Serviceability is the third and most nuanced ceiling. It answers: given your income and all your committed expenses, what's the biggest loan you can actually repay?
The stress-test rate
APRA's Prudential Practice Guide APG 223 (August 2022) requires all authorised deposit-taking institutions to assess mortgage applications at a minimum buffer of 3.0 percentage points above the loan's interest rate. GetReal applies this buffer to its current interest rate assumption. See Interest rate assumptions for the rate used.
Monthly surplus
The available surplus is net income minus all committed monthly expenses. GetReal calculates this as:
Maximum serviceable loan
The maximum loan is solved iteratively (binary search): find the highest loan amount where monthly repayments at the stress rate do not exceed the monthly surplus.
APRA's macro-prudential framework directs lenders to limit high-DTI lending. While APRA does not set an absolute hard cap, mainstream lenders have adopted 6× gross income as their practical ceiling. GetReal applies 6× as its upper bound.
What counts as debt
| Debt type | How it's counted |
|---|---|
| Existing mortgages | Outstanding balance at time of application |
| Credit cards | Total approved limit across all cards — not the balance |
| Car loans / personal loans | Outstanding balance |
| HECS/HELP | Excluded from DTI since 30 September 2025 |
| BNPL (Afterpay, etc.) | Outstanding balance — included if declared |
LVR determines how much of a property's value a lender will fund. The higher the LVR, the more risk the lender takes on — so they charge LMI as protection above 80%, and cap the LVR outright at a maximum.
| Property type | Owner-occupier max LVR | Investor max LVR |
|---|---|---|
| House | 95% | 90% |
| Townhouse | 95% | 90% |
| Apartment | 90% | 80% |
These are practical market maximums at mainstream lenders. Individual lenders may apply stricter limits — particularly for apartments in high-density postcodes, small apartments under 50m², studio apartments, or in regional areas with limited comparable sales.
LMI capitalisation: When LMI is added to the loan balance (capitalised), GetReal ensures the resulting LVR including LMI does not exceed the ceiling. A 95% LVR ceiling means the total loan including capitalised LMI is at most 95% of the purchase price.
LMI is not optional at high LVR — it's a lender requirement. It protects the lender, not the borrower. The premium can be paid upfront or capitalised into the loan (added to the balance). GetReal assumes capitalisation, as this is the most common approach.
Additionally, stamp duty is payable on the LMI premium itself in most states. GetReal applies state-based LMI stamp duty rates from the lmi_stamp_duty_rates table.
LMI rates
GetReal uses indicative LMI rates from the Home Loan Experts rate table (May 2026), stored in the lmi_rates Supabase table. Rates are segmented by LVR band (15 bands) and loan amount band (5 bands from under $300k to over $1M).
| LVR band | Loan <$300k | $300k–$500k | $500k–$750k | $750k–$1M | Over $1M |
|---|---|---|---|---|---|
| 80.01–85% | 0.70% | 0.70% | 0.76% | 0.76% | 0.76% |
| 85.01–90% | 1.43% | 1.43% | 1.43% | 1.43% | 1.43% |
| 90.01–95% | 2.77% | 2.77% | 3.09% | 3.09% | 3.09% |
Indicative only. Actual rates depend on lender and insurer. Genworth and QBE are the two main LMI providers in Australia.
stamp_duty_brackets Supabase table.
Stamp duty is calculated before deposit can be applied — it comes out of your savings first. This is why it matters so much to the deposit ceiling calculation: a $50,000 stamp duty bill on a $900,000 purchase is $50,000 that doesn't become deposit.
General formula (all states)
State-by-state rates (standard, non-FHB, established)
| State | $500k | $800k | $1.0M | $1.5M |
|---|---|---|---|---|
| NSW | $17,990 | $31,090 | $40,090 | $68,215 |
| VIC (standard) | $21,970 | $43,070 | $55,000 | $82,500 |
| VIC (owner-occ ≤$550k) | $5,765 | $43,070 | $55,000 | $82,500 |
| QLD | $8,750 | $24,525 | $34,525 | $59,525 |
| WA | $17,765 | $28,453 | $37,453 | $63,453 |
| SA | $21,330 | $38,730 | $50,730 | $81,730 |
| TAS | $18,247 | $30,247 | $38,247 | $63,247 |
| ACT | $12,800 | $25,200 | $34,200 | $60,200 |
| NT | $23,929 | $38,422 | $49,007 | $82,543 |
Figures are approximate, rounded, non-FHB, established property. Use the stamp duty calculator for precise figures.
