A home sale is two events. The developer counts the booking. The State counts the registration. They are months apart, recorded by different parties, in different units, over different geographies — and they are not the same number. This brief maps both records and explains why they cannot be netted.
Listed developers reported record FY26 pre-sales bookings. Maharashtra's registration office recorded 1,50,254 Mumbai property registrations in CY 2025, a 14-year high. The booking and the registration are two stages of the same purchase — and the two numbers do not subtract.
Maharashtra's Inspector General of Registration recorded 1,50,254 property registrations in Mumbai in CY 2025 — a 14-year high — generating ₹13,487 crore of stamp-duty revenue, up 11 percent year-on-year and itself a 14-year high. FY25 stamp duty was ₹12,899 crore; December 2025 alone saw 14,447 registrations and ₹1,263 crore of duty. The State's FY26 stamp-duty revenue target is ₹63,500 crore.
The IGR record is the legal record of property changing title. It counts registered conveyance documents and the stamp-duty revenue collected on them. Two things it is not: it is not an aggregate rupee-value of transacted property — IGR publishes a count and a revenue figure, not a summed transaction value — and it is not developer-attributed; the records are masked, and no public dataset tags a registration to a named builder.
One more boundary matters. "Mumbai" in this tracker is Mumbai municipal limits — not the wider Mumbai Metropolitan Region — and it covers all property: residential and commercial, new sales and resales. It is explicitly a resale-inclusive count. Listed developers are a minority slice of total Mumbai registrations.
Listed developers reported record FY26 pre-sales. Pre-sales is a developer-defined operating metric — bookings, counted at the allotment-letter stage. The State registration is a conveyance — counted months or years later, after construction. They are two markers on the same purchase, and the timeline below shows how far apart they sit.
Listed developers reported record FY26 pre-sales bookings: Godrej Properties ₹34,171 crore (+16%), Prestige Estates ₹30,024 crore (+76%), Macrotech / Lodha ₹20,530 crore (+16%), DLF ₹20,143 crore, Oberoi Realty ₹5,447 crore (+4%). Every one of those numbers is verbatim from a filed disclosure, and every one is a legitimate, RERA-era industry-standard metric.
Pre-sales is counted at Stage 1 of the timeline — the booking, when an allotment letter is issued. The IGR record is counted at Stage 3 — the registered conveyance, after construction. The two markers describe the same purchase at two stages, separated by the construction and conveyance calendar. The pre-sales metric is the forward number; the registration record is the backward number; the space between them is time.
It is tempting to subtract one record from the other — to take developer pre-sales, take State registrations, and read the difference as something. The two numbers do not support that arithmetic. Three independent mismatches make any single "gap" figure non-constructible from public data.
Geography. Developer pre-sales is booked pan-India. Godrej, DLF, and Prestige book most of their pre-sales rupees outside Maharashtra — and no developer discloses a state-level split of bookings. The Maharashtra IGR record is, by definition, inside Maharashtra. The two cover different maps.
Unit. Pre-sales is a rupee value of bookings. The IGR record is a count of registered documents plus the stamp-duty revenue on them — not a summed rupee-value of transacted property. A rupee-for-rupee comparison is not constructible because one side of it is not published.
Population. IGR registrations are all-property and resale-inclusive — most registrations are not new-build at all, and listed developers are a minority slice of the total. Developer pre-sales is new-build only, from listed developers only. The two count different populations of transactions.
Geography, unit, population — three reasons the developer metric and the State record are non-comparable by construction. M-17 does not net them. It maps both, and explains the difference.
| Dimension | Developer pre-sales | State registration (IGR) |
|---|---|---|
| What it counts | Rupee value of bookings | Count of registered conveyances + stamp duty |
| The event | Allotment letter + threshold payment | Sale deed executed + registered |
| When in the timeline | Stage 1 — forward indicator | Stage 3 — backward record |
| Geography | Pan-India · no state split disclosed | Inside Maharashtra |
| Population | New-build only · listed developers only | All property · new + resale + commercial |
| Recorded by | The developer | The State |
Because the two records count different events, they answer different questions — and a reader gets the right answer only by reading the right record.
The pre-sales metric tells a reader about developer sales momentum. A booking is a forward indicator: it is revenue the developer will recognise as construction completes. A reader who wants to know "how much are developers selling, and is it accelerating" reads pre-sales — and reads it across developers, who all define it similarly.
