Maldivian property earns less than the debt that would buy it
Seven ten-year strategies modeled on the honest evidence base return 1–9% nominal in the base case — below every quoted lending rate. Debt-service coverage sits under 1.0 across the board, leverage subtracts value in four of the five financeable strategies, and the only routes that clearly pencil are inherited land and USD-indexed rent.
What a valuation can honestly rest on
The Maldives has no transaction-price register, no house-price index and no rent index [OFFICIAL-absence, Land Act 1/2002 registries publish no prices]. A valuation exercise here cannot produce point estimates; it can produce defensible ranges by category, each tagged with what kind of evidence supports it. That is what this paper does — and then it pushes the ranges through ten-year cash-flow models to ask the question valuations in Malé are never asked: does the income the asset can actually earn justify the price being asked for it?
The reconciliation logic matters more than any single row. For new-build apartments, the July 2026 asking band of MVR 3,150–4,330/sqft (USD 204–281/sqft at the 15.42 peg; roughly USD 2,200–3,000/sqm) [ASKING, portal sample n=9] sits close to replacement cost: mid-grade construction at MVR 2,500–3,500/sqft [OWN-CALC from the Hiyaa official anchor escalated at the MBS +13.53% construction-price rise, Mar 2024–Mar 2025 [OFFICIAL]] plus a land share and a developer margin brackets the ask. Apartment prices are defensible on cost — this is a construction-cost-driven market, not a land-speculation one. For Malé land, the opposite: portal asks of MVR 7,200–9,400/sqft and the single sold-marker at 9,412 [ASKING/LISTING, Jul 2026] are the levels evidence supports, while the broker-quoted MVR 12,000–20,000/sqft ceiling [INDUSTRY-EST, no methodology] and the government’s ~2018–19 administered sale at a flat 15,000 [MEDIA, vintage unverified] are quotes no income stream can carry — our residual-land-value work in paper 9 found redevelopment yields of only ~4.2–4.9% gross on cost at those levels [OWN-CALC]. HDC’s Hulhumalé Phase 2 tender floors falling from MVR 5,500–6,500 (Aug 2022) to 4,000–4,500/sqft (Sep 2024) [OFFICIAL floors] are the one place the state itself marked land down.
Our reasonable-range verdicts, category by category: Malé prime residential land MVR 7,000–9,500/sqft supported, 12,000+ quoted but unsupported; Hulhumalé land 4,000–6,500 (official floors; one private ask at 10,650 is aspirational); open-market apartments 3,150–4,330/sqft; HDC policy-priced stock 2,770–3,060 (not a market comp — it is ~15–30% below market by design); Thilafushi industrial freehold 1,700/sqft floor [OFFICIAL, clearing bids unpublished]; outer-atoll land under 700/sqft [ASKING]. Every figure asking, administered or listing tier.
The ratio toolkit — and where it breaks
Matched asking-price/asking-rent pairs (D09, July 2026 portal sample) give the only yield evidence that exists:
| Segment | Gross asking yield | Est. net yield | Price-to-rent (years of gross rent) | Tier |
|---|---|---|---|---|
| Hulhumalé Ph1 2BR furnished | 6.8–8.4% (price basis contested) | ~5.5–6.7% | 11.9–14.7 | OWN-CALC on ASKING/LISTING |
| Hulhumalé new-build 2BR | 4.6–6.8% | ~3.0–5.3% | 14.7–21.7 | OWN-CALC on ASKING |
| Malé 3BR unfurnished | 5.7–6.9% | ~4.2–5.4% | 14.5–17.5 | OWN-CALC on ASKING/LISTING |
| Malé 3+1 furnished, USD-quoted | 8.1–8.5% at peg | ~6.5–7.0% | 11.8–12.3 | OWN-CALC on ASKING/LISTING |
| All-segment band | 4.5–8.5% (centre ~6–7%) | ~2.5–6.5% | 11.8–22.2 | OWN-CALC |
Three limits of these ratios. First, there is no cap rate in the institutional sense — a cap rate is a transaction-derived price of income, and no verified income-property transaction is publicly observable in the Maldives. What we quote is asking-on-asking arithmetic; the ratio is more robust than either level (bias partly cancels), but it is not a traded price. Second, net is an assumption: no vacancy statistics or operating-cost surveys exist, so the 1.5–3.0 point drag from gross to net is our stated estimate [OWN-CALC], not data. Third, the currency of the rent changes everything. Five of fourteen sampled rental listings quote in USD [LISTING]. At the official peg those rents convert at 15.42; at the parallel rate of ~MVR 20.7 (July 2026 [MEDIA]) they are worth ~34% more rufiyaa. A landlord collecting USD 2,600/month holds an asset yielding 8.5% at peg-converted rent — and roughly 11.3% in effective MVR terms at the parallel rate [OWN-CALC]. USD-indexed rent is the single largest legal return enhancer in this market, and it is available only to owners of stock that expatriate and corporate tenants will pay dollars for.
