India's stalled-projects story is usually told in national aggregates, 6.29 lakh delayed homes, SWAMIH's 61,000 rescues. Maharashtra lets you zoom all the way in: the regulator publishes the actual lists, lapsed registrations by year, suspensions, deregistrations, revocations, with names and numbers. The lists are the state's distress map, updated continuously, and almost nobody reads them. This piece is about reading them.
Key takeaways
- MahaRERA publishes the primary distress data: year-wise lapsed lists (2019 through 2024 and onward), suspension and abeyance lists, deregistered and revoked projects.
- The baseline is higher than anyone assumes: 26.1 percent of published Maharashtra registrations read Lapsed, De-Registered or Revoked, and no district is below 18 percent.
- So a catchment's distress share means nothing on its own. Compare it to its district, where the median is 28.4 percent and the range runs to 39.7.
- Correct the vintages for age: recent cohorts look healthy because a young registration has not had time to lapse, not because the market fixed itself.
- The lists convert 'is this area safe?' from a feeling into a lookup.
What the lists are
Four documents, all public, all downloadable from the regulator: lapsed projects by year, registrations whose validity expired without completion or extension, the largest and most telling list; suspended or in-abeyance projects, the enforcement half-states; deregistered projects, the orderly exits under Order 42/2023; and revoked or void registrations, the terminations for cause. Each row is a project that promised a completion date to the public record and could not keep the registration alive.
Their analytical value is the address column. Aggregate the lists by taluka or pin code and Maharashtra's distress geography appears: which corridors absorbed their launch waves and which choked, which vintage years produced the failures now surfacing, and which localities carry so much dead stock that their surviving projects compete against ghosts.
What the map actually looks like
The lists describe a condition, and the first question anyone should ask before reading one is what normal looks like. It is much higher than most readers expect.
Across 55,913 published Maharashtra registrations, 14,584 (26.1 percent) currently read Lapsed, De-Registered or Revoked. Roughly one registration in four on the public record is dead. That is the baseline every catchment figure has to be read against, and it is the single most useful number in this article.
Source: ReraGenie analysis of 55,913 published MahaRERA projects, as updated on 21 September 2026, counting a registration as dead where its filed status reads Lapsed, De-Registered or Revoked. Districts with fewer than 200 published projects are excluded.
The whole state sits between 18 and 40 percent, and the spread between the best and worst district is 21.6 points. That reframes the exercise: the question is never whether a catchment has lapsed registrations in it, because they all do. It is whether it has more than its neighbours, and by how much.
Note also that the suspension list overlaps the lapsed one rather than sitting beside it. The abeyance flag is carried by 5,922 projects, and 5,175 of those are on registrations that already read Lapsed. Adding the four lists together therefore double counts; read them as one population with flags, not four populations.
An analogy: the flood-plain map
Before buying a house, sensible people check the flood map, not because their plot flooded, but because water respects geography, not optimism. Project distress behaves like water: it follows the low ground of weak demand, over-supplied corridors and lender retreats, and it revisits the same basins. The lapsed lists are the flood map of Maharashtra real estate, drawn by the regulator from actual failures, and consulting them before a purchase is exactly as radical as checking where water goes.
Reading the map as a buyer
- Compute your catchment's distress share, then compare it to 26 percent. Dead registrations as a fraction of the catchment's total, the same geometry as the micro-market method. Do not carry an intuition about what a bad number looks like into this: a catchment at a fifth is doing better than every district in the state, and one at a quarter is simply average. What earns attention is a catchment running well above its own district, which is a comparison you can only make by computing both.
- Read the vintages, and correct for age before you conclude anything. This is the step that misleads, because the raw numbers look like a clean recovery and are not. Dead share by registration year runs 30 percent for 2017, peaks at 42 percent for 2020, then falls to 22 percent for 2023, 7 percent for 2024 and 1 percent for 2025. Read naively that says the market fixed itself. It does not. A registration lapses when its validity runs out without completion or extension, so a 2025 registration has not had time to die yet. The young cohorts are not healthy, they are young. What the vintage table genuinely shows is that the 2019 to 2021 cohort is the damaged one, and any catchment whose stock is concentrated there deserves the stricter read regardless of how its recent registrations look.
