Underwriting a developer from their MahaRERA record: a credit checklist

Real estate credit files are thorough about the borrower entity and thin about the sponsor's behaviour. The financials get spread, the security gets valued, the title gets searched, and the question that actually decides recovery, how does this promoter behave when a project goes wrong, gets answered by reputation. Since 2017, that question has had a documentary answer. Every registered project the promoter has touched carries a public record of promises, revisions, certificates and disputes. This is the origination checklist that reads it.

Why conduct data beats reputation#

The lending logic first. Real estate exposure fails through the sponsor far more often than through the market: money moves between projects, timelines slip serially, and disputes consume the completion capital. None of that is visible in the borrowing entity's financials, because the borrowing entity is usually a fresh SPV with no history at all. What the register offers is conduct data at the sponsor level: sworn, dated, and spanning every project the group has registered since 2017. The same record buyers are told to check before booking, the builder's track record, read at portfolio width with a credit question. MahaRERA once planned to grade projects on much the same filed inputs, and its shelved grading matrix still works as a rubric for a credit desk.

The six-dimension checklist#

Dimension 6 is the one most often skipped. A promoter's capacity to finish your project is a function of every other project they must finish first. The register gives the full load: units, booked percentages, promised dates. In a market carrying heavy inventory, that load is the difference between a sponsor and a hostage; Pune alone carried 57,879 unsold units into H2 2026, up 19 percent year on year (Knight Frank, India Real Estate, H1 2026), and someone is funding every one of them.

Worked example: scoring a Rs 40 crore proposal#

Illustrative numbers throughout. A mid-size Pune promoter group seeks Rs 40 crore of construction finance through a fresh SPV. The register, aggregated across four entities, shows nine registrations since 2018: five completed, four live.

Score the six dimensions. Delivery: the five completed projects promised an average 36-month build and delivered at a median of 44 months, a slip of plus 8 months, with the worst at plus 19. Extensions: three of four live projects carry section 7 extensions, all citing approval delays, two filed in the last 18 months. Filing discipline: one live project missed the April 2026 QPR deadline and appears in the show-cause list of 8,212 (MahaRERA enforcement sweep, May 2026). Litigation: two projects disclose cases, one an allottee delay matter, one a title dispute on land the group does not propose to mortgage to you. Warrants: none, a genuine positive given that MahaRERA's warrant ledger shows Rs 792 crore ordered and Rs 268.87 crore recovered across 1,291 complainants by November 2025 (MahaRERA data via Business Standard, December 2025). Portfolio load: 1,840 unsold units across the four live projects against a trailing sales pace of about 60 units a quarter, roughly 30 quarters of inventory at current velocity.

Now translate the score into terms, because a record that does not move the term sheet was never underwriting. The plus-8 median slip prices as tenor: a facility maturing on the promoter's promised date is mispriced by their own history, so the maturity carries an eight-month buffer and the interest reserve is sized to it. The extension pattern prices as covenants: extension applications anywhere in the group become notification events, and a second extension on the financed project a review event. The show-cause appearance prices as conditions precedent: filings current everywhere before first draw. The portfolio load prices as structure: releases gated to the financed project's own certified progress, never to group cash flow, because 30 quarters of group inventory is 30 quarters of competing claims on the sponsor's liquidity.

The read: not a decline, but not the file the relationship manager described. Price the median slip into the tenor, covenant the QPR currency and extension notifications, take the portfolio load to committee honestly, and size against the group's finishing capacity rather than the SPV's project math. The whole annexure took one afternoon, because every input was filed.

Reading the delivery record honestly#

The six dimensions produce numbers; underwriting judgment is in the adjustments, and three matter enough to spell out.

Survivorship. The register shows the projects the group registered; it does not show the land deals they walked away from or the launches they shelved. A promoter with three completed projects and no live ones is not necessarily disciplined; they may simply be small. Scale the record against the group's claimed history and check that the registration application's self-declared past projects reconcile with what the pitch deck claims. Gaps between the two lists are questions, and occasionally answers.