Ten-year models A–G: design
We model seven strategies an investment committee, lender or Maldivian family would actually consider, over 2026–2036, in low/base/high scenarios. All outputs are [OWN-CALC]. The scenarios are internally consistent outcome bundles — low pairs weak rents with soft exits and high costs; high pairs the reverse — not probability-weighted forecasts. All figures are nominal MVR at the official peg; a step devaluation (a live underwriting scenario per the macro paper) would rescale every MVR number and is deliberately held outside the model rather than hidden inside it.
| Common assumption | Low | Base | High | Anchor |
|---|---|---|---|---|
| Rent growth, nominal per year | 0% | 2.5% | 5% | wage-constrained: HIES incomes flat since 2019; WB shows pooling to afford rents [OFFICIAL] |
| Vacancy, residential | 10% | 6% | 3% | no vacancy data exists — assumption [OWN-CALC] |
| Operating costs, share of collected rent | 15% | 15% | 15% | assumption; no survey exists |
| Exit value | year-11 gross rent capitalized at 8.0% | at 6.5% | at 5.5% | today’s asking-yield band 4.5–8.5%, centre ~6.5 [OWN-CALC] |
| Construction cost, models D/E | MVR 3,500/sqft | 3,000 | 2,500 | Hiyaa anchor + MBS construction inflation [OFFICIAL→OWN-CALC] |
| Tax | 15% flat on net rent and on nominal sale gains | — | — | Income Tax Act 25/2019 top bracket; simplification |
| Acquisition friction | 1% | — | — | court-executed transfer; no stamp duty exists [OFFICIAL] |
| Debt, models A/B/C | BML 7.5%, 20 years, 70% LTV | — | — | bank product page [OFFICIAL(bank), snippet-captured] |
| Debt, models D/E | 11.5%, 25 years, 70% of build cost | — | — | HDFC construction range 11.00–11.75% [OFFICIAL(bank)] |
Model-specific deviations: C (Hulhumalé Phase 2) carries its own weaker rent path (−2%/+1.5%/+4% growth, 15/10/5% vacancy, exits at 8.5/7.0/5.5%) because administered stock shadows its rental market; G (Thilafushi) uses lease rates of MVR 5.0/6.0/7.5 per sqft per month [MEDIA/LISTING legacy-lease anchors], 10% costs, and industrial exits at 7.0/5.5/4.5%. F and G are modeled unleveraged — no Maldivian bank product finances raw-land purchase for individuals on disclosed terms.
Results: the return stack, and what debt does to it
| Model | Base entry outlay | Entry yield | IRR low | IRR base | IRR high | Leveraged IRR base | Cash-on-cash yr 1 base | DSCR base | Simple payback |
|---|---|---|---|---|---|---|---|---|---|
| A · Malé 3BR apartment let | MVR 3.52m | 6.2% gross | 0.8% | 6.0% | 11.0% | 5.5% | −5.9% | 0.73 | ~20y |
| B · Hulhumalé Ph1 2BR furnished | 3.18m | 8.4% gross (contested) | 2.8% | 8.7% | 15.0% | 11.4% | −0.8% | 0.98 | ~16y |
| C · Hulhumalé Ph2 new-build | 3.95m | 4.8% gross | −5.8% | 1.2% | 7.3% | −9.0% | −10.1% | 0.53 | ~27y |
| D · Redevelop inherited Malé land | 48.0m | 6.8% on cost | −1.2% | 6.6% | 14.6% | 0.1% | −10.9% (stab.) | 0.62 | ~20y |
| E · Buy Malé land + build | 72.2m | 4.5% on cost | −6.3% | 1.2% | 8.4% | −6.5% | −4.1% (stab.) | 0.62 | ~27y |
| F · Hold vacant Malé land | 9.51m | no income | −0.1% | 2.1% | 5.2% | n/a | n/a | n/a | n/a |
| G · Thilafushi industrial land, leased | 8.59m | 4.2% gross | −3.6% | 2.6% | 11.4% | n/a | n/a | n/a | ~27y |
All IRRs are 10-year, post-tax, nominal MVR [OWN-CALC]. Entry outlays include 1% friction and initial capex; D and E outlays vary by scenario with construction cost (D: 40–56m; E: 59–86m; base shown). USD at peg: model A entry ≈ USD 228,000; model D ≈ USD 3.1m; model E ≈ USD 4.7m.