- Rate the survivors by the company they keep. A healthy-looking project in a high-distress basin faces suppressed pricing, wary lenders and buyers, and whatever local condition drowned its neighbours. Its filings deserve the stricter read: current quarters, certificate trail, extension history.
- Check your shortlist against the lists by name. The two-minute lookup that catches expired registrations still being marketed, the commonest abuse the lists exist to end.
Rohit (illustrative, as ever) ran the geography before his second Maharashtra purchase: two shortlisted corridors in the same district, similar pricing, similar brochures. Against a district running 27 percent, corridor one came in below it and its dead stock was almost entirely 2018 and 2019 vintage, a basin that had already flooded and drained. Corridor two ran well above the district, and the part that decided it was that several of its dead registrations were 2022 and 2023 vintage, which is early to die. Its "bargain" pricing suddenly had a denominator. He bought in corridor one, at a price the flood map had already explained.
The lists also power the state's cleanup narrative honestly: revival mechanisms, transfers to a new promoter, which need two-thirds of the buyers' consent, allottee-association takeovers and deregistrations all begin from this inventory. The two largest cleanups, Amrapali's and Jaypee's, had no such inventory to start from and were reconstructed in court instead, which is the subject of what the big stalled-project collapses looked like in the filings. A locality's lapsed list shrinking year over year is genuine recovery; one merely growing quieter is not, and the difference is visible in the year-wise files.
One thing the map does not carry
The register publishes a registration's status. It does not publish the date that status changed, so nothing above can tell you WHEN a project lapsed, only that it currently reads lapsed and what year it was registered. Those are different questions, and the difference matters for step two: "dead stock of 2019 vintage" means registered in 2019, not failed in 2019.
MahaRERA's own year-wise lists are the other half, because they are organised by the year of the lapse. Use the published lists for timing and the register for vintage, and do not let one stand in for the other.
Where to run this free
The district pages carry every registration with its current status, so a district's dead share is countable by hand, and each project page states its own status plainly rather than showing every registration as healthy. That is most of the lookup in step four, at no cost.
The Rs 499 buyer report is for the part a status field cannot answer: whether the promoter in front of you has dead registrations elsewhere in the state, which is the flood history of everyone involved rather than of the plot. On a purchase in a high-distress catchment that is the question worth paying for. Sign up on ReraGenie before buying in a basin.
The one-line summary
The regulator publishes the flood map, and the water level is higher than anyone assumes: one registration in four across Maharashtra is dead, no district is under 18 percent, and the 2020 cohort is the damaged one. Compute your catchment's share against its district rather than against a feeling, correct the vintages for age before reading recovery into them, and hold survivors in bad basins to a stricter standard.
Methodology and sources
- Status counts, district shares and vintage: ReraGenie analysis of 55,913 published MahaRERA projects, as updated on 21 September 2026. A registration counts as dead where its filed status reads Lapsed (14,003), De-Registered (500) or Revoked (81), totalling 14,584, or 26.1 percent.
- The abeyance flag is orthogonal to status rather than a status of its own: 5,922 projects carry it and 5,175 of those also read Lapsed, so the lists overlap and must not be summed.
- District figures cover the 22 districts with 200 or more published projects. A district is not a catchment. These are the baseline a 2 to 3 kilometre catchment should be compared against, and a catchment can and does deviate from its district by far more than districts deviate from each other.
- Vintage is the year of a project's filed registration date, not the year it lapsed, which MahaRERA does not publish. The falling dead share in recent cohorts is an age effect and is labelled as one wherever it appears.
- National delayed-home and SWAMIH figures are from the linked article and its own cited sources, not from the register.
Evaluating a project right now?
The ReraGenie buyer report reads every filing for one project and sets out the red flags, the checks that came back clear and what to verify before you book, each fact with its filing date. Rs 499, one time.
See the buyer report