NT — formula-based calculation
NT uses a formula rather than bracket tables for purchases up to $525,000:
First home buyer concessions
Every state offers some form of FHB relief. Thresholds and taper rates are stored in the stamp_duty_concessions table and applied automatically when you indicate you're a first home buyer in the calculator.
| State | FHB exemption threshold | Taper top |
|---|---|---|
| NSW | $800,000 | $1,000,000 |
| VIC | $600,000 | $750,000 |
| QLD | $700,000 | $800,000 |
| WA | $450,000 | $600,000 |
| SA | No duty concession (grant only) | — |
| TAS | 50% discount on established homes | — |
| ACT | Full exemption (income-tested, Home Buyer Concession Scheme) | — |
| NT | $18,601 rebate (phaseout starts at $650k) | $723,000 |
The change means HECS no longer affects your DTI ceiling (Ceiling 2). But it still affects your serviceability ceiling (Ceiling 3), because the ATO automatically withholds HECS repayments from your salary above certain income thresholds — reducing the take-home pay available to service a mortgage.
Compulsory HECS repayment rates (2025–26)
| Repayment income threshold | Repayment rate |
|---|---|
| Below $54,435 | Nil |
| $54,435 – $62,738 | 1.0% |
| $62,739 – $66,602 | 2.0% |
| $66,603 – $70,619 | 2.5% |
| $70,620 – $74,856 | 3.0% |
| $74,857 – $79,347 | 3.5% |
| $79,348 – $84,107 | 4.0% |
| $84,108 – $89,154 | 4.5% |
| $89,155 – $94,503 | 5.0% |
| $94,504 – $100,174 | 5.5% |
| $100,175 – $106,185 | 6.0% |
| $106,186 – $112,556 | 6.5% |
| $112,557 – $119,309 | 7.0% |
| $119,310 – $126,467 | 7.5% |
| $126,468 and above | 8.0% |
GetReal estimates monthly HECS withholding from gross income and deducts it from take-home pay in the serviceability calculation. If you enter your actual take-home pay (already net of HECS), GetReal does not double-count — HECS withholding is only applied if you declare a HECS debt and the tool is estimating your net income.
↗ Source: ATO — HELP debt repayment thresholdsHEM covers basic living costs: food, utilities, transport, clothing, and everyday spending. It excludes rent, school fees, and private health insurance — these are added separately. GetReal segments HEM by household type, number of dependants, and location (metropolitan vs regional).
| Household type | Dependants | Metro / month | Regional / month |
|---|---|---|---|
| Single | 0 | $2,480 | $2,150 |
| Single | 1 | $3,050 | $2,700 |
| Single | 2 | $3,480 | $3,100 |
| Single | 3+ | $3,950 | $3,520 |
| Couple | 0 | $3,680 | $3,200 |
| Couple | 1 | $4,200 | $3,700 |
| Couple | 2 | $4,680 | $4,150 |
| Couple | 3+ | $5,200 | $4,650 |
Indicative figures based on JMD Mortgages HEM estimates (March 2026). Actual lender HEM benchmarks vary and are not publicly published.
Metro vs regional: GetReal determines location type (metropolitan or regional) from the buyer's postcode, using ABS ASGS Edition 3 classifications. Metropolitan = all eight GCCSA capital cities. Regional = everything else.
All three serviceability ceilings are sensitive to the interest rate assumption. GetReal uses a market-average variable rate rather than any lender's advertised rate, because the tool is designed to give a realistic estimate across lenders, not to reflect any specific product.
| Rate component | Current assumption |
|---|---|
| Representative variable rate | 6.49% p.a. |
| APRA serviceability buffer | + 3.00% |
| Stress-test rate | 9.49% p.a. |
| Loan term assumed | 30 years, principal and interest |
The calculator allows an interest rate override if you have a specific rate in mind — for example, a pre-approval offer from a lender. When overridden, the stress-test rate is still the entered rate + 3%.
The score is not a count of available listings. It's a signal of how often properties matching your full combination of criteria appear in this market. A tight score on any single factor — especially budget — dominates the result.
Score calculation
Budget factor (NSW)
For NSW suburbs, GetReal counts exactly how many of the last 13 months' sold properties fall at or below the budget. This is a direct calculation from individual records — no estimation involved.
Budget factor (VIC)
VIC budget scoring uses estimated curves — see VIC methodology below.
Bedroom and bathroom factors (NSW)
Where agency-sourced enrichment data is available (Ray White, McGrath, LJ Hooker), bedroom and bathroom distributions are calculated from actual sold records. For un-enriched properties, national distribution estimates are used as a fallback. Current NSW enrichment coverage: approximately 12% of records.
Grade labels
| Grade | Score range | What it means |
|---|---|---|
| Highly Realistic | 75–100% | Strong supply. Regular opportunities. |
| Realistic | 55–74% | Good supply. Patience needed. |
| Competitive | 35–54% | Limited supply. Act quickly or compromise. |
| Tight | 15–34% | Very limited supply. Consider relaxing one criterion. |
| Very Difficult | 5–14% | Rare supply. Most criteria are binding. |
| Unrealistic | 0–4% | This combination almost never appears. |
NSW has over 146,000 individual sale records, which makes it possible to measure the actual distribution of sale prices around the median — how many sales fall at 80% of median, 90%, 100%, 110%, and so on. This distribution shape turns out to be consistent across NSW with two meaningful variables: price bracket and market depth.
Price brackets
| Bracket | Median price range |
|---|---|
| 1 | Under $500,000 |
| 2 | $500,000 – $800,000 |
| 3 | $800,000 – $1,200,000 |
| 4 | $1,200,000 – $1,800,000 |
| 5 | Over $1,800,000 |
Market depth
Suburbs with fewer than 30 sales per year (thin markets) cluster more tightly around the median — buyers compete hard for the limited stock. Active markets (30+ sales) have wider price dispersion. GetReal uses two depth tiers: active (≥30 sales/year) and thin (<30 sales/year). Three tiers were tested; high-volume and medium-volume suburbs were statistically indistinguishable.
Annual sales count — important note
The annual sales count for VIC suburbs is the rolling 12-month figure from the Q4 2025 VGV report — meaning calendar year 2025, not the trailing 12 months from today. GetReal says "in 2025 (VGV annual data)" not "in the last 12 months" for this reason.
Curve lookup
| Source | Used for | Update frequency | Licence |
|---|---|---|---|
| NSW Valuer General — Bulk PSI | NSW individual sale records (146,000+) | Daily | Open — CC BY |
| Victorian Valuer General — Quarterly report | VIC suburb medians + sales volumes | Quarterly | Open |
| Ray White sold listings (API) | NSW bedroom/bathroom enrichment | Weekly | Scraping — best effort |
| McGrath sold listings (scrape) | NSW bedroom/bathroom enrichment | Weekly | Scraping — best effort |
| LJ Hooker sold listings (API) | NSW bedroom/bathroom enrichment | Weekly | Scraping — best effort |
| State revenue offices (all 8) | Stamp duty brackets and concessions | On legislative change | Public |
| Melbourne Institute — HEM | Living expense benchmarks | Quarterly (licensed) | Licensed — indicative figures used |
| Home Loan Experts — LMI rates | LMI premium estimation | As published | Indicative |
| RBA — Cash Rate Target | Interest rate assumption base | Periodic manual review | Public |
| ABS ASGS Edition 3 | Postcode metro/regional classification | Static (2021) | CC BY |
| ATO — HECS repayment schedule | Estimated HECS withholding by income | Annual (July) | Public |
| Date | Change | Section |
|---|---|---|
| Aug 2026 | Methodology page created. All formulas consolidated from deposit-faq.html and faq.html. | All |
| Sep 2025 | HECS/HELP excluded from DTI calculation per APRA direction effective 30 September 2025. DTI section updated. | DTI, HECS |
| Aug 2025 | LMI rates updated to May 2026 Home Loan Experts table. | LMI |
| Jan 2025 | Initial methodology documented. NSW individual records pipeline live. VIC curve estimation methodology established. | All |