Three findings dominate. First, base-case returns cluster between 1% and 9% nominal — and the weighted-average lending rate in the banking system is ~11.3% (MMA WALR, 11.29% at March 2026, via the construction-economics dossier) [OFFICIAL, MMA], the cheapest mortgage 7.5%, construction money 11–11.75% [OFFICIAL(bank)]. Property in Greater Malé, bought at asking prices and let at asking rents, earns less than the price of the money that would buy it, in every strategy, in the base case.
Second, leverage therefore subtracts value. Debt-service coverage at entry is below 1.0 in every financeable base case (0.53–0.98): the building does not pay the loan; the owner’s salary does. Gearing cuts the base-case IRR in four of the five financeable models (A: 6.0% to 5.5%; C: 1.2% to −9.0%; D: 6.6% to 0.1%; E: 1.2% to −6.5%), and year-one cash-on-cash is negative in all five — even model B, the strongest, runs −0.8% before turning positive as rents grow. Only in B (and in the high scenarios generally) does the exit gain overcome the negative carry, and B’s advantage rests on a contested price basis: at the mid-2026 market-guide price of MVR 2.5–3.5m for a Phase 1 2BR [ASKING] the model returns 8.7%; at the 3.91m structured-portal ask [ASKING], the same flat returns 5.7% unleveraged and leverage stops helping. The 3.91m ask is a Phase 2 new-build comp, used here as an upper bound rather than a Phase 1 like-for-like. We carry both readings; the conflict is unresolvable without transaction data. This is also a statement about the banks: Maldivian housing credit is salary lending secured on property, not property lending — exactly consistent with underwriting practice (EMI capped at ~50% of net income [OFFICIAL(bank)]) and with mortgage depth of only ~5.4% of GDP [OWN-CALC on MMA May 2026].
Third, the two routes that clearly pencil are the two the market cannot scale. Model D (6.6% base, 14.6% high) works only because inherited land enters at zero cost basis — and inherited Malé land is precisely the asset trapped in co-ownership deadlock, where one objecting heir blocks everything [OFFICIAL, Land Act ss.17, 19, 25, 31]. Model B works on USD-adjacent furnished rent and a cheap entry that may not exist at scale. Everything freely purchasable at asking prices — C, E, G, F — returns 1–3% base: below Maldivian deposit rates in some years, and far below the sovereign’s own borrowing cost.
Model notes
A — Malé 3BR let (entry MVR 3.485m, Aafaluge-type new-build [ASKING]; rent 16,500–20,000 unfurnished [LISTING]): the honest middle of the market. 6.0% base IRR, 20-year payback. Works as a store of value with income; fails as a leveraged investment.
B — Hulhumalé Phase 1 2BR furnished (rent 18,000–25,000 [LISTING/INDUSTRY-EST]): the best risk-adjusted market entry if the 2.5–3.5m price guide is real. Phase 1 is the one sub-market with tenant depth (finished, lift-served, expatriate-acceptable stock) and the strata-titled, MRPS-financeable product banks will actually fund. Furnishing capex of MVR 150,000 plus a year-5 refresh is in the model.
C — Hulhumalé Phase 2 new-build (entry 3.91m from-price [ASKING]): the weakest strategy in every scenario, and the one being marketed hardest. Phase 2’s rental ceiling is administered: Hiyaa flats at MVR 4,984/month all-in and FDC at 9,000–10,500 [OFFICIAL] anchor tenant expectations, an informal sublet market undercuts private landlords, and thousands of further administered units are in the pipeline. A negative-rent-growth low case is not a stress test here; it is a plausible outcome.
D — Redeveloping inherited Malé land (2,000 sqft / 186 sqm plot, 16,000 sqft / 1,486 sqm gross build, 12 units): 6.8% yield on cost, 6.6% base IRR — the best unleveraged base return available, purely because the land is free. The model prices no consent risk: assembling heir agreement is the binding constraint (paper 9), construction debt at 11.5% destroys the return (leveraged base 0.1%), so this is a strategy for families with patient equity.
E — Buying the land and building the same building: base 1.2%. The only difference from D is paying MVR 19–30m for land (9,500–15,000/sqft) — which converts the best strategy into one of the worst. The asking price of Malé land is the developer’s margin, transferred to the seller.