The IGR registration record tells a reader about completed title transfer and State revenue. A registration is a backward-looking record: a conveyance actually executed, stamp duty actually paid. A reader who wants to know "how much property changed hands legally, and how much revenue the State collected" reads IGR.
Neither is the other. Neither is wrong. A 14-year high in Mumbai registrations and record developer pre-sales are two corroborating signals of a strong residential market — read together, as two readings of the same strength, not subtracted as if one were a correction to the other. The booking-to-registration lag is the structural fact between them: in a rising market, the forward number runs ahead of the backward number by sequence alone.
A developer booking and a State registration are two records of two events in one purchase. The booking is counted by the developer, at the allotment-letter stage, as a rupee value, pan-India. The registration is counted by the State, at the conveyance stage, as a document count and stamp-duty revenue, inside one State, resale-inclusive. They are separated by the construction and conveyance calendar, and they are constructed so differently that they cannot be subtracted. Mirror Brief 17 makes one claim: read each record for the question it answers, and do not net one against the other. Every developer figure in this brief is verbatim from a filed disclosure and is a legitimate post-RERA metric. The brief decomposes two records; it alleges nothing about either.
None of these are recommendations. They are decompositional implications — how to read two records that are routinely conflated.
After mapping the State registration record and the listed-developer pre-sales filings, four things remain undisclosed in public data. Mirror Brief 17 does not claim to fill them — and stays at aggregate market level because of the first.
One. Per-developer Maharashtra registration counts. IGR records are masked — no public dataset attributes a registration to a named builder. M-17 therefore makes no per-developer registration statement of any kind.
Two. Aggregate transacted rupee-value in the IGR record. IGR publishes a document count and stamp-duty revenue — not a summed transaction value. A rupee-for-rupee comparison with pre-sales cannot be constructed from public data.
Three. The Maharashtra share of each pan-India developer's pre-sales. No developer discloses a state-level split of bookings.
Four. Pune district full-year CY2025 registration count — not surfaced in accessible public sources. M-17 anchors on the Mumbai municipal record, where Knight Frank's monthly tracker provides a verified series.
All URLs verified live by RAOSCAFF Investment Researcher Phase 0 source viability check 2026-05-21. Full URL list in FACTS.md § I. New figures added to the Cross-Brief Number Registry on M-17 publish.
Research approach. Mirror format — anchor on the State registration record (Maharashtra IGR, via Knight Frank's monthly Mumbai tracker) and the listed-developer pre-sales filings, and explain the measurement-universe divergence between them. No primary collection. No per-developer registration attribution — that data does not exist in public form.
Source standards. Every figure traces to the IGR record, a Knight Frank tracker release, or a listed-developer FY26 filing — fetched live 2026-05-21 by Investment Researcher Phase 0. Cross-brief numbers reuse the RAOSCAFF Cross-Brief Number Registry byte-identical (M-16 listed-developer pre-sales: DLF ₹20,143 crore · Oberoi ₹5,447 crore · Prestige ₹30,024 crore).
Construction. FACTS.md is the source-of-truth file. The hero is a booking-to-registration lifecycle timeline SVG. Phase 0 caught and corrected one fabrication risk: the originally-scoped clean numeric "gap percentage" between pre-sales and registrations was dropped — three universe mismatches (geography, unit, population) make any single gap figure non-constructible from public data. The editorial spine is the measurement-universe divergence: two events, two records, non-comparable by construction.
Limitations. Per-developer Maharashtra registration counts do not exist publicly — IGR records are masked. IGR publishes a document count and stamp-duty revenue, not a summed transaction value. No developer discloses a state-level pre-sales split. The brief stays at aggregate market level and anchors on the verified Mumbai municipal series.
Editorial position. Predict-not-recommend. Defamation-disciplined, with three binding red lines observed throughout: no per-developer attribution, no "gap percentage" netting non-aligned universes, no loaded framing. RAOSCAFF holds no advisory relationship with any developer named. Every developer figure is verbatim from a filed disclosure and is a legitimate post-RERA operating metric. The brief decomposes two records and alleges nothing about either. RAOSCAFF does not recommend purchase, sale, or holding of any property or security. The reader's decision is the reader's.