Cycle adjustment. A delivery slip earned across 2020 and 2021 is not the same datum as one earned in a normal market: the force majeure window MahaRERA granted during Covid moved every promoter's dates, and the force majeure rules define what counted as legitimate. Date-stamp every slip before scoring it. A promoter whose only extensions cluster in the pandemic quarters has a different conduct profile from one whose extensions arrived in 2019 and again in 2025, even if the median slip is identical.

Litigation quality. Counting cases is screening; reading them is underwriting. An allottee delay complaint, a title suit on project land, and a financial creditor's proceeding are three different risks wearing one disclosure label. The register flags the existence; the litigation-history method grades the substance. Nothing in the checklist substitutes for reading the two or three cases that matter.

These adjustments are also where the register disciplines the lender, not just the borrower: a desk that scores mechanically will decline good sponsors with pandemic-shaped records and approve small groups with short clean histories. The data is the floor of the analysis, not the ceiling.

When there is no record to read#

The checklist assumes a track record exists. Measured across the register, most sponsors do not have one. Of 42,513 distinct promoters behind 55,682 published Maharashtra projects on our August 2026 read, 86.2 percent have registered exactly one project and 94.3 percent have registered one or two. Only 181, four in every thousand, carry ten or more. Two-thirds of all published projects, 65.8 percent, sit under a promoter with no second registration anywhere in the state.

The naive reading of that is wrong, and worth stating before a desk acts on it. A single-registration promoter is not necessarily a first-time developer: a fresh SPV per project is ordinary structuring, and the operating group behind it may carry decades of delivery under names the register will not connect for you. That is exactly why step 1 is group mapping rather than an entity lookup, and why it decides whether the other five dimensions produce anything at all.

Where the group genuinely is new, the gap is real. Only 39.4 percent of promoters carry even one completed registration, so for the majority the delivery dimension is empty. Score it as absent, never as clean. An unpopulated field is not a good field, and the difference between those two readings is the difference between pricing risk and ignoring it.

Three structural limits sit underneath all of this. The register begins in 2017, so a sponsor's pre-RERA decades are only ever self-declared. It holds no price, so collections cannot be inferred from booking counts. And it lags a quarter, which is acceptable at origination and matters at monitoring. Against a book where commercial real estate credit grew 16.2 percent year on year in January 2026 (RBI, sectoral deployment of bank credit data, January 2026), the number of thin-record sponsors arriving at credit desks is rising rather than falling.

The Bhandarkar family office (illustrative, as our stories always are) ran this checklist on two structured-debt proposals last quarter, both sponsors carrying respected brand names. One group's record read exactly like its pitch: median slip of two months, filings current everywhere, one old and closed litigation. The other showed four extensions across three cities and a deregistered project under a sister entity that the pitch deck's track-record slide had quietly omitted. Same asking rate. The pricing conversation that followed was different, which is the entire point of the exercise.

The same record works one level up, for a family office choosing between funds rather than deals: reading a real estate AIF's sponsor record turns this checklist into limited partner diligence.

The manual cost is the honest barrier: aggregating a group across entities is work the portal does not support, because MahaRERA records each SPV separately and no parent. ReraGenie shortens the rest of it. Each developer page lists every registration under a filed promoter name with its status and measures its completions against the dates it promised, so assembling the group is a matter of reading a few named pages rather than hundreds of filings; the grouping itself stays your call. The Rs 2,999 project analysis sets a single named project against its micro-market, its operators and a risk register, with every figure cited to its filing. Before the next term sheet, pull the entities' records and read them the way you read account conduct.

The one-line summary#

The SPV has no history; the sponsor has nine projects of it, sworn and public. Map the group, score the six dimensions, and let the register's version of the track record set the terms the pitch deck must answer to.