F — Holding vacant Malé land (1,000 sqft at the 9,412 sold-marker level ≈ MVR 9.4m / USD 610,000): base 2.1%. The finding is not the return; it is that the return is achieved with zero effort, zero risk of tenant default, and zero holding cost — no property tax, no vacancy tax exists [OFFICIAL-absence]. Against model E’s 1.2% for enormous effort and construction risk, doing nothing dominates building. That single comparison is the deepest explanation of underdeveloped prime plots in Malé.
G — Thilafushi industrial freehold, leased (5,000 sqft at the MVR 1,700/sqft floor [OFFICIAL]): base 2.6%. At legacy lease rates of ~MVR 5–7.5/sqft/month [MEDIA/LISTING], the state’s posted land price already capitalizes the port-relocation and Thilamalé-bridge thesis; the 11.4% high case requires both to deliver. Buying at the floor is buying the infrastructure forecast, not the current income.
Benchmarking: what instructs, what misleads
The Maldives yields like a frontier market — roughly 6–7% gross against 2–3% in Singapore and Hong Kong and ~5.7–5.8% in Malta and Cyprus [LISTING-grade crowd data; Maldives = our own calc] — and the models above show why the spread is not a bargain: it is compensation for illiquidity (buyers must be Maldivian citizens [OFFICIAL]), FX-convertibility risk (a 31–34% parallel premium [MEDIA]), tenure and registry friction, and zero institutional exit (the Maldives is unrated in JLL’s GRETI transparency index — invisible to global capital [INDUSTRY-EST]).
Comparators that instruct. Colombo is the price anchor: Malé new-build asks ~USD 2,600/sqm at peg versus Colombo’s ~2,170 [ASKING both sides] — near-parity despite Maldivian income per head being ~3× Sri Lanka’s [OFFICIAL, WB 2024]. Both are shallow-credit markets where capital-city prices decouple from wages; Colombo 2022 is also the live rehearsal of the disorderly-FX tail. Malta is the institutional counterfactual — near-identical population, land area and density, but EU-grade institutions produce 40%+ mortgage depth and 5.7% yields with actual liquidity; the Malé–Malta gap is an institutions gap, not a geography gap. Mauritius supplies the policy template for ring-fenced foreign products (US$375k minimum ticket + residency) against which any Pearl-Residence-type Maldivian scheme should be priced — at a discount, given a lease-not-freehold product and a weaker governance score (CPI 39 vs 48 [OFFICIAL-adjacent]). Seychelles (2008 default, float, recovery within ~3 years) bounds the devaluation tail as severe repricing, not annihilation. Barbados and Fiji show pegs managed through crisis with controls; Hong Kong teaches exactly one thing — leasehold tenure is no barrier to liquidity when registries and courts are credible.
Comparators that mislead. Singapore/Hong Kong price levels (USD 15,000–25,000/sqm) embed financial-centre incomes, 40–90% household-credit depth and exit liquidity — “Malé is cheap versus Singapore” is the most seductive wrong conclusion this table offers, because Malé lacks every load-bearing input of the Singapore price. Numbeo-style crowd ratios omit five of our ten comparators and produce artifacts where thin (Sri Lanka’s 55.4 price-to-income is a post-crisis sample artifact [LISTING-grade, flagged]); any Maldives-versus-world affordability chart built on them is built on air. “Tourism-dependent peers” as a class understates the Maldives as the extreme point — 84.7% of exports versus Fiji’s 49.7% next-closest [OFFICIAL, WB 2019]. Golden-visa markets price EU/US retiree demand the Maldivian constitution excludes. And global construction-cost indices flatter the Maldives’ nominal USD 1,000–1,600/sqm build cost while missing the parallel-FX surcharge and per-island logistics that make it one of the expensive places in the set to build relative to income.
What we don’t know
Every input above the OWN-CALC line is asking, listing or administered tier: no model output can be validated against a single verified transaction, and asking-based IRRs inherit asking-price optimism at entry and an assumed capitalization at exit. There are no vacancy, operating-cost or achieved-rent statistics; our 6%/15% base assumptions are stated guesses. The exit market itself is untested — the buyer pool is citizens-only by law, and we cannot observe how long a MVR 48m redeveloped building would take to sell, or at what discount. Model B’s entry price rests on two conflicting sources we carry unresolved. Model C’s oversupply calendar (administered completions by year) is not published. Model D’s heir-consent probability is unquantifiable. The models are nominal-MVR at the peg; a step devaluation would rescale every figure, and USD-indexed rents would reprice sharply upward in MVR — directionally favouring owners of USD-lettable stock, but we decline to model a devaluation date. Finally, incomes: the last survey is HIES 2019, so every affordability-linked rent ceiling in the low scenarios is anchored seven years back. These are the gaps a transaction register, a rent index and a vacancy survey would close — none exists as of July 2026.