Same Unit, Different Story: Why Our Repeat-Sales Index Shows 6% Growth vs ADREC's 71%

ADREC's official residential sale price index has risen from a base of 100 in Q1 2020 to 171.0 by Q1 2026 — a cumulative 71% appreciation, with apartments alone up 20% year-over-year as of June 2026. Meanwhile, Daar's own repeat-sales index — built directly from ADREC's transaction-level CSV data using a Case-Shiller-inspired methodology — puts cumulative appreciation over the same period at just 6%. Both numbers come from the same underlying government data. Both are correct. They are simply measuring two different things, and the gap between them is itself one of the more interesting signals in the Abu Dhabi market right now.

ADREC's official index is a cross-sectional median: every quarter, it looks at whatever mix of units sold and tracks the median price movement of that basket. This is the standard, internationally-recognized approach, and it's the right tool for answering 'what is the market price today.' But it has one well-known weakness: when the mix of what's transacting changes — specifically, when expensive new-build inventory enters the sales pool — the median moves even if not a single existing owner's unit changed in value. A market that sells mostly AED 800K apartments one year and mostly AED 3M waterfront villas the next will show a rising median regardless of whether any individual unit appreciated.

The repeat-sales index strips that composition effect out by construction. It only counts a price change when the *same unit* — matched on project, layout, and floor area — sells twice: once as a primary (off-plan or first) sale, and later as a secondary (resale). The measured return is the actual price change that specific owner experienced, annualized and pooled across every matched pair in the quarter (1,508 matched pairs in Q3 2026 alone, filtered to sensible hold periods of 6 months to 10 years and annualized returns between -30% and +50% to exclude likely data errors).

The scale of the gap — 71% cumulative vs 6% — is large enough that composition shift is clearly the dominant explanation, not a rounding difference. It lines up with everything else in this report: H1 2026 alone saw two entirely new districts (Ramhan Island, Fahid Island) go from zero transaction history to a combined AED 7.3 billion, and the >AED 28,000/sqm price band expanded 3.3x to 46% of transaction value. Every one of those is new, more expensive supply entering the transaction pool — exactly the kind of mix-shift that inflates a cross-sectional median without requiring any existing owner's unit to gain a single dirham in value.