assumptions. Where our system cannot fit a car from observed listings it substitutes a segment-wide prior. Those rows never appear in a ranking that names a car.
- Real depth, in the right age range. At least 6 model-year cohorts, 200 listings, a 6-year span, and — critically — at least 4 cohorts and 200 listings aged 0–6 years.
- No boundary-clamped curves. Our estimator limits implausible rates. A curve resting on that limit is a boundary, not a measurement, and was discarded.
- Agreement between two methods. Published only where the fitted curve and the observed price ladder came within 15 percentage points of each other.
The notable exclusions
| SUV | Listings | Why it is not ranked |
|---|---|---|
| Proton X50 | 2,408 | Anchor-free fit quality of 0.79 against our 0.80 floor — one threshold, missed by 0.01. Only six model years exist to fit. |
| Toyota Harrier | 2,440 | Part of the fitted curve rests on the estimator's rate limit. A boundary is not a measurement. |
| Nissan X-Trail | 1,339 | Cleared the gates, but too little near-new depth to build an independent price ladder. Single-method only. |
| Perodua Ativa | 1,068 | Five model years of history. A five-year retention claim would be extrapolation. |
| BMW X1 | 969 | Boundary-clamped, like the Harrier. |
| Toyota Fortuner | 703 | Cleared the gates, but single-method only — no corroborating ladder. |
| Porsche Cayenne | 534 | Contested: 54.0% on the curve against 77.6% on the ladder, a 23.6-point disagreement. |
Across the market as a whole, the 39 models where both methods could be computed showed a rank agreement of Spearman 0.68. We publish both estimates for every row so you can see the spread rather than trusting a single decimal.
06What to actually do with this
A retention table is a shortlisting tool, not a verdict. Here is how the twelve rows above translate into decisions.
- If you want the cheapest SUV to own outright: the Perodua Aruz loses the least in ringgit (RM2,539 a year), pays the least road tax (RM120), and has a five-star crash result behind it. Nothing else in the table is close on cost.
- If you want a larger SUV without the Continental penalty: the Honda CR-V and Mazda CX-5 carry the deepest evidence in the ranking — 2,439 and 2,765 listings respectively — and both sit in the mid-60s on retention.
- If you want a German SUV, buy it old. A five-year-old BMW X5 asks RM151,140 against RM359,925 near-new — 42% of the money for the same vehicle. That is an excellent trade if, and only if, you budget honestly for out-of-warranty repairs and RM1,637 a year in road tax.
- Check the free-service clock on any Mazda under five years old. It is a transferable asset with an expiry date, and it is worth real money on the negotiation.
- In every case: the model-level figure is a starting point. Year, variant, mileage, service history and condition decide the number that actually matters.
Frequently asked questions
Which SUV holds its value best in Malaysia?
In our 14 August 2026 analysis the Mazda CX-30 ranked first, retaining an estimated 72.8% of its value over five years, backed by 668 marketplace listings. The Perodua Aruz followed at 71.4% and the Honda CR-V at 69.3%. These are estimates from asking-price patterns across model-year cohorts, and each is published with its sample size.
Why is the Proton X50 not in the ranking?
It failed one of our publication gates by 0.01. The X50's anchor-free fit quality came out at 0.79 against a 0.80 floor, largely because only six model years of it exist to fit a curve across. That is a statement about what we can measure, not a finding about how the X50 depreciates. We expect it to qualify once another model year of data exists.
Do SUVs pay more road tax than sedans in Malaysia?
Below roughly 2,300cc, yes. Malaysia assesses SUVs, MPVs, vans and pickups on a separate non-saloon rate table, and the two schedules cross at about that displacement. A 2.0-litre SUV pays RM440 a year against RM380 for a 2.0-litre saloon, but a 3.0-litre SUV pays RM1,637 against RM2,130 for the equivalent saloon.
Are Continental SUVs a bad buy in Malaysia?
They are expensive to buy new and inexpensive to buy used, which is the same fact stated twice. The BMW X5, X3 and X4 all retain under 52% at five years, and the X5 loses about RM41,757 a year. If you are the second owner that transfer works in your favour — provided you price out-of-warranty repair risk honestly rather than assuming it away.
How many SUVs did you actually measure?
We resolved a body type for 470 model cohorts, of which 117 SUV cohorts produced a valid depreciation curve. Fifteen cleared our depth and fit-quality gates and twelve were corroborated by a second independent method. The remaining 102 were excluded for thin samples, poor fits, boundary-clamped curves, or disagreement between the two methods.
Sources and references
- Carvaly: Understanding your Carvaly valuation
- Carvaly Market Intelligence: Malaysia Used-Car Market Report · July 2026
- Mazda Malaysia: Peace of Mind: 5 Years Warranty + 5 Years Free Service Maintenance
- Mazda Malaysia: Mazda Extended Service Maintenance
- Paul Tan's Automotive News: Proton X50 versus X70 and Honda HR-V: Servicing Costs Over Five Years
- Honda Malaysia: Honda Malaysia Fuel Pump Product Recall
- ASEAN NCAP: Perodua Aruz Safety Rating
- JPJ Malaysia: Motor Vehicle Licence Rate Calculation Guide
- Malaysian Automotive Association (MAA): Production and Sales for May